Anarchy and Culture:
Insights from the Anthropology of War
Jack Snyder
The end of the Cold War has given rise to hopes among many international relations scholars and public activists that a dramatic transformation in world politics is now unfolding. They contend that changes in norms, ideas, and culture have the power to tame the historically war-prone nature of international anarchy. 1 This analysis and the prescriptions that follow from it exaggerate the autonomy of ideas and culture in shaping behavior in anarchy. A rich body of research on war by anthropologists suggests that ideas and culture are best understood not as autonomous but as embedded in complex social systems shaped by the interaction of material circumstances, institutional arrangements, and strategic choices, as well as by ideas and culture. Cultural prescriptions that ignore these multifaceted interactions will provide a poor road map to guide strategies of global change.
Those who foresee substantial opportunities to transform the war-prone international system into a realm governed by benign norms contend that "anarchy is what states make of it." 2 In their view, culture, defined as shared knowledge or symbols that create meaning within a social group, determines whether behavior in the absence of a common governing authority is bloody or benign. If more benign ideas and identities are effectively spread across the globe through cultural change and normative persuasion, then "ought" can be transformed into "is": Support for warlike dictators can be undermined, perpetrators of war crimes and atrocities can be held accountable, benign multicultural identities can be fostered, and international [End Page 7] and civil wars will wane. 3 These academic concepts have a potent counterpart in the international human rights approach of activist organizations. 4
In contrast, skeptics about such transformations argue that anarchy, whether among states coexisting in a self-help system or among contending groups inside collapsed states, gives rise to an inescapable logic of insecurity and competition that culture cannot trump. 5 These skeptics fear that a transformative attempt to supersede self-help behavior amounts to reckless overreaching that will create backlashes and quagmires. Ironically, in this view, the idealist vanguard of the new world order will need to rely increasingly on old-fashioned military and economic coercion in a futile effort to change world culture for the better. 6
This is a debate of compelling intellectual and practical import. It lays bare the most fundamental assumptions about the nature of world politics that underpin real policy choices about the deployment of the vast military, economic, and moral resources of the United States and other wealthy democracies.
However, some of the leading voices in this debate, both in academic and broader public settings, overlook the decisive interplay between situational constraints and the creation of culture. Prophets of transformation sometimes assert that politics in anarchy and society is driven by "ideas almost all the way down." They dismiss as negligible what Alexander Wendt calls "rump" material constraints rooted in biology, the physical environment, or other circumstances unalterable through changes in symbolism. 7 For them, "agency" by political actors committed to social change consists primarily in working to alter prevailing principled ideas, such as promoting the norm of universal jurisdiction in the case of crimes against humanity. In contrast, working for improved outcomes within existing constraints of material power, for example, by bargaining with still-powerful human rights abusers, does not count for them as true "agency"; rather it is mere myopic "problem solving" within constraints. 8 Conversely, when prophets of continuity discuss culture at all, they treat it as a largely unchangeable force that may have some effect in constituting the units competing for security but that has at most a secondary effect on strategic interactions between those units, which are driven mainly by the logic of the anarchical situation. 9 [End Page 8]
This is an unnecessarily truncated menu of possibilities for imagining the relationship between anarchy and culture. Ironically, in light of the ambitiously activist agenda of the proponents of cultural approaches to international relations, their one-dimensional approach limits agents to a peculiarly circumscribed set of tools for promoting political change. A more promising approach would integrate the material, institutional, and cultural aspects of social change, drawing on the insights of theories of complex systems.
Robert Jervis reminds us that the elements of complex systems, such as international anarchy, are highly interconnected and consequently the behavior of the system as a whole cannot be understood just by examining its separate parts. 10 In a tightly coupled system, a change in one of its aspects, such as norms or ideas, is unlikely to have simple, linear effects. The consequences of any change can be predicted only by considering its interaction with other attributes of the system. For example, whether the spread of the concept of national self-determination promotes peace or war may depend on the material and institutional setting in which it occurs. Negative feedback may cancel out a change that is at odds with the self-correcting logic of the system as a whole. Conversely, in unstable systems, positive feedback may amplify the effects of small changes. More complicated feedback effects may also be possible, depending on the nature of the system. Actions in a system may have different consequences when carried out in different sequences. In social systems, outcomes of an actor's plans depend on strategic interactions with the choices of other independent decision makers. For example, projects for cultural change are likely to provoke cultural counterprojects from those threatened by them. Even in "games against nature," changes in behavior may transform the material setting in ways that foil actors' expectations. For all these reasons, system effects are likely to skew or derail transformative efforts that focus narrowly on changing a single aspect of social life, such as norms and ideas.
All of these system effects are relevant to understanding the effect of culture on conflict in anarchy. As I describe later, anthropological research on war shows that ideas, norms, and culture are typically interconnected with the material and institutional elements of anarchical social systems in ways that produce the full panoply of Jervis's system effects. In such systems, efforts to promote cultural transformation need to take into account the material and situational preconditions that sustain these developments; otherwise they are likely to produce unintended consequences. Underestimating situational constraints is just as dangerous and unwarranted as reifying them.
Testing the Effects of Culture:
Insights from the Anthropology of War
Current debates about anarchy and culture have been carried out largely at the level of abstract philosophy and visceral morality. Ultimately, however, the impact of [End Page 9] culture on war in anarchy is an empirical question. What evidence should be examined? To assess the claim that behavior in an anarchical system is what the units and their culture make of it, the obvious methodological move is to vary the culture of the units or of the system as a whole and then assess the effect on behavior.
Reasonably enough, some scholars who see anarchical behavior as culturally constructed examine contemporary changes, such as the peaceful end of the Cold War, the emergence of the democratic peace, and the purported current strengthening of human rights norms. 11 In assessing such recent developments, it is difficult to distinguish the hopes of transitional moments from enduring trends. These kinds of tests, while not irrelevant, are not well designed to disentangle the effects of autonomous changes in ideas and culture from the effects of self-justifying U.S. hegemonic power, an ideological pattern that was quite familiar in the old world order.
Other scholars try to show that the progenitor of the contemporary international system--the historical European balance-of-power system among sovereign states--was itself a by-product of ideas, such as the Protestant Reformation or analogies between sovereignty and individual property rights. 12 The implication is that whatever has been established by ideas can also be dismantled by ideas. However, it is not a simple task to disentangle the effects of war, state formation, and ideological change on the emergence of the competitive states system. 13
Arguably, a comparison of the European system with behavior in other anarchical state systems offers a methodologically cleaner way to vary culture and assess its effects. However, when cultural constructivists do look at behavior in anarchies in cultural settings radically different from our own, they sometimes fail to exploit obvious opportunities for focused comparison. For example, Iain Johnston's prominent book Cultural Realism shows how the strategic wisdoms of the anarchical ancient Chinese Warring States system were passed down to future generations to constitute a warlike strategic "culture." His adherence to a cultural account of Chinese strategic practices remains untroubled by the fact that these ideas and practices are similar to those of the anarchic European balance-of-power system, the ancient Greek city-states, and the ancient Indian states system described by Kau tilya, a set of cultures diverse in almost every way except their strategic behavior. 14 At a first approximation, it would seem from this evidence that state behavior in anarchy is not fundamentally altered by variations in culture. This is not to deny that cultural differences may have influenced the meaning the actors imputed to their military behavior, some of the goals for which they fought, and some political [End Page 10] features of these anarchical systems. Nonetheless, the evidence from historical state systems strongly suggests that the situational incentives of anarchy have significantly shaped strategic behavior in ways that transcend culture.
Constructivists have paid less attention to another body of evidence ideally suited to assessing the effects of variations in culture on behavior in anarchy. For decades, anthropologists have been amassing a theoretically rich, empirically substantial, and methodologically self-aware body of statistical and case-study research on the relationship between war and culture in stateless societies and pre-industrial anarchic systems. 15 Many of the causal factors and processes they examine will seem strikingly familiar to students of modern international relations--for example, security fears, economic rivalry between groups, economic interdependence, the institutionalization of cooperative ties across political units, the popular accountability of decision makers, and the nature of identities and cultural symbolism of the political units and of the anarchic system as a whole.
Notwithstanding the familiarity of these categories, the kinds of societies anthropologists of war study differ vastly from contemporary, industrialized, bureaucratized societies, and thus research findings on the anthropological history of war cannot simply be read off and applied to debates about the construction of culture in today's "new world order." Indeed, a central part of the constructivist claim is that the spread of a new democratic culture may be on the verge of making obsolete all those old cultural patterns, whether those of the Cold War, the ancient Chinese Warring States, or warring villages in the Venezuelan jungle. 16 Moreover, evidence based on technologically primitive societies, some of which lack the minimal economic resources needed for assured survival, may load the dice in favor of explanations based on material pressures. However, following the arguments of Durkheim or Weber, one could also argue that this type of evidence is biased in favor of cultural explanations on the grounds that social solidarity in such societies is achieved more through cultural rituals than through differentiated, rational-legal institutions.
In some respects, the anthropological record provides a particularly good test for constructivist hypotheses about culture and anarchy. A common culture and dense social ties across politically autonomous units characterize many of the anarchical societies anthropologists have studied. These cases are ideally suited to evaluating whether such ties can overcome the Hobbesian dangers of anarchy.
The main value of reviewing this anthropological research is not to report any universally definitive findings on anarchy, culture, and war. There is no reason to believe that future anarchical systems will be limited only to the patterns observed in the past. Rather, I am interested in expanding our thinking about the varied ways in which situational constraints in anarchy may relate to changing patterns of culture. Just as Clifford Geertz's writings on Bali have stimulated constructivist [End Page 11] thinking about social relations in more familiar settings, so too might a look at anthropological research on warfare help international relations scholars step outside the limiting categories we sometimes take for granted.
A review of these anthropological findings suggests that attempts to explain behavior in anarchy solely in terms of either cultural or material causes are equally unconvincing. 17 More promising approaches posit different ways to integrate material, institutional, and cultural elements. One such approach nests these elements hierarchically, with the material situation establishing constraints or problems that require institutional or cultural solutions. 18 Yet this hierarchical approach, a carry-over from the work of earlier functionalists like Bronislaw Malinowski, is compelling only when material pressures are overwhelming and cultural responses line up with them in a tidy fashion, which is seldom the case. A second approach treats material, institutional, and cultural factors as fully independent causal variables that interact to shape outcomes. 19 While this straightforward approach may be satisfactory for understanding some problems, anthropological case studies of warfare often depict complex feedback relationships among material, institutional, and cultural patterns. Culture shapes material goals and capabilities, and material tasks and tools shape culture and institutions. 20 Capturing these reciprocal dynamics calls for yet a third approach, a systems perspective. Such an approach does not necessarily envision simple homeostatic systems in which war is the endlessly stable equilibrium. Rather, anthropologists of war commonly depict evolutionary systems in which the historical trajectory counts, equilibrium is elusive, and material conditions and cultural patterns affect each other in complex, changing ways.
In developing these arguments, I first define basic concepts. Second, I assess a variety of hypotheses that attempt to reduce the explanation of war behavior in anarchy to one of three types of causes: material-environmental, institutional, or symbolic-cultural. Third, I discuss three approaches that try to integrate material, institutional, and cultural factors in explanations of war behavior in anarchy: hierarchical nesting, interacting variables, and systems approaches. Finally, I discuss the democratic peace as a material-institutional-cultural system and address strategies for transforming behavior in contemporary anarchies.
Behavior in Anarchy:
Basic Definitions and Analytic Concepts
The research program on war in the field of anthropology seeks to explain variations in the degree of organized political violence between autonomous political groups as a function of variations in (1) material, environmental, technological, and [End Page 12] situational constraints, including the strategic implications of anarchy; (2) social institutions; and (3) culture. This literature assumes a variety of basic motives driving behavior, especially the physical survival of individuals and their success in genetic propagation, but sometimes it posits additional social and psychological needs of various kinds. It examines a number of potential causal mechanisms linking individual motives and constraining structures to violent outcomes, such as selection through competition, strategic interaction, institutionalization through converging expectations, socialization through myths and rituals, and persuasive communication. Ethnographic case studies reconstruct causal processes and cultural meanings. Sometimes case studies provide snapshots of a particular moment in time; sometimes they trace developments over time. Statistical analyses employ standardized databases, such as the Standard Cross-Cultural Sample of the characteristics of 186 pre-industrial societies during the nineteenth or twentieth century. 21 Before discussing the findings of this literature, its basic terms need to be defined and its epistemological assumptions justified.
War
Malinowski, a founder of modern anthropology, defined war as "the use of organized force between two politically independent units, in pursuit of [each unit's] policy." 22 Similarly, a recent statistical study based on the Standard Cross-Cultural Sample defines war as "socially organized armed combat between members of different territorial units." 23 Such studies offer no quantitative cutoff point comparable to the Correlates of War project's benchmark of 1,000 battle deaths for distinguishing warfare from lesser forms of violence, but they do normally distinguish qualitatively between warfare organized by recognized group leaders and unsanctioned violence by certain in-group members against certain out-group members. 24 Wars included in the Standard Sample often kill a higher proportion of the unit's population than do modern wars. 25 These studies and databases distinguish warfare between politically autonomous units sharing a common language or culture and warfare between groups having different languages or cultures. [End Page 13]
Anarchy
As in international relations research, in the anthropological literature political autonomy means the absence of an external enforcement authority wielding power over the unit. Such autonomous units, whether villages or nation-states, may have social and normative ties linking them, but as long as the units enjoy autonomy of political decision making in deciding, for example, how to apply such norms to a particular situation, they stand in an anarchical relationship to each other. Thus, most anthropological studies of war are conceived in a way that structurally parallels the concern of international relations scholars with warfare between autonomous states in an anarchic setting. As Kenneth Waltz notes, when structural conditions permit or encourage it, "many different sorts of organizations fight wars, whether those organizations be tribes, petty principalities, empires, nations, or street gangs." 26
Sometimes identifying the valid units of these anarchical systems can be problematic. Some of the most belligerent societies, such as those in South American jungles, live in politically autonomous villages, each loosely led by a headman. However, these villages sometimes split into groups along kinship lines. Even when they do not split, they normally maintain cross-cutting kinship ties with rival villages, and this often entails prohibitions on individuals' killing their own kinfolk during intervillage combat. 27 Kinship lineages may have a decision-making structure that is no less authoritative than that of the villages, making it difficult to determine the basic unit of the anarchical system. Just as international relations scholars typically treat territorial units, states, as the basic units in anarchy-- rather than ethnic groups or other cross-cutting loyalty networks--so too do anthropologists tend to treat residential villages as their units in anarchy, while also recognizing that nonterritorial lineages can play a decisive role, for example, in mobilizing power at higher levels of aggregation, such as the tribe. Conceptually, the crucial criterion should be not territoriality per se but which type of unit exhibits the greatest decision-making autonomy.
Culture and Institutions
Like many anthropologists, I follow Clifford Geertz in defining culture as a system of symbols that creates meaning within a social group. 28 This definition generally accords with standard usage among social constructivists in the field of international [End Page 14] relations, who see culture (along with two of its components, norms and identity) as a central concept in their ideas-based accounts of politics. 29 However, some anthropologists studying war define culture in a more encompassing way that includes not only language and symbols but also social institutions and prevailing patterns of behavior. Indeed, anthropological research on war and culture had its beginnings when culture was commonly defined in this broad way, and this broad definition is resurfacing in some circles. 30 In that tradition, even recent statistical studies of culture and warfare, such as Marc Ross's, are apt to define culture in the old way as "the particular practices and values common to a population living in a given setting," including "both beliefs and norms about action and behaviors and institutional practices." 31 Below, I alert the reader when I refer to anthropological studies that subsume institutions under the heading of culture.
An advantage of the Geertzian definition, which limits culture to meaningful symbols, is that it permits scholars to investigate the relationship between ideas and institutions empirically, rather than defining them as integrally related under the umbrella heading of culture. 32 This is important for scholars of international relations, who have developed institutional theories of behavior in anarchy (whether based on domestic institutions like the democratic peace or on international institutions like multilateralism) that are generally seen as distinct from cultural theories. Thus, most international relations scholars will want to distinguish institutional hypotheses from cultural hypotheses when reviewing anthropological research on war. To that end, I define an institution as a conventional pattern of behavior around which expectations converge. 33 The holding of periodic elections, the coordination of military action by orders from a general staff, the habit of patrilocal residence of sons, and the practice of cross-cousin marriages are examples of institutions. War itself is an institution only if, as some anthropologists argue, patterns of engagement with the enemy are governed to a significant extent by conventions, in the manner of rule-governed sporting events, rather than by unregulated strategic optimizing behavior. 34
Even with this definition, it is not always easy to draw a sharp line between institutions and cultural symbols in concrete social settings. Both institutions and culture are part of the process by which people coordinate their behavior in groups. Symbols (ideas) may serve as focal points to guide the coordination of expectations in cases where purely pragmatic strategic calculations could yield multiple equilibria or no equilibrium. 35 Moreover, institutions tend to be shored up by symbols [End Page 15] (norms) that convince people to behave in accord with institutionalized practices for moral reasons, and not simply for the practical benefits of coordinating behavior with the expectations of others. Conversely, norms often require institutionalized enforcement. For example, Human Rights Watch stresses the importance of backing the concept of rights with the development of strong international judicial institutions that have a capacity to enforce them. 36 Thus, I am not claiming that the causes and consequences of culture and institutions are not intertwined. Nonetheless, the conceptual distinction in principle remains clear: Culture is a system of symbols; institutions are conventional, repeated behavior.
The distinction between cultural symbols and institutions serves the analytic purpose of clarifying the role of culture as an explanatory variable. This is somewhat ironic, since anthropologists like Geertz who champion the symbol-centered definition of culture are often opposed to explaining social life in terms of causal variables. In part, this is because such scholars are more interested in understanding the meaning social life has for participants in it than in explaining or predicting behavior, so treating culture as a variable would be pointless in terms of their research objectives. In contrast, constructivist scholars in international relations adhere to the symbolic or idea-based conception of culture in large part because it allows the posing of nontautological propositions about the effect of ideas on behavior. 37
Another methodological problem arises in defining the appropriate unit of analysis for cultural propositions about behavior in anarchy. In principle, behavior in anarchy might be shaped by the separate cultures of the politically autonomous units, by the shared culture of the anarchical society as a whole, or by both. Ethnographic case studies may discuss either or both. Quantitative anthropological literature typically defines "external" war as organized violence between culturally distinctive societies, whereas "internal" war takes place between politically autonomous units, such as villages or bands, that are culturally the same. 38 This creates the opportunity to study whether commonalities of culture affect the prevalence of warfare between groups. However, it does tell us whether warfare is more common in anarchical settings than in hierarchical ones, because even the internal conflicts are between autonomous units.
The Material and Environmental Setting
It is difficult to draw a sharp line between explanations for war that focus on material or environmental constraints and those that focus on the groups' institutions or culture. Basic biological needs and features of the physical environment impose [End Page 16] material constraints that may affect the prevalence of warfare, yet human institutions and culture may exert powerful effects on the ecological setting in which social competition and cooperation arise. The strategic environment affects people's decisions about war through the information they have about it, and information is influenced not only by the objective availability of data but also by decisions about what information to seek out and how to interpret it.
Also ambiguous is the category "anarchy," which is often considered one of the material or environmental explanations for war. However, its very definition is institutional: It is the lack of authoritative institutions to govern social relationships among autonomous groups. Yet the absence of such institutions means that the relationships are played out in an environment where the material capabilities of the units and the material constraints on their strategies are the ultimate arbiters.
Finally, so-called material explanations for war sometimes hinge on prevailing military or economic technologies, which are self-evidently products of human ingenuity, institutions, and ideas. Wendt, for example, identifies technology, including the "mode of production," as one of the "rump material" factors that may to some degree shape behavior in anarchy, whereas he locates "the social relations of production" squarely in the nonmaterial zone. 39
While remaining mindful of the problematic nature of the dividing line between material-environmental influences and institutional-cultural influences, I propose two rules of thumb for making this distinction:
1. Whether the material-environmental constraint or capability arose for some reason unrelated by intention or immediate functional necessity to the phenomenon under investigation. For example, the rise of sedentary agriculture as a technology and mode of production seems to have increased the frequency of warfare somewhat, but it was not developed in order to fight wars, so it counts as an exogenous material-situational factor; in contrast, insofar as the state was developed precisely in order to fight wars better, it counts as an institutional factor.2. Whether social actors can escape the material-environmental constraint in the short run by institutional or cultural innovations. Reciprocal fears of surprise attack in anarchy, whether between Amazonian hunting bands or modern states, can often be mitigated by phased reciprocal demobilization and mutual monitoring, combined with explanations of intentions. 40 In this respect, anarchy may indeed be what the actors make of it. In contrast, it may be much harder to devise a credible commitment not to exploit a massive impending shift in relative power, since any assurances will be viewed as cheap talk unless the institutional and cultural bases for trust have been [End Page 17] established well before the emergence of the trend in relative power. 41 If so, actors may be captives of the logic of a situation that their cultural ingenuity cannot appreciably change in time to avert preventive war.
Hypotheses from Anthropological Research on War and Peace in Anarchy
In the following sections, I review prominent hypotheses on behavior in anarchy from the anthropological literature, laying out their motivational assumptions and causal mechanisms and highlighting relevant research findings. I address in turn material-environmental, institutional, and cultural explanations of war and peace. 42 I assess reductionist theories that attempt to explain war behavior solely in terms of these types of variables or that include other types in only a subordinate way. After showing the shortcomings of the reductionist strategies, I turn to a more complete examination of the interrelationships among causal factors of these different types.
Material-Environmental Hypotheses
Anarchical environments may be conducive to war for a variety of material reasons. International relations scholars have paid particular attention to the exceptionally dramatic possibility that mutual security fears in anarchy may be sufficient to cause war between groups with no other reason to fight. 43 However, there are other material causes of war in which anarchy serves as only a permissive cause or as an exacerbating factor. Groups in anarchy may fight with each other in a zero-sum competition over scarce resources that they both need to survive, because no superior authority prevents this and because institutional mechanisms needed to implement less costly methods of distributing and using resources are missing. Groups in anarchy may also fight over material and biological objectives not immediately necessary for survival yet highly valued across most cultural settings, such as opportunities to improve members' ability to procreate or gain material wealth. In addition, they may fight to maintain their reputations for fierceness and bravery, not because these qualities are arbitrarily valued in the culture, but because they are advantageous for competing successfully in anarchy.
Whether mutual fear is the primary cause or merely a permissive cause, once the cycle of warfare has begun, security fears may intensify and sustain conflicts beyond the point that the material stakes would have otherwise warranted. Thus, according to the realist perspective, groups in anarchy will resort to war recurrently as long as [End Page 18] security is scarce and war is feasible as a method of obtaining resources. Groups that do not seize opportunities to improve their strategic position through warfare will, in this view, tend to be selected out by the unforgiving process of competition with more warlike groups.
The findings of anthropological research on behavior in anarchy are quite consistent with these hypotheses. Indeed, most anthropologists conducting systematic research on this topic consider material-environmental constraints and incentives to be the necessary starting point for any analysis. 44 Over 90 percent of the pre-industrial societies sampled in one prominent quantitative study experienced wars between autonomous units within the cultural group ("internal" wars) at least once per decade, unless they had been pacified through conquest by an outside power, and about half fought almost constantly. About 70 percent fought frequent wars against other cultural groups ("external" wars). 45
This extremely high frequency of warfare in anarchy is not in itself proof of realist hypotheses. Other theories might also predict frequent fighting. For example, social identity theory, based on laboratory experiments showing an almost universal tendency to form instant in-group preferences and out-group stereotypes, predicts rivalry in anarchy for the psychological reason that seeing one's own group as superior bolsters an individual's self-esteem. 46 In contrast, the near-universality of warfare in anarchy does not speak well for theories that rest on culture "almost all the way down." 47 It is difficult to see why so many different cultures would wind up with war-fueling symbolic systems unless some nearly universal features of situation, psychology, or biology were pushing the development of culture in this direction. To solve this awkward problem, Wendt tries to appropriate social identity theory for the cultural approach, but since this is a theory about alleged universal psychological tendencies, not about culture, it fits uneasily with the basic ontology the cultural theorists propound. 48
Anthropological case studies lend further support to the salience of security fears as a source of conflict in anarchy. Anecdotally, descriptions of life in some of the most warlike societies read like caricatures of the historical European balance-of-power system on a particularly bad day. Incentives for preventive attack in the face [End Page 19] of shifts in relative power, the need to maintain the credibility of deterrence, and fine-grained calculations of alliance power and reliability are common themes. Such Hobbesian analyses are prevalent among scholars who have studied primitive warfare in such culturally and ecologically diverse settings as highland New Guinea, the Amazon jungle, and Central Africa. 49 Some participants in these anarchic rivalries tell the anthropologists that they feel "trapped" in a war system they consider bad but strategic exigency makes it impossible to escape from the system. 50
But do these patterns arise from the mere fact of anarchy, or are other exacerbating circumstances needed to bring them about? Research based on the Standard Cross-Cultural Sample suggests that the experience of an unpredictable ecological disaster leading to extreme food shortage is the strongest predictor of an increased likelihood of war. 51 However, a simple model of competition for scarce material resources does not adequately capture the underlying causal dynamic. Societies that fear unpredictable food shortages, even though they have not experienced one in the past twenty-five years, are about as war-prone as those that actually suffered shortages in the recent past. Moreover, routine, predictable shortages of food due to periodic drought, pests, or rising population appear not to increase the chance of war involvement, once the effects of unpredictable food disasters are taken into account. 52 It is debated whether actual land shortages correlate with the likelihood of war, but there is general agreement that perceived land shortages do. 53 Foraging bands (hunters and gathers), which are on average less vulnerable to unpredictable food crises than are settled agricultural groups, tend to fight wars somewhat less frequently. 54 However, this may reflect the more egalitarian social institutions typical of foragers rather than the direct effects of food shortages, and also the fact that foragers' mobile lifestyle makes the accumulation of wealth inconvenient, so there is less to fight over. 55
Statistical studies also show that fear and mistrust of outsiders correlate strongly with war-proneness. Some argue that this is a function of cultural attributes, but others claim that it reflects real variations in the intensity of the material insecurities facing the society. Since the experience of natural disaster is positively correlated with both mistrust and war, it seems clear that material insecurity is at least in part an underlying cause. 56 However, there is also evidence that the culture of mistrust takes on a life of its own apart from material circumstances. Societies experiencing ecological disasters are likely to have folktales involving capricious violence, [End Page 20] though these narratives do not feature food shortages per se. Thus, material threats seem to create a general cultural propensity that becomes conceptually disconnected from its underlying cause. 57
Some practices in anarchy that at first glance seem exotic and "cultural" turn out on closer examination to serve a material and strategic purpose. For example, the Yanomamo invite newly allied villages to feasts to cement the pact and establish its terms. During these events, men from the respective villages take turns slugging each other in the chest as a test of strength and toughness. The outcome of the competition determines the terms of trade in exchanges of women and goods between the villages and also the allocation of rights to use hunting lands. 58 Sometimes the feasts are a ruse, and the stronger village massacres its guests. This coercive bargaining behavior, despite its partly conventional nature, can be interpreted largely in terms of material-environmental influences on violence.
Material and environmental factors affect the warlikeness of groups not only by shaping their strategic calculations but also by selection through competition. Even proponents of cultural explanations of war and peace generally acknowledge that belligerence toward outside groups was functional among pre-industrial peoples. Warlike cultures tended to displace non-warlike ones. The latter were wiped out, incorporated, pushed into inhospitable ecological niches, or forced to develop their own warlike states in order to survive. Outlier cases for cultural theories, which show a mismatch between benign internal cultural attributes and warlike external behavior, are typically those that were forced to adapt their foreign policies to an increasingly rough neighborhood. 59
In short, material and environmental hypotheses about the causes of war, including Hobbesian propositions about security dilemmas in anarchy, seem roughly consistent with anthropological findings. However, there is strong evidence that at least partially independent social, cultural, or perceptual conditions interact with material and environmental factors in shaping the pattern of warfare.
Material motives of subunits. The unit of analysis in studies of pre-industrial warfare is typically the politically autonomous community. Because of the small size of the units, it is easy for ethnographers to trace the connections between the micro-motives of individuals and the strategies adopted by their groups. In some cases, the material objectives of the group are essentially the same as the benefit that accrues to the typical members of the group. Survival of the unit and survival of the members may go hand in hand. In other cases, however, war strategies of the group may benefit a segment of the group's members at the expense of the rest--for example, a stronger lineage at the expense of a weaker one, men at the expense of women, leaders at the expense of followers. Brian Ferguson argues that the "master variable" explaining the war-proneness of a group is the "material self-interest of [End Page 21] those who decide military policy," which may or may not coincide closely with the material interest of the average group member. 60 Thus, unlike many realist scholars of modern international relations, "realist" anthropologists of war are less wedded to the fiction that it is the autonomous political unit that seeks to survive and prosper rather than the individuals and coalitions it comprises. They observe, as in the case of the Yanomamo, that units may fission when disadvantaged subgroups enjoy exit opportunities, as calculated in terms of adequate group size for self-defense, available allies, or secure locations to start a new village. 61
Understood in this fashion, a material-environmental view of behavior in anarchy leads naturally to an interest in internal institutional arrangements that affect who decides on war, who benefits from it, and how tensions between members' motives affect the formation and survival of groups. Anthropological research finds that the material basis of a society strongly affects the society's likelihood of adopting a particular set of domestic institutions. Sedentary agriculture, for example, is associated with the rise of the state and with increasing inequalities of wealth and status, both of which, according to many if not all studies, increase the unit's frequency of war involvement. 62
When do material explanations work best? The anthropological literature suggests that material and environmental factors are among the better predictors of war in pre-industrial anarchies. Nonetheless, strong qualifications must be introduced before we can draw conclusions about the relevance of straightforwardly realist theories to politics in anarchies in our own day.
The direct effects of material-environmental factors on pre-industrial warfare seem to have been strongest for societies living close to the margin of survival, where the combined effects of economic scarcity and reciprocal fears in anarchy were unusually compelling. We can compare politics in these situations of extreme security compulsion with the contemporary behavior of predatory warlord groups in failed states. 63 As in Somalia at the beginning of the 1990s, in the midst of widespread famine and government collapse, insecurity is especially intense where economic security and security from attack are both scarce. 64 This holds true both for pristine pre-industrial societies and for those in contact with economically advanced societies. Military rivalry triggered by competition for Western manufactured goods among the Yanomamo strongly parallels today's power struggles among would-be tinpot dictators for privileged access to the international gem trade, World Bank loans, and humanitarian aid. 65 [End Page 22]
In societies with a more complex division of labor, the causes of war are more likely to be found in institutional arrangements than in the direct effects of ecological scarcity. 66 Thus, patterns of anarchical behavior in institutionally underdeveloped settings should not be unreflectively extrapolated to the analysis of anarchical relations among rich, stable states. By the same token, solutions that work in rich, developed societies should not be extrapolated to economically poor, institutionally thin ones.
Institutional Hypotheses
Behavior in anarchy may vary with the institutional arrangements inside the units, or with the strength and nature of the institutions that regularize links between the units. Literature on the anthropology of war has examined institutional hypotheses that are strikingly familiar to students of relations between modern states. Although many of these studies see a significant causal role for institutions, they also find that the significance of institutions in shaping conflict in anarchy needs to be interpreted in light of environmental and cultural factors.
Institutional patterns within units may determine which interests are most powerful, how accountable the leaders are, and how easily the group can mobilize its capabilities for armed combat. Thus internal institutions that allow elites to reap the gains of conquest, insulate elites from accountability, privilege military specialists, and facilitate military mobilization should make war more likely. Institutional patterns between the units may affect the predictability and transparency of the units' interests and policies in ways that affect the likelihood of identifying mutually acceptable bargains. In this view, effective institutions that regularize relations between units should reduce the likelihood of war, except when improved cooperation between some units comes at the expense of other units. Offensive leagues may facilitate conquest, for example, and cartels may collude to bring rivals to their knees.
Institutionalized patterns of participation and accountability. Both quantitative and qualitative anthropological research support the hypothesis that groups with widespread participation in decision-making processes and elite accountability have been less war-prone. 67 This finding applies principally to patterns of "internal" war and peace, that is, to relations between autonomous political units within a homogeneous cultural group. As with the democratic peace hypothesis about modern states, these are units with widespread political participation that are not fighting each other. In contrast, political units and societies characterized by insulated leadership, hierarchy, and specialized military roles are more war-prone. For example, Carol Ember and Melvin Ember illustrate their statistical findings on [End Page 23] these institutional effects by contrasting the warlike Marshallese Islanders, who were ruled by a stratified hereditary elite with absolute life-and-death powers over their subjects, to the peaceful Cuna Indians of Panama, whose elected chiefs could be sacked at any time by popular demand. 68
Although this finding pertains on its face to the consequences of variations in political institutions, material circumstances and culture are also relevant to the analysis. In order to produce strong statistical results that "account for around 70 percent of the cross-cultural variation in frequency of internal war," Ember and Ember controlled for variations in the strategic setting, such as geographical isolation, and for the tendency of units to split up as a result of internal disputes. 69 Thus groups like the Yanomamo that are participatory but fission-prone are also war-prone. Since fissile groups wind up fighting each other after the breakup, it is important to know the reasons for the tendency to fission in the first place. Arguably, Ember and Ember are smuggling the undertheorized fission variable, which may reflect cultural differences in the management of conflicts of all types, into what seems to be an institutional hypothesis. Finally, the fact that the peacefulness of high-participation groups mainly affects the frequency of "internal" wars among culturally similar units suggests that culture, and not simply institutional structure, may be at work in some way.
Institutionalized patterns of kinship and trade. Even more research has focused on the consequences for warfare of variations in institutionalized patterns of kinship and economic exchanges. For example, societies in which war is entirely absent are often "unsegmented"--that is, kinship ties are understood solely from the perspective of each individual. Lacking the concept of lineages, people in such societies hold individuals, not groups, accountable for their own transgressions. As a result, individual instances of violence and retribution do not escalate into feuding and warfare at the group level. However, even unsegmented societies fight wars in the face of serious ecological scarcity. 70
The pattern of kin ties also affects a group's ability to sustain and mobilize fighting alliances. For example, when sons reside in the same village as their fathers, local armed conflict is much more likely. However, the causal direction is in dispute: Some scholars say that the institution of patrilocal residence facilitates the formation of brotherly aggressor forces who pick fights with neighboring villagers, whereas others say that in rough neighborhoods, parents keep their sons nearby for enhanced protection. 71
Conversely, when sons live in their wives' villages, the chance of success in external war (if not necessarily its frequency) is enhanced. The reason is that a wider net of social ties allows the mobilization of a larger coalition against external foes. [End Page 24] For example, the brideprice and residence customs of the Central African Nuer tribe created strong social debts and kinship ties between individuals in distant villages, which could be tapped to mobilize Nuer alliances to outnumber the hapless neighboring Dinka in any dispute. Dinka social institutions lacked these extensive networks, so for the Dinka collective action in wartime was harder to accomplish. As a result, the Dinka, a people otherwise at roughly the same level of technology and resources as the Nuer, were gradually pushed back by the inexorable Nuer depredations. This is an institutional explanation in the sense that it rests on social relationships forged by conventional patterns of behavior around which expectations of diffuse reciprocity converge. However, unless one takes the functionalist view that Nuer brideprice habits were somehow caused by the need to mobilize allies for war, this explanation rests also on the fortuitous normative (which is to say cultural) conviction that relations with in-laws ought to be managed through the prevailing pattern of feasting and uxorilocal residence. 72
Like theorists of interdependence and institutionalized cooperation in the field of international relations, many anthropologists have argued that the institutionalization of social or economic ties to out-groups is likely to mitigate the chance of violent conflict with them. Some argue that this is accomplished through marriage to potential enemies. As Claude Lévi-Strauss put it, "Mankind has understood very early that, in order to free itself from the wild struggle for existence, it was confronted with the very simple choice of 'either marrying-out or being killed-out.' " 73 However, some statistical findings suggest that exogamy is a double-edged sword: It creates bonds of reciprocity, but it also gives people in neighboring villages issues to fight about, such as the nonpayment of brideprice obligations. Some studies suggest that economic exchange unlinked to kinship obligations is a better way of forging reciprocal ties that mitigate conflict, at least with respect to "internal" wars between units that share a common culture. 74 Others argue that exploitative marriage and trade ties increase war, whereas fair exchanges reduce it. 75 What constitutes a fair exchange is likely to be determined by cultural standards.
Ethnographers point out that exchanging goods in order to institutionalize relations with potentially dangerous out-groups typically involves exchanging goods of mainly symbolic value, as in the ritualized "kula ring" exchanges of the Trobriand Islanders. 76 Since symbolic exchange is a sacred signal of an intention to enter into a norm-governed social relationship, exchanging symbolic goods may be more effective than the exchange of utilitarian goods as a means to cement relations in anarchy. Or to put this in rationalistic terminology, pragmatic exchanges based on [End Page 25] simultaneous reciprocity reveal less about the other's character type than do exchanges based on diffuse or delayed reciprocity. Once again, institutionalized patterns of behavior interact with cultural understandings to affect the probability of war. 77
Cultural Hypotheses
Many anthropologists look to a culture's ideas, myths, and symbols to understand a society and the behavior of those within it. Cultural symbolism constitutes the repertoires of action people can imagine, the meaning they ascribe to actions, and the values and desires they hold. As Geertz says, "not only ideas, but emotions too, are cultural artifacts." 78
From one vantage point, a people's cultural symbols and ideas can be viewed as problem-solving tools that help them get what they need and in turn shape what they want. 79 Ideas and symbols play a crucial role in helping people maintain group solidarity, coordinate behavior with others, bargain effectively, manage conflicts of interest, and predict the consequences of their actions. Causal theories, rules of thumb, conventions, norms, and myths all play a role in these essential tasks. While some of this problem-solving mental life may follow familiar rules of scientific inference through processes like trial and error, much of it is conventional, ritual, arbitrary, and symbolic. 80 Such symbolism is valuable not only to provide people with reassuring notions about unknowable aspects of life, death, and cosmology but also as a set of costly, hard-to-fake signals of commitment to the group and its norms. Even materially poor societies commonly put a huge effort into conventionalized displays of solidarity, morality, role-initiation, hospitality, status, or intimidation, taking pains to ensure that myth-supported rituals are carried out with no shirking and at a high level of transparently emotional commitment. 81 One reason this works is that people internalize symbolic messages and are socialized to want what the culture prescribes.
The nature of the group's symbolic life is sure to be reflected in the group's war behavior, and vice versa. Ferguson remarks that "war is a virtual magico-religious magnet," because it "involves, in extreme form, virtually all of the circumstances which have been invoked to explain religion and magic," such as collective action, social control, group survival, and hazardous unknowns, not to mention "tragedy, injustice, and immorality" and therefore questions of ultimate meaning. 82 Warfare occupies the terrain of cultural symbolism. However, this very fact creates a problem in testing hypotheses about the effect of the content of a culture's [End Page 26] symbolism on its war behavior. Insofar as situations of personal and group insecurity heighten and shape the symbolic character of politics, it could be the experience of war that calls forth certain kinds of myths and values rather than the other way around. Therefore, a good methodological rule is to look for connections not only between war behavior and overtly warlike symbols but also between war and other aspects of the groups' cultural life not directly functionally related to war.
Those offering cultural explanations for war in anarchy face two tasks: to explain why war has been so common in anarchy and to explain variations in war's frequency or intensity. They provide some telling insights into the causes of variation, but they find explaining war's near-ubiquity more challenging.
Explaining the pervasiveness of warlike cultures. Proponents of the view that behavior in international relations is determined by culture, ideas, and identity have been inclined to fall back on psychological explanations such as social identity theory to account for the pervasiveness of violent rivalry between in-groups and out-groups in anarchy. This is a theoretically unsatisfying move because it leaves culture little to explain in its own right: Once it is posited that the need for self-esteem leads directly to in-group favoritism and out-group denigration, the development of rivalrous collective identities and cultures seems pre-ordained. Moreover, when proponents of this view make the move to argue that individualist cultures and democracies are less prone to this bias than are collectivist cultures, it is hard to see how individualist cultures could arise in the first place if social identity theory is so universal and powerful. 83
A second line of argument fits better with the underlying premises of theories that highlight cultural identity. In this view, groups in anarchy fight so frequently because maintaining the identity of each leads to or even requires the enmity of the other. 84 Group solidarity depends on costly investments in an elaborate set of practices that police boundaries, sanction shirkers, and motivate contributors to collective action. These practices are backed up by a host of cultural symbols, including religious myths and ethical norms, that underwrite the kind of behavior and attitudes that make the group cohere and set it apart from other groups. "Durkheim taught that in religious worship society adores its own camouflaged image," says Ernest Gellner. "In a nationalist age, societies worship themselves brazenly and openly." 85 These symbolic systems that explicitly or implicitly extol the in-group chafe on contact with rival cultures, say Peter Berger and Thomas Luckmann: "The appearance of an alternative symbolic universe poses a threat because its very existence demonstrates that one's own universe is less than inevitable." 86 Fredrik Barth reports that New Guineans experience "shock and fear" [End Page 27] on learning that nearby tribes' sacred rituals sometimes invert the symbolism of their own. 87
Even so, this argument fits awkwardly with the anthropological finding that "internal" war among units of the same culture is more frequent than war between societies with different cultures. For example, although Amazonian and Melanesian villages do engage in out-group denigration and in-group chauvinism, ascribing this to cultural differences is difficult, since neighboring villages normally share the same myths, symbols, and ethical beliefs. Thus a cultural reductionist would need to argue that groups maintain a common culture of symbiotic out-group rivalry in order to sustain the solidarity of the separate village units.
This is exactly what anthropologist Simon Harrison concludes from his fieldwork on Melanesian warfare. He notes that villages are highly problematic as political units in the part of New Guinea he studied, because clans whose membership cuts across village boundaries have a strong functional basis as trading networks and a strong cultural basis in kinship norms. Thus clan loyalties vie with village loyalties in the organization of fighting alignments and economic relationships. This is particularly troubling for men, says Harrison, because women play a key role in kin-based trading relationships, which they can use to pursue their independent economic objectives. 88 Consequently men have an interest in reinforcing the priority of village-based loyalties and in policing the territorial boundaries of the village. They do this, he argues, by forming non-kin-based secret societies of the village's mature men, who carry out deadly but highly ritualized warfare in the event of encroachments or offenses by members of other villages. 89 This ritualized killing, which entails the only cannibalism permitted in a culture that otherwise abhors it, implicates male members of all the clans in the village in a collective transgression that reinforces village unity, in Harrison's view. 90
Harrison argues that his interpretation turns Hobbes on his head: "In Melanesia it is not so much groups that make war, but war that makes groups. That is to say, war is part of the way in which groups having claims or interests in resources assert their existence in the first place. It is through conflict that these groups separate themselves out from each other and constitute themselves as distinct entities capable of competing for resources." 91 Moreover, he argues, the war-making process accomplishes these ends by its symbolic effects, not just by its strategic consequences: "What they were trying to do in war and ritual was not simply use power against outsiders and act against them at the level of force, but also to act ideologically against the perpetual moral accountability to outsiders which implicitly threatened the idea of the community as a political entity." 92 [End Page 28]
Harrison's interpretation is impressive, and its basic line of argument might be highly generalizable once particulars are adjusted to take into account different social settings. However, its theoretical implications are unclear. Instead of turning Hobbesian or materialist-functionalist explanations on their heads, it might be seen as reinforcing them: In the self-help situation of anarchy, where warfare is an ever-present possibility, a self-interested group having a comparative advantage in the use of violence (in this case, co-residential men) fights other groups (in this case, other villages of the same culture, other villages of a different culture, and nonterritorial kin-groups) to establish security, economic advantages, and prestige defined in terms of power. The fact that this is accomplished with a heavy overlay of highly dramatic symbolism does little to demonstrate that cultural rather than material-strategic elements should take precedence in the causal hierarchy. When viewed in terms of my methodological rule for establishing the independence of cultural effects, nothing in Harrison's story shows that the outcome was driven by more general cultural patterns that were prior to or functionally unrelated to warfare.
In short, attempts to use culture to explain why war has been so prevalent in anarchy run into serious theoretical and empirical problems. Given that many cultural systems feature warlike symbolism, this seems more readily explained as a by-product of more basic causes: either the inherent insecurity of life in anarchy, as the realists have argued, or perhaps some innate psychological bias in the way in-groups perceive out-groups, independent of culture.
Cultural explanations of variations in war making. Culture is better suited to explaining variations in the frequency of war in anarchy than to explaining its overall prevalence. Whether culture is conceived as a symbolic system or as a set of norm-laden practices, anthropologists find that the frequency and intensity of war correlate with a generalized culture of mistrust, fear, and harshness in social relationships. Ember and Ember find, for example, that fear of nature and fear of others are among the strongest correlates of war. 93 Socialization to distrust seems to be more deeply engrained than socialization to aggression itself. When war-prone societies are pacified by outside powers, aggressive attitudes dissipate, but attitudes of generalized mistrust tend to remain. 94 Put differently, aggression seems to be more situational, mistrust more dispositional or cultural.
On the plane of ideas and symbols, myths in warlike societies often feature tales of violence, threat, and fearful imagery. For example, Yanomamo mythology centers on fierce jaguars with quasi-human traits and on the putatively historical self-destruction of the community through an out-of-control episode of murder and rape. 95 Yanomamo villages in the most conflict-prone regions believe that their [End Page 29] people originated from blood falling from the wounded moon, but Yanomamo in less conflictual areas lack that myth. 96
Moreover, such cosmological myths may be closely tied to norms governing everyday behavior. Those Yanomamo areas with sanguinary myths are also at the epicenter of the waiteri cult of fierceness, which motivates and regulates social practices like ritualized tests of strength and aggressiveness at feasts, governs status relations within the village through the distinction between men who have killed and those who have not, determines opportunities to marry, justifies male-female violence within the village and raiding of women from other villages, motivates men to participate in warfare, and destabilizes alliance politics. In turn, cultural expectations of armed rivalry and glorification of fighters affect the material circumstances of conflict through female infanticide and the consequent incentive to obtain mates by raiding.
In a rigorous multiple-methods study of the relationship between war and culture in anarchy, Ross shows that the overall level of conflict, including warfare, is best predicted by the harshness of a society's child-rearing practices, the socialization of children to generalized mistrust of others, and the level of male gender-identity conflict. This last variable manifests in such forms as aloof relations with fathers, severe rites of passage to manhood, and male ambivalence in relationships with women, exemplified by New Guineans' extreme menstruation taboos and separate housing for men and women. 97 (Note that Ross's study adopts a broad definition of culture, which includes not just symbols and ideas but also institutionalized practices reflecting those ideas.) Social structure, including kinship patterns, affects whom people will fight, but psychocultural variables determine how much they fight, Ross says.
However, some of Ross's statistical correlations may be consistent with a reverse causal interpretation: War may shape culture rather than the other way around. Societies that face intense security threats may have harsh coming-of-age rites in order to steel boys to face the dangers of their tough neighborhoods. Similarly, war may explain New Guinea highlanders' ambivalence toward their wives, who usually come from rival clans. As they put it, "We marry the people that we fight," and therefore wives are considered a potential fifth column. 98 Moreover, Ross does not statistically control for ecological scarcity, though in narrative discussions he acknowledges this factor as an important cause of war.
Ferguson's case-study research similarly observes that war-prone societies often have harsh, fearful, vindictive cultural practices, but he is more skeptical than Ross about the causal direction. Revenge rationales, he argues, are highly malleable. Groups and individuals will take revenge on those who are weaker but will let bygones be bygones with respect to stronger transgressors. Likewise, he argues, allegations of witchcraft are not causes of conflict but rather are a mode of [End Page 30] expressing a conflict that arises for some other, more pragmatic reason. 99 However, even if pragmatic factors are important, why does hostility take the form of witchcraft, and what independent effects do these beliefs and practices have in increasing or mitigating armed violence?
The few societies that experience little or no war tend to have benign patterns of social relations among group members, but not always. At the benign end of the continuum, the peaceful Mbuti pygmies have dense, mutually supportive social networks, collective child-care, rituals that stress nurturing relations with people and nature, anticompetitive norms, high levels of in-group trust, no hierarchy, no wealth, and forest homes that provide them something of a haven from their Bantu neighbors. They do have disagreements, but the effects of these are contained. 100 In contrast, some other societies take very different routes to peace. The peaceful Cayapa hunters, who live in dispersed jungle households in highland Ecuador and Colombia, are the biggest outliers in Ross's data set. They have harsh, capricious childhood socialization practices, and adults are highly suspicious of one another--so much so that they avoid everyone who is not kin. 101 This is reminiscent of Rousseau's prescription for overcoming the danger of war in an anarchical system of states: Remain isolated and materially poor so that nobody bothers to come calling. 102 In contrast, the tribes of the Upper Xingu region of Brazil maintain peace through a myth-reinforced system of mutual deterrence under high levels of social and economic interdependence. They fear people from neighboring villages, including their witchcraft, and therefore prudently avoid showing anger toward them. As part of a self-conscious strategy of controlling impulses to aggression, they avoid eating spicy food or red meat, which they fear will make them choleric and bloody-minded. Myths drive home the dangers that lie in store for people who lose self-control and become like the wild, animalistic, aggressive tribes who live in the nearby hinterlands. 103
In short, the practice of war in anarchy is intimately bound up with cultural symbols. This is not surprising, because some of the main roles of culture--to strengthen group solidarity, to mark off group boundaries, and to motivate contributions to collective action--are especially salient in time of war. However, the direction of causality running between war and culture is complex and hard to establish and in many cases may be reciprocal. In most circumstances, understanding strategic choices in anarchy requires knowing both the situation actors face and the cultural equipment they use to devise and discuss responses to the dangers of anarchy. [End Page 31]
Theoretically Integrating Material, Institutional, and Cultural Elements
Anthropologists of war deal with these problems of causal complexity in three ways. Some array material, institutional, and cultural factors in a hierarchy of nested constraints, an approach that enriches monocausal reductionism but shares many of its limitations. A second approach treats material, institutional, and cultural factors as fully independent variables and studies the effects of their interactions. A third approach views anarchies as historically developed systems of action constituted jointly by material, institutional, and cultural factors. I assess each of these approaches in the following sections.
Nested Constraints
Emile Durkheim believed that "social life, in all its aspects and in every period of its history, is made possible only by a vast symbolism." 104 However, he also believed that the form these symbols take depends in part on aspects of the material setting, reflected in the division of labor in society, and that the "collective consciousness" of a group mirrors its social relationships. Thus, while Durkheim hardly wanted to reduce social symbolism in a crude way to brute material facts, he nonetheless nested the symbolic in the material and institutional in ways that gave causal priority to the latter. 105 Many significant social theorists from Marx to Malinowski to Gellner have taken this path. Some prominent students of the anthropology of war have likewise chosen to nest material, institutional, and cultural elements in roughly this way.
Ferguson, for example, depicts the causes of war as a "nested hierarchy of progressively more limiting constraints." 106 The broadest constraints are set by what he calls infrastructure--that is, demography, ecology, and technology. To the list of infrastructural characteristics, we might add the constraints that follow from the fact of anarchy. These factors create a set of propensities that may directly affect the likelihood of war and also indirectly affect the chance of war by shaping constraints at other levels. The next broadest level of constraints for Ferguson is the structural--that is, the structure of the economy, the pattern of kinship, and political institutions. These factors may likewise load the dice for or against war directly, or they may do so indirectly by affecting what Ferguson calls the superstructure--that is, psychology and beliefs. In this scheme, the likely configuration at each nested layer is bounded but not strictly determined by influences from the prior one.
Ferguson's decision to start with the material is not arbitrary. It is based on his empirically grounded claims that material constraints have causal priority. He admits that different cultures may respond somewhat differently to similar material [End Page 32] challenges, such as contact with militarily and economically superior societies, depending in part on their cultural repertoire of concepts and tactics. Yet these variations may simply reflect rationally defensible differences in preferences over strategy rather than arbitrary cultural biases. Even within the same group, Ferguson notes, different leaders may urge different solutions: fight, flight, assimilation, conciliation, or "fundamentalist" revival. Moreover, Ferguson and Neil Whitehead argue, "in all these cases, existing cultural patterns are reshaped and employed practically, and in ways that show substantial cross-cultural uniformity." 107 Thus, while the material does not fully subsume the cultural, according to Ferguson's view, it is possible to understand the role that culture plays only by taking into account the material parameters that structure the situation to which the culture is responding.
In contrast, proponents of cultural theories of warfare, such as Harrison, attempt to reverse the explanatory nesting, starting with culture and showing how it constitutes patterns of social relations and even material arrangements. 108 In principle, rigorous testing may in a given case be able to establish which nesting sequence fits the evidence best. This analytical strategy will break down, however, whenever there is reciprocal causality among material conditions, social institutions, and culture. In analyzing Yanomamo warfare, for example, it may make little sense to ask whether its intensity is due to the strategic incentives of anarchy or to the waiteri cultural complex of fierceness or to bloody myths of origins or to kinship institutions that encourage the fissioning of villages, since these are all co-constituting and mutually reinforcing. Even if we accept Ferguson's argument that competition for manufactured trade goods triggers an intensification of the normal level of conflict, this external "material" factor produces its distinctively violent consequences only because of its interaction with the existing material-situational-institutional-cultural system. In such circumstances, nesting theories must give way to theories based on interaction effects and/or a systems approach.
Interaction Effects of Independent Variables
Much quantitative and comparative anthropology treats material, institutional, and cultural factors as fully independent causal variables, which interact to shape outcomes. Sophisticated research designs do not simply add up the separate effects of these variables; rather, they study how the presence or absence of one variable may alter or even reverse the effect of another. 109 This method can sometimes be convenient and powerful, but I have shown that studies using this approach may be blind to the effects of feedback processes, reciprocal causality, and the interaction [End Page 33] of interdependent strategic choices. Such studies assume the causal independence of variables that in a complex system may in fact be exerting reciprocal causal effects on each other. For example, studies of this type conclude that fraternal co-residence and harsh coming-of-age rituals increase the incidence of local warfare, but in fact war may also be the cause of these correlations, not simply a consequence. The deep functional interdependence of war, material capability, institutions, and culture suggests that feedback effects are rife in this arena. If so, the convenience of simple interaction models may sometimes be a false friend; systems models may capture causal relationships in a more valid way.
Material-Institutional-Cultural Systems of Action
In this view of the social construction of reality in anarchy, it is not just the material setting or the institutional setting or the cultural setting that confronts the individual and the group as a taken-for-granted inevitability; it is the whole of the multifaceted anarchical system. When acting in or analyzing a system, it is insufficient to think in terms of the effects of separate variables, because feedback processes often counteract expected consequences. 110 Because of the complex causal interconnections among their elements, systems have emergent properties that differ from the independent properties of their parts.
No single part of a system suffices to define the system and its behavior. Even anarchy itself, though it may load the dice probabilistically in favor of war, does not predetermine action in the system. Nor does any specific institutional fact or cultural feature determine the system's behavior independent of the larger context of the system as a whole. Indeed, the effect of each element can be understood only in the context of the rest of the system. For example, witchcraft beliefs may deter aggression in some self-help systems and intensify it in others. Likewise, the ecological fact of food scarcity can in itself produce war, trade, or state building, depending on the other potentialities of the social system and choices that its actors make. 111 Without knowing how any particular system operates as a whole, changing one variable (if that were possible) might not lead to the expected results: Reducing scarcity might simply make the units in a warlike system more capable of sustaining the fight; making units more trusting in a warlike system might get them wiped out.
Some anthropologists have conceived systems effects as creating a homeostatic equilibrium in which the culture of warfare plays a regulatory role. For example, Roy Rappaport argued prominently that in the absence of central authority, traditional rules surrounding ritualized warfare in New Guinea sustained an ecological equilibrium ratio of land, people, and pigs. The "regulatory function" of war and associated feasting, alliance, and land redistribution practices, he claimed, "helps to maintain an undegraded environment, limits fighting to frequencies that do not [End Page 34] endanger the existence of the regional population, adjusts man-land ratios, facilitates trade, distributes local surpluses of pig in the form of pork throughout the regional population, and assures people of high-quality protein when they most need it." 112
However, even former proponents have recanted this type of hyperfunctionalist equilibrium theory. While taking land and distributing pork to allies is a consequence of some fighting, this does not appear to correlate closely enough with the population density of people or pigs to explain the frequency and location of warfare. Norms typically forbid the redistribution of land as a result of fighting, so it is difficult to argue that norms and behavior exist in a seamlessly self-reinforcing equilibrium. 113 Some New Guinea peoples with the strongest norms regulating violence are among the most warlike, because their norms require endless rounds of retaliation. 114 The latter may constitute a sort of equilibrium, but it does not appear to be a functional one.
Although functionalist equilibrium theories of war have not stood up well in the face of evidence, other kinds of system theories may fare better in capturing the reciprocal interaction of material, institutional, strategic, and cultural factors in Melanesian warfare. Such theories conceive the system not as closed and homeostatic but as open, historically path-dependent, and strategic. Bruce Knauft, for example, portrays the evolution of New Guinea warfare in terms of mutually interacting ecological, institutional, cultural, and strategic developments, characterized by both positive and negative feedbacks. In this account, the causal chain starts with increasing population density, which leads to increasing forest clearage, declining wild food sources, increasing pig-raising, introduction of the sweet potato to feed pigs, and further population and pig increases. These in turn create increased opportunities for conflict over land and pigs, the substitution of wealth for people in social exchanges, including homicide compensation and bridewealth payments, and the increasing role of authoritative political leaders ("big-men") in the mediation of disputes, the coordination of military and diplomatic strategy, and the regional integration of economic exchange. 115 Although rising population density plays a role in the causal process leading to increased war, there is no simple correlation between them because of the varying adjustments in the complex social ecology that may emerge in different localities.
In this model, the ecological competition for scarce resources intensifies a second-order competition for security in the absence of a central authority. However, system outputs cannot be simply read off from these material and situational parameters, because cultural and political tools are created or adapted to manage the resulting mixed-motive situation, which includes incentives to use force to gain unilateral advantages as well as incentives to cooperate in a division of labor and to limit the damage of unbridled conflict. Reflecting these mixed motives, communities invoke rules against taking land in a fight and fighting within a blood-related clan, [End Page 35] yet they also invoke rules requiring retaliation for uncompensated offenses. These rules may be at odds with each other, and applying them requires interpretation. This creates room for calculating which interpretation is strategically advantageous for the group. "Big-men" emerge as the strategic optimizers and normative interpreters whose calculations steer behavior at the micro-level, weighing the balance of forces, managing the risk of two-front wars, and guarding the group's normative reputation. As Mervyn Meggitt puts it, "percipient Mae men are aware that the identity of one's current enemies and allies is largely dependent on the interaction of two kinds of variables: (1) the moral constraints ascribed to the norms of kinship and of exchange, which themselves are interpreted variably in light of pragmatic considerations, and (2) the material stimuli of population pressures and land resources, which of course are never wholly stable." 116
Thus, these decision makers weigh complex-system effects that simultaneously encompass material, institutional, and cultural elements. Although feedback processes are at work, they do not produce a stable equilibrium in the manner theorized by Rappaport. The system comprises some elements of positive feedback, such as the self-amplifying dynamic promoting the intensification of the pig economy, and also some elements of negative feedback, such as the role of retaliatory norms. Overall, however, the effects of feedback and other causal interactions are more complex and cannot be characterized in terms of a simple equilibrium. Rather, a host of indeterminate, exogenous, and arbitrary elements keep the system in flux: the uneven growth of political units; uncertainties about the best strategy; the unpredictability of the application of norms; the impetuosity of individuals; and exogenous shocks from nature or from outsiders, especially colonizers or the postcolonial state. Although the system may have no stable equilibrium, we must take into account system effects involving interactions between material-situational and institutional-cultural factors in order to understand the trajectory of change and the calculations of the actors in the system. 117
The Contemporary International System and the Democratic Peace
What does anthropological research on culture and anarchy suggest about ways to understand the contemporary international system? The most common typologies of international systems in the existing political science literature tend to focus on one of the system's dimensions--the material, the institutional, or the cultural. Focusing on material power distributions, Waltz proposes bipolar, multipolar, and now unipolar types of systems; focusing on ideas, culture, and identity, Wendt proposes Hobbesian, Lockean, and Kantian types; focusing on political institutions, Bruce [End Page 36] Russett contends that democratic and authoritarian anarchies behave differently. 118 In a nesting approach to integrating material, institutional, and cultural factors, such typologies organized around a supposed taproot factor might be adequate. However, if anarchical systems are jointly constituted by all three of these kinds of elements and are not reducible to any one of them, a useful typology cannot be based on the presupposition of linear, nested, causal relations among the elements.
Indeed, it seems most plausible that the contemporary international system is shaped by the interaction of elements of all three kinds, not simply by one of them and not simply by adding their separate effects. The incidence, types, and outcomes of war in the current system are surely influenced both by the system's unipolarity and by the fact that the hegemonic power and most of the secondary powers are well-institutionalized, liberal democracies. If John Ikenberry's argument about the bargaining effectiveness of a democratic hegemon is correct, democracy may stabilize and institutionalize the unipolar distribution of power. 119 Insofar as the hegemon uses its power to promote democracy and respect for human rights, the unipolar distribution of power may in turn affect the culture of the system, though perhaps not in quite so linear a fashion as human rights advocates would hope. This may influence international behavior in the system both by increasing the number of democracies and by shaping the type of democracies that emerge. Insofar as the democratic peace rests in part on mutual perceptions of a common basis of legitimacy, it matters whether the democratic hegemon succeeds in promoting the spread of liberal, inclusionary, civic democracies--rather than ethnically exclusionary democracies, which might wind up fighting over the mistreatment of each others' minorities. 120 In all these ways, assessing the future of the international system requires thinking in terms of mutual feedbacks among material, institutional, and cultural elements.
Even if we set aside the effects of unipolarity, the basic workings of the democratic peace rely on supports in the material, institutional, and cultural domains. The absence of war between mature democracies depends on the material motivation of the average member of society to avoid needless death and impoverishment (goals widely if not universally shared across cultures); political institutions that predictably empower the median voter; and a set of cultural symbols sanctifying civil rights, free speech, and electoral legitimacy in ways that underpin those institutions, facilitate peaceful bargaining, and establish a nonthreatening, "in-group" identity among democratic states. The democratic peace works best when these material, institutional, and cultural elements are all in place. 121 [End Page 37]
Moreover, democracy itself has material preconditions. Adam Przeworski finds that transitions to democracy in countries with per capita annual incomes above $6,000 (in 1985 constant dollars) are almost always successfully consolidated, whereas when the amount is below $1,000, democratic transitions almost always suffer reversals (with very few exceptions, such as India). 122 These economic levels may to some degree be proxies for closely related factors such as literacy and the development of a middle class. Between those levels, consolidation seems to depend on a number of institutional preconditions, such as the strength of the rule of law and the development of civil society organizations. 123 The fact that these material and institutional preconditions often arise along with symbols and ideas supportive of democracy does not mean that democratic culture can somehow substitute for them. Not surprisingly, Western jawboning in favor of free speech, fair elections, and human rights has borne little fruit in countries that lack these preconditions. 124 There is no cultural shortcut to a global democratic peace.
The functioning of the democratic peace system as a whole cannot simply be reduced to the properties of its constituent parts, namely, democracy, free trade, and multilateral organization. Before World War I, democracies were rather likely to get into militarized disputes with each other. 125 In that era, moreover, economic interdependence per se did little or nothing to reduce the risk of war between states; indeed, mutual economic vulnerability may have increased the chance of war. During the 1930s, liberal international organizations did nothing to promote peace. But put all three of these ingredients together in a system in which these elements predominate, and the democratic peace emerges. 126
Even so, it is premature to characterize the democratic peace as having reached a stable equilibrium, "the end of history." China, Russia, Iran, Indonesia, Egypt, and Nigeria are potentially powerful and consequential states that have yet to consolidate democracy. Despite (or perhaps even because) of the attempts of liberal powers to promote democracy, free trade, and human rights, states undergoing democratic and market transitions may be at greater risk for ethnic conflict and international aggression, especially if they are nationalistic great powers that resent external pressure, resist liberal traditions, and lack rational-legal state institutions. Today's unipolar democratic peace embodies nonlinear feedback effects across its material, institutional, and cultural elements, but it would be premature to say that it has settled down into a stable equilibrium. In such a system, where change is a certainty but some trajectories of change may not be for the better, systems effects must be taken into account in judging the likely consequences of efforts to promote transformation. [End Page 38]
Acting to effect change in a system. Scholars and activists who claim that changing the ideas, norms, and culture of an anarchical system can transform it are not wrong; they are just oversimplifying. An anthropological anecdote is instructive. Carole Robarchek and Clayton Robarchek, anthropologists who have studied the Waorani of the upper Amazon, the most murderous society on Earth, tell the inspiring tale of a pair of Waorani women who returned to their tribe in 1956 after a long sojourn among the Waorani's foes. They bear the news that these neighbors are not, as had been believed, a pack of untrustworthy cannibals. This information and interpretations triggered a reassessment of the Hobbesian assumptions underpinning the Waorani worldview. As a result, Robarchek and Robarchek claim, the Waorani began behaving more peacefully, not only toward the outsiders but also within the anarchical Waorani society itself. 127 Despite the nightmarish weight of history pressing upon this Hobbesian culture, a creative and resolute vanguard was able to use information and ideas to exploit openings for the possibility of change.
But a vanguard does not make history as it pleases. It makes the history that historical possibilities enable. Among the Waorani, the window for change was opened by missionaries who accompanied the returning tribeswomen and by the availability of trade goods that made a change in behavior seem materially attractive. Cultural entrepreneurs of political ideas must not succumb to idealistic voluntarism; they need to understand the material and institutional potentials of the system they are trying to change.
If those promoting cultural change do not take into account systemic interconnections, their efforts may not only fail but also be harmful. As Jervis reminds us, in a complex system the shortest distance between two points is not necessarily a straight line. For example, he says, "people jump from the premise that we are seeking a color-blind society to the conclusion that government policies should be color blind," but in a world where "many people and groups in fact do respond to others in terms of skin color, compensatory color-aware policies . . . may bring . . . us closer to a society in which color plays a much smaller role." 128
Similarly, proponents of international cultural transformation are often too direct and nonstrategic in their approach to achieving their objectives. For example, Thomas Carothers notes that the typical "strategy" of democracy-assistance efforts is to generate a checklist of the attributes of mature democracies and then mount parallel programs to try to install each of them in the targeted country right away. 129 This approach is flawed because it pays insufficient attention to interaction effects between these efforts, issues of sequencing and preconditions, strategic responses from resistant actors, and other negative feedback effects. 130
Jawboning for a democratic transition and accountability to international human rights norms can be risky in settings that lack even the rudiments of the rule of law [End Page 39] or the material resources needed to sustain an independent civil society and media. In such settings, transformative projects may unleash a populist form of mass politics at an inopportune, premature moment, when elites threatened by such changes can exploit social turmoil by playing the ethnic card. Thomas Risse and Kathryn Sikkink acknowledge that there may be a backlash phase in response to efforts to press dictators to accept human rights norms, but they portray this as a transitory delaying tactic. 131 In fact, coercive persuasion that skips over the needed material and institutional preparatory steps may trigger costly setbacks that could have long-term consequences. Untimely voting demanded by the international community in such places as Burundi in 1993 and East Timor in 1999 has led directly to hundreds of thousands of deaths and refugees. 132 Arguably, it has also deepened ethnic and social cleavages and tainted democratic remedies.
The risk of such backlashes may be increasing. The "third wave" of democratization, already eroding in some places, consolidated democratic regimes mainly in the richer countries of Eastern Europe, Latin America, southern Africa, and East Asia. 133 A fourth wave would have to take on harder cases: countries that are poorer, more ethnically divided, and starting from a weaker base of governmental institutions and citizen skills. In facing this challenge, culturally creative activists and thinkers may play an important role in changing behavior in the international system, but they must do so within a context that is structured by the system's material and institutional possibilities.
Conclusion:
The Potential for Transforming Contemporary Anarchy Through
Culture
Anthropological research on culture and war casts doubt on one-dimensional conceptions of behavior in anarchy. The radical constructivist notion that anarchy is nothing more than ideas, culture, and identity "almost all the way down" is just as misleading as the hyper-realist notion that the unchangeable situation of anarchy will always make life nasty, brutish, and short, regardless of anarchy's institutional and cultural content. Those who seek to transform the culture of contemporary anarchy need to work within an existing material and institutional setting that may enable, derail, or pervert efforts to promote change. Efforts to force the pace of change risk unintended consequences that could wind up hindering change and increasing its costs.
If world politics is seen as composed of nothing but ideas and culture, and if culture is malleable and socially constructed, then the barriers to realizing our ideals may seem insubstantial. Sophisticated international relations scholars like Wendt deny that they are pure voluntarists or subjectivists who think that ideas can [End Page 40] transform the world by a wave from the magic idea-wand. Indeed, his social theory of international politics can be read in part as an account of the ways that ideas and identities become locked into a cultural structure that is hard to change. 134 Nonetheless, Wendt's voluntarism threatens to overwhelm his structuralism. He asserts that "anarchy as such is an empty vessel and has no intrinsic logic; anarchies only acquire logics as a function of the structure of what we put inside them." 135 "Although there is no 1:1 correspondence between positions in the idealism-materialism debate and beliefs about the ease of social change," he says, "showing that seemingly material conditions are actually a function of how actors think about them opens up possibilities for intervention that would otherwise be obscured." 136
Martha Finnemore and Kathryn Sikkink state the transformative moral goal of theory even more directly: "For decades now, [international relations] research has been divorced from political theory on the grounds (implicitly, if not explicitly, articulated) that what 'is' in the world and what 'ought to be' are very different and must be kept separate, both intellectually and in policy. However, contemporary empirical research on norms is aimed precisely at showing how the 'ought' becomes the 'is.' " 137
Understanding the relationship between "is" and "ought" should indeed be
a central task of contemporary international relations scholarship. While
undertaking that task, however, it is important that we avoid taking
a one-dimensional, voluntarist view of behavior in anarchy that could
foster the kinds of outcomes that principled scholars and serious-minded
practitioners want to avoid. There is no way to change the world other
than working within it.
Jack Snyder is Robert and Renée Belfer Professor of International Relations in the Department of Political Science and the Institute of War and Peace Studies, Columbia University, New York, New York. He can be reached at jls6@columbia.edu.
Notes
I am grateful to Jacqueline Berman, Carol Ember, Melvin Ember, Brian Ferguson, Stacie Goddard, Peter Gries, Ted Hopf, Robert Jervis, Benjamin Judkins, Elizabeth Kier, Ronald Krebs, Paul MacDonald, Daniel Nexon, Marc Ross, Bruce Russett, Leslie Vinjamuri, Kenneth Waltz, and Alexander Wendt, and to participants in seminars at MIT, Ohio State University, Rutgers-Newark University, and Yale University for extremely useful comments and suggestions.
1. See Wendt 1999; and Finnemore and Sikkink 1998.
2. Wendt 1992. All sides of this debate define anarchy as a situation in which no overarching authority has enforcement power over politically autonomous actors.
3. See Finnemore and Sikkink 1998, 916; Risse and Sikkink 1999; Wendt 1999, 141, 377-78; and Ruggie 1998, 199 et passim. These scholars adhere to the constructivist approach in the study of international politics, but not all constructivists are so clearly wedded to this transformative political agenda. For more qualified views, see Katzenstein 1996, 536-37; and Owen 1997, 232-35.
4. Human Rights Watch 2000.
5. See Mearsheimer 2001 and 1994/95; and Posen 1993.
6. Layne and Schwartz 1999. For a more positive assessment of the role of military power in support of a constructivist agenda of global transformation, see Ruggie 1998, 240-55.
7. Wendt 1999, 90 and chap. 3; also underplaying material constraints are Risse and Sikkink 1999, 35-36.
8. Cox 1986, 208.
9. See Posen 1993; and works on ethnic partition cited by Byman and Van Evera 1998, 49-50, including the somewhat more constructed yet largely realist view of the role of culture in ethnic conflict presented by Kaufmann 1996. For an intermediate view, see Keohane 2000.
10. Jervis 1997.
11. In addition to the works already cited, see Koslowski and Kratochwil 1995.
12. See Philpott 2001; and Ruggie 1998, 147.
13. See Tilly 1990; and Spruyt 1994.
14. Johnston 1995. Reus-Smit compares different anarchical systems, but the variation in outcome that he seeks to explain is the style of diplomatic interaction, not the likelihood of war. Reus-Smit 1999. For a comparative analysis that stresses state building and strategic choices rather than culture, see Hui forthcoming.
15. For surveys of this literature, see Reyna and Downs 1994; Ferguson and Whitehead 1992b; and Haas 1990.
16. Wendt 1999, 323.
17. Robarchek and Robarchek 1998, 128, 137.
18. Ferguson 1994.
19. Kelly 2000.
20. See Sahlins 1976a, 74 (on Malinowski), and chap. 2, 5.
21. Ember and Ember 1997, 3. Other databases, such as Murdock's "World Ethnographic Sample," include as many as 565 units, but most studies of war use more select samples in which the units are adequately documented. See Tefft 1975, 699; and Keeley 1996, 28-29.
22. Malinowski 1937, 142, quoted in Carneiro 1994, 6.
23. Ember and Ember 1994, 190.
24. One-sided ambushes and organized feuding between kinship groups residing in different territorial units are sometimes included in the definition of war and as a practical matter are counted as war in the Standard Cross-Cultural Sample (Kelly 2000, 54). For theoretical and practical justifications of subsuming feuding under warfare, see Kelly 2000, 5-7; and Meggitt 1977, 38. For a definition that differentiates them, see Otterbein 1994, xix-xx.
25. Ember and Ember 1994, 190. See also Meggitt 1977, 108-10.
26. Waltz 1979, 67. Street gangs must pay attention to the likelihood of police intervention, so they are not strictly in an anarchical situation.
27. See Harrison 1993, 47; and Chagnon 1968, 109-59. Chagnon's professional ethics and the reliability of his data have recently come under severe criticism in Tierney 2000. One of Tierney's charges is that Chagnon's distribution of trade goods intensified the warfare among the villages he was studying, a point that was integrated into scholarly interpretations earlier by Ferguson 1995. I occasionally mention Chagnon's interpretations since they have occupied a significant place in the anthropological literature on war, but I do not rely on them to support my own arguments.
28. For the nuances of definitional debates about the concept, see Ortner 1999, 7-11.
29. Johnston's definition directly follows Geertz. Johnston 1995, 222-23. Wendt uses the comparable term "shared knowledge." Wendt 1999, 141.
30. Shweder 2000, 7-9.
31. Ross 1993, 21. He updates his definition of culture in Ross 1997.
32. Johnston 1995, 221-22.
33. See Powell and DiMaggio 1991; and Schotter 1981, 9. In contrast, Wendt calls institutions a "specific form" of culture. Wendt 1999, 141.
34. Harrison 1993, 126-27.
35. Garrett and Weingast 1993.
36. Human Rights Watch 2000.
37. See Johnston 1995, 35-39; and Katzenstein 1996, 65-68.
38. Ross codes "internal" and "external" based on the participants' subjective judgment of cultural similarity; "local" conflict is Ross's term for fighting within a territorial community. Ross 1993, 75, 81-89. See also Tefft 1975, 694.
39. Wendt 1999, 109.
40. For example, see Chagnon 1997, 218-21. Meggitt reports that New Guineans entering potentially hostile territory for the purpose of ceremonial exchanges bring
along numerous armed men for safety and a few women to signal benign intent. Meggitt 1977, 45.
41. See Jervis 1978; and Fearon 1995. On the centrality of fears of preemptive and preventive attack in New Guinea warfare, see Gordon and Meggitt 1985, 147; and Knauft 1999, 144.
42. Ferguson calls these the infrastructural, structural, and superstructural factors. Ferguson 1992, 201.
43. Jervis 1978.
44. Ferguson 1994, 85-111.
45. Ember and Ember 1997, 5. Kelly argues that Ember and Ember's liberal coding rules inflate these percentages; however, Otterbein applies a more restrictive definition to a somewhat different data set and reports findings comparable to those of Ember and Ember. See Kelly 2000, 123; and Otterbein 1994, 184. See also Tefft 1975, 701.
46. Mercer 1995, 229-52. For an argument that intergroup comparisons do not have to lead to conflict, see Brewer 2001. Although the psychology of self-esteem is offered as the underpinning for social identity theory, it is easy to imagine rational or functional considerations that might also lead people to make conflict-promoting distinctions between in-group members and out-groups. However, the experiments leading to the formulation of social identity theory were designed to eliminate rational and functional incentives. A rational-functional account of such biases would lead back toward material-environmental explanations of war in anarchy.
47. Wendt 1999, 90.
48. Ibid., 241-42, 322-23, 350-51.
49. See Meggitt 1977; Knauft 1999, 94, 135, 144, 154, 166; Keeley 1996, 5-8; and Sahlins 1968, 4-13.
50. Gordon and Meggitt 1985, 13.
51. Ember and Ember 1992.
52. Ember and Ember 1997, 8; see also Ferguson 1984; Sanday 1981, 174; and Carneiro 1990, 191.
53. See Meggitt 1977, 14, 34, 36, 183-84; Ember 1982; Vayda 1989; Gordon and Meggitt 1985; and Knauft 1999, 124-25.
54. Ember and Ember 1997, 7, 10-13.
55. Ferguson 1997, 336.
56. Ember and Ember 1994, 192-93.
57. See Cohen 1990; and Ember and Ember 1994, 193-95.
58. See Ferguson 1995, 42-43; and 1992, 219-20.
59. Ross 1993, 94, 142-43; see also Ferguson 1994, 103; and Gregor 1990, 107.
60. Ferguson 1997, 336.
61. See Ember and Ember 1994, 200-203; and Chagnon 1968, 146-47, 150.
62. For a critique, see Ross 1993, chap. 3.
63. See Reno 1998; and Kahl 1998.
64. Laitin 1999, 146-80.
65. Compare Ferguson 1995 to Reno 1998.
66. Ferguson 1994, 102-103.
67. See Ember, Ember, and Russett 1992; and Russett 1993, chap. 5.
68. Ember and Ember 1994, 202-203; but see also Ross 1993, 43-44.
69. Ember and Ember 1994, 202.
70. Kelly 2000, 43-51, 101, 143.
71. See Otterbein 1994, 174-75; and Ferguson and Whitehead 1992a, 10.
72. See Kelly 1985; and Otterbein 1994, 176. For a statistical study that rejects the hypothesis that extensive kinship ties increase the probability of external war, see Tefft 1975, 702.
73. Lévi-Strauss 1971, 350, quoting Edward Tylor.
74. Tefft 1975, 703. See also Kelly 2000, 61.
75. Pospisil 1994, 123-25. A major cause of fighting among the extremely violent Gebusi of New Guinea is the accusation of sorcery coinciding with a failure to reciprocate in sister exchange marriage. Knauft 1987, 477.
76. Ziegler 1990.
77. Ross 1993, 40-44.
78. Quoted in Sahlins 1976b, 14.
79. Swidler 1986.
80. See Malinowski 1984, 17, 29; and Sahlins 1976b.
81. See Knight 1999, 228-31; and Bloch 1989, 21-24.
82. Ferguson 1990, 46.
83. Mercer 1995.
84. Wendt 1999, 274-75. For relevant research findings, see Stein 1976.
85. Gellner 1983, 56.
86. Berger and Luckmann 1967, 100.
87. Barth 1987, 9.
88. Harrison 1993, 8, 47.
89. Ibid., 61, 75-77.
90. Ibid., 89.
91. Ibid., 18.
92. Ibid., 150.
93. Ember and Ember 1994, 194.
94. Ibid., 192.
95. Sanday 1981, 49-50.
96. Ferguson 1992, 224.
97. See Ross 1993, 60-64, 77-84, 125-26; and Meggitt 1977.
98. Knauft 1999, 163.
99. Ferguson 1992, 223-24.
100. Ross 1993, 7-8, 89-92.
101. Ibid., 136-38.
102. Hoffmann 1965, 80-81.
103. Gregor 1990.
104. Durkheim [1915] 1965, 264.
105. Barnard 1999.
106. Ferguson 1992, 201.
107. Ferguson and Whitehead 1992a, 18. See also Ferguson 1997, 336.
108. Harrison 1993.
109. Jervis lists such interaction effects as one of the characteristics of complex systems, but many research designs attempt to study such interactions without adopting the other aspects of a systems approach. Jervis 1997, 39-44.
110. Jervis 1997.
111. Cohen and Service 1978.
112. Rappaport 1968, 224.
113. See Vayda 1989; and Meggitt 1977, 28.
114. Knauft 1999, 138.
115. Ibid., 127-28, developing arguments by Modjeska 1982.
116. Meggitt 1977, 43; also 8-9, 12, 42, 58.
117. LiPuma 2000.
118. See Waltz 1979, chap. 7-8; Wendt 1999, chap. 6; and Russett and Oneal 2001.
119. Ikenberry 2001.
120. Russett and Oneal discuss whether the recent Kashmir fighting, which pits Pakistan's religious principles of citizenship against India's territorial ones, should count as a war between democracies. Russett and Oneal 2001, 48.
121. See Russett 1993; and Owen 1997.
122. Przeworski et al. 2000, chap. 2.
123. Linz and Stepan 1996, 7-15.
124. For a balanced assessment, see Carothers 1999.
125. Gowa 1999, 44-67.
126. Russett and Oneal 2001.
127. Robarchek and Robarchek 1998, 156.
128. Jervis 1997, 37.
129. Carothers 1999.
130. On sequencing and system effects, see Jervis 1997, 37-39.
131. Risse and Sikkink 1999.
132. Lund, Rubin, and Hara 1998.
133. Diamond 1996.
134. Wendt 1999, 137. Wendt further qualifies his views in Wendt 2000.
135. Wendt 1999, 249. For a critique, see Copeland 2000.
136. Wendt 1999, 371; see also 369, 377-78.
137. Finnemore and Sikkink 1998, 916.
References
Barnard, Alan. 1999. Modern Hunter-Gatherers and Early Symbolic Culture. In The Evolution of Culture, edited by Robin Dunbar, Chris Knight, and Camilla Power, 50-70. New Brunswick, N.J.: Rutgers University Press.
Barth, Fredrik. 1987. Cosmologies in the Making. Cambridge: Cambridge University Press.
Berger, Peter, and Thomas Luckmann. 1967. The Social Construction of Reality. New York: Anchor.
Bloch, Maurice. 1989. Ritual, History, and Power. London: Athlone.
Brewer, Marilynn. 2001. Ingroup Identification and Intergroup Conflict: When Does Ingroup Love Become Outgroup Hate? In Social Identity, Intergroup Conflict, and Conflict Reduction, edited by Richard D. Ashmore, Lee Jussim, and David Wilder, 17-41. New York: Oxford University Press.
Byman, Daniel, and Stephen Van Evera. 1998. Why They Fight: Hypotheses on the Causes of Contemporary Deadly Conflict. Security Studies 7 (3):1-50.
Carneiro, Robert. 1990. Chiefdom-Level Warfare as Exemplified in Fiji and the Cauca Valley. In The Anthropology of War, edited by Jonathan Haas, 190-211. Cambridge: Cambridge University Press.
___. 1994. War and Peace. In Studying War: Anthropological Perspectives, edited by S. P. Reyna and R. E. Downs, 3-28. Amsterdam: Gordon and Breach.
Carothers, Thomas. 1999. Aiding Democracy Abroad. Washington, D.C.: Carnegie Endowment for International Peace.
Chagnon, Napoleon A. 1968. Yanomamo Social Organization and Warfare. In War: The Anthropology of Armed Conflict and Aggression, edited by Morton Fried, Marvin Harris, and Robert Murphy, 109-59. Garden City, N.Y.: Natural History Press.
___. 1997. Yanomamo. Fort Worth, Tex.: Harcourt Brace.
Cohen, Alex. 1990. A Cross-Cultural Study of the Effects of Environmental Unpredictability on Aggression in Folktales. American Anthropologist 92 (2):474-81.
Cohen, Ronald, and Elman Service, eds. 1978. Origins of the State. Philadelphia: Institute for the Study of Human Issues.
Copeland, Dale C. 2000. The Constructivist Challenge to Structural Realism. International Security 25 (2):187-212.
Cox, Robert. 1986. Social Forces, States, and World Orders. In Neorealism and Its Critics, edited by Robert O. Keohane, 204-54. New York: Columbia University Press.
Diamond, Larry. 1996. Is the Third Wave Over? Journal of Democracy 7 (3):20-37.
Durkheim, Emile. [1915] 1965. The Elementary Forms of Religious Life. Reprint, New York: Free Press.
Ember, Carol R., and Melvin Ember. 1992. Resource Unpredictability, Mistrust, and War. Journal of Conflict Resolution 36 (2):242-62.
___. 1997. Violence in the Ethnographic Record: Results of Cross-Cultural Research on War and Aggression. In Troubled Times: Violence and Warfare in the Past, edited by Debra L. Martin and David W. Frayer, 1-20. Amsterdam: Gordon and Breach.
Ember, Carol R., Melvin Ember, and Bruce Russett. 1992. Peace Between Participatory Polities: A Cross-Cultural Test of the "Democracies Rarely Fight Each Other" Hypothesis. World Politics 44 (4):573-99.
Ember, Melvin. 1982. Statistical Evidence for an Ecological Explanation of Warfare. American Anthropologist 84 (3):645-49.
Ember, Melvin, and Carol R. Ember. 1994. Cross-Cultural Studies of War and Peace: Recent Achievements and Future Possibilities. In Studying War: Anthropological Perspectives, edited by S. P. Reyna and R. E. Downs, 185-208. Amsterdam: Gordon and Breach.
Fearon, James D. 1995. Rationalist Explanations for War. International Organization 49 (3):379-414.
Ferguson, R. Brian. 1984. Introduction: Studying War. In Warfare, Culture, and Environment, edited by R. Brian Ferguson, 1-82. Orlando, Fla.: Academic Press.
___. 1990. Explaining War. In The Anthropology of War, edited by Jonathan Haas, 26-55. Cambridge: Cambridge University Press.
___. 1992. A Savage Encounter. In War in the Tribal Zone, edited by R. Brian Ferguson and Neil Whitehead, 199-228. Santa Fe, N.M.: School of American Research.
___. 1994. The General Consequences of War. In Studying War: Anthropological Perspectives, edited by S. P. Reyna and R. E. Downs, 85-111. Amsterdam: Gordon and Breach.
___. 1995. Yanomami Warfare: A Political History. Santa Fe, N.M.: School of American Research.
___. 1997. Violence and Warfare in Prehistory. In Troubled Times: Violence and Warfare in the Past, edited by Debra Martin and David Frayer, 321-56. Amsterdam: Gordon and Breach.
Ferguson, R. Brian, and Neil Whitehead. 1992a. The Violent Edge of Empire. In War in the Tribal Zone, edited by R. Brian Ferguson and Neil Whitehead, 1-30. Santa Fe, N.M.: School of American Research.
Ferguson, R. Brian, and Neil Whitehead, eds. 1992b. War in the Tribal Zone. Santa Fe, N.M.: School of American Research.
Finnemore, Martha, and Kathryn Sikkink. 1998. International Norm Dynamics and Political Change. International Organization 52 (4):887-917.
Garrett, Geoffrey, and Barry Weingast. 1993. Ideas, Interests, and Institutions: Constructing the European Community's Internal Market. In Ideas and Foreign Policy, edited by Judith Goldstein and Robert O. Keohane, 173-206. Ithaca, N.Y.: Cornell University Press.
Gellner, Ernest. 1983. Nations and Nationalism. Ithaca, N.Y.: Cornell University Press.
Gregor, Thomas. 1990. Uneasy Peace: Intertribal Relations in Brazil's Upper Xingu. In The Anthropology of War, edited by Jonathan Haas, 105-24. Cambridge: Cambridge University Press.
Gordon, Robert J., and Mervyn J. Meggitt. 1985. Law and Order in the New Guinea Highlands. Hanover: University of Vermont.
Gowa, Joanne. 1999. Ballots and Bullets: The Elusive Democratic Peace. Princeton, N.J.: Princeton University Press.
Haas, Jonathan, ed. 1990. The Anthropology of War. Cambridge: Cambridge University Press.
Harrison, Simon. 1993. The Mask of War: Violence, Ritual, and the Self in Melanesia. Manchester, U.K.: Manchester University.
Hoffmann, Stanley. 1965. The State of War. New York: Praeger.
Hui, Victoria Tin-bor. Forthcoming. War and State Formation in Ancient China and Early Modern Europe. Cambridge: Cambridge University Press.
Human Rights Watch. 2000. Human Rights Watch World Report 2001. New York: Human Rights Watch.
Ikenberry, G. John. 2001. After Victory: Institutions, Strategic Restraint, and the Rebuilding of Order After Major Wars. Princeton, N.J.: Princeton University Press.
Jervis, Robert. 1978. Cooperation Under the Security Dilemma. World Politics 30 (2):167-214.
___. 1997. System Effects. Princeton, N.J.: Princeton University Press.
Johnston, Alastair Iain. 1995. Cultural Realism: Strategic Culture and Grand Strategy in Chinese History. Princeton, N.J.: Princeton University Press.
Kahl, Colin H. 1998. Population Growth, Environmental Degradation, and State-Sponsored Violence: The Case of Kenya, 1991-1993. International Security 23 (2):80-119.
Katzenstein, Peter J., ed. 1996. The Culture of National Security. New York: Columbia University Press.
Kaufmann, Chaim. 1996. Possible and Impossible Solutions to Ethnic Civil Wars. International Security 20 (4):136-75.
Keeley, Lawrence H. 1996. War Before Civilization. New York: Oxford University Press.
Kelly, Raymond C. 1985. The Nuer Conquest. Ann Arbor: University of Michigan Press.
___. 2000. Warless Societies and the Origin of War. Ann Arbor: University of Michigan Press.
Keohane, Robert. 2000. Ideas Part-way Down. Review of International Studies 26 (1):125-30.
Knauft, Bruce M. 1987. Reconsidering Violence in Simple Human Societies: Homicide Among the Gebusi of New Guinea. Current Anthropology 28 (4):457-500.
___. 1999. From Primitive to Postcolonial in Melanesia and Anthropology. Ann Arbor: University of Michigan Press.
Knight, Chris. 1999. Sex and Language as Pretend Play. In The Evolution of Culture, edited by Robin Dunbar, Chris Knight, and Camilla Power, 228-47. New Brunswick, N.J.: Rutgers University Press.
Koslowski, Rey, and Friedrich V. Kratochwil. 1995. Understanding Change in International Politics: The Soviet Empire's Demise and the International System. In International Relations Theory and the End of the Cold War, edited by Richard Ned Lebow and Thomas Risse-Kappen, 127-66. New York: Columbia University Press.
Laitin, David. 1999. Somalia: Civil War and International Intervention. In Civil Wars, Insecurity, and Intervention, edited by Barbara F. Walter and Jack Snyder, 146-80. New York: Columbia University Press.
Layne, Christopher, and Benjamin Schwartz. 1999. For the Record. The National Interest 57 (fall):9-16.
Lévi-Strauss, Claude. 1971. The Family. In Man, Culture, and Society, edited by Harry Shapiro, 333-57. New York: Oxford University Press.
Linz, Juan, and Alfred Stepan. 1996. Problems of Democratic Transition and Consolidation. Baltimore, Md.: Johns Hopkins University Press.
LiPuma, Edward. 2000. Encompassing Others: The Magic of Modernity in Melanesia. Ann Arbor: University of Michigan Press.
Lund, Michael, Barnett Rubin, and Fabienne Hara. 1998. Learning from Burundi's Failed Democratic Transition, 1993-1996. In Cases and Strategies for Preventive Action, edited by Barnett R. Rubin, 47-92. New York: Century Foundation.
Malinowski, Bronislaw. 1937. Culture as a Determinant of Behavior. In Factors Determining Human Behavior, edited by Harvard Tercenary Conference of Arts and Sciences, 133-68. Cambridge, Mass.: Harvard Tercenary Publications.
___. 1984. Magic, Science, and Religion, and Other Essays. Westport, Conn.: Greenwood.
Mearsheimer, John. 1994/95. The False Promise of Institutions. International Security 19 (3):5-49.
___. 2001. The Tragedy of Great Power Politics. New York: W. W. Norton.
Meggitt, Mervyn J. 1977. Blood Is Their Argument: Warfare Among the Mae-Enga Tribesmen of the New Guinea Highlands. Palo Alto, Calif.: Mayfield.
Mercer, Jonathan. 1995. Anarchy and Identity. International Organization 49 (2):229-52.
Modjeska, Nicholas. 1982. Production and Inequality. In Inequality in New Guinea Highlands Societies, edited by Andrew Strathern, 50-108. Cambridge: Cambridge University Press.
Ortner, Sherry B. 1999. Introduction. In The Fate of "Culture": Geertz and Beyond, edited by Sherry B. Ortner, 1-13. Berkeley: University of California Press.
Otterbein, Keith F. 1994. Feuding and Warfare. Amsterdam: Gordon and Breach.
Owen, John M., IV. 1997. Liberal Peace, Liberal War. Ithaca, N.Y.: Cornell University Press.
Philpott, Daniel. 2001. Revolutions in Sovereignty: How Ideas Shaped Modern International Relations. Princeton, N.J.: Princeton University Press.
Posen, Barry. 1993. The Security Dilemma and Ethnic Conflict. Survival 35 (1):27-47.
Pospisil, Leopold. 1994. I Am Very Sorry I Cannot Kill You Anymore: War and Peace Among the Kapauku. In Studying War: Anthropological Perspectives, edited by S. P. Reyna and R. E. Downs, 113-26. Amsterdam: Gordon and Breach.
Powell, Walter W., and Paul J. DiMaggio, eds. 1991. The New Institutionalism in Organizational Analysis. Chicago: University of Chicago Press.
Przeworski, Adam, Michael Alvarez, José Cheibub, and Fernando Limongi. 2000. Democracy and Development. Cambridge: Cambridge University Press.
Rappaport, Roy A. 1968. Pigs for the Ancestors. New Haven, Conn.: Yale University Press.
Reno, William. 1998. Warlord Politics and African States. Boulder, Colo.: Lynne Rienner.
Reyna, S. P., and R. E. Downs, eds. 1994. Studying War: Anthropological Perspectives. Amsterdam: Gordon and Breach.
Reus-Smit, Christian. 1999. The Moral Purpose of the State. Princeton, N.J.: Princeton University Press.
Risse, Thomas, and Kathryn Sikkink. 1999. The Socialization of International Human Rights Norms into Domestic Practice. In The Power of Human Rights: International Norms and Domestic Change, edited by Thomas Risse, Stephen C. Ropp, and Kathryn Sikkink, 1-38. Cambridge: Cambridge University Press.
Robarchek, Clayton, and Carole Robarchek. 1998. Waorani: The Contexts of Violence and War. New York: Harcourt Brace.
Ross, Marc Howard. 1993. The Culture of Conflict. New Haven, Conn.: Yale University Press.
___. 1997. Culture and Identity in Comparative Political Analysis. In Comparative Politics, edited by Mark Irving Lichbach and Alan S. Zuckerman, 42-80. Cambridge: Cambridge University Press.
Ruggie, John Gerard. 1998. Constructing the World Polity. London: Routledge.
Russett, Bruce. 1993. Grasping the Democratic Peace. Princeton, N.J.: Princeton University Press.
Russett, Bruce, and John R. Oneal. 2001. Triangulating Peace: Democracy, Interdependence, and International Organizations. New York: W. W. Norton.
Sahlins, Marshall D. 1968. Tribesmen. Englewood Cliffs, N.J.: Prentice-Hall.
___. 1976a. Culture and Practical Reason. Chicago: University of Chicago Press.
___. 1976b. The Use and Abuse of Biology. Ann Arbor: University of Michigan Press.
Sanday, Peggy Reeves. 1981. Female Power and Male Dominance. Cambridge: Cambridge University Press.
Schotter, Andrew. 1981. The Economic Theory of Social Institutions. Cambridge: Cambridge University Press.
Shweder, Richard A. 2000. Rethinking the Object of Anthropology (and Ending Up Where Kroeber and Kluckhohn Began). Items and Issues 1 (2):7--9.
Spruyt, Hendrik. 1994. The Sovereign State and Its Competitors. Princeton, N.J.: Princeton University Press.
Stein, Arthur. 1976. Conflict and Cohesion. Journal of Conflict Resolution 20 (2):143-72.
Swidler, Ann. 1986. Culture in Action: Symbols and Strategies. American Sociological Review 51 (2):273-86.
Tefft, Stanton. 1975. Warfare Regulation: A Cross-Cultural Test of Hypotheses. In War: Its Causes and Correlates, edited by Martin A. Nettleship, R. Dale Givens, and Anderson Nettleship, 693-712. The Hague: Mouton.
Tierney, Patrick. 2000. Darkness in El Dorado: How Scientists and Journalists Devastated the Amazon. New York: W. W. Norton.
Tilly, Charles. 1990. Coercion, Capital, and European States, A.D. 990-1990. Cambridge: Basil Blackwell.
Vayda, Andrew P. 1989. Explaining Why Marings Fought. Journal of Anthropological Research 45 (2):159-78.
Waltz, Kenneth. 1979. Theory of International Politics. Reading, Mass.: Addison-Wesley.
Wendt, Alexander. 1992. Anarchy Is What States Make of It: The Social Construction of Power Politics. International Organization 46 (2):391-425.
___. 1999. Social Theory of International Politics. Cambridge: Cambridge University Press.
___. 2000. On the Via Media: A Response to the Critics. Review of International Studies 26 (1):165-80.
Ziegler, Rolf. 1990. The Kula: Social Order, Barter, and Ceremonial Exchange. In Social Institutions: Their Emergence, Maintenance, and Effects, edited by Michael Hechter, Karl-Dieter Opp, and Reinhard Wippler, 141-70. New York: Aldine and de Gruyter.
