The Economy of the Roman World as a Complex Adaptive System. Testing the Case in Second to Fifth Century CE Sagalassos

COLLECTION LATOMUS Fondée par M. RENARD en 1939 Continuée par J. DUMORTIER-BIBAUW et C. DEROUX Dirigée par D. ENGELS VOLUME 350 Paul ERDKAMP and Koenraad VERBOVEN (eds.) Structure and Performance in the Roman Economy Models, Methods and Case Studies ÉDITIONS LATOMUS BRUXELLES 2015 Table of Contents Preface ..............................................................................................................7 Koenraad VERBOVEN Models or Muddles? What are Theories Good for in Ancient Economic History? ............................................................................................................9 PART I. THEORIES AND MODELS ...............................................................15 Paul ERDKAMP Structural Determinants of Economic Performance in the Roman World and Early Modern Europe. A Comparative Approach ...................................17 Koenraad VERBOVEN The Knights who Say NIE. Can Neo-Institutional Economics Live up to Its Expectation in Ancient History Research? ............................................33 George GRANTHAM A Search-Equilibrium Approach to the Roman Economy .............................59 Jeroen POBLOME The Economy of the Roman World as a Complex Adaptive System. Testing the Case in Second to Fifth Century CE Sagalassos ..........................97 PART II. MODELS AND METHODS APPLIED ..........................................141 Wim BROEKAERT Recycling Networks. The Structure of the Italian Business Community on Delos .......................................................................................................143 Dennis KEHOE Poverty, Distribution of Wealth, and Economic Growth in the Roman Empire ..........................................................................................................183 Annalisa MARZANO Villas as Instigators and Indicators of Economic Growth ............................197 Anne KOLB Epigraphy as a Source on Ancient Technology ............................................223 Jeroen POBLOME The Economy of the Roman World as a Complex Adaptive System. Testing the Case in Second to Fifth Century CE Sagalassos 1 May I Introduce Myself? The website of the research network “Structural Determinants of Economic Performance in the Roman World” claims1 : “The study of ancient economies entered a new phase in the 1990s. Economic archaeology and natural science research provide new data on economic performance. Neo-institutional and development economics offer new theoretical frameworks. Comparative and ‘longue durée’ analyses gained central importance.” There is a lot that should excite Roman archaeologists in these lines and in the rest of the research network’s programme, especially those who, as in my case, have a background in the study of material culture 2 and the fields of artisanal production organization,3 long-term regional development 4 and interdisciplinary 1 http://www.rsrc.ugent.be/sdep J. POBLOME / D. MALFITANA / J. LUND, Tempus fugit. “FACTA” manent. Editorial Statement in FACTA. A Journal of Roman Material Culture Studies 1, 2007, p. 13-20; J. POBLOME / D. MALFITANA / J. LUND, Scherben bringen Glück. HEROM’s Editorial Statement in HEROM. Journal on Hellenistic and Roman Material Culture 1, 2012, p. 7-22. 3 J. POBLOME, The Potters of Sagalassos Revisited in M. FLOHR / A. WILSON (eds.), Beyond Marginality. Craftsmen, Traders and the Socioeconomic History of Urban Communities in the Roman World, Oxford, in press. 4 J. POBLOME, Word versus Dirt. History and Archaeology Applied to Proto-Historical Sagalassos in T. BOIY / J. BRETSCHNEIDER / A. GODDEERIS / H. HAMEEUW / G. JANS / J. TAVERNIER (eds.), The Ancient Near East, a Life! Festschrift Karel Van Lerberghe (Orientalia Lovaniensia Analecta 220), Leuven, 2012, p. 457-470; J. POBLOME / D. BRAEKMANS / M. WAELKENS / N. FIRAT / H. VANHAVERBEKE / F. MARTENS / E. KAPTIJN / K. VYNCKE / R. WILLET / P. DEGRYSE, How Did Sagalassos Come to Be? A Ceramological Survey in M. TEKOCAK (ed.), Studies in Honour of K. Levent Zoroğlu, Antalya, 2013, p. 527-540; J. POBLOME, Shifting Societal Complexity in Byzantine Asia Minor and Dark Age Pottery in LRCW4. 4th International Conference on Late Roman Coarse Ware, 2 98 JEROEN POBLOME archaeology.5 However, despite the excitement, the use of archaeology in the reconstruction of past economic patterns, structures and behaviour is not unproblematic or straightforward. In most cases, especially when performing fieldwork, ‘ifὅ’ and ‘maybeὅ’ are an inherent part of archaeological epistemology. An important part of my time on campaign, for instance, is spent in the pottery shed. This is often the place where the gap between the archaeological material and the wider object of study (in this case the Roman economy) becomes very clear and wide. By using pottery templates, functional, typological, chronological, taphonomical and quantified details are gathered from sherds from excavated and surveyed loci, resulting in integrated and systemized information on tens of thousands of fragments of broken pottery. Over the years, this practice has taught me two things. Firstly, that the use and meaning of numbers and totals of things found by archaeologists is never straightforward. For instance, locus SA-2012-PQ2-00041-0171, excavated in Eastern Suburbia at ancient Sagalassos (SW Turkey) contained 24 sherds of cooking vessels, weighing 659 g in total, and 831 sherds of the local tableware for serving and consuming food and drink, weighing 30,075 g in total. But, what does this actually mean in terms of real life in old Sagalassos? Each archaeological case needs translation to give meaning to ‘factὅ’. Such translation can be provided by theories, concepts and models. One such model will be explored in this paper. The second thing I learned in my pottery shed is that those real people in classical antiquity, particularly the Romans, threw away an awful lot of stuff. So, either these were all very clumsy and dirty people (but I am prepared to positively discriminate against that idea) or a lot of things for day-today consumption were available to most citizens in the Roman world. Inevitably, my work with numbers and trying to match ancient production with consumption drove me, eyes wide shut, towards that notorious academic battlefield of the study of the ancient economy. The first thing to do when in battle is to know your position. This helps to avoid unfortunate events such as, for instance, friendly fire. Considering the position of archaeology in the intellectual battlefield of reconstructing the past, the Cambridge historian and medieval numismatist Philip Grierson expressed his Cooking Ware and Amphorae in the Mediterranean. Archaeology and Archaeometry. Mediterranean: A Market without Frontiers (Thessaloniki, 7-10 April 2011) (BAR International Series), Oxford, Archaeopress, 2014, p. 623-642. 5 B. NEYT / D. BRAEKMANS / J. POBLOME / J. ELSEN / M. WAELKENS / P. DEGRYSE, LongTerm Clay Raw Material Selection and Use in the Region of Classical/Hellenistic to Early Byzantine Sagalassos (SW Turkey) in Journal of Archaeological Science 39, 2012, p. 12961305; D. BRAEKMANS / P. DEGRYSE / J. POBLOME / B. NEYT / K. VYNCKE / M. WAELKENS, Understanding Ceramic Variability: An Archaeometrical Interpretation of the Classical and Hellenistic Ceramics at Düzen Tepe and Sagalassos (Southwest Turkey) in Journal of Archaeological Science 38, 2011, p. 2101-2115. A COMPLEX ADAPTIVE SYSTEM 99 opinion as follows: “It has been said that the spade cannot lie, but it owes this merit to the fact that it cannot even speak.”6 Clearly, it would be uncouth to reduce Philip Grierson’s appreciation of archaeology to only one sentence, but as an active participant on the scene of historical archaeology, I sometimes wonder whether there is not some form of truth to this statement. I say this not because I, or any other fellow archaeologist for that matter, believe that we should be ashamed of the empirically driven epistemological trajectories of our archaeological discipline, but because examples of meaningful integration of historical and archaeological lines of thought are hard to come by. The research ὀetwὁὄkΝ “Structural Determinants of Economic Performance in the Roman World” should therefore be considered as a genuine opportunity to find crossdisciplinary synergies. Before the latter are achieved, however, we need to explain some basic points in order to understand each other’s scientific rationale. The main aim of this paper, therefore, is to present the archaeological heuristic framework I use in my approach to the ancient economy. 2 Complex Complexity Admittedly, with its use of machinery labelled “Inductively Coupled Plasma Mass Spectrometer”(to analyse the isotope balance in elements such as Strontium and Neodymium in order to determine geological provenances of raw production materials in antiquity) or of a technique called “Gas Liquid Chromatography”(to determine the presence and nature of chemical residue in archaeological artefacts to determine their function) archaeology—more specifically archaeological science in this case—may sound complicated. The important message to bring across, however, is that archaeology is not complicated, but complex. To most people, an airplane is complicated: it has lots of bits and parts that need to work together for it to operate. Contrary to a complex system, however, each airplane is supposed to function within a relatively narrow and preferably predictable range of possibilities. Archaeology on the other hand, as well as history, is complex, with a wide diversity of subjects, theories, methods, practices and epistemological traditions.7 Archaeology is characterized by its ability to zoom in and out on an issue, considering its micro-environment as well as its contribution to broader disciplinary currents, giving rise to diverging research questions and paradigms.8 As a matter of fact, archaeology harbours the fairly unique potential 6 P. GRIERSON, Commerce in the Dark Ages: a Critique of the Evidence in Transactions of the Royal Historical Society 9, 1959, p. 129. 7 See the arrays discussed in J. BINTLIFF (ed.), A Companion to Archaeology, Oxford, 2004 and I. HODDER (ed.), Archaeological Theory Today, Cambridge, 2012. 8 T. INSOLL, Archaeology. The Conceptual Challenge, London, 2007. 100 JEROEN POBLOME to explore past societal development by combining the dimensions of materiality and cognition with time and space.9 What this means in simple terms, is that besides the traditional archaeological ‘ifὅ’ and ‘maybeὅ’, there are also always more sides to an issue that you have not considered yet. For instance, ground-breaking as the latest historical approaches to the study of the economies of the Greco-Roman world may be,10 emphasizing how ancient economic structures fostered a potential of moderate growth on a regional scale, their innovative message has not yet been fully embraced in archaeology. More extensive and interdisciplinary archaeological case-studies on long-term regional development are still very much needed in order to map and compare aspects of societal development11 and improve our understanding of what “moderate growth” could mean. Moreover, although a generation of interdisciplinary archaeological research in the region of ancient Sagalassos has established a pattern of regional growth in the intensity of exploitation of agricultural assets and raw materials, settlement patterning and demographic densities and we deduce from this that societal complexity increased between archaic and Roman Imperial times,12 how did this affect the life experience of an average citizen in Roman Sagalassos, who did not know he was so-called better off than his archaic predecessor? Also, ‘region’ is a difficult concept to define, already in antiquity 13 but clearly also in archaeology.14 For instance, local communities within what is traditionally regarded as a region in antiquity could follow different trajectories.15 There is an implicit tension in archaeology in approaching regions and places16 that makes the development of meaningful comparisons of societal and economic 9 R. W. PREUCEL / L. MESKELL, Knowledges in L. MESKELL / R.W. PREUCEL (eds.), A Companion to Social Archaeology, Oxford, 2004, p. 3-22. 10 W. SCHEIDEL / I. MORRIS / R. SALLER (eds.), The Cambridge Economic History of the Greco-Roman World, Cambridge, 2007; W. SCHEIDEL (ed.), The Cambridge Companion to the Roman Economy, Cambridge, 2012. 11 E.g. J.-F.BERGER / L. NUNINGER / S. VAN DER LEEUW, Modelling the Role of Resilience in Socioenvironmental Co-Evolution. The Middle Rhône Valley between 1000 BC and AD 1000 in T.A. KOHLER / S.E. VAN DER LEEUW (eds.), The Model-Based Archaeology of Socionatural Systems, Santa Fe NM, 2007, p. 41-59. 12 POBLOME et al., How Did Sagalassos Come to Be? [n. 4]. 13 D. DUECK, Geography in Classical Antiquity, Cambridge, 2012. 14 R. TABOR, Regional Perspectives in Archaeology. From Strategy to Narrative (BAR International Series 1203), Oxford, 2004. 15 For example, Tanagra, Thespiae and Koroneia in Roman Boeotia as discussed in J. POBLOME / P. BES / R. WILLET, Thoughts on the Archaeological Residue of Networks. A View from the East in S. KEAY (ed.), Rome, Portus and the Mediterranean (Archaeological Monographs of the British School at Rome 21), London, 2012, p. 393-401. 16 G. REGER, Regions Revisited. Identifying Regions in a Greco-Roman Mediterranean Context in FACTA. A Journal of Roman Material Culture Studies 1, 2007, p. 65-74. For a useful, geographical introduction: A. HEROD, Scale, London, 2011. A COMPLEX ADAPTIVE SYSTEM 101 development in the past difficult at best. Yet, developing higher level synthesis views on regional trajectories is essential in order to understand the workings of the Roman economy in greater detail, going beyond GDP guesstimates or macroscale analyses.17 Apart from archaeological epistemologies being complex, working in historical archaeology also means that the subject matter is complex. Basically, whereas archaeology in general approaches the analysis of social complexity and its evolution in the long-term, historical archaeology is mostly concerned with specific periods and regions where the functioning of society itself had become complex. Historical archaeologists should consciously avoid aiming to establish a homogeneous, evolutionary account of social and regional development (from chiefdom to polis to state). Instead they should make the particularities and inconsistencies in the archaeological record of their particular study regions or domains contribute to debates in social, economic and regional archaeology. In trying to understand the Roman economy it is important to have an improved understanding of social complexity, and how this is tied to regionally specific pathways.18 Although, through time, society has become more complex, “complexity should not be conceived as the ultimate goal of social evolution."19 Rather than evolution, we need to understand social complexity. Different communities in the Roman Empire experienced evolving complexity in different ways, depending on how inequalities were established and contested. Social complexity and its development occurred on many levels, on many scales and in many contexts forming a system that cannot be explained by reducing it to its component parts.20 3 Heuristic Frameworks The more complex things are the clearer the mind-map and intellectual concepts need to be. Combining and contrasting scales of analysis or ‘multi-scalarity’ is a current buzzword in archaeological circles.21 In order to approach the issues raised above, we can deploy conceptual triangles as heuristic tools (Fig. 1). 17 W. SCHEIDEL, In Search of Roman Economic Growth in JRA 22, 2009, p. 46-70; A. WILSON, Indicators for Roman Economic Growth: a Response to Walter Scheidel in JRA 22, 2009, p. 71-82. 18 E.g. P. A. J. ATTEMA / G.-J. BURGERS / P. M. VAN LEUSEN, Regional Pathways to Complexity. Settlement and Land-Use Dynamics in Early Italy from the Bronze Age to the Republican Period (Amsterdam Archaeological Studies 15), Amsterdam, 2010. 19 R. CHAPMAN, Archaeologies of Complexity, London, 2003, p. 7. 20 R. BENTLEY / H. MASCHNER, Complex Systems and Archaeology. Empirical and Theoretical Applications, Salt Lake City, 2003. 21 I. HODDER, Contemporary Theoretical Debate in Archaeology in HODDER Archaeological Theory Today [n. 7], p. 1-14. 102 JEROEN POBLOME Figure 1. Overlapping heuristic triangles The first conceptual triangle is focused on the thesis that regions develop in sustainable and economically sound ways (or not), as a result of the interaction over a long time-span between society and the physical environment. The interaction between nature and society constitutes regions socially. Regions are considered not as geographical units but as social constructs reproduced in the particular, localised cultural practices of individuals embedded in social and natural relationships, and these practices are repeated over various spatiotemporal scales.22 It is the long-term production of the region that provides clues for sustainable social and economic development,23 possibly resulting in moderate growth. In order to make this archaeologically operational, a second heuristic triangle distinguishes the various spatial scales at which to unravel the complexity of these 22 Often the study of material culture provides a useful proxy for approaching regions and regional identities, e.g. S. HALES / T. HODOS (eds.), Material Culture and Social Identities in the Ancient World, Cambridge, 2010. 23 For inspiration, see R. A. DODGSHON, Society in Time and Space. A Geographical Perspective on Change (Cambridge Studies in Historical Geography 27), Cambridge, 1998; F. BERKES / J. COLDING / C. FOLKE (eds.), Navigating Social-Ecological Systems. Building Resilience for Complexity and Change, Cambridge, 2003. A COMPLEX ADAPTIVE SYSTEM 103 processes: the scale of the household and the individuals composing it,24 the scale of the community and its vertical interrelations with ecological and regional contexts,25 and the scale of the horizontal networks that reach beyond communities in context, and connect them to the wider society, in this case the Roman Empire.26 Finally, after space, time also needs to be defined. When considering past regional developments, the basic idea developed by the Annales School of historiography that different historical processes work at different temporal scales, is often put forward as an alternative to linear interpretations of cultural change and social evolution.27 The temporal scales of the Annales School are: the long-term, describing very slow-moving processes such as environmental changes, the medium-term, referring to social and structural history such as forms of social or economic organization, and the short-term, pertaining to events or individuals, a traditional focus of political history. History is constituted by unique combinations of the short, medium and long-term, whose processes run concurrently but at different wavelengths. It is the task of the historian and archaeologist to present the evidence for processes at the different timescales, and then analyse retrospectively how these interacted to create unique and unpredictable outcomes. Time is not a fixed structure in which changes simply take place, but is as multi-layered as these changes, and is moulded by them as much as it moulds them.28 This approach is part of new research programmes that were only recently launched.29 This paper, therefore, will necessarily remain limited to a preliminary exploration and partial application of these conceptual triangles. 4 Higher Analytical Scales When considering the longue durée as a chronological unit and empires as spatial units and social constructs, one way or the other, higher analytical scales suggest a certain degree of determinism. The homepage of the Orbis website,30 featuring 24 E.g. K. BOWES, Houses and Society in the Later Roman Empire, London, 2010. For a non-Roman example, N. MAC SWEENEY, Community Identity and Archaeology. Dynamic Communities at Aphrodisias and Beycesultan, Ann Arbor, 2011. 26 D. J. MATTINGLY, Imperialism, Power, and Identity. Experiencing the Roman Empire, Princeton NJ, 2011. 27 M. PLUCIENNIK, Social Evolution, London, 2005. 28 G. LUCAS, The Archaeology of Time, London, 2005. 29 The CORES network (http://iap-cores.be/): Comparing Regionality and Sustainability in Pisidia, Boeotia, Picenum and NW Gaul between Iron and Middle Ages (1000 BC - AD 1000) and BOF/GOA 13/04 on Approaching Patterns of Nature-Society Interactions in Regional Development. An Interdisciplinary Dialogue between Past and Present in the Region of Sagalassos. 30 http://orbis.stanford.edu/ 25 104 JEROEN POBLOME the Stanford Geospatial Network Model of the Roman World states that: “Spanning one-ninth of the earth’s circumference across three continents, the Roman Empire ruled a quarter of humanity through complex networks of political power, military domination and economic exchange. These extensive connections were sustained by pre-modern transportation and communication technologies that relied on energy generated by human and animal bodies, winds, and currents. Conventional maps that represent this world as it appears from space signally fail to capture the severe environmental constraints that governed the flows of people, goods and information. Cost, rather than distance, is the principal determinant of connectivity” [my stress]. The very title of this book and of the research network ‘structural Determinants of Economic Performance in the Roman World’ suggests a similar view, for instance, with the ecological component being approached as a structural determinant, although generating both structural constraints and possibilities. The scope of this research network is limited, however, compared to popular science authors such as Jared Diamond in his study on how societies collapse.31 Diamond works with a five-point framework of possible contributing factors to collapse. His first factor “involves damage that people inadvertently inflict on their environment”, disturbing the balance between the fragility of landscapes and their resilience. Diamond’s second consideration is climate change, resulting from “changes in natural forces that drive climate and that have nothing to do with humans”, but that provoke huge and sometimes impossible management issues for societies. The third and fourth factors are hostile neighbours, who are always ready to exploit your own weaknesses, and decreased support by friendly neighbouring trading partners. The last set of factors involves the responses of societies to their problems. Diamond’s analytical framework tries to balance these factors, but in considering ancient society, he sees one overriding factor: demography. Basically, he takes a Malthusian approach. Population increase pushed up labour input per hectare and levels of land productivity, but also resulted in lower marginal returns per added unit of labour input, leading ultimately to ecological damage, demographic collapse and a reduction or loss of societal complexity. Considering our own global society, ecological problems play a much greater role in Diamond’s work. But these too add to a framework that appears to be fairly deterministic, teleological and mostly external to society and human agency. It is as if people are allowed some degree of choice at remedying conditions, but mostly fail to anticipate, perceive, attempt to or manage to solve problems. Of course Jared Diamond cannot change the course of history for the case studies he presents and conjure up a happy ending where there is none, but I would like to come back to the Malthusian appreciation of demographic patterns. 31 J. DIAMOND, Collapse. How Societies Choose to Fail or Survive, London, 2005. A COMPLEX ADAPTIVE SYSTEM 105 Indeed when considering active, not-yet-collapsing pre-industrial societies, the so-called low-equilibrium trap, as recently discussed in the context of the Roman economy,32 provokes a similar sentiment. The low-equilibrium trap is a way of expressing how underdeveloped economies cannot generate sufficient turnover, which is a necessary condition to induce investment, innovation and technological progress, from which systematic per capita growth can be realized.33 Economies caught in the low-equilibrium trap cannot grow, as “in the long term, limited increases in output will raise surpluses less than population size and the latter will eventually offset intermittent productivity gains.”34 Somehow, the macro-level of analysis is no source for optimism. 5 Dealing with Change To be sure, Jared Diamond, Walter Scheidel and the like are rightfully protagonists in this debate as they manage to develop highly original syntheses in interpretative frameworks that give meaning to the particularities and deficiencies of specific case-studies. Even though it is at no point their intention to convey deterministic messages and they basically aim for qualitative historical analysis of past communities as well as provide explanations for change, I wonder whether we cannot allow for more open-endedness in this story. If there is one structural determinant in the archaeological record, for instance, it is its constant change especially at the lower scales of the conceptual triangles, not only in physical terms, but also in how to understand it or to model it so as to better understand the past.35 This paper proposes to consider whether aspects of complex systems theory, representing a conglomerate of converging theories from ecology and the social sciences, might be applicable in archaeology, and more specifically in modelling trajectories of regions as dynamic socio-ecologic systems,36 based on the heuristic framework of the mentioned conceptual triangles. This is done in partial, provisional and preliminary terms, considering the initial stages of the associated research projects and the attested difficulties in reaching meaningful 32 W. SCHEIDEL, Approaching the Roman Economy in SCHEIDEL, Companion [n. 10], p. 12-16. 33 R. R. NELSON, A Theory of the Low-Equilibrium Trap in The American Economic Review 46, 1956, p. 894-908. 34 W. SCHEIDEL, Demography in SCHEIDEL et al., Cambridge Economic History [n. 10], p. 55-56. 35 G. LUCAS, Understanding the Archaeological Record, Cambridge, 2012. 36 T. A. KOHLER / S. E. VAN DER LEEUW, Historical Socionatural Systems and Models in KOHLER / VAN DER LEEUW [n. 11], p. 1-12. 106 JEROEN POBLOME interdisciplinary comparisons of environmental, climatic and socio-cultural patterns of change.37 Interestingly, forests and human societies can both be considered good examples of what are called complex self-organizing adaptive systems.38 In contrast to other (complex) systems, these are open and adapt to new challenges and problems. Complex adaptive systems are non-linear in trajectory, not predictable in behaviour, yet self-organised in the sense that they enhance coevolution between entities in the system, improving their performance and strengthening complexity. The properties and behaviours of complex systems can only be attributed to the system as a whole, not to any of its particular parts. They are therefore considered as emergent properties. Ecological and social systems are not only complex, they are also integrated as social-ecological systems, as approached in the first conceptual triangle.39 The question is whether modelling regions as dynamic and complex adaptive socio-ecologic systems can make a useful contribution to research on the Roman economy. The focus on sustainability and resilience in the study of socialecological systems seems most compatible with interdisciplinary approaches of past regional development. In this respect, assuming change and explaining stability, rather than the other way round, is a most interesting perspective from an archaeological point of view.40 For the social-ecological system of the Roman Empire to develop in sustainable ways, and for its economy to be successful at no matter which analytical scale, energy was needed.41 Access to energy made the Roman Empire and the development of its regions resilient, adaptive and capable to solve problems, such as (among others) those caused by the level of its societal complexity. These characteristics of social-ecological systems are based on the fact that energy can move and when this happens the resulting flow of energy can sometimes be used to change things in the physical world; energy can produce ‘work’. In simple terms, Roman farmers had a healthy breakfast in order to have sufficient energy to work their lands, the produce of which supplied sufficient 37 M. O. BALDIA / T. K. PERTTULA / D. S. FRINK (eds.), Comparative Archaeology and Palaeoclimatology. Socio-Cultural Responses to a Changing World (BAR International Series 2456), Oxford, 2013. 38 C. S. HOLLING, Understanding the Complexity of Economic, Ecological and Social Systems in Ecosystems 4, 2001, p. 390-405. 39 F. BERKES / J. COLDING / C. FOLKE, Introduction in BERKES et al. [n. 32], p. 5-9. 40 S. VAN DER LEEUW, Land Degredation as a Socionatural Process in R. J. MCINTOSH / J. A. TAINTER / S. K. MCINTOSH (eds.), The Way the Wind Blows. Climate, History and Human Action, New York, 2000, p. 190-210. 41 J. A. TAINTER / T. F. H. ALLEN / A. LITTLE / T. W. HOEKSTRA, Resource Transitions and Energy Gain: Contexts of Organization in Conservation Ecology 7(3), 2003, p. 4: http://www.consecol.org/vol7/iss3/art4; T. HOMER-DIXON, The Upside of Down. Catastrophe, Creativity and the Renewal of Civilization, London, 2006, p. 36-42. A COMPLEX ADAPTIVE SYSTEM 107 energy for landlords or urban councils to capitalize and invest in urban building projects. Mapping flows of energy as well as how much work a given system contains is the province of thermodynamics. In this way, thermodynamics can be considered as very appropriate to capture the global properties of complex systems, such as the Roman economy, without getting lost in details. This paper has no intention to archaeologically emulate thermodynamic laws on energy conservation and increasing entropy, or quantitative measures of the amount of thermal energy not available to do work. But archaeology can learn from this debate.42 Despite the fact that energy cannot be created or destroyed, it does inevitable degrade, making it progressively less useful for work. Whereas the thermodynamic laws pertain mostly to closed systems, complex systems such as the Roman economy are open. Such systems interact with other systems within the socio-ecological framework, by requiring high quality energy to maintain its societal complexity and upon work, releasing low quality energy such as waste. The discipline of ecology is currently developing the concept of exergy to apply the Second Law of Thermodynamics to open systems,43 such as the complex adaptive socio-ecological systems under discussion. By definition, the exergy of a system or resource is the maximum amount of useful work that can be obtained from this system or resource when it is brought to equilibrium with the surroundings through reversible processes. The exergy concept traditionally finds application in technical process analysis, typically employed to find inefficiencies, or in environmental, economic and sustainability analyses of industrial systems. Basically, the flow and loss of exergy through specific parts of a process are mapped, in order to identify opportunities for improving process efficiency by modifying parts of a process that incur maximum loss of energy. For instance, whilst the energy flow diagram of a power plant would identify the large quantity of rejected heat as an opportunity for improving the system, the exergy analysis puts much less stress on this point because of the low temperature of the rejected heat, and thus a low ability to do work. Improving the steam generation system, however, results in much less overall loss of the system, including rejected heat, and much higher exergy generation. Roman society was an open system receiving external exergy fluxes, mainly as solar radiation. This complex adaptive system used part of that external exergy to increase its internal exergy levels, such as those contained in the institutions 42 J. BINTLIFF, The Paradoxes of Late Antiquity: A Thermodynamic Solution in Antiquité Tardive 20, 2012, p. 69-73; J. A. TAINTER, Energy and Sociopolitical Collapse in C. J. CLEVELAND (ed.), Encyclopedia of Energy, San Diego, 2004, p. 529-543. 43 J. DEWULF / H. VAN LANGENHOVE / B. MUYS / S. BRUERS / B. BAKSHI / G. F. GRUBB / D. M. PAULUS / E. SCIUBBA, Exergy: Its Potential and Limitations in Environmental Science and Technology in Environmental Science & Technology 42(7), 2008, p. 2221-2232. 108 JEROEN POBLOME and structure of society, aimed at improving order in the system through selection and learning processes. In general, ecosystems and societies with higher exergy levels are more successful in dissipating external exergy flows; it means that they are better buffered and thus have higher stability, order and complexity.44 Along these lines, the concept of moderate economic growth in Roman antiquity could be translated into the question of whether some of its communities and regions, and possibly even the Empire itself, were successful in buffering exergy and matching its exergy needs with those of the supporting ecosystems, combining sustainability and resilience. Archaeologists and ancient historians have learned to recognize scenarios of growth and have developed methodologies to analyse and explain these patterns.45 But knowing the power of numbers in reconstructing the past, I wonder whether we are not sometimes confused by their meaning. Instead of measuring growth in regional systems, should we not be approaching the sustainability and resilience of regions and communities? Indeed, it is only when these manage to increase their exergy levels without substantially reducing that of their ecosystem or connected communities, that the pattern of development can be called sustainable. A very important notion is that such systems are not in equilibrium— maintaining their operational level requires constant exergy input. Our poor old friend Sisyphos is an ideal metaphor for this condition, forever trying but failing to push his immense boulder up the slopes of Tartaros. But societies are even worse off than Sisyphos. In order to constantly maintain internal exergy levels and head off problems in case they do not, quite often societies increase their levels of complexity, resulting in higher operational costs and, at some point in the balance, in diminishing returns potentially reaching marginal levels.46 In other words, providing for constant exergy needs implies adaptable strategies and changes in society. In complex adaptive systems, therefore, change is the norm. To an archaeologist, for whom documenting change in material culture or in stratigraphic layers is standard business, the accommodation of change in the concept of complex adaptive systems should sound more attractive than grand 44 B. MUYS, Sustainable Development Within Planetary Boundaries: A Functional Revision of the Definition Based on the Thermodynamics of Complex Social-Ecological Systems in Challenges in Sustainability 1, 2013, p. 41-52. 45 K. HOPKINS, Taxes and Trade in the Roman Empire in JRS 70, 1980, p. 101-125; R. B. HITCHNER, Olive Production and the Roman Economy. The Case for Intensive Growth in the Roman Empire in M.-C. AMOURETTI / J.-P. BRUN (eds.), La production du vin et de l’huile en Méditerranée (Bulletin de Correspondance Héllenique Suppl. 26) Athens, 1993, p. 499-505; R. SALLER, Framing the Debate over Growth in the Ancient Economy in J. G. MANNING / I. MORRIS (eds.), The Ancient Economy. Evidence and Models, Stanford CA, 2005, p. 223-238; R. SALLER, Human Capital and Economic Growth in SCHEIDEL, Companion [n. 10], p. 71-86. 46 J. A. TAINTER, The Collapse of Complex Societies, Cambridge, 1988, p. 91-126. A COMPLEX ADAPTIVE SYSTEM 109 meta-narratives of social evolution or low-equilibrium traps. Even more so as complex systems literature approaches change in social, economic and ecological systems in the same way, which presents an untapped potential for more integrated interpretation of past phenomena in interdisciplinary archaeology. Moreover, change is configured very close to the human experience. The source and role of change in complex adaptive systems has been conceptualized under the heading of panarchy theory.47 Coined in contrast to ‘hierarchy’ in its original meaning of a set of sacred rules, the term ‘panarchy’ stands for a framework of natural rules, with symbolic reference to the Greek god of nature, Pan. The central conceptual tool in panarchy is the adaptive cycle (Fig. 2). This combines the factors of the rising/declining potential of systems with their degree of connectedness and their rising/declining resilience, and sees complex adaptive systems as typically evolving through variable cycles of growth (r), stability (K), cἳtἳὅtὄὁphicΝ ὅhiftΝ (α)Ν ἳὀdΝ ὄeὁὄgἳὀiὐἳtiὁὀΝ (Ω).48 During the growth phase, the system’s potential and connectedness increase, while its resilience gradually declines. At the top of the curve the system collapses, resulting in diminished connectedness and potential. There is a gradual build-up towards catastrophic shifts as a result of internal or external stress factors, but the exact time and space of regime shift is extremely hard to predict. Constrained breakdown can result in reorganization of the system, leading to a new equilibrium which might be very different fὄὁmΝtheΝ pὄeviὁuὅΝὁὀeέΝἙὀΝthiὅΝ wἳyΝ “theΝ ἳdἳptiveΝcycleΝ emἴὄἳceὅΝtwὁΝ opposites: growth and stability on one hand, change and variety on the other.”49 Another attractive consideration, which fits in well with archaeology, is that adaptive cycles never exist in isolation but are nested in a hierarchy of slow large and small fast adaptive cycles. This nested hierarchy of adaptive cycles represents a panarchy, potentially spanning a large spatial range, from soil bacteria to the entire planet, and an equally vast temporal range, from seconds to geological epochs.50 This notion is highly compatible with the concept of multi-scalarity in archaeological analyses,51 laying out how different archaeological phenomena can be explained at one level but not necessarily pertain to others. Archaeological analysis needs to combine all scales. In survey archaeology, for instance, the Annales perspective has become one of the dominant frameworks to explain changes in the surface record (the ‘conjoncture’), as this follows from the 47 L. GUNDERSON / C. S. L. HOLLING (eds.), Panarchy: Understanding Transformations in Human and Natural Systems, Washington DC, 2002. See also the website of the Resilience Alliance: http://www.resalliance.org/. 48 HOLLING, Understanding [n. 38], p. 390-405. 49 HOLLING, Understanding [n. 38], p. 395. 50 C. S. L. HOLLING / G. PETERSON, Sustainability and Panarchies in GUNDERSON / HOLLING, Panarchy [n. 47], p. 63-102. 51 I. HODDER , Contemporary Theoretical Debate in Archaeology in HODDER [n. 7], p. 9-11. 110 JEROEN POBLOME interplay between the ‘histoire événementielle’ of historical sources, the more stable background of the landscape (‘longue durée’) and the ‘mentalitéὅ’ of individuals and societies.52 History is thus made up of unique combinations of the short-, medium- and long-term, whose processes run at different wavelengths, but concurrently. Going back to panarchy, an interesting observation regarding adaptive cycles working together is that total collapse, in a Malthusian sense, can only happen when the various cycles are at the very same high point in their fore loops or are aligned at the same phase of vulnerability, approaching catastrophic shift. If there is no convergence of cycles, systems change, evolve and adapt. Collapse can happen, but change is more likely. Furthermore, the fact that different adaptive cycles operate at different levels, scales and speeds prevents a situation in which panarchy functions as a single deterministic system with only one outcome scenario. As a result, adaptive cycles can work as heuristic tools to describe socioeconomic complexity, with archaeological phenomena as proxies for the potential, connectedness and resilience of a given society. Archaeological regions, for instance, could be seen as panarchies, with linked adaptive cycles represented by households, communities and empires, and regional development traced according to the Annales perspective in order to establish shifting balances in social-ecological systems and the sustainability of regions. Clearly, as a heuristic tool, mapping adaptive cycles allows for breathing life and human agency into complex systems. Not many archaeological studies in this domain have been published, however.53 The final reason why concepts of thermodynamics and complex adaptive systems hold potential is that they create a methodological framework for comparison between adaptive cycles. Through the use of proxy data, these become measurable and comparable. The issue of diachronic or inter-regional comparison remains fundamentally unresolved in archaeology. As the concept of exergy revolves around work, both natural and social proxy data can be considered, implying that the full agenda of interdisciplinary archaeology can be 52 J. BINTLIFF (ed.), The Annales School and Archaeology, Leicester, 1991. See for instance C. L. REDMAN / A. P. KINZIG, Resilience Theory, Society, and the Longue Durée in Conservation Ecology 7(1), 2003, p. 14: http://www.consecol.org/vol7/iss1/art14. For a comparable perspective applying aspects of thermodynamics, see BINTLIFF [n. 42], p. 69-73. 53 A COMPLEX ADAPTIVE SYSTEM 111 brought to bear. More than ever, the successful application of this model depends on sound source criticism and genuine interdisciplinary collaboration. Touching base again with the Roman economy, recent efforts at higher level synthesis were mostly focused at the higher scale of empire. Notwithstanding the solvability of this research, the integration of other analytical scales remains a Figure 2. Graphic representation of the adaptive cycle scholarly requirement. Making the concept of adaptive cycles operational at those lower levels of analysis, such as regional communities and households, could make a contribution to the debate, in order to make variability in development in spatial and chronological terms visible as well as allow for human agency. The Roman economy is not an independent sphere, but intimately connected with nature and its exergy flows and the path dependency, resilience and creativity of its social communities. The economic sphere formed part of a dynamic and dialectic web of relationships with demography, culture, technology, politics, religion and society, in accordance to Heraclituὅ’ adage panta rhei. 112 JEROEN POBLOME 6 Sagalassos between the Second and Fifth Centuries CE I will now explore these concepts, building on a range of already published ma terial related to the productive landscape of a case-study with which I grew familiar over the years: the archaeological site and region of Sagalassos (SW Turkey) (ancient Pisidia). Recently, a first attempt was made to approach loss of societal complexity in Byzantine Dark Age Sagalassos within the framework of complex adaptive systems.54 In this paper, the period between the second to fifth centuries CE will be looked into.55 As far as its natural environment is concerned, the archaeological site of Sagalassos is located in the western part of the Turkish Taurus Mountains. The ἳὀcieὀtΝtὁwὀΝiὅΝtuckedΝἳwἳyΝiὀΝἳΝlἳὄgeΝἴeὀdΝὁfΝtheΝχğlἳὅuὀΝdἳğıΝὄἳὀgeΝ(c. 1800 m)Ν(ἔigέΝἁ),ΝfὁὄmiὀgΝtheΝὅpectἳculἳὄΝcὄeὅtΝtὁΝtheΝὀὁὄthΝὁfΝtheΝὅiteΝwithΝtheΝχkdἳğΝ (2271 m) dominating the northeast end of the range. From its position at the top of a V-shaped valley incised in the mountains, the ancient town (1490-1600 m) overlooked the lower areas to the east and south. The valley acts as a permanent creek, draining several springs feeding the middle courseΝὁfΝtheΝχğlἳὅuὀΝÇἳyı,ΝἳΝ permanent small river in the valley south of Sagalassos and a tributary of the ancient river Kestros, part of which formed the eastern border of the territory of Roman imperial Sagalassos.56 The mountain landscape directed connectivity. Most of the research area, corresponding to the 1200 km² territory of Sagalassos in Roman Imperial times, is actually a series of interconnected mountain basins. The Burdur Plain represents the largest tract of flat, fertile lands within the Roman territory, at c. 30 km from the town. Natural corridors from the Bay of Fethiye and Antalya reach the latter plain, connecting into the Isparta Plain. The Augustan Via Sebaste followed the east corridor, representing, together with the Kestros Valley, the major lines of communication from the south coast to the interior57 (Fig. 4). Pronounced winter precipitation and summer dryness characterize the climate in the vicinity of Sagalassos as Mediterranean, with a shorter dry season and lower temperatures in all seasons compared to coastal zones and significantly cold 54 J. POBLOME, Shifting Societal Complexity [n. 4]. TheΝ teὄmὅΝ ‘lἳteΝ ἤὁmἳὀ’Ν ἳὀdΝ ‘eἳὄlyΝ ἐyὐἳὀtiὀe’Ν ἳὄeΝ uὅedΝ ἳὅΝ ὄelἳtiveΝ chὄὁὀὁlὁgicἳlΝ indicators of periods, mostly referring to the later 3rd to the first half of the 5th century CE for late Roman and the later 5th to 7th centuries CE for early Byzantine. 56 E. PAULISSEN / J. POESEN / G. GOVERS / J. DE PLOEY, The Physical Environment at Sagalassos (Western Taurus, Turkey). A Reconnaissance Survey in M. WAELKENS / J. POBLOME (eds.), Sagalassos II. Report on the Third Excavation Campaign of 1992, Leuven 1993, p. 229-231. 57 S. MITCHELL, Anatolia. Land, Men, and Gods in Asia Minor, 1. The Celts and the Impact of Roman Rule, Oxford, 1993, p. 70-79. 55 A COMPLEX ADAPTIVE SYSTEM 113 winters with a high number of frost-thaw cycles.58 The environmental and climatic conditions sustained the regional vegetation of the area, forming part of an Oro-Mediterranean vegetation belt with deciduous oak forests found below coniferous forests. Comparable to what was happening in many other communities in the Roman Empire, the Antonine era was a period of bliss for ancient Sagalassos. The programme of urban monumentalisation peaked with the spectacular Antonine Nymphaeum on the Upper Agora, as well as the completion of the enormous Imperial Baths and the scenic Theatre, worthy of Edward Gibbon’s famous praise of the period.59 Judging mainly from the epigraphic record, the region of Pisidia seems to have escaped the worst effects of the third century CE crisis and even ὅἳwΝ“theΝiὀteὄeὅtὅΝὁfΝtheΝἤὁmἳὀΝὅtἳteΝἳὀdΝὁfΝtheiὄΝὄegiὁὀἳlΝὅuἴjectὅΝiὀΝἢiὅidiἳΝ…Ν defiὀitivelyΝἳὀdΝcὁὀὅciὁuὅlyΝἳligὀed,ΝἴὁthΝἳtΝtheΝpuἴlicΝἳὀdΝpὄivἳteΝlevelέ” 60 The Potterὅ’ Quarter of Sagalassos, where Sagalassos red slip ware was produced on a large scale, was also one of the few proficient production centres in the Roman East to continue its production activities throughout this period of crisis. 61 Recently, pollen analysis has picked up contemporary signals of stress of a different nature, however. Also, the second half of the fourth century CE introduction of amphora production in the surroundings of Sagalassos has been considered as part of an agricultural attempt at reconversion,62 resulting in the local successes during the Theodosian dynasty. 63 58 PAULISSEN et al. [n. 55], p. 231-233; M. VERMOERE, Holocene Vegetation History in the Territory of Sagalassos (Southwest Turkey). A Palynological Approach (Studies in Eastern Mediterranean Archaeology 6), Turnhout, 2004, p. 8. 59 E. GIBBON, History of the Decline and Fall of the Roman Empire, London, 1993 [1776-1788], p. 90. 60 S. MITCHELL, Greek Epigraphy and Social Change. A Study of the Romanization of South-West Asia Minor in the Third Century AD in XI Congresso Internazionale di Epigrafia Greca e Latina, Rome, 1999, p. 421. 61 J. POBLOME, Mixed Feelings on Greece and Asia Minor in the Third Century AD in D. MALFITANA / J. POBLOME / J. LUND (eds.), Old Pottery in a New Century. Innovating Perspectives on Roman Pottery Studies, Catania, 2006, p. 189-212. 62 J. POBLOME / M. CORREMANS / P. BES / K. ROMANUS / P. DEGRYSE, It Is Never Too Late… the Late Roman Initiation of Amphora Production in the Territory of Sagalassos in I. DELEMEN / S. ÇOKAY-KEPÇE / A. ÖZDIZBAY / Ö TURAK (eds.), Euergetes. Festschrift für Prof. Dr. Haluk Abbasoğlu zum 65. Geburtstag, Antalya, 2008, p. 1001-1012. 63 M. WAELKENS / I. JACOBS, Sagalassos in the Theodosian Period in I. JACOBS (ed.), Production and Prosperity in the Theodosian Age, Leuven, 2014. 114 JEROEN POBLOME Figure 3. View from the Akdağ towards ancient Sagalassos. The town was laid out on the plateaus in the centre of the image The previous paragraph reads like a traditional historical synthesis of development, albeit with some regional accents. But perhaps this line of thought should allow for a degree of complexity? How do Antonine and Theodosian bliss compare for the community of ancient Sagalassos? I intend to explore the regional complex adaptive system in preliminary ways, considering the interdisciplinary indicators available in publications related to the productive landscape, bearing in mind the heuristic framework discussed above. The main question is how nature-society interactions sustained and/or limited subsistence strategies and community building from the second to the fifth centuries CE in the study region of the 1200 km² territory of Roman Sagalassos. The available indicators are positioned within adaptive cycles, evaluating their rising/declining potential with their degree of connectedness and their rising/declining resilience through time, space and scale in consideration of path dependent assets. Palynological, archaeobotanical, archaeozoological and artisanal data are highly relevant when considering the productive landscape and their analysis provides information on the potential A COMPLEX ADAPTIVE SYSTEM 115 of the regional nature-society interactions, on presence and changes in flora and fauna, and indirectly on environmental and climatological conditions, and human impact and exploitation potentially affecting connectivity and resilience. 7 Pollen, Precipitation and Plants In this context, some of the published pollen sequences are relevant, collected within the χǧlἳὅuὀΝἳὀdΝÇἳὀἳklıΝVἳlleyὅ,ΝtheΝἕὄἳvgἳὐΝἴἳὅiὀΝἳὀdΝtheΝἐeὄeketΝ marshes, as the crow flies at respectively 3 and 8 km south, and 15 and 37 km southwest of the town of Sagalassos and within its Roman Imperial territory. 64 The general position of Anatolia, forming a bridge between continents, the rough regional topography of the Taurus Mountains, as well as the long history of human presence in the landscape resulted in a complex history of changes in climate, vegetation and human impact. A traditional focus of palaeo-ecological and palaeo-climatological research is the so-called ἐeyşehiὄΝἡccupἳtiὁὀΝἢhἳὅe, 65 characterized by signs of intense human impact such as revealed by increases in arboriculture, identified at various locations in the Eastern Mediterranean between roughly 1000 BCE and 800 CE. ἔὁllὁwiὀgΝtheΝὅtἳὄtΝὁfΝtheΝἐeyşehiὄΝἡccupἳtiὁὀΝἢhἳὅeΝiὀΝtheΝὄegiὁὀ,ΝtheΝpὁlleὀΝ records show an additional increase in anthropogenic indicators during the fiὄὅtΝceὀtuὄyΝἐἑέΝἙὀΝtheΝχǧlἳὅuὀΝVἳlleyΝὁἳkΝwὁὁdlἳὀdΝiὀcὄeἳὅedΝἳlὁ ng with secondary anthropogenic indicators, being non-cultivated plants which fare better in association with human activities such as agriculture, while in 64 VERMOERE Holocene Vegetation [n. 58]; J. BAKKER / E. PAULISSEN / D. KANIEWSKI / V. DE LAET / G. VERSTRAETEN / M. WAELKENS, Man, Vegetation and Climate during the Holocene in the Territory of Sagalassos, Western Taurus Mountains, SW Turkey in Vegetation History and Archaeobotany 21, 2011, p. 249-266; J. BAKKER / E. PAULISSEN / D. KANIEWSKI / J. POBLOME / V. DE LAET / G. VERSTRAETEN / M. WAELKENS, Climate, People, Fire and Vegetation: New Insights into Vegetation Dynamics in the Eastern Mediterranean since the 1st Century AD in Climate of the Past 9, 2013, p. 57-87; D. KANIEWSKI / E. PAULISSEN / V. DE LAET / K. DOSSCHE / M. WAELKENS, A High-Resolution Late Holocene Landscape Ecological History Inferred from an Intramontane Basin in the Western Taurus Mountains, Turkey in Quaternary Science Review 26, 2007, p. 2201-2218; D. KANIEWSKI / E. PAULISSEN / V. DE LAET / M. WAELKENS, Late Holocene Fire Impact and Post-Fire Regeneration from the Bereket Basin, Taurus Mountains, Southwest Turkey in Quaternary Research 70, 2008, p. 228-239. 65 W. VAN ZEIST / H. WOLDRING / D. STAPERT, Late Quaternary Vegetation and Climate in Southwestern Turkey in Palaeohistoria 17, 1975, p. 55-143. 116 JEROEN POBLOME Bereket and Gravgaz signals of the increased presence of olive trees and cereals were picked up. 66 In general, the valleys and the lower hill slopes are considered to have yielded cereals. Bread wheat formed the main part of the diet throughout the study period. Palynological and archaeological evidence allows for postulating that olive cultivation also made a contribution to the regional portfolio, as indicated for the Gravgaz Valley and the slopes surrounding the ὀὁὄtheὄὀΝpἳὄtΝὁfΝtheΝÇἳὀἳklıΝVἳlleyέ 67 In addition to the pollen samples, stone weights from lever-and-weight presses are mostly found in the western and central parts of the town’s territory. 68 These simple presses could have been used for the processing of olives as well as grapes. 69 Olive trees, however, weὄeΝὀὁtΝὅὁΝcὁmmὁὀΝiὀΝtheΝχǧlἳὅuὀΝVἳlleyέΝἘeὄe,ΝwἳlὀutΝcultivἳtiὁὀΝplἳyedΝ a role, as did patches of oak woodland and, on the higher slopes, for instance north of Sagalassos, coniferous forests were maintained. M. Vermoere cὁὀὅideὄedΝtheΝcὁὀtiὀuedΝpὄeὅeὀceΝὁfΝἴὁthΝfὁὄeὅtΝtypeὅΝἳὅΝἳὀΝiὀdicἳtiὁὀΝὁfΝ“ἳΝ kiὀdΝ ὁfΝ ecὁlὁgicἳlΝ cὁὀὅciὁuὅὀeὅὅ” 70 on behalf of the community of Sagalassos. In the wider region, storax was possibly also tapped for its resin while a semi-parasitic insect on kermes oaks produced vermilion dye. 71 Pollen analysis indicated the cultivation of grapes near Gravgaz and also on the southern slopes near Sagalassos. 72 From the middle of the third century CE onwards, cereal and olive cultivation is proportionally less well represented in the Gravgaz basin, but neither cultivation disappeared before the seventh century CE. At the same time, more pollen linked with pine and moist deciduous trees is picked up, as well as of plants associated with meadows. During the later third century CE, the environment became moister. 73 At Bereket, olive cultivation seems to have come to an end in the first half of the fourth century CE at the latest, while open steppe vegetation became more typical for this valley. 74 In this way, the 66 VERMOERE, Holocene Vegetation History [n. 58], p. 108-111. Holocene Vegetation History [n. 58], p. 156-164. 68 VERMOERE, Holocene Vegetation History [n. 58], p. 239-276. 69 R. I. CURTIS, Food Processing and Preparation in J.P. OLESON (ed.), The Oxford Handbook of Engineering and Technology in the Classical World, Oxford, 2008, p. 381383. 70 VERMOERE, Holocene Vegetation History [n. 58], p. 114. 71 VERMOERE, Holocene Vegetation History [n. 58], p. 108-114. 72 VERMOERE, Holocene Vegetation History [n. 58], p. 160, 186. 73 BAKKER et al., Climate of the Past [n. 64], p. 70. 74 BAKKER et al., Climate of the Past [n. 64], p. 70; E. KAPTIJN / J. POBLOME / H. VANHAVERBEKE / J. BAKKER / M. WAELKENS, Societal Changes in the Bereket Valley during the Roman Imperial Period. Results from the Sagalassos Territorial Archaeological 67 VERMOERE, A COMPLEX ADAPTIVE SYSTEM 117 eἳὄlyΝeὀdΝὁfΝtheΝἐeyşehiὄΝἡccupἳtiὁὀΝἢhἳὅeΝiὀΝtheΝἐeὄeketΝἴἳὅiὀΝmὁὄeΝὁὄΝleὅὅΝ coincided with a gradual decrease of human impact on the environment in and around the Gravgaz basin and relatively moister conditions. Although the latter could have been beneficial for intensive agricultural activity, as attested elsewhere in the Roman East, 75 in the cases of Gravgaz and Bereket, and probably alὅὁΝtheΝχǧlἳὅuὀΝVἳlley,ΝexpἳὀdiὀgΝmἳὄὅheὅΝἳὀdΝwetlἳὀdὅΝmἳyΝhἳveΝ reduced the availability of arable land. It is unclear whether the same was hἳppeὀiὀgΝiὀΝὁtheὄΝἳὄeἳὅΝὁfΝἥἳgἳlἳὅὅὁὅ’ΝteὄὄitὁὄyέΝWetlἳὀdὅΝἳὄeΝἳctuἳllyΝfἳiὄlyΝ exceptional landscape phenomena in the study area, but outside of these landscape pockets pollen is insufficiently preserved. Geomorphological ἳὀἳlyὅiὅ,ΝὁὀΝtheΝὁtheὄΝhἳὀd,ΝiὀdicἳteὅΝthἳtΝtheΝἐὸğdὸὐΝὄiveὄΝvἳlleyΝἳccumulἳtedΝ large quantities of fine sediments, mainly from the first century CE onwards. The sediment eroded from the surrounding hill slopes, on which grassland and forests were converted into agricultural land. During the period concerned in thiὅΝpἳpeὄ,ΝtheΝlἳckΝὁfΝmἳjὁὄΝdiὅtuὄἴἳὀceὅΝiὀΝtheΝὄiveὄ’ὅΝὅedimeὀtἳtiὁὀΝcὁuldΝ be indicative of maintained agricultural practices until around 600 CE. 76 In general, the continuity of the pollen signal hints at rational exploitation schemes of the forests, cultivated tree species and cereal cultivation into early Byzantine times, as the relatively moist conditions persisted until the middle of the seventh century CE. 77 Survey 2008 (Southwest Turkey) in Anatolian Studies 63, 2013, 75-95. 75 Y. HIRSCHFELD , A Climatic Change in the Early Byzantine Period? Some Archaeological Evidence in Palestine Exploration Fund Quarterly Statement 136, 2004, p. 133-149; M. DECKER, Tilling the Hateful Earth. Agricultural Production and Trade in the Late Antique East, Oxford, 2009, p. 8-11. 76 K. ϊ’ἘAEN, Fingerprinting Late Holocene Sediment Fluxes in an Eastern Mediterranean Mountain Catchment, Unpublished Ph.D. Thesis, University of Leuven, 2012, p. 156; B. DUSAR, Late Holocene Sediment Dynamics in a Mediterranean Mountain Environment, Unpublished Ph.D. Thesis, University of Leuven, 2012, p. 170. 77 V ERMOERE [n. 57], p. 171-190; M. V ERMOERE / S. SIX / J. POBLOME / P. DEGRYSE / E. P AULISSEN / M. WAELKENS / E. SMETS, Pollen Sequences from the City of Sagalassos (Pisidia, SW Turkey) in Anatolian Studies 53, 2003, p. 161-173. 118 JEROEN POBLOME Figure 4. General geographical overview of the study region A COMPLEX ADAPTIVE SYSTEM 119 In terms of potential, the evolution of the regional vegetation patterns can be considered against a long-term background. In that sense, the discussed centuries cleἳὄlyΝ fὁὄmΝ pἳὄtΝ ὁfΝ theΝ ἐeyşehiὄΝ ἡccupἳtiὁὀΝ ἢhἳὅe,Ν withΝ ἳὄἴὁὄicultuὄeΝ ἳὀdΝ secondary anthropogenic indicators typically represented in the regional pollen profiles. Even if this phase comes to an early end in the Bereket basin, the resulting vegetation pattern in this locality would still be different from how further climatological changes from the middle of the seventh century CE onwards would affect the regional vegetation cover. Moreover, recent analysis of archaeological survey evidence revealed that occupation continued in the Bereket basin but depended on shifts in the local subsistence strategies with a more important role for pastoralism.78 At the general level, the main difference within the discussed period was an increase in moisture conditions, as indicated from the pollen analysis in the Bereket and Gravgaz basins. In specific areas this may have resulted in an increase in marshes and wetlands, but considering that other landscape elements are also represented in the study region, it is an open question whether and how the regional agricultural potential was affected from the later third century CE onwards, as was the case in other regions of the Roman East.79 All in all, a heavy human hand is noticeable in the vegetation patterns of managing cereal and olive cultivation. Also the maintenance of oak woodlands and the continued presence of coniferous woods, insofar as these are not over-represented in the pollen diagrams as a result of the long transport of this pollen in open landscapes, could reflect rational measures against soil erosion apart from aiming ἳtΝhἳὄveὅtiὀgΝtheΝfὁὄeὅtΝpὄὁduceέΝἙὀΝthiὅΝὄeὅpect,ΝtheΝἐeyşehiὄΝἡccupἳtiὁὀΝἢhἳὅeΝ represented optimal potential for the regional vegetation patterns and resilient human use and maintenance of these in the medium term. ἑὁὀὅideὄiὀgΝ cὁὀὀectivity,Ν ὅiὀceΝ theΝ ὅtἳὄtΝ ὁfΝ theΝ ἐeyşehiὄΝ ἡccupἳtiὁὀΝ ἢhἳὅeΝ different landscape pockets seem to have been characterized to some degree by different vegetation patterns, invoking local exchange. Contemporary agricultural management policies did not erase these differences and if the moister conditions were general in the study region, this basically implied differences shifting. Cereal and olive cultivation stopped in at least part of the Bereket basin, saw a reduction in the Gravgaz basin, while olive cultivation had never been common iὀΝtheΝχğlἳὅuὀΝVἳlley,ΝἴutΝpὁὅὅiἴlyΝcὁὀtiὀuedΝmἳiὀlyΝiὀΝtheΝweὅteὄὀΝἳὀdΝceὀtὄἳlΝ 78 KAPTIJN et al., Societal Changes [n. 73]. A. IZDEBSKI, Why Did Agriculture Flourish in the Late Antique East? The Role of Climate Fluctuations in the Development and Contraction of Agriculture in Asia Minor and the Middle East from the 4th till the 7th c. AD in Millennium. Jahrbuch zu Kultur und Geschichte des ersten Jahrtausends n. Chr. 8, 2012, p. 291-312; J. BINTLIFF, The Complete Archaeology of Greece. From Hunter-Gatherers to the 20th Century AD, Oxford, 2012, p. 351-368; S. MITCHELL, Olive Cultivation in the Economy of Roman Asia Minor in S. MITCHELL / C. LATSARI (eds.), Patterns in the Economy of Roman Asia Minor, Swansea, 2005, p. 83-113; S. MITCHELL, A History of the Later Roman Empire, Oxford, 2007, p. 329370; DECKER, Tilling the Hateful Earth [n. 75]. 79 120 JEROEN POBLOME parts of the territory, insofar as the press weights can be attributed to this period and olive or grape pressing. To be sure, vines and walnut had always been linked to specific localities. In general, the continued management of the vegetation patterns displayed path dependency where possible and rational policies at exploitation throughout. Archaeological excavation at Sagalassos itself confirms these observations. Upon excavation, occupational deposits are wet-sieved in order to retrieve macrobotanical remains. The botanical analysis revealed the general importance of naked wheat, especially during the Roman Imperial period. From late Roman and especially early Byzantine times onwards, barley, which is more resistant to poor soils, gained in importance, while millet was introduced and some naked wheat was still consumed. In general, through time a larger variety of pulses, vegetables, fruits and nuts were consumed at Sagalassos, with grapes and figs as the most common varieties, as attested in the latrine underneath the Roman Baths.80 Walnut as well was very abundant in Roman times. Unfortunately, most late Roman deposits were excavated quite early in the excavation history of Sagalassos, when wet-sieving was not yet in operation. In 2012, the remains of a modest house, to the east of the Neon-Library, which had burned down with its original content around 500 CE, were excavated. Wet-sieving resulted in a concentration of well-preserved botanical remains: whole fruits or fragments of almonds, grapes and apple/pear, along with a supply of hulled barley, free threshing wheat, grass pea and lentil. These results confirmed the increasing importance of barley in cereal consumption in these centuries, as already observed in the archaeobotanical studies based on over 800 litres of sediment from different locations at the site. In general, this type of analysis confirms the continuation of arboriculture and the importance of fruits and nuts in the human diet between Roman Imperial and early Byzantine times. In late Roman times, macro-remains related to fruit trees are slightly less attested in the excavated deposits, while pulses and barley seem to increase to some degree. Especially the latter could be an indication of resilience illustrating a shift towards crops with increased resistance to bad conditions, probably grown in more marginal areas. In this respect, the botanical analysis is in correspondence with the pollen analysis, indicating a well-organized anthropogenic landscape in which most pockets were rationally exploited.81 80 J. BAETEN / E. MARINOVA / V. DE LAET / D. DE VOS / M. WAELKENS, Faecal Biomarker and Archaeobotanical Analyses on Sediments from a Public Latrine Shed New Light on Ruralisation in Sagalassos, Turkey in Journal of Archaeological Science 39, 2012, p. 1143-1159. 81 I would like to thank Dr. Elena Marinova of the Leuven Centre for Archaeological Sciences for commenting on the botanical analysis, as well as kindly making recent, unpublished results available. A COMPLEX ADAPTIVE SYSTEM 121 8 Animal Bones, Isotopes and Diachronic Trends The analysis of faunal remains adds to this picture of potential, connectivity and resilience. Between early Roman Imperial and early Byzantine times, domestic animals or livestock are the most represented category of animal remains found during the excavations in the town of Sagalassos. The combination of the suitability of the different animals to the environment, the work or secondary products the animals could produce and the diet preferences of the local community resulted in goat as the most represented species, closely followed by sheep, cattle and pig. The excavated assemblages of animal bone display some diachronic trends, noting path dependency on ovicaprines, some increase of pig during the first centuries CE and a more marked increase of cattle in late Roman times. Sheep and goat representation would increase again in early Byzantine times.82 In terms of exergy, the meat yield is actually a better indicator than counts and proportions of animal bone. Cattle produced proportionally more meat than pig, and pig more meat than sheep and goat. In this way, beef clearly formed the majority of consumed meat between early Roman Imperial and early Byzantine times, varying between 70% and even 90% in the second half of the fourth century CE. Sheep and goat contributed at most 15% of meat on the market, whereas the degree of consumption of pork remained stable throughout the centuries.83 Recent carbon and nitrogen stable isotope ratio analyses of human bone collagen indicates that the protein portion of the human diet was in general mainly based on C3 plants, such as wheat and barley, and domesticated animals.84 Animal protein was not an occasional, but a regular part of the diet at ancient Sagalassos. The isotope ratio analysis programme also included livestock.85 As bone collagen mainly reflects the origin of the protein portion of the diet averaged over the animal’s lifetime, animal husbandry and feeding practices can be deduced and followed through time. During the study period, after the moderate increase in representation of pig bones in the first centuries CE, their proportion would remain more or less stable into Byzantine times. In contrast to the other livestock, pigs were mainly kept for the production of meat, as indicated by their young 82 The first half fourth century CE turning point mentioned by B. DE CUPERE, Animals at Ancient Sagalassos. Evidence of the Faunal Remains (Studies in Eastern Mediterranean Archaeology 4) Turnhout, 2001, p. 139, is linked to the fill inside the Neon-Library, which has since been re-dated to the second half of that century POBLOME et al., Old Pottery [n. 61]. 83 DE CUPERE, Animals at Ancient Sagalassos [n. 82], p. 145-146. 84 B. T. FULLER / B. DE CUPERE / E. MARINOVA / W. VAN NEER / M. WAELKENS / M. P. RICHARDS, Isotopic Reconstruction of Human Diet and Animal Husbandry Practices during the Classical-Hellenistic, Imperial, and Byzantine Periods at Sagalassos in American Journal of Physical Anthropology 149, 2012, p. 165. 85 FULLER et al., Isotopic Reconstruction [n. 84], p. 157-171. 122 JEROEN POBLOME slaughter age, and the degree of consumption of pork remained more or less constant throughout the centuries. In this respect, the herding of pigs is an indicator of continued potential for the local economy. Traditionally, these animals are considered to have been pastured in patches of oak, beech and chestnut woodland and also fed with beans and grain.86 Bea De Cupere linked pig herding with the attested patches of oak woodland, as present for instance in the χğlἳὅuὀΝVἳlley,ΝὄepὄeὅeὀtiὀgΝiὀteὄlocking interests in maintaining and managing these natural elements in the medium term.87 Interestingly, the isotope analysis indicates nearly no variation in the pigὅ’ diet through time, confirming aspects of resilience. Their diet was mostly based on C3 plants and the animal protein portion likely on human refuse.88 Microwear and hypoplasia analysis of pig teeth, measuring abrasion and development stress, indicates that these animals were most probably free range during most of their lives, and fattened up, possibly in pig sties, prior to slaughter with a soft non-abrasive diet, possibly based on urban waste.89 Additionally, heavy metal analysis of lead and copper, reflecting pollution resulting from human occupation and industrious activities, contained in soils, sampled throughout the territory of Roman Sagalassos,90 were matched with a set of pig bones. Anthropological anomalies in soil geochemistry were cὁὀceὀtὄἳtedΝiὀΝtheΝvἳlleyὅΝὁfΝχğlἳὅuὀΝἳὀdΝÇἳὀἳklı,ΝὅὁuthΝὁfΝἥἳgἳlἳὅὅὁὅέΝσὁtΝὁὀlyΝ were high concentrations of the same heavy metals noted in the analysed pig bones, but the analytical results remained high throughout the study period, implying that pigs were herded at Sagalassos and in its polluted environment throughout the study period.91 From the second half of the fifth century CE onwards, pigs seem to have declined in size possibly indicating less successful breeding and/or keeping strategies,92 even though the habitat of these pigs did not change. Until then, the pig raising strategies were in balance with environmental strategies, revealing aspects of sustainable practices. 86 K.D. WHITE, Roman Farming, London, 1970, p. 318. DE CUPERE, Animals at Ancient Sagalassos [n. 82], p. 137-144. 88 FULLER et al., Isotopic Reconstruction [n. 84], p. 167. 89 S. VANPOUCKE / I. MAINLAND / B. DE CUPERE / M. WAELKENS, Dental Microwear Study of Pigs from the Classical Site of Sagalassos (SW Turkey) as an Aid for the Reconstruction of Husbandry Practices in Ancient times in Environmental Archaeology 14(2), 2009, p. 137-154. 90 P. DEGRYSE / P. MUCHEZ / B. DE CUPERE / W. VAN NEER / M. WAELKENS, Statistical Treatment of Trace Element Data and Ancient Animal Bone: Evaluation of Roman Byzantine Environmental Pollution in Analytical Letters 37, 2004, p. 2819-2834. 91 H. VANHAVERBEKE / P. DEGRYSE / B. DE CUPERE / W. VAN NEER / M. WAELKENS / P. MUCHEZ, Urban-Rural Integration at Ancient Sagalassos (SW Turkey). Archaeological, Archaeozoological and Geochemical Evidence in Archaeofauna 20, 2011, p. 73-83. 92 VANPOUCKE Dental Microwear Study [n. 89]. 87 A COMPLEX ADAPTIVE SYSTEM 123 For the proportionally largest meat provider, cattle, some more diachronic trends are observable. Fairly large meadows, or the fodder thereof, and sufficient water supply were needed to raise cattle. The first diachronic trend the isotope programme indicates, is that the proportion of C4 plants became more important in the diet of cattle, while C3 plants continued to play a part and the protein portion of the diet remained relatively unchanged. Ben Fuller et al. consider, therefore, that cattle could have been raised on different farms with different food strategies and environments.93 In this respect, the identification of the C4 plants is of interest. As the archaeobotanical analysis of wet-sieved excavation deposits at Sagalassos indicates that millet was only introduced from the early Byzantine period onwards, other such food sources must have played a role in earlier centuries. C 4 plants related to wet conditions are considered a likely candidate by Ben Fuller and his team, linking cattle with a C4 isotopic signature with pastures near rivers and wetlands or their fodder being collected from such localities. Palynological analysis shows that in specific habitats marshes and wetlands were expanding from the late Roman period onwards. Although the prior use of these lands in agriculture came to an end, reuse of some of these areas for livestock may have been an option and could demonstrate resilient behaviour. C4 plants also fare relatively better under grazing pressure, so their increased importance in the cattle diet could be seen as an indicator for the general intensification of land use over time, especially in the late Roman period when the cattle bone proportion in the excavated assemblages at Sagalassos was at its height. In general, cattle at Sagalassos were slaughtered at an older age, implying their important prior role as working animals in the field and for transport, and milk providers. This aspect, together with the growing importance of cattle, especially in late Roman times, has been considered to reflect the potential of the regional economy.94 To be sure, grain and olive cultivation also are important indicators of stability in the productive landscape. The increased proportion of cattle as working animals in the fields also hints at the continued role of agriculture in the region. Another important diachronic trend derives from the heavy metal analysis of lead and copper pollution in soils and cattle bone.95 During the High Empire, the cattle consumed at Sagalassos were mainly came fὄὁmΝtheΝχğlἳὅuὀΝἳὀdΝÇἳὀἳklıΝ Valleys, following the elevated heavy metal content of the bones. In late Roman times, however, the average pollution levels in cattle bones drop, implying that the cattle which were consumed at Sagalassos came from different, less polluted areas. As this is also the period of maximum representation of cattle bone at Sagalassos and cattle reflect intensive agricultural practices and economic 93 FULLER et al., Isotopic Reconstruction [n. 84], p. 167. at Ancient Sagalassos [n. 82], p. 141. 95 DEGRYSE et al., Statistical Treatment [n. 90]; VANHAVERBEKE et al., Urban-Rural Integration [n. 91]. 94 DE CUPERE Animals 124 JEROEN POBLOME potential, these combined late Roman trends could be read as an indicator for generally successful agricultural practices and management now also in areas ἴeyὁὀdΝtheΝ χğlἳὅuὀΝἳὀdΝÇἳὀἳklıΝVἳlleyὅέΝἘἳὀὀelὁὄeΝVἳὀhἳveὄἴekeΝ et al. have considered the late Roman pattern as meaningful to indicate a better integration of parts of the territory into the urban economy.96 The evolution of the pollution signals in the cattle bone could be read in this way. But in pre-late Roman times Sagalassos was also dependent on other goods which were not sufficiently available in its polluted first degree catchment basin, but are less visible in the archaeological record, such as olive oil and wine,97 whereas other parts of the territory were clearly doing well also in pre-late Roman times, such as the Bereket basin. Considering the diachronic trend in the palynological evidence that different landscape pockets seem to have been characterized to some degree by different vegetation patterns, invoking local exchange, the shift in provenances of the consumed cattle at Sagalassos could perhaps be understood along similar lines. As a matter of fact, from the middle of the fifth century CE onwards most consumed cattle were again being raised within the most polluted zone. Also sheep and goats were mostly slaughtered at adult age, implying their initial usefulness in supplying dairy products, wool and hair. Their contribution as urban meat supplier was small in comparison to cattle. Sheep are traditionally considered to require different habitats from goats, preferring pastures or arable land to rough wooded areas.98 The diet of the goats found at Sagalassos was almost exclusively based on C3 plants. Before the early Byzantine period a small C4 component was identified in their diet as well, but less important compared to that of the sheep. Ben Fuller et al. consider these isotopic results an indication that goats were allowed to graze more freely, in different areas than the other livestock, possibly on the forested mountain slopes around Sagalassos.99 As with cattle, during most centuries goats grazed in the polluted zone surrounding Sagalassos, with dropping degrees of heavy metal pollution in their bones and associated different and more distant provenance in the late Roman period.100 Through time, most sheep were fed on C3 plants, while some animalὅ’ diet contained also C4 plants or fodder. The isotopic analysis is indicative of subtle diachronic changes in the way sheep were fed, and an interesting association between cattle and sheep on pastures in similar habitats was suggested, especially during High Imperial times. The isotopic signatures are more mixed in late Roman times with sheep possibly shifting throughout the year between cattle and 96 VANHAVERBEKE et al., Urban-Rural Integration [n. 91]. Considering the fact that in late Hellenistic and Roman Imperial times a range of imported wine amphorae are attested in excavation deposits at Sagalassos, perhaps mostly olive oil was another local product. 98 WHITE Roman Farming [n. 86], p. 304-306; 313. 99 FULLER et al., Isotopic Reconstruction [n. 84]. 100 VANHAVERBEKE et al., Urban-Rural Integration [n. 91]. 97 A COMPLEX ADAPTIVE SYSTEM 125 goat grazing areas. In this way, each type of domestic animal had a specific diet in these centuries, reflecting degrees of agricultural specialization and potential. In early Byzantine times, the isotopic analysis indicates that sheep and goats shared the same grazing areas. Although no heavy metal analysis was performed on sheep bones, mostly local provenance is suggested for the High Imperial centuries and the early Byzantine ones, based on the association between sheep and respectively cattle and goats. Considering the mixed late Roman signals shared between sheep on the one hand and cattle and goat on the other, and the fact that the latter two types of animals had low signals of heavy metal pollution in their bones in the same period, a more distant provenance for sheep can also be postulated in those days. This deduction seems to be confirmed by unpublished ἀί1ἀΝ ὅuὄveyΝ wὁὄkΝ ὁὀΝ theΝ flἳὀkὅΝ ὁfΝ theΝ χğlἳὅuὀΝ ἳὀdΝ χkdἳğΝ mὁuὀtἳiὀὅ,Ν wheὄeΝ shelters for pastoralism datable to the Roman Imperial period fell into disuse. In this way, the diachronic pattern shown in the ovicaprine bones sustained regional potential, connectivity through local exchange and resilience in shifting grazing locations. As mentioned, the proportion of sheep and goat bones would increase again in early Byzantine Sagalassos, but that process seems to be part of a different phase in nature-society interactions.101 Considering the long-term perspective, the isotope ratio analysis programme provided evidence for the fact that, during the Classical/Hellenistic period, the values of livestock mostly cluster together indicating that sheep, goats, cattle and pigs were herded together in the same general area or fed on similar foods, most probably providing subsistence to the self-sustaining communities at Düzen Tepe and contemporary Sagalassos. During Roman Imperial times this pattern clearly became more complex, with more specific habitats for each livestock species, albeit still mostly within the general surroundings of the town of Sagalassos. The highest degree of specialization in animal keeping is attributable to late Roman times, while signals of some loss of specialized practices and/or complexity can be associated with early Byzantine times.102 9 Other Lines of Production An earlier study103 showed that the start of local amphora production during the second half of the fourth century CE was associated with an attempt at intensification of part of the agricultural production. Both reflect a rational decision making process and policy of investment on the part of the landholders. 101 POBLOME, How Did Sagalassos Come to Be LRCW4 [n. 4]. et al., Isotopic Reconstruction [n. 84]. 103 POBLOME et al., It Is Never Too Late [n. 62]. 102 FULLER 126 JEROEN POBLOME The ophiolitic/flysch clays of these so-called Fabric 4 amphorae were traced to theΝceὀtὄἳlΝpἳὄtὅΝὁfΝtheΝχğlἳὅuὀΝVἳlley.104 Since the amphorae were presumably made on the farming estate(s) where they were to be filled, their contents can be cὁὀὅideὄedΝ tὁΝ ὄeflectΝ ἳgὄicultuὄἳlΝ pὄὁductiὁὀΝ chὁiceὅΝ iὀΝ theΝ χğlἳὅuὀΝ VἳlleyέΝ Initially, because the typology of the local vessels resembles other contemporary and popular amphora series in the late Roman East, an original function as wine containers was suggested. A programme of residue analysis on an early Byzantine collection of Fabric 4 amphorae, however, showed that apart from Dionysoὅ’ favourite drink, they had also been used for olive oil and walnut oil.105 Walnut cultivation is commonly represented in palynological and macrobotanical test results. But, based on pollen analysis, olive cultivation had so far not been ἳttὄiἴutedΝἳὀΝimpὁὄtἳὀtΝὄὁleΝiὀΝtheΝχğlἳὅuὀΝVἳlleyέΝἙὀΝἳὀyΝcἳὅe,ΝtheΝἳdditiὁὀΝὁfΝ amphora production in the course of the fourth century CE in a region where pottery manufacturing was endemic is a further sign of specialization in craft production. Possibly the diversity of their contents implies the same for agricultural production. The application of data distribution techniques on Sagalassos red slip ware, or the locally produced tableware, attested both in urban excavated deposits and survey collections, demonstrated that, in general terms, the High Imperial centuries represented the largest artisanal output, with a declining role for the local tableware during the third and fourth centuries CE and a revival of the industry in the next two centuries.106 Even though less well represented, the continued production of Sagalassos red slip ware throughout the third century CE in a tableware production landscape where most wares with wide distribution patterns in the Roman East witness a serious reduction or even interruption of production, has been considered important. It seems to indicate that Sagalassos may not have suffered from the contemporary crisis in the empire.107 During the fourth century CE, a new line of Sagalassos red slip ware was launched, possibly together with the Fabric 4 amphorae and a series of mould-made products, such as oil lamps and figurines as well as so-called oinophoroi or wine flasks. The launching of a new design of tableware has recently been considered to depend on general and mixed aspects of regional well-being.108 What is remarkable in the 104 NEYT et al., Long-Term Clay Raw Material Selection [n. 5]. K. ROMANUS / J. BAETEN / J. POBLOME / S. ACCARDO / P. DEGRYSE / P. JACOBS / D. DE VOS / M. WAELKENS, Wine and Olive Oil Permeation in Pitched and Non-Pitched Ceramics: Relation with Results from Archaeological Amphorae from Sagalassos, Turkey in Journal of Archaeological Science 36(3), 2009, p. 900-909. 106 J. POBLOME / R. WILLET / N. FIRAT / F. MARTENS / P. BES, Tinkering with Urban Survey Data. How Many Sagalassos-es Do We Have? in P. JOHNSON / M. MILLETT (eds.), Archaeological Survey and the City, Oxford, 2013, p. 146-174. 107 POBLOME, It Is Never Too Late [n. 62]. 108 R. WILLET, Red Slipped Complexity. The Socio-Cultural Context of the Concept and 105 A COMPLEX ADAPTIVE SYSTEM 127 case of Sagalassos, is that not only a new typological set of its tableware was introduced on the market, but that the local pottery industry also specialized in further product diversification with the production of mould-made wares in the so-called Coroplast workshop.109 Oinophoroi are an important product within this collection. In contrast to the contemporary tableware, these products saw some supra-regional distribution.110 Considering that their iconography was mostly based on the wine-god Dionysos and that they are a late representative of a tradition of similar wine flasks in Asia Minor,111 the start of the production of oinophoroi in the Potterὅ’ Quarter of Sagalassos could be seen as another indicator for specialization in artisanal production, but possibly also in agricultural produce. Other craft activities are attested or presumed at Sagalassos, including wool dying, metal production, bone and antler cutting as well as glass blowing. Most evidence for bone cutting has been collected in the second half of the fourth century CE fill inside the destroyed Neon-Library.112 However, as (fragments of) bone objects are a traditional find in most contexts at Sagalassos, it is unclear whether this specific deposit of bone working refuse has specific chronological importance. The latter does seem to be the case for glass working, however. Whereas in earlier centuries Sagalassos had mostly depended on imported glass vessels, chemical and archaeological analysis confirmed the existence of a local glass workshop working with imported chunks from the second half of the fifth century CE onwards.113 In this sense, the trend in artisanal production does not necessarily imply increases in output in late Roman times, but the production landscape seems to have become more specialized compared to earlier or later centuries. Part of the increased specialization in craft production seems to have been associated with similar processes in the field of agricultural production. Use of Tableware in the Roman East, Unpublished Ph.D. thesis, University of Leuven, 2012. 109 E. MURPHY / J. POBLOME, Technical and Social Considerations of Tools from RomanPeriod Ceramic Workshops at Sagalassos (Southwest Turkey): Not Just Tools of the Trade? in JMA 25, 2012, p. 69-89. 110 P. TALLOEN / J. POBLOME, What Were They Thinking Of? Relief Decorated Pottery from Sagalassos: A Cognitive Approach in Mélanges de l’École Française de Rome in Antiquité 117, 2005, p. 55-81. 111 U. MANDEL, Kleinasiatische Reliefkeramik der mittleren Kaiserzeit. Die ‘Oinophoren’– Gruppe und Verwandtes (Pergamenische Forschungen 5), Berlin, 1988. 112 DE CUPERE, Animals at Ancient Sagalassos [n. 82], p. 147-159. 113 V. LAUWERS, The Glass of Sagalassos. Towards a Geochemical and Typochronological Interpretation, Unpublished Ph.D. thesis, University of Leuven, 2008, p. 212. 128 JEROEN POBLOME 10 Counting Sites and People In order to give meaning to the attested processes of specialization, these are best projected against the background of the regional settlement pattern and demography, especially considering the mentioned effects of the Malthusian lowequilibrium trap. Within its territory, ancient Sagalassos remained the most complex and the only urban settlement. During the Severan dynasty the construction of new public monuments came to a halt. Major building activities are not attested afterwards until the end of the fourth century CE, when the Imperial Baths were renovated. Shortly afterwards, a new town wall and a number of churches were built, heralding a new phase in urban development.114 Surrounded by its necropoleis, the city area measured 37.5 hectares. After an Augustan phase of expansion, this estimated area seems to have remained stable during the study period. Also the extent of the local Potterὅ’ Quarter of 3.5-4 hectares seems to have remained the same, following its establishment in Augustan times and subsequent growth until Flavian times. Both in the case of the town and in that of the Potterὅ’ Quarter developments cannot be accurately measured, but are extrapolated from the excavated and surveyed parts. Based on the estimated area of 25.22 hectares for the residential quarters and a population density of 100 to 150 people per hectare, the size of the urban population can be estimated as between 2,500 and 3,750 people.115 Within the study period, unfortunately, the available data do not allow the reconstruction of a pattern of demographic change for the town of Sagalassos. Therefore, the fact that it remained the only urban settlement in the study region is more important than the exact total estimate or its evolution. It is beyond the scope of this paper to discuss the evolving urban landscape in detail. Instead, we will focus on water provisioning, as its functioning is illustrative of urban potential and its impact on urban-rural connectivity. Considering its long-term potential, at least from the third century BCE onwards, ancient Sagalassos became the most prominent settlement in the wider region.116 By early Roman Imperial times, Sagalassos had established a territory of approximately 1,200 km², drawing in resources and potential from the wider region. This resulted in spectacular urban building programmes, in which, amongst others, monuments related to water played a prominent role. The late 114 M. WAELKENS, Sagalassos. Erste Stadt Pisidiens, Freund und Bundgenosse der Römer in Antike Welt: Zeitschrift für Archäologie und Kulturgeschichte 2011, p. 62-71; M. WAELKENS, Sagalassos in R. S. BAGNALL et al. (eds.), The Encyclopedia of Ancient History, Oxford, 2013, p. 6007-6009. 115 WILLET, Red Slipped Complexity [n. 108], p. 183. 116 P OBLOME, How Did Sagalassos Come to Be? LRCW4 [n. 4]; WAELKENS, Sagalassos [n. 114]. A COMPLEX ADAPTIVE SYSTEM 129 Hellenistic Fountain House, the Flavian, Hadrianic and Severan Nymphaea on the Lower Agora, as well as the Antonine Nymphaeum on the Upper Agora are prime examples. Also the Augusto-Tiberian Bath Building, superseded by the enormous Imperial Bath Complex, testifies to the important role of water in the town of Sagalassos. In antiquity, water-related monuments must have formed some of the more characteristic features of the urban landscape of Sagalassos. They no doubt held great symbolic importance for the local community and became part of their cultural identity.117 Water was no doubt also an essential commodity in ordinary daily life activities, and served a crucial function in sustaining local craft industries, of which the pottery industry perhaps played the most important role.118 Apart from the engineering skills required in antiquity to locate, construct and maintain these water-related monuments, as well as sustain daily life and craft industries, impressive technological knowledge was required to find, capture, tap and guide sufficient amounts of fresh water towards the city of Sagalassos. In addition to the local fresh water springs, such as the one feeding the late Hellenistic Fountain House, remains of various aqueducts were found in the mountains to the west and east of the town.119 These aqueducts brought water into town, feeding the life and culture that were typical for the splendour of ancient Sagalassos. Fieldwork in 2012, mainly focused on preserved remains on the flἳὀkὅΝ ὁfΝ theΝ χğlἳὅuὀΝ ἳὀdΝ χkdἳğΝ mὁuὀtἳiὀὅΝ tὁΝ theΝ eἳὅtΝ ὁfΝ theΝ tὁwὀ, added in important respects to the earlier studies of the aqueducts of Sagalassos. It resulted in the discovery of the source feeding the eastern aqueducts, the cistern capturing and distributing this resource, as well as many newly discovered preserved sections of the channel, allowing a better understanding of the route of the main eastern aqueduct and its points of bifurcation, feeding different parts of the town (Fig. 5). The source was located on the north-eastern flanks of the Akdağ, near the locality aptly called BaşpıὀἳὄΝἳtΝ1ι1ίΝmέΝἑleἳὄly,ΝdetἳiledΝiὀveὅtigἳtiὁὀΝὅhὁuldΝ 117 J. RICHARD, In the Elites' Toolkit. Decoding the Initiative and Reference System behind the Investment in the Architecture and Decoration of Roman Nymphaea in Facta: A Journal of Roman Material Culture Studies 5, 2011, p. 65-100; J. RICHARD, Water for the City, Fountains for the People. Monumental Fountains in the Roman East: an Archaeological Study of Water Management (Studies in Eastern Mediterranean Archaeology 9), Turnhout, 2012; F. MARTENS / J. RICHARD / M. WAELKENS, The Roman Baths at Sagalassos (SW-Turkey): A Preliminary Study of the Research Potential for a Reconstruction of Its Water Management System in P. FRICKE / G. SCHMIDT (eds.), Proceedings of the ‘Internationales Frontinus-Symposium zur Technik- und Kulturgeschichte der antiken Thermen’, Leuven, 2009. 118 POBLOME, The Potters of Sagalassos [n. 3]. 119 E. OWENS, The Aqueducts of Sagalassos in M. WAELKENS / J. POBLOME (eds.), Sagalassos III. Report on the Fourth Excavation Campaign of 1993, Leuven, 1995, p. 91113. 130 JEROEN POBLOME be continued to map this water provisioning system in more detail, but from the 2012 results the entire length of the eastern aqueducts can be measured at c. 24.5 km,ΝcὁmpletelyΝeὀciὄcliὀgΝtheΝhigheὄΝflἳὀkὅΝὁfΝtheΝ χkdἳğ,Ν withΝἳΝ diffeὄeὀceΝiὀΝ height between source and the eastern edge of Sagalassos of around 100 m. Parts of the channel were worked into limestone outcrops on the flanks of the χkdἳğΝ ἳὀdΝχğlἳὅuὀΝmὁuὀtἳiὀὅ,ΝὅὁmetimeὅΝtὁΝἳΝdepthΝὁfΝὅeveὄἳlΝmeteὄὅ,ΝwithΝἳὀΝἳveὄἳgeΝ width of 0.4 m and watermarks at the average height of 0.5 m. Earlier calculations of the water supply arrived at impressive estimates of 530 litres per second or a total of nearly 45,000 m³ of water per day.120 The 2012 discovery should allow reconsideration of these calculations, taking the actual source into account. Apart from the eastern aqueduct being a testimony to the technical ingenuity of the time, its concept also symbolizes a landscape of power, potential and connectivity, whereby the town of Sagalassos managed to capture and guide one of the more important water sources in the entire region for its civic benefits. Future study is required to accurately date the development of these engineering works, their maintenance and also their moment of abandonment. Considering the waterrelated monuments that were provided by this system, however, it must have been operational during the entire study period of this paper, providing an important new insight for understanding the wider outreach of the town. More work is also required to understand how, upon taking the eastern aqueduct into use, the lower slopes along its ‘tracé’ ἳὀdΝ eὅpeciἳllyΝ theΝ χğlἳὅuὀΝ VἳlleyΝ weὄeΝ ὅtillΝ pὄὁvidedΝ withΝ ὅufficieὀtΝ irrigation water in order to sustain their intensive agricultural use, but clearly the functioning of the aqueduct was an important structural determinant in this respect. 10 8 6 4 2 -25 10 45 80 115 150 185 220 255 290 325 360 395 430 465 500 535 570 605 640 675 0 Figure 5. Gaussian distribution graph of datable pottery from the Ağlasun Valley intensive survey campaigns, dated between the end of the first century BCE and the end of the seventh century CE (n=1595 sherds) 120 A. STEEGEN / K. CAUWENBERGHS / G. GOVERS / M. WAELKENS / E. J. OWENS / P. DESMET, The Water Supply to Sagalassos in M. WAELKENS / L. LOOTS (eds.), Sagalassos V. Report on the Survey and Excavation Campaigns of 1996 and 1997, Leuven, 2000, p. 646. 131 Figure 6. The eastern aqueduct, from the Akdağ mountain into Sagalassos A COMPLEX ADAPTIVE SYSTEM ἑὁὀὅideὄiὀgΝ ὅettlemeὀtΝ pἳtteὄὀὅΝ iὀΝ theΝ χğlἳὅuὀΝ Vἳlley,Ν ὄeceὀtΝ ceὄἳmὁlὁgicἳlΝ evaluation of the results of intensive survey campaigns between 1999 and 2004, which mostly covered the central parts of the valley, demonstrated that discrete concentrations of material in the terrain which could be interpreted as farms were fairly easily recognizable for the Classical/Hellenistic and Byzantine periods.121 Strikingly, late Hellenistic and early Roman Imperial material was mostly missing 132 JEROEN POBLOME iὀΝtheΝχğlἳὅuὀΝVἳlley. I would like to propose to see this as the result of a process of demographic nucleation within burgeoning Sagalassos. When considering the totals of pottery fabrics and types recovered from all sites in the central parts of the latter valley datable to the Roman centuries (Fig. 6), a first peak in data representation is situated around the middle of the second century CE and the second and main peak in the first half of the sixth century CE. Although situated in the intermediate period with low data representation, a small peak is recognizable in the second half of the fourth century CE. Compared to the collected urban excavation and survey pottery data (Fig. 7), the two peaks of highest data representation in respectively the second and sixth centuries CE are recognisable in both graphs, albeit in reverse order with a more important signal ὁfΝὁccupἳtiὁὀΝiὀΝtheΝχğlἳὅuὀΝVἳlley in the first half of the sixth century CE. Figure 7. Gaussian distribution graph of selected Sagalassos urban excavation pottery data (dotted grey line; n=53,910 sherds) compared to urban survey data (dark grey line; n=23,464 sherds) The slope towards the smaller second century CE peak is also slower in the χğlἳὅuὀΝVἳlleyΝcὁmpἳὄedΝtὁΝἥἳgἳlἳὅὅὁὅ,ΝimplyiὀgΝἳὀὁtheὄΝdegὄeeΝὁfΝdevelὁpmeὀtΝ of the valley compared to the town, possibly as a result of the process of urban nucleation. Another contrast is that the growth towards the sixth century peak in theΝχğlἳὅuὀΝVἳlleyΝiὅΝὅtὄὁὀgeὄΝcὁmpἳὄedΝtὁΝtheΝtὁwὀ,ΝwithΝἳΝfifthΝceὀtuὄyΝupὅwiὀgΝ of occupation in the valley. As the residential areas in the town of Sagalassos were considered stable throughout the study period of this paper, with associated 121 POBLOME et al., How Did Sagalassos Come to Be? LRCW4 [n. 4]. A COMPLEX ADAPTIVE SYSTEM 133 demographic stability, it does seem to be the case that there were more farms in theΝχğlἳὅuὀΝVἳlleyΝiὀΝtheΝfifthΝἳὀdΝfiὄὅtΝhἳlfΝὁfΝtheΝὅixthΝceὀtuὄieὅΝἑEέΝχlthὁughΝ the link between sherds and people is never straightforward, the fact that the archaeological material is associated with a slight increase in number of sites and densities of on-site material could be an indirect indication that there were slightly mὁὄeΝpeὁpleΝiὀΝtheΝχğlἳὅuὀΝVἳlley in these particular centuries. At least in the surveyed, central part of the valley, this period of increased number of farms is more or less contemporary with the period of cattle, goats and by extrapolation also sheep being kept outside of the immediate polluted zone around Sagalassos. χὅΝ ἳΝ cὁὀὅequeὀce,Ν theΝ lἳteΝ ἤὁmἳὀΝ χğlἳὅuὀΝ VἳlleyΝ fἳὄmὅΝ weὄeΝ mὁὅtΝ pὄὁἴἳἴlyΝ focusing on other crops, possibly used to fill the local Fabric 4 amphorae, the clays for which were traced to the same landscape pockets. In additiὁὀΝtὁΝtheΝὄeὅultὅΝiὀΝtheΝχğlἳὅuὀΝVἳlley,ΝexteὀὅiveΝὅuὄveyiὀgΝcἳmpἳigὀὅΝ in the wider territory of Sagalassos indicated that those lands witnessed the highest overall number of sites ever during the late Roman period.122 More than double the number of sites could be attributed to these centuries compared to early Imperial times, while site numbers drop considerably in early Byzantine times. Although the available totals are supportive of a contemporary demographic increase, caution is warranted as the total number of sites compared to the extent of the territory is fairly low while also the methodology of extensive surveying cannot sustain fine-grained reconstructions. Settlements were located in the plains, on the lower slopes and for the first time since Hellenistic times again on less accessible, albeit unfortified mountain sites,123 indicative of intensive exploitation of the landscape. Villages and farms were located at regular intervals in the landscape, functioning as localities for the collection and processing of agricultural produce. In contrast to early Imperial times, little evidence for elite involvement or presence in the countryside is available. Caution is again required, however, since in most cases this can be acceptably demonstrated only by the presence of monumental tombs and sarcophagi. Such markers are no longer common in late Roman times, reducing elite visibility considerably. Another diachronic trend, which is perhaps easier to evaluate based on the available evidence, is that between early Imperial and early Byzantine times the number of villages steadily grew and the number of farms was simultaneously reduced.124 122 VANHAVERBEKE et al., Urban-Rural Integration [n. 91], p. 75. H. VANHAVERBEKE / M. WAELKENS, The Chora of Sagalassos: The Evolution of the Settlement Pattern from Prehistoric Until Recent Times (Studies in Eastern Mediterranean Archaeology 5), Turnhout, 2003, p. 241-283. 124 H. VANHAVERBEKE / F. MARTENS / M. WAELKENS / J. POBLOME, Late Antiquity in the Territory of Sagalassos in W. BOWDEN / L. LAVAN / C. MACHADO (eds.), Recent Research on the Late Antique Countryside (Late Antique Archaeology 2), Leiden, 2004, p. 247-279. 123 134 JEROEN POBLOME Hannelore Vanhaverbeke et al.125 identified farms as sites extending over 2,500 m² on average, while villages were at least four times this size and typically formed a nucleated type of settlement. So although caution is still called for, it seems that we can assume some degree of population increase in parts of the territory of Sagalassos. In an attempt to compare the demographical potential of the town and territory of Sagalassos with the output of the Sagalassos red slip ware, Rinse Willet recently engaged in the estimation of the regional rural population.126 Methods to establish rural populations are mostly based on ratios of urban population totals or estimated amounts of people per km². The former traditional methods arrive at very low totals for the territory of Sagalassos, as their premise is in most cases the model of more populated Classical poleis. For the latter, recently a ratio of 15 people/km² was proposed for Anatolia,127 which results in a total population for the territory of Sagalassos of 18,000 people. Considering that the applied ratio is on the low side even for modern countries, this total is possibly an underestimation. Intensive surveying campaigns in parts of the territory as well as the study of agricultural carrying capacities of the territory should shed more light on this issue in future. At the moment, diachronic estimates are not possible given the nature of the evidence, but as the rural settlement pattern was at its densest during late Roman times, we should expect this period to represent the highest diachronic demographic total. Apart from its palynological record, the high altitude valley of Bereket revealed another important trend related to connectivity within the territory of Sagalassos. This valley is one of the more remote points of the study region, at about 37 km southwest from Sagalassos as the crow flies. In roughly the same period as the mentioned interruption of the intensive cultivation of olives and grain, possibly in favour of increased pastoralism, a change in supply of tableware was noted at the archaeological sites in the Bereket Valley. Before late Roman times, Sagalassos red slip ware was the most common type of tableware represented, which seems acceptable considering the output of the local production centre at Sagalassos and its distance to the Bereket Valley. In late Roman times, the supply of Sagalassos red slip ware is only a trickle of its former self, however, while the valley is provided with red slipped wares from yet unknown production centre(s). Also amphora sherds were discovered at the surface of the sites in the Bereket Valley which are typologically similar to the mentioned Fabric 4 amphorae, but their fabrics were different, suggesting a 125 VANHAVERBEKE et al., Late Antiquity [n. 124], p. 258-259. Red Slipped Complexity [n. 108], p. 191-195. 127 B. FRIER, Demography in A. BOWMAN / P. GARNSEY / D. RATHBONE (eds.), The Cambridge Ancient History, XI, Cambridge, 2000, p. 812. 126 WILLET, A COMPLEX ADAPTIVE SYSTEM 135 different origin for these vessels compared to the ones attested at Sagalassos.128 While the ceramological evidence seems to indicate a somewhat larger distance between the regional pole of attraction and the Bereket Valley, this is also the period when the cattle and goat bones, and the sheep by association, were supposed to be brought to Sagalassos for consumption from larger distances,129 when pastoralism is presumed to have become more important at Bereket,130 and when natural conditions and increased pressure sustained the growth of C4 plants in wetlands in the general study region.131 Although there is no precise allocation in time and place nor proof of causality for these suggested changes in connectivity, the archaeological results from the Bereket Valley do indicate that patterns of exchange and relations between Sagalassos and the different parts of its territory need not have stayed constant over time, while changes in subsistence strategies and/or ceramic suppliers most probably reflect rational choices on behalf of the communities involved, trying at all times to make the most from changing situations. 11 General Trends People, or rather manpower, are everything to a pre-industrial society and economy. Although the available evidence only allows for a crude approximation of numbers, a difference in trends between the second and fifth centuries CE can be cautiously forwarded. Mainly the nature and the distribution of rural sites seem to suggest some demographic increase by late Roman times. The evolution of population totals cannot be reconstructed and the link between the facts of the archaeological record and demographical reconstruction remains a difficult one, but even if future research would prove growth to be minimal, regional population totals seem not to have declined. If growth was realized, this seems to have happened mostly in the rural context. This trend is not dissimilar to the evidence presented by Michael McCormick,132 with many regions in the Roman East having a stable or slightly increasing population, especially in rural contexts. This ὄuὄἳlΝ cὁὀὀectiὁὀΝ iὅΝ ὁfΝ impὁὄtἳὀceΝ ἳὅἈΝ “theΝ cὁuὀtὄyὅideΝ wἳὅΝ theΝ demὁgὄἳphicΝ wellspring of society, since deaths always outstripped births in pre-modern cities …Ν ἙὀΝ ἳΝ pὄe-modern economy, the extent of land farmed was the first and 128 KAPTIJN et al., Societal Changes [n. 74]. et al., Urban-Rural Integration [n. 91]. 130 BAKKER et al., Climate of the Past [n. 64]. 131 FULLER et al., Isotopic Reconstruction [n. 84]. 132 M. MCCORMICK, Origins of the European Economy. Communications and Commerce, AD 300-900, Cambridge, 2001, p. 30-38. 129 VANHAVERBEKE 136 JEROEN POBLOME primordial economic fact determining food production and therefore wealth at its most basic level.”133 Considering the potential past demographical patterns, pre-industrial societies are typically pulled between two general trends. On the one hand, there is the mentioned low-equilibrium trap,134 lowering expectations about the possibility of sustained economic growth, with limited resources as a binding constraint for development, income and population growth, especially in non-industrial societies. In this scenario, communities could attempt some reorganization, displaying resilience and creativity, but such attempts mostly resulted in a gradual shift from high-exergy return on investment to low-exergy returns, reflecting or inducing lower social complexity. The decreased role of grain and olive cultivation in the Bereket and Gravgaz basins, whether provoked by changes in the levels of precipitation or resulting from rational decisions to change subsistence strategies, could be read as a move away from high-exergy, export oriented crops such as grain and olive oil, and a contemporaneous shift towards low-exergy livestock cultivation, representing less inherent risk of failure but at the same time less return on investment. In this way, there may have been more people in the Sagalassos countryside, but their average income was dropping resulting from changing subsistence strategies, eventually resulting in the attested regional demographical reduction in early Byzantine times. A sign of the times could be the lower output rates and reduced pattern of distribution of Sagalassos red slip ware reflecting the reduced economic drive towards generating highexergy income from maintaining distant exchange and export networks. Another indication could be provided by the increasing gap between rich and poor which characterized much of late antiquity, and is best symbolized by the construction of the lavish palatial mansion at Sagalassos in the course of the second half of the fourth century CE.135 Developments such as these illustrate how economic surplus, in case there was any, could be siphoned off in the hands of the happy few, resulting in outward signs of well-being but masking patterns of social inequality. Such scenarios, developed from Thomas Malthuὅ’ (1766-1834) ideas and work, still inspire economic historians. As we know this is also the case with the Roman Empire. As with most models and theories, when placed in context, aspects and assumptions can be criticized. K.G. Persson in his discussion of the economic history of Europe between 600 CE and the present summarized three 133 MCCORMICK, Origins [n. 132], p. 31. Approaching [n. 32]. 135 I. UYTTERHOEVEN / H. KÖKTEN / M. CORREMANS / J. POBLOME / M. WAELKENS, Late Antique Private Luxury. The Mosaic Floors of the ‘Urban Mansion’ of Sagalassos (Ağlasun, Burdur – Turkey) in JÖAI 82, 2013, 373-407. 134 SCHEIDEL, A COMPLEX ADAPTIVE SYSTEM 137 main issues with Malthusian theory.136 First, Malthusian thinking underestimates permanent effects induced by technological progress, even low-key improvements in agricultural yields, beyond the traditional acknowledgment of short-term influence on income levels and population growth. Secondly, the economy of such communities is traditionally characterized as fairly closed, neglecting potential and connectivity generated from local, regional and wider specialization and resulting exchange. Finally, Malthusian models work with underdeveloped fertility strategies for households, focussed mostly on how increases in income resulted in higher numbers of children, but underestimating how families could be forward-looking in their appreciation that quality of childcare (education, nutrition) mattered too, as well as maintaining given standards of living for the entire family. Besides Malthusian models, the ideas and work of Adam Smith (1723-1790) have been developed into an alternative analytical framework.137 According to these views, the conditions for economic growth per capita in a pre-industrial economy are the creation of gain from specialization or division of labour, applying knowledge based on experience and trial and error, and from trade based on differences in resource and climate. As a matter of fact, this point of view can be understood along similar lines as the development of optimal exergy buffering strategies by complex adaptive systems. Specialization allows producers to improve their skills, increasing efficiency of practices and labour productivity, creating sufficient margin to increase output and engage in exchange with producers specialized in other goods. The limit for specialization or division of labour is the level of aggregate demand in an economy. As a result, population increase will sustain division of labour as long as this growth is also associated with increasing aggregate income, allowing full-time specialization. The geographical area determining aggregate demand is open and in tandem with the socio-political order of the day, in this case ranging from the scale of the Roman Empire to the regional unit of ancient Sagalassos and its territory. Changes to the socio-political structure of society can reduce demand, making specialization superfluous, and invoking a process of technological regress. K.G. Persson ὅtipulἳtedΝthἳtἈΝ“ὅὁciἳlΝὁὄdeὄ,ΝpὁpulἳtiὁὀΝgὄὁwth,ΝtὄἳὀὅpὁὄtΝὀetwὁὄkὅ,ΝmἳὄketὅΝἳὀdΝ money are prerequisites for growth based on increased division of labour and trade.”138 Science rarely produces rich insights by choosing between two so-called rival models. In this case as well the challenge is rather to understand how the study region developed a balance between Malthusian and Smithian forces or between 136 K. G. PERSSON, An Economic History of Europe. Knowledge, Institutions and Growth, 600 to the Present, Cambridge, 2010, p. 42-59. 137 PERSSON, Economic History [n. 136], p. 21-41. 138 PERSSON, Economic History [n. 136], p. 28. 138 JEROEN POBLOME negative and positive demographic effects in order to result in constant abovesubsistence income. It seems to me that the latter condition is met in our case-study, especially in late Roman times. To start with, it is important that the polis of Sagalassos remained functioning, representing a concentration of people which were in most cases not active in food production. Even though farmers may have represented a certain proportion of the urban population, they were not the typical inhabitants. In general, townsfolk are estimated to represent about 10% of the total population in antiquity.139 Therefore, the very existence of towns proves that the urban dwellers managed to get hold of part of the agricultural surplus. They did not do this occasionally, but year in, year out. In this way, farming produce and surplus entered the towns in a systematic way. As a result, farmers had to create a surplus for their own survival purposes and had a guaranteed market potential in their towns. Surplus agricultural production and its exchange were systematic features of ancient society. Unlike in capitalist societies, however, the farming strategy remained traditional and aimed at minimum risk and maximum options of survival. Having a little of everything was considered better than profit-seeking. Surplus therefore was systematic but also mostly smallish and production was based on small-scale units. We should go one step further, however. Indeed, sometimes harvests went wrong, climatic conditions changed or some cities outgrew their hinterlands. Ancient society and its towns actually had no institutional way of dealing with these problems of distribution. In general, imports reached the market as a result of private initiative. In effect, climate and geography in combination with social and political conditions ensured that there would be a substantial medium-range movement of staple foodstuffs. This medium range exchange of goods is another systematic feature of ancient society. It involved a lot of trade, but also reciprocity between various estates of one owner, as well as with his friends, between imperial estates and between lands owned by churches and monasteries. Redistribution played another major role in supplying institutionally benefiting parties. Also in the case of medium-range exchange the tendency is to restrict risks and loss-on-investment as much as possible. Exchange in antiquity follows a model of delegated, and whenever possible, shared risks. It is very much these tendencies and structural features of ancient society which explain why the comparison of the archaeological record between towns or regions mostly looks like an unintelligible mix. The hotchpotch of distribution and consumption patterns is not dependent on models of direct trade, rather on untraceable models of dependency in exchange and overlap of directions and exchange mechanisms by multiple parties and institutions, fit to be deconstructed as a panarchy. 139 SCHEIDEL, Demography [n. 34]. A COMPLEX ADAPTIVE SYSTEM 139 In its own sort of way Roman Imperial Sagalassos was a player in this field, generating potential, resilience and connectivity. This paper did not so much want to dissect the High Imperial conditions, but wanted to engage in diachronic comparisons, applying the mentioned heuristic framework and ideas generated by the debate on complex adaptive systems. Clearly, there are some similarities between second century CE Sagalassos and its late Roman counterpart, such as the continued tendency of shifting subsistence strategies provoked by nature or man in the different landscape pockets of its ancient territory. This aspect of flexibility and change has been characterized as a feature of complex adaptive systems. There are also differences between High Imperial and late Roman Sagalassos, however, and perhaps more than the similarities, the points of divergence can be considered to be of historical value. Indeed, the two major tendencies of presumed rural population growth and attested increased specialization of the productive landscape may be meaningfully combined, in providing sufficient foundation for recognizing growing aggregate demand as well as increased per capita income. The low-equilibrium trap seems not to have worked in case of late Roman Sagalassos. Moreover, the positive demographic evolution and degree of specialization associated with late Roman times can be considered to reflect successful strategies at exergy buffering, resulting in increased stability, order and complexity in society. Finally, a range of indicators have been mentioned which, especially in late Roman times, represented a balance between contemporary exergy needs and the supporting ecosystem. This equilibrium has been defined as a precondition for approaching regional sustainability and developing moderate growth on a regional scale. The heuristic framework of complex adaptive systems has been helpful in unravelling and placing the various aspects of this reconstruction, as well as concluding that there is a degree of increase in the regional potential and connectivity in late Roman Sagalassos and its territory. This conclusion has as a potential consequence, however, that the regional system could have become less resilient over time, as its growing dependency on specialization in the productive landscape made it also more vulnerable to processes of internal and external change. The Justinianic Plague possibly represented such an external shock to the system, amongst others. It is an open question in which way the so-called third century CE crisis and fourth century CE degree of loss of the role of Sagalassos as regional pole of attraction140 hampered the gradual build-up of the regional demographical pool as well as the potential, connectivity and resilience of the study area. The notion of panarchy that crisis can be defined when various adaptive cycles converge on the point of collapse does not seem to be compatible with the current understanding of the medium-term development of the region in the third and fourth centuries 140 POBLOME et al., Tinkering [n. 106]. 140 JEROEN POBLOME CE, however, at least not when regarded from the particular angle of the productive landscape. Although I hope to have demonstrated the usefulness of the heuristic framework, including that of complex adaptive systems, in approaching the development of nature-society interaction, regional analysis and diachronic comparison, clearly the exercise is not finished. This paper focused mostly on aspects of production in rural and urban contexts or the generation of means at the regional level, mostly based on material available in publication. Important indicators such as the regional agricultural carrying capacity or forestry yields are still being developed, for instance, and are expected to contribute in important ways to the methodological development as well as to the application of the model. Also, ways need to be examined on how better to measure or define the effects of the different indicators or adaptive cycles onto one another. Future papers should also aim at deeper integration of the indicators and heuristic frameworks, aspiring to investigate regional sustainability from the following formula: (climate + vegetation + landscape) + (agriculture + animal husbandry + artisanate) + (society) + (time). Finally, in order to reconstruct an encompassing socio-ecological system in the past aspects of expenditure of means as well as governance also need to be looked into. In the meantime, however, it should be safe to presume that although it must have been fun to live in Sagalassos during the Antonine period, so much praised by Edward Gibbon, especially if you were rich, the Theodosian age seems to have provided more guarantees at success for every member of society. Acknowledgements The research for this paper was supported by the CORES network of the Belgian Programme on Interuniversity Poles of Attraction (http://iap-cores.be/), the Research Fund of the University of Leuven (GOA 13/04), Project G.0562.11 of the Research Foundation Flanders (FWO) and the Hercules Foundation (AKUL/09/16). J. Poblome was appointed Francqui Research Professor for 2011-2014. I should like to thank Wim Van Neer, Bea De Cupere, Elena Marinova, Peter Talloen, Rinse Willet and Patrick Degryse for having commented on earlier drafts of this paper, in the true spirit of academic camaraderie.