5 Territorial and Nonterritorial Routes to Power: Reconciling Evolutionary Ecological, Social Agency, and Historicist Approaches

We argue that evolutionary and ecological models of territorial behavior are useful to archaeologists holding varied theoretical positions. First, we explain how evolutionary and ecological approaches complement, rather than conflict with, social agency and historicist approaches. Second, we review and expand upon models from evolutionary ecology and their application to the ethnographic and archaeological records. This review reveals that territorial behavior spans a continuum from defense of selected resources, to control of “home ranges” or spheres of influence, to complete defense of a proscribed geographic area. Third, we emphasize that archaeologists and modelers should explicitly define the demographic scale of territorial behavior under consideration, given that resource defense by large groups requires solutions to collective action problems. Finally, we suggest that the economic defensibility logic underlying ecological models of territoriality applies to any resource type, not just territory. Furthermore, social and political power often require successfully defending spatial territory. [evolutionary ecology, economic defensibility, territoriality, human behavioral variation, archaeology] T he chapters in this volume effectively counter the notion that social power can be reduced to control over fixed, discrete, and contiguous territory. In their introduction, VanValkenburgh and Osborne use historical and archaeological examples to argue that social power does not always directly involve control over geographic space. They argue further that the origin and implementation of social power is culturally, historically, and institutionally particular. Together, these arguments seem to imply that models of territoriality developed in evolutionary ecology are not so useful to archaeologists. In our chapter, we maintain that these models of territoriality are useful tools for analyzing the exercise of power over people and places. In fact, we suggest this is the case because of particularly human characteristics such as institutions and agency, not despite them. Our argument is fourfold. First, we explain how social agency, historicist, and ecological approaches to social behavior complement one another. Ecological approaches can and should incorporate the social agency, history, and chance featured in standard social science approaches. Second, we explain the principle of economic defensibility that underlies ecological models of territoriality. Ecological conditions, broadly construed, interact with technological and sociopolitical factors to shape economic defensibility, leading to predictions that anthropologists and archaeologists can and have tested successfully, though much remains to ARCHEOLOGICAL PAPERS OF THE AMERICAN ANTHROPOLOGICAL ASSOCIATION, Vol. 22, Issue 1, pp. 72–86, ISSN 1551-823X, online ISSN 1551-8248. C © 2013 by the American Anthropological Association. All rights reserved. DOI: 10.1111/apaa.12004. Territorial and Nonterritorial Routes to Power 73 be done. Third, we show that the economic defensibility of a resource depends in part on the size of the group that is necessary to defend it. For this reason, resources indefensible by individuals or smaller groups may be defensible by larger groups, but only if group members can successfully cooperate in defending them. Defense of territory within larger groups entails solutions to collective action problems, and thus may require multifaceted social strategies and yield complex social dynamics, as other approaches suggest. Finally, we generalize the logic of economic defensibility and show that it can apply to any type of resource, not just territory. In doing so, we bring the argument back to the land, so to speak, showing that social and political power is often a byproduct, however indirect, of successfully defending resources. Yet territorial routes to power are more numerous, complex, and variable than suggested by the standard models contested in this volume. Throughout this chapter, we use examples from the ethnographic and archaeological literature, including ones presented in other chapters in this book, to support our argument. Evolutionary Ecological Models Are Compatible with Agency, History, and Chance Evolutionary Ecology and Behavioral Contingency: Room for Agency “Evolution” has an image problem with many social scientists. Laland and Brown (2011) present a balanced account of the controversies surrounding the use of evolutionary theory to understand human behavior. They note that some researchers think it presumes the existence of rigid, genetically programmed patterns of behavior that have no roles for agency and cultural variation. Others associate it with notions of “stages” of social evolution, from simple to complex. Both images are inaccurate and outdated. Kelly (2013) provides a concise history of how evolutionary thought in anthropology transformed from a unilineal to a multilineal model, beginning with the work of Julian Steward and continuing through the development of human behavioral ecology to the present. Furthermore, recent work showing how humans dramatically alter the adaptive landscape to suit their evolved goals, a process sometimes termed niche construction, makes the theory and methods of evolutionary ecology potentially applicable to all human societies, including postindustrial ones (Broughton et al. 2010; Laland and Brown 2006; Smith and Wishnie 2000). To demonstrate that evolutionary ecological models are compatible with social agency, we examine the assumption of behavioral rigidity. Many models of evolution by natural selection do assume for simplicity that there are fixed, genetically determined behavioral strategies. Although this approach is sufficient to understand some patterns of human diversity, such as the distribution of blood antigens in response to selection within various disease ecologies, it is insufficient for analyzing complex behavior. Researchers must also consider three other important processes. First, various learning mechanisms allow faster adjustment to selection pressures than do genetic adaptations. While learning itself is not an evolutionary process, the underlying mechanisms that support it are housed in the brain, which like any organ in the human body has been subject to many generations of evolution. The evolution of human cognition may entail variability selection, which is the selection of adaptations suitable to a wide range of environmental conditions, such as the highly variable environments in which early humans evolved (Potts 1998). Second, another adaptation to environmental variability among humans is social learning. In humans, social learning involves the imitation of complex tasks, the transmission of large volumes of information using symbolic communication (language), and the imitation of successful or prestigious individuals (Richerson and Boyd 2005). High reliance on social learning creates its own evolutionary dynamic (i.e., cultural evolution) and alters the selective environment for genetic variants (i.e., gene–culture coevolution). Finally, the aforementioned niche construction involves the feedback between an organism’s effect on its environment and its adaptive fit. In the present volume, Greene and Lindsay (Chapter 4) discuss political transformation in Late Bronze Age Armenia, which included increased investment in built environments. They argue that these built environments mediated transformations in the social order, which may be an example of niche construction. All three of the processes defined above provide ample room for agency because they depend on people making choices. Choices in turn depend on beliefs and preferences, which may be inherited culturally, partly shaped by evolved learning mechanisms, produced through gene–culture coevolution, and so on. In many cases, we do not need to know which of these processes shaped the behavioral patterns we seek to understand, since the various mechanisms will often produce the same final outcome. Thus, we can simply use evolutionary models (such as the economic defensibility model of territorial behavior discussed below) to derive testable hypotheses about these patterns. No particular degree of determinism, genetic or otherwise, is required for this strategic approach to adaptation. Rather, it assumes 74 Benjamin Chabot-Hanowell and Eric Alden Smith that agents inherit or invent conditional strategies that respond to varied social and environmental constraints and opportunities with the appropriate (adaptive) tactic. That said, any evolutionary analysis should recognize the constraints on adaptation. In the next section, we discuss these constraints and link them by analogy to key features of historical particularism. Constraints on Adaptation: Room for History Phylogenetics is the study of evolutionary relationships between evolutionary lineages, such as species (Felsenstein 2004). Phylogeneticists distinguish between traits shared by common descent on the one hand and those shared through parallel or convergent evolution on the other. Similarly, Boas and other historical particularists focus on how populations influence one another while maintaining characteristic features through time (Perry 2003). The inheritance of cultural traits through imitation is analogous (though not identical) to the inheritance of genetic material (Mesoudi 2007; Mesoudi et al. 2004), and imitation across cultural boundaries (often termed diffusion) is analogous to gene flow between populations. Thus any general model of human behavior must account for historical relationships, else we risk drawing spurious correlations between traits and environments (Gray et al. 2007; Mace and Holden 2005; Mace and Pagel 1994). Increasingly, evolutionary anthropologists have adopted formal phylogenetic methods to resolve the issue of spurious correlations, termed Galton’s problem. Historical particularists also emphasize the uniqueness of culture histories. A good evolutionist should not disagree. As evolutionary biologist Ernst Mayr (1983) noted, evolution is opportunistic, acting on whatever variation it encounters. The Sun Dance among Plains Indians provides an example of the importance of history to cultural evolution. Once a signal of the bravery and dedication of prospective warriors, the Sun Dance is now a marker of commitment to ethnic identity (Jorgensen 1972). Another reason history is important is that multiple adaptive solutions exist for the same selection pressure. Which solutions emerge depends in part on phylogenetic constraints, and the adoption of one solution places constraints on future evolution (Mayr 1983). Related to the historical constraints on evolution is the constraint of genotype cohesion (Mayr 1983). Change in one trait may affect other traits for a variety of reasons, a process called coevolution. The relationships between traits are also the product of an organism’s evolutionary history. In cultural and historical contexts, the structure of a “cultural recipe” may influence the mode in which it is transmitted and resulting evolutionary outcomes (Mesoudi and O’Brien 2008). Such recipes might include the modular, hierarchical set of skills necessary to construct and use tools (e.g., material selection and procurement, construction, hand–eye coordination). In sum, while evolutionary models do not always account for historical particularity, researchers must and often do extend these models for this purpose when applying the models to real cases. Furthermore, comparative studies drawing on evolutionary theory can strengthen rather than weaken our understanding of how general and particular processes are linked. Accidents and Stochastic Processes: Room for Chance In evolution, is there room for surprises or are behavioral and physiological outcomes perfectly determined and predictable? Applied mathematicians and others continue to argue that evolutionary researchers should pay close attention to random processes (Jagers 2010). Meanwhile, historians often question the difference that small changes would make to history (Mason 2009). What if Napoleon had not (apocryphally) suffered from hemorrhoids, which precluded him from surveying the battlefield of Waterloo? What if Churchill had been killed on the Western Front during World War I? If a massive asteroid had hit Earth 1.8 million years ago and annihilated all large-bodied animals including hominids, neither Napoleon nor Churchill would have existed. How does modern evolutionary theory account for such chance events? On the one hand, it is no accident that globally catastrophic events like asteroid impacts would more likely than not lead to the extinction of large-bodied species like hominids (Cardillo et al. 2005). On the other hand, it is also possible that small changes to the experiences of influential figures would make little difference to large-scale historical outcomes because history arises from a multitude of factors, a recognition that dates back in anthropology at least to Kroeber’s Configurations of Cultural Growth (1944). In such cases, evolutionary reasoning suggests that outcomes commonly attributed to chance or unique events, including those instigated or experienced by influential social agents, may be in some sense determined. Contrastingly, outcomes thought to be caused by specific historical events and figures may be more accurately understood as the product of myriad interacting factors that can, as we shall see, be better understood using evolutionary logic. The rise and fall of empires throughout prehistory and history is a strong example (Turchin 2003). From this perspective, evolutionary theory must have “room Territorial and Nonterritorial Routes to Power 75 for surprises” because evolution does not proceed along a single pathway. In order for variants to be selected, they must better fit prevailing conditions than other extant variants. Accidents shape the course of evolution by creating essentially random shifts in selection pressures, such as climatic variation. Moreover, evolutionary theorists have long recognized the importance of randomness in evolution with two of the four classic forces of evolution, mutation and drift. The introduction of mutations is in some sense random (though not entirely; see Moxon et al. 1994; Nowak 2006a; Pagel et al. 2007; and Rainey and Moxon 2000 for examples of directional selection on mutation rates). In the cultural domain, “mutations” can result from errors in transmission of cultural information, or from innovations. Importantly, the fact that mutations introduced by innovations may be directed to particular goals (i.e., reflect agency), sometimes even adaptive ones, shifts but does not fundamentally derail the evolutionary dynamics (Boyd and Richerson 1985; Henrich et al. 2008). As for evolutionary drift (random error in the sampling of traits from generation to generation), when combined with selection, it can lead to different outcomes than predicted by a simple deterministic model (Taylor et al. 2004). A consequence of the balance between drift and selection is that there are no guarantees for the ascendancy of even highly beneficial traits. On the other hand, drift can increase the chance that adaptively superior traits will replace alternatives that are stabilized at lower adaptive peaks on the fitness landscape (Griffiths et al. 2004; Lande 1985; Wade and Goodnight 1991; Wright 1988). For example, Powell et al. (2009) combined models of random copying error and the introduction of innovations to estimate the critical population density sufficient for the emergence of the modern Homo sapiens toolkit, demonstrating the importance of population density and size (key determinants of the drift–selection balance) to the proliferation of innovations. In sum, the evolutionary process is not purely deterministic even if, for simplicity, the models that describe it often are. Accidents and stochasticity play a major role in population dynamics and thus the emergence of behavioral strategies. Combined with the room for agency and history, room for chance in the evolutionary process narrows the gap between evolutionary, social agency, and historicist perspectives. Evolution is not simple, as evolutionary biologists and social scientists well appreciate. However, many evolutionary models strategically treat evolution as simple and deterministic in order to gain insights into evolutionary processes or outcomes. We now turn to one such strategically simple model, the economic defensibility model of territoriality. Evolutionary Ecological Models of Territoriality Strategic Territoriality: The Logic of Economic Defensibility Nearly 50 years ago, the biologist Jerram Brown published a short paper called “The Evolution of Diversity in Avian Territorial Systems” (Brown 1964). Brown’s immediate goal, as telegraphed in his title, was to explain why some bird species are strongly territorial, others are not, and some species in fact alternate between territorial and open-access regimes. Further variation exists in the immediate benefits of controlling a resource: for many species, it is local patches of food that are defended, while for others nesting sites or roosts are the object. Brown pointed out that controlling the resources in a territory has benefits, but comes at a cost: time and energy (and potentially risk of injury) spent monitoring an area, advertising one’s presence, and deterring intruders. His simple but profound argument was that natural selection would only favor defending territories if the benefits of doing so exceeded these costs—if the net benefits (measured in fitness or its correlates) were positive. This principle has since been called economic defensibility. Brown gave this principle empirical meaning by linking it to the spatiotemporal distribution of resources—specifically, their density and their predictability. Dense resources are more defensible (are more likely to repay the costs of territorial defense) because the area the territory holder must defend to control access to a given resource is smaller, entailing less time and effort in monitoring. Predictable resources are more defensible for two reasons: the area that must be defended to encompass them is easier to locate, and the income (resource consumption) from the area is higher and more reliable. Lest readers think this argument entails environmental determinism, consider that the costs and benefits of both territory defense and resource acquisition are also dependent on the resource user’s capabilities (e.g., birds can fly, can advertise territory residence with song, and so on); in the human case, these capabilities include technology and other culturally variable attributes, including social norms and institutions. Although dated in its particulars, Brown’s simple argument has helped biologists explain the diversity of avian territorial systems, just as his title had promised. It harnessed the logic of natural selection (the tradeoffs involved 76 Benjamin Chabot-Hanowell and Eric Alden Smith in gaining some benefit at some cost, and their links to fitness) to ecological variation. The economic defensibility model has proved remarkably durable and has been validated in hundreds of studies of spatial behavior in a wide variety of species (Davies and Houston 1984; Dubois and Giraldeau 2005). Dyson-Hudson and Smith (1978) published the first anthropological application of the model. That paper employed qualitative assessments of the economic defensibility model in various ethnographic and ethnohistoric contexts to draw three broad conclusions: (1) territorial behavior (exercise of spatial ownership claims, controlling access to resources) is facultative and varies strategically (i.e., people are at least as clever and flexible as birds); (2) this strategic behavior corresponds to variation in resource density and predictability across space and over time, as predicted by the economic defensibility model; and (3) within the same social system (even the same household), territorial strategies may be applied to some resources but not to others. We now briefly discuss these points, summarizing relevant ethnographic and archaeological data. Variation in Human Resource Defense The basic expectations derived from the economic defensibility model are summarized in Table 5.1 (after DysonHudson and Smith 1978; see also Cashdan 1992). Note that these expectations and their predictors are stated in ordinal terms (e.g., low versus high density, intermediate defensibility); rigorous tests would require precise quantitative measures rarely found in the ethnographic and archaeological data on land use (for a recent effort to do so, plus a formalization of the economic defensibility model, see Baker 2003). Nevertheless, the qualitative evidence is quite extensive and (in our view) convincing. Variation across space The economic defensibility model predicts that if density and predictability are high enough, territorial systems (property rights in land) will be favored. There are many cases in which steep gradients in resource density or predictability correlate with marked shifts in land use. For example, over a vast stretch of the Pacific coast of North America (from the Salish of Puget Sound to the Tlingit of what is now Southeast Alaska), dense seasonal runs of salmon fostered permanent villages, territorial claims to salmon streams by corporate kin groups (Donald and Mitchell 1994), and chronic warfare (focused not only on control of resource sites but also on seizing property and slaves). Yet a short distance inland, all along the inland side of the coastal ranges in the Plateau and Subarctic areas, resources were much lower in density and also generally less predictable, and the indigenous societies had traditional land use patterns stressing communal access rights (the exceptions being favored salmon-fishing spots at falls and rapids, which were sometimes owned by kin groups). The contrast between the coast and the interior Pacific Northwest did not match linguistic (and hence presumed cultural historical) divisions: the coastal areas in particular were linguistically diverse, and in some cases language groups spanned the coast–interior divide (e.g., Salishan and Athapaskan languages). Finer-scale variation in land use within ethnolinguistic areas is perhaps more convincing evidence of economic defensibility in action. The Eskimoan (Yup’ik and Iñupiat) peoples of coastal Alaska strongly defended territorial boundaries (Andrews 1994; Burch 1980). Yet after some Iñupiat spread eastward across the Canadian arctic about one millennium ago (in what archaeologists term the Thule expansion; McGhee 1984), the much lower resource density encountered there led to relaxed boundaries approaching open-access land use. Similar variation in territoriality and related aspects of land use are found in the ethnolinguistically homogenous Great Basin. Most of the Great Basin is very arid, resulting in low resource density and predictability. The indigenous Shoshone and Paiute were highly mobile and had few access rules for land use. Yet in well-watered areas such as the Owens Valley, relatively dense and predictable resources were matched by decreased mobility and forms of land ownership that were clearly territorial (Bettinger 1983; Steward 1938; Thomas 1981). Ambrose and Lorenz (1990), Field (2005), and Kennett and Clifford (2004) present similar archaeological analyses of variation in territorial strategies. Steep gradients in resource density and predictability can have profound implications for broad aspects of political economy. Many scholars have pointed out that stratified social systems have initially arisen in areas with such gradients, such as fertile floodplain valleys surrounded by arid regions (Carneiro 1970) and rainforests with patchily distributed potable water resources (Lucero 2002). The link to economic defensibility may be quite direct: where kin groups or other coalitions are able to control the resource patches, subordinates have few options, and stable levels of exploitation can increase (Boone 1992; Smith et al. 2010). Again, note that this is not environmental determinism: resource density and predictability are in part functions of social and technological variables (e.g., agricultural techniques, labor division, social stratification), and social agency (in seeking or resisting political and economic domination) is central to the process. Territorial and Nonterritorial Routes to Power 77 Table 5.1. Resource Density, Predictability, Defensibility, and Resultant Land Use Resource density Resource predictability Economic defensibility Predicted land use Low Low Low High mobility, dispersed population Low High Intermediate Home range system High Low Intermediate Mobility, information sharing High High High Geographically stable territoriality Note: Modified from Dyson-Hudson and Smith 1978.