Egoism (self-help)

Evolutionary theory offers a powerful explanation for the trait of egoism (by which we mean the nonpejorative definition of “self-regarding, prompted by self-interest”).⁸⁶ Given competition for limited resources and threats from predators and the environment, an individual organism is primed to seek its own survival and—the Darwinian bottom line—reproductive success. The organism has to ensure that its physiological needs (for food, water, shelter, and so on) are satisfied so that it can survive and reproduce. The imperative for survival in a hostile environment also requires that an individual organism places its survival, especially in a time of danger or stress, above the survival of others. The genes of egoistic individuals survive and spread at the expense of those that fail to effectively put their own interests first. Of course, cooperation and helping behaviors are common in nature, but such behaviors persist only where they help the genes causing that behavior to spread. Genes obviously do not “want” or “try” to spread, but the machinery of natural selection means that self-serving alleles will generally increase in relative frequency in the population over time, at the expense of alleles that are neutral or self-sacrificing for no return benefit.

Evolutionary theorists now recognize, following William Hamilton’s concept of “inclusive fitness,” that egoism has complexities. Darwin himself envisioned these nuances, even though he did not know the biological mechanisms at work. Hamilton used genetic models to show that, while individual organisms are egoistic, they should be less so in their behavior toward genetic relatives, especially in parent-offspring and sibling relationships.^{Reference Hamilton87,Reference Hamilton88} This decrease in egoism is because close relatives share many of the same genes—one-half for siblings and parents, one-quarter for aunts, uncles, and grandparents, and one-eighth for cousins. An individual’s Darwinian fitness therefore includes the success of related others (hence the phrase inclusive fitness). The preeminent evolutionary theorist J.B.S. Haldane thus quipped that he would give his life to save two of his brothers (each sharing half of his genes) or eight of his cousins (each sharing one-eighth of his genes).^{Reference Haldane89,90} Inclusive fitness provides a biological basis for the common intuition that individuals favor those who are close genetic relatives.^{Reference Betzig91}

Evolutionary theory accounts for egoism and explains why cooperation can extend to the family or close kin group but remains difficult between unrelated individuals. Cooperation among unrelated individuals is possible but only as the result of interactions that help genes replicate in the long run, through mechanisms such as reciprocal altruism, indirect reciprocity, and signaling.^{Reference Nowak92} Even cooperation and helping behavior, therefore, are strategies that increase an individual’s Darwinian fitness—indeed, that is precisely why they evolved.^{Reference West, El Mouden and Gardner93} In nature, genetic egoism is the basis of natural selection.

Dominance (power maximization)

Evolutionary theory can also explain dominance. Like egoism, the desire to dominate is a trait of human nature (which, as with egoism, we stress does not necessarily apply to every individual or situation but is a statistical tendency underlying behavior). As evolutionary economist Robert Frank has explained, “Evidence suggests that we come into the world equipped with a nervous system that worries about rank. Something inherent in our biological makeup motivates us to try to improve, or at least maintain, our standing against those with whom we compete for important positional resources.”^{Reference Frank94} In the context of evolutionary theory, dominance usually means that particular individuals in a social group have priority of access to resources in competitive situations.^{Reference Milinski, Parker, Krebs and Davies95} A wide variety of animals exhibit a form of social organization called a dominance hierarchy, in which members of a social group each have a status rank descending from the alpha male down through all the other individuals to the lowliest subordinates. A dominance hierarchy is created competitively, often violently, and is maintained forcefully, but it can serve to prevent or reduce conflict within a group because it establishes a pecking order that is generally respected.

As with all things in nature, dominance hierarchies vary considerably. Hierarchies may be weak or strong, and “alpha” males may sire nearly all offspring or just more than others. However, what is striking is the prevalence and potency of dominance in social organization, despite variations in the specifics. Dominance behavior occurs in thousands of taxonomic groups ranging from fish and reptiles to birds and mammals.

Of particular note regarding the impact of dominance on human behavior are the roles of both phylogeny (a species’ ancestral lineage) and ecology (its adaptations to local conditions). With regard to phylogeny, most primates and all the great apes (the group to which humans belong) have strong social dominance hierarchies, and humans are no exception—dominance hierarchies have been extensively documented among humans in a wide variety of settings and eras.^{Reference Wason96,Reference Mazur and Booth97,98} With regard to ecology, dominance hierarchies are a common form of social organization in the kind of ecological settings in which humans evolved (social groups with competing interests, variation in power, and finite resources).

The strength of dominance hierarchies in humans is debated and varies empirically, but such hierarchies are always evident in some form or other. Even where dominance hierarchies are actively suppressed, such as in more egalitarian small-scale societies, suppression itself is evidence of the competition for status that simmers beneath the surface.^{Reference Boehm99} In historical and contemporary societies, competition for power is all too obvious, and the quest for status, dominance, and leadership is ancient and ever present.^{Reference Mazur100,101,Reference Ludwig102,Reference King, Johnson and Van Vugt103}

Males of most mammal species are particularly competitive with each other over females. Indeed, the competition for mates is subject to a special type of evolutionary selection process—sexual selection, as opposed to standard natural selection. Reproductive access to females tends to be highly skewed, with a few males responsible for a large proportion of progeny. This is because a single male can reproduce with multiple females, whereas females can usually only reproduce and rear the offspring of one male at a time, with a long delay before becoming reproductively available again. In some species, reproductive access is settled by coercion, in which the strongest male defeats rivals to dominate a harem. In other species, males cannot coerce females, but the females are choosy about with whom they mate, leading to selection pressures for males to demonstrate or signal their quality as attractive partners. In either case, it is females rather than males that are the limiting factor in sexual competition, making male competition for available females intense. Sexual selection has led to costly biological adaptations, such as fighting, the growth of heavy “weapons” (e.g., antlers), risky courtship displays, or adornments that signal genetic quality (e.g., gigantic tails). These types of adaptations not only consume precious time and energy but can also decrease survival in other, nonreproductive domains of life (for example, the plumage of male peacocks limits their ability to fly). Thus, the power of sexual selection can lead to the evolution of traits that actually damage survival in order to achieve superiority over other males.^{Reference Lincoln, Short and Balaban104,Reference Trivers and Campbell105} Reproduction trumps survival in evolution.

Sexual selection is typically responsible for the hierarchical nature of group-living animal species, including humans, as males fight for rank and the reproductive benefits in brings. As we would expect, this leads to sex differences in the desire for status. Both laboratory experiments and real-world observations have identified empirical differences between men and women in a range of social behaviors, not least that men tend to have relative-gains, or “zero-sum” motivations (wanting to get ahead at the expense of others), whereas women tend to favor payoff-maximization, or “variable-sum” motivations (content to do well even if others also do well in the process).^{Reference Lopez, McDermott and Petersen106,Reference Ellis, Hershberger, Field, Wersinger, Pellis, Hetsroni and Geary107,Reference Taylor, Klein, Lewis, Gruenewald, Gurung and Updegraff108,Reference Van Vugt and Spisak109}

It is well established that dominance and status-seeking behaviors in humans are based on many of the same biochemical and neurological processes as in other mammals, such as the secretion and uptake of testosterone and serotonin, which modify status-challenging behavior.^110,111 However rational we may like to think we are, our judgments and decision-making are nevertheless influenced by cognitive mechanisms and biochemicals circulating in our bodies and brains that relate to dominance behavior.^{Reference McDermott112,Reference Damasio113,114,115}

Dominance hierarchies need not only benefit those at the top. An organized social structure can help promote the harvesting of resources, coordinate group activity, and reduce within-group conflict. Although alphas in the hierarchy tend to have the highest reproductive success, other males may benefit from group membership by gaining protection from other groups, or by biding their time for a chance to challenge the alpha male when they become strong enough or old enough. Ethological studies have shown that hierarchical dominance systems within a primate group’s social network can reduce overt aggression, although aggression increases again when the alpha male is challenged.^{Reference Knauft116,Reference Flack, Girvan, de Waal and Krakauer117,Reference de Waal118}

To summarize, a species that lives communally could have two broad forms of social organization. The group can accept organization with some centralization of power (dominance hierarchies), or it can engage in perpetual conflict (“scramble competition”), which incurs costs in terms of time, energy, and injuries, as well as depriving the group of many benefits of a communal existence, such as more efficient resource harvesting.¹¹⁹ Among social mammals, and primates in particular, dominance hierarchies have emerged as the primary form of social organization.

Thus far, we have emphasized a state of “anarchy” in evolutionary history, in which there was no overarching power to provide protection from predators, rivals, or other threats. Egoism and dominance arose as strategies that provided solutions to achieving survival and reproduction in this environment. However, dominance hierarchies were in some sense a mechanism by which this anarchy could be suppressed—at least within the group—to the benefit of all group members since they share at least some common interests (such as avoiding conflict). This collective benefit points to the special and much more significant role of anarchy at a higher level—anarchy between groups. As we show in the next section, competition between groups is especially significant for human evolution, and for international politics, precisely because it is at the intergroup level where anarchy reigns supreme and is much harder to suppress.

Ingroup/outgroup bias (fear of others)

Egoism and dominance are important mechanisms for attaining security, but also important is attaining security from members of other groups. In fact, interactions with people from other groups—unrelated others that may not have shared interests—may be especially costly or lethal compared to within-group interactions. Evolutionary theory provides an important framework for understanding the ingroup/outgroup distinction commonly noted by anthropologists, sociologists, and political scientists, and perhaps most prominently by psychologists.^{Reference Kurzban, Neuberg and Buss120,121} Of the many biases identified in the so-called cognitive revolution in psychology, the ingroup/outgroup bias is one of most pervasive, pernicious, and powerful. The key finding is that humans quickly adopt an “us” (ingroup) versus “them” (outgroup) worldview. We are positively biased toward our own groups and negatively biased toward other groups. Psychologists argue that the ingroup/outgroup distinction develops from a need for social identity. Identification with a specific group provides individuals with meaning and purpose, encouraging them to become part of a community with common interests, values, and goals.^{Reference Hewstone, Rubin and Willis122,Reference Fiske123,Reference Sidanius and Pratto124} One also knows what one is not—the outgroup, which is stereotyped and homogenized as the “Other.” Among the many different possible ingroup categories, the most common and significant include family, friends, age, sex, class, ethnicity, politics, religion, and nationality.

The ubiquity and strength of the ingroup/outgroup bias across history and across human cultures suggest it is an ingrained aspect of human nature, and evolutionary theory explains why such a mechanism would evolve.^{Reference Alexander125,Reference Hammond and Axelrod126,Reference Choi and Bowles127} First, considerable evidence from both archeological and ethnographic research on preindustrial societies points to intensive intergroup conflict in our past.^{128,129,130,131,132,133} As we noted earlier, around 15 percent of male populations in indigenous small-scale societies died in warfare (and, in some such societies, war-related mortality rates were considerably higher).^134,135 War also remains a significant influence on the social organization and physical distribution of these societies even when they are not actually at war. In human history and prehistory, other human groups were commonly the most dangerous threat in the environment, and the ingroup/outgroup bias is likely to have helped the ingroup remain cohesive, avoid and be wary of outgroups, solve the collective action problem in emergencies, and kill outsiders.^{136,137,Reference Tooby, Cosmides and Høgh-Olesen138}

Second, the ingroup/outgroup bias offers a rapid heuristic to weigh the various threats when encountering other humans. Will the outsider be a threat to oneself or to one’s family? Will an outsider compete for the current or future resources that the insiders need to survive or expand? Will a male from the outgroup present competition for mates, or will his presence threaten the ingroup male’s position in the extended family or group? In environments where resources are highly contested, outgroup fear can become extreme. For example, Wrangham recounts that among the Inuit of the Arctic, “unfamiliar men would normally be killed even before questions were asked.”¹³⁹

Such wariness of individuals from other groups is paralleled among animals. For example, among wolves, lions, and chimpanzees, when members of rival groups are found alone, they are extremely vulnerable and risk being killed.^140,141,142 We discussed intergroup killing in chimpanzees earlier, but the pattern is notable among social carnivores, too: Studies of undisturbed wolf populations in Alaska have found that 39 to 65 percent of adult deaths were due to intergroup killing.^{Reference Mech, Adams, Meier, Burch and Dale143}

Of course, the ability to assess threats is much more complex in humans than it is in other animals, and human intelligence gives us a greater repertoire of behavior. Thus, humans may consider other variables, such as the possibility of future trade or cooperation, when assessing outgroups. Still, humans exhibit an instinctive fight or flight response, just as other animals do, which operates below (and faster than) conscious awareness.^{Reference Axelrod144,Reference Cronk and Leech145} Befriending or cooperating with a stranger may have benefits, but those benefits are uncertain and limited whereas the costs of trusting an outsider may be lethal. It is therefore no surprise, as psychologists have argued, that evolution has favored a bias to be fearful of strangers to avoid the costlier error.^{Reference Haselton and Nettle146,Reference Rozin and Royzman147}

Summary: The legacy of evolution

In this section, we have presented standard biological arguments that egoism, dominance, and ingroup/outgroup bias are deeply rooted behavioral adaptations common among mammals in general and primate species in particular. Of course, human behavior is not a direct extension of the behavior of other animals, but, as we have explained, the ecological setting in which our own species evolved made these same traits as or even more important for humans. Egoism, dominance, and ingroup/outgroup bias are widespread because they increased survival and reproductive success compared with other strategies and were therefore favored by natural selection. The cognitive mechanisms underpinning the three traits were established in an environment very different from the one in which humans now live, but they persist because our brains, biochemistry and nervous systems, which evolved over many millions of years, have remained the same despite the rapid sociological and technological advances of the last few centuries. The fact that these evolved behaviors are not always beneficial today does nothing to undermine their evolutionary logic or empirical presence.

Note that we did not pick the traits of egoism, dominance, and ingroup/outgroup bias out of a hat. As well as being the key behavioral traits identified by Mearsheimer, self-interest, social stratification, and groupish behavior are three of the most prominent behavioral features of social animals. That natural selection should have drawn out the same three traits as Mearsheimer may seem a remarkable coincidence. However, because anarchy is a problem both in nature and in international politics, it is no coincidence at all. Rather, we suggest it is an example of what biologists call “evolutionary convergence”—similar traits arising in different settings because they are good solutions to a common problem.

One reason why an evolutionary explanation of egoism, dominance, and ingroup/outgroup bias is useful is because alternative explanations for these empirical traits have failed. Egoism, dominance, and ingroup/outgroup distinctions have previously been attributed to variables such as culture, economics, or religion.^148,149 For example, Karl Marx and his followers identified egoism as a result of capitalism and called for its suppression and the triumph of class consciousness. Likewise, many other religious and utopian theorists attribute egoism, dominance, and ingroup/outgroup bias to special, or at least changeable, circumstances. The fact that all prior utopian visions have failed to come to fruition does not prove that current ones will fail too. However, the persistence of these three traits across domains and over time casts doubt on arguments like these, and strongly counts in favor of an evolutionary explanation instead.

We recognize that humans are influenced by culture, norms, rational calculation, and moral principles. Nevertheless, overwhelming evidence shows that people also behave in ways that can be predicted from the biological knowledge outlined above. The impact of these biological factors on social and political behavior will vary depending on context. However, our contention is that significant aspects of political behavior could be given a stronger foundation if we acknowledge the powerful and basic biological principles that are chronically ignored in the political science literature, as well as the conditions under which they become exacerbated or suppressed.

Novel ultimate cause and theoretical foundation

Our theory advances offensive realist arguments without seeking an ultimate cause in the anarchic international state system. Offensive realism based on evolutionary theory makes the same predictions for state behavior, but the ultimate causal mechanism is different: human evolution in the anarchic, dangerous, and competitive conditions of the late-Pliocene and Pleistocene eras. Evolutionary theory explains why humans are egoistic, strive to dominate others, and make ingroup/outgroup distinctions. These adaptations in turn serve as a foundation for offensive realism—what Mearsheimer independently identified as self-help, power maximization, and fear.

The central issue raised by our theory is what causes states to behave as offensive realists predict. Mearsheimer argues that anarchy is the fundamental cause of such behavior. The fact that there is no world government compels the leaders of states to take steps to ensure their security, such as striving to have a powerful military, forging and maintaining alliances, and acting aggressively when necessary. This is what neorealists call a self-help system: Leaders of states are forced to take these steps because nothing else can guarantee their security in the anarchic world of international politics.

We argue that evolutionary theory also offers a fundamental cause for offensive realist behavior (see Table 1). Humans evolved in a state of nature where competition for resources and dangers from other humans and the environment were great. Humans survived (and note that several other Hominin species did not) by virtue of evolved behavioral traits—among them egoism, dominance, and the ingroup/outgroup bias—which were adaptations to competitive ecological conditions. These traits help to explain why humans (including political leaders) will behave, in the proper circumstances, as offensive realists expect them to behave. In sum, evolutionary theory offers realist scholars a natural-scientific behavioral foundation for offensive realism. Human evolution explains why people seek control over resources, why some of us (particularly males) will seek to dominate others in order to achieve and maintain a privileged position in a dominance hierarchy, and why we are suspicious and wary of other groups.¹⁵⁰ Since the leaders of states are human, they too will be influenced—perhaps especially influenced—by this evolutionary legacy as they react to the actions of other states and decide how to respond.^{151,Reference McDermott152,Reference McDermott153}

Table 2.

Behavior under anarchy in different domains.

Note: In all cases, we argue, holding onto the status quo is not always good enough to ensure survival. Instead, the best strategy is a constant effort to maximize power to stay ahead of rivals. Even if this strategy is never successful, it motivates individuals to achieve the maximum possible. We invoke anarchy in all situations in the table because, while our core argument is that evolved dispositions (egoism, dominance, groupishness) give rise to offensive realist behavior today even in the absence of anarchy, these evolved dispositions will be more prominent and influential where regulation is lax. Under an iron fist, even the most egotistical, dominating, and xenophobic actor will be prevented from realizing his or her goals.

Table 3.

Who wants power? A comparison among alternative realist theories.

Extensions of offensive realism to other domains

The second contribution of our theory is that it offers an explanation of the behavior of humans in a wide variety of contexts extending beyond international politics. Our evolutionary theory of offensive realism is unlimited in time, explaining behavior from the ancestral environment to the present day, whereas offensive realism is conventionally inapplicable prior to 1648, when the Treaty of Westphalia established the European state system. Our theory is also unlimited in domain, explaining behavior wherever there are human actors and weak external constraints on their actions, from ancestral human groups, ethnic conflict, and civil wars to domestic politics, free markets, and international relations. The theory might thus be extended to explain the behavior and actions of many phenomena: the Roman Empire, warfare among Papua New Guinean or Native American tribes, the European conquest of South America, the race for the American west and the failed Mormon and Confederate secessions, the imperialist scrambles over African colonies, institutions like the medieval Catholic Church, commercial organizations from the East India Company to Coca-Cola, the struggles of rival ethnic groups the world over, and the ruthlessness of electoral campaigns.

Table 2 illustrates the range of domains to which an evolutionary theory of offensive realism applies. Note that the table captures key patterns, not universal behavior. Clearly, not all individuals or businesses or states act the same way all the time or in all circumstances. Behavior varies considerably, just as standard offensive realism predicts for states, and countervailing forces would sometimes mitigate power-maximization strategies—although the very need for and difficulties of those countervailing forces help to demonstrate the fact that offensive realist behavior remains an underlying problem. Indeed, part of the beauty of evolutionary approaches is their ability to predict sources of variation—the socio-ecological conditions under which we should expect to see humans acting (in this case) more fearful and more self-interested, and pursuing more power maximization, rather than less. Older versions of evolutionary theory sometimes presented strategies and behaviors as fixed or “hard wired.” Modern biology stresses the contingent, context-dependent nature of behavioral adaptations, which generates finer predictions for when we should expect to see different types of behavior.^{Reference Davies, Krebs and West155} This is an important point to which we will return.

Wherever actors are left to compete with each other by relying on their own devices (whether in the human evolutionary past or today), we predict that actors within those systems will exhibit similar behavior, not least self-help behavior to maximize power. Offensive realists can thus explain more than the behavior of states or great powers. If anchored on evolutionary theory, offensive realism allows new insights to elucidate why individuals and substate groups are self-interested, vie for power, and fear each other, and it can explain political behavior and war that occurred long before the creation of the modern state system in 1648.

Explaining leaders’ lust for power

The third contribution of our theory is that it identifies a more explicit role for leaders (see Table 3). Mearsheimer’s offensive realism argues that states gain power to ensure security. That certainly may be, as he attempts to demonstrate. However, an overtone of this argument is that power or domination is distasteful for leaders—that they tolerate it only for the sake of their state’s security. Leaders are forced to maximize power when perhaps they would rather cooperate or share power with others. By contrast, our theory posits that a tendency toward offensive realist behavior, however modulated by other tendencies, would have conferred a fitness advantage in the environment in which humans evolved and should thus have led to dispositions to seek and like power. Indeed, a wide range of empirical evidence from psychology and neuroscience suggests instead that humans, especially men, not only want to be leaders but also enjoy the pursuit of power (as well as its material fruits).^{156,Reference Robertson157,158} The force of this motivation is frequently revealed in victors’ expressions of the satisfaction of conquest. As Chinggis Khan is purported to have said: “The greatest happiness is to vanquish your enemy, to drive him before you, to see his cities reduced to ashes.”¹⁵⁹ Although not usually expressed in such stark terms, the pleasure of competition and victory has been widely recognized as a feature of human nature from classical times to the present day, and success in competitive interactions and the domination of others are known to increase testosterone and dopamine responses in men—the so-called victory effect.¹⁶⁰ Such dominance behavior is, we suggest, exaggerated among leaders because they are generally ambitious and competitive, and usually male. Moreover, the very acquisition and exercise of power itself is known to inflate dominance behavior further.¹⁶¹

Scope conditions: Human behavior is contingent, not fixed

We recognize that a challenge to the theory of offensive realism is the empirical mix of cooperation and conflict in the real world. Offensive realists and other theorists of international relations may see more or less of each. However, an evolutionary perspective is particularly useful here because it predicts that behavior is contingent, not fixed. Individuals may follow generalized decision rules, but these rules give rise to different behaviors in different contexts. Individuals fight when benefits are expected to exceed costs (on average), and not otherwise. Chimpanzees, for example, will attack others when they have a numerical advantage, but they will retreat if they are outnumbered.^{Reference Wilson, Britton and Franks162} This behavior makes perfect sense from an evolutionary perspective, because a decision-making mechanism that takes account of the probability of winning will spread at the expense of a decision-making mechanism that does not. As such, an evolutionary account does not necessarily expect animals, humans, or states to act as offensive realists all the time and in all situations. Offensive realism also does not have such expectations. Rather, as Mearsheimer points out, states do best if they expand only when the opportunity for gains presents itself—they “try to figure out when to raise and when to fold.”¹⁶³ Evolution has been doing this for a long time.

Our point is therefore not that humans are naturally “good” or naturally “bad” at all times and in all circumstances, but rather that people have evolved mechanisms for egoism, dominance, and groupishness that are activated and amplified in certain settings. Under conditions of anarchy, when there is the threat of predation and resource competition (as in many eras and locales in history), cognitive and physiological mechanisms of egoism, dominance, and groupishness are triggered. Where these conditions are tempered, such as in the modern peaceful democracies of Western Europe, these cognitive and physiological mechanisms are likely to be more subdued. For example, Western Europeans feel relatively secure (at least while the United States provides for their security). In general, humans cooperate where we can (e.g., within groups or within alliances deriving mutual benefit), but the anarchy of international relations is a hostile environment that, like the one in which humans evolved, tends to trigger our egoism, dominance, and group bias. In short, our theory is one of behavioral ecology—human and animal behaviors are not constants, but are contingent strategies that become engaged or elevated in order to best seek payoffs depending on the particular circumstance or environment. This insight has important implications for international politics because it suggests that we can potentially create—at least in principle—environments that take account of our human nature so we can turn them to our advantage, such as designing institutions that elicit cooperative rather than conflictual tendencies.^{Reference Keohane164,Reference Stein165}

Meet the family: Chimpanzees versus bonobos

Debates about evolved human propensities have often centered on whether human behavior more closely resembles the behavior of common chimpanzees or that of bonobos (pygmy chimpanzees that live in central Africa and are somewhat less aggressive than common chimpanzees). Such comparisons are not central to our argument. As we have stressed, the human traits of egoism, dominance, and ingroup/outgroup bias are adaptations to the ecological conditions prevalent in human evolution. We do not assume that humans and our primate cousins simply inherited these traits wholesale from a common ancestor. Clearly, when it comes to the many distinctive physiological and behavioral changes humans have undergone, ecology has been as or more important than phylogeny (hence, the field of evolutionary anthropology focuses on hunter-gatherer analogues, not nonhuman primate analogues). The particular socio-ecological setting in which humans evolved meant that egoism, dominance, and groupishness were important behavioral adaptations, irrespective of the traits found in related species.

Nevertheless, in evolutionary biology, the attribution of traits to common ancestry (a species’ phylogenetic history) can be important too. Therefore, to the extent that it matters, let us address the bonobo-chimpanzee issue briefly here, because certain phylogenetic and socio-ecological factors suggest that we are more like chimpanzees than bonobos.

The key observation is that bonobos are less aggressive than chimpanzees. However, it is important to make clear that humans did not descend from either species. All three species descended from an (unknown) common ancestor. Some evidence suggests that the separation between common chimpanzees and bonobos was quite recent, occurring perhaps only 0.86 million to 0.89 million years ago, although it remains possible that the separation occurred much earlier, between 1.5 million to 2.5 million years ago.^{Reference Won and Hey166} Either way, humans separated from our common ancestor with both chimpanzee species long before, about 5 million to 6 million years ago. In short, on the basis of the family tree, there is little reason to assume that humans should be more or less like bonobos or chimpanzees.

What is more important is the ecological differences and similarities that we shared with the two species. Humans and chimpanzees shared some features of their socio-ecological environment, such as spatially and temporally variable food resources, which required that individuals leave the protection of the group to forage in open or bordering areas, often alone or in small groups, subjecting them to greater risks of predation or ambush from conspecifics.^{Reference wrangham, Pilbeam, Galdikas, Briggs, Sheeran, Shapiro and Goodall167} In contrast, the ecology of bonobos has been relatively benign. They have enjoyed an absence of competition from gorillas (bonobos only live south of the Congo River, while gorillas only live on the north side of the river), high-quality foliage for food, and dense forest, which reduced vulnerability to ambush and thus, it is thought, the utility of aggression in males.^168,169 Accordingly, bonobos may not be a good model for understanding human behavior, for reasons of both phylogenetic history and shared ecology. Chimpanzees do at least have some important ecological similarities to humans. We reiterate the point above, however, that it is the context of our own evolution as hunter-gatherers in the socio-ecological conditions of the Pleistocene era that offers the crucial evidence on human behavioral adaptations.

What about the prevalence of cooperation in nature?

Given the considerable cooperation evident in the natural world, one might think that evolution provides a foundation for cooperative behavior rather than selfish, power-maximizing behavior. The fact is that evolution explains and predicts both (under the relevant circumstances). Indeed, there is a considerable literature on animal and human adaptations for cooperation.^{Reference Fehr and Fischbacher170,Reference Dugatkin171,172} However, while cooperation is frequent and widespread, this empirical observation does nothing to dent the evolutionary logic that cooperation helps the helper—it evolved to occur only where it brings return benefits.^{Reference Trivers173,174} This is precisely why the cooperation literature has remained so heated. Theorists have had to explain how cooperation could occur in the face of significant individual self-interest, the difficulties of collective-action, and the free-rider problem.^{Reference Boyd175,Reference Olson176,Reference Ostrom177} Special conditions are needed for cooperation to emerge and remain stable among unrelated individuals.^{178,Reference Sigmund179} Typically, those special conditions are ones that make helping advantageous to the genes responsible for the behavior.

Similarly formidable obstacles to cooperation exist in international relations. Cooperation is extremely hard to achieve and requires special conditions. Even optimists acknowledge that remarkable mechanisms and institutions are required to generate and sustain cooperation, and the identification and implementation of these conditions occupies many of our colleagues.^{180,Reference Milner181,182 183} The European Union, to give one flagship example, is often put forward as a vision of the future, demonstrating that democratic states are willing to subordinate self-interest for a greater good, and that war is becoming obsolete. However, the European project was set up precisely to respond to centuries of European powers competing and fighting for power at great cost. We are also yet to see how European states will cooperate or compete when the U.S. security umbrella is removed. The recent crises of the Euro and migration have shown in stark terms that individual states continue to exploit the opportunity to free-ride on others if they can, and even the most powerful states, such as Germany, France, and the United Kingdom, have been reluctant or unwilling to make sacrifices to protect other states. When the stakes are high and one’s livelihood or survival is threatened, the traits of egoism, dominance, and fear of outgroups come to the fore—a conclusion we can draw from any number of conflicts in the Balkans, Northern Ireland, Ukraine, Syria, Iraq, Nigeria, India, and elsewhere.

If intergroup conflict is so important, might evolution favor the most cooperative groups?

Evolution is sometimes argued to operate on groups rather than individuals (“group selection”). This idea is important because, if individuals are working for the good of the group rather than only for themselves, then groups composed of more-cooperative individuals may do better than less-cooperative groups, meaning that genuinely altruistic traits (sacrificing one’s own interests for the good of others) can spread in the population.^{Reference Wilson184,Reference Wilson and Sober185} However, there are several reasons why this possibility does not affect our argument.

First, group selection is a controversial hypothesis, which has been rejected by many prominent evolutionary biologists.¹⁸⁶ While selection at the level of groups is possible in principle, it requires special conditions to overcome what are generally agreed to be the much more powerful forces of competition and selection acting on individuals, and these forces are always in play whether groups are in competition with each other or not. Therefore, even the strongest advocates of group selection, such as David Sloan Wilson, argue that, in any given case, it remains an empirical question as to whether or not the selection pressures acting at the level of individuals are outweighed by selection pressures potentially acting at the level of the group (so called multi-level selection). While biological group selection in humans is possible in theory, there have not been any published empirical examples.

Second, even if group selection does occur, it can only increase altruism within groups. Between groups, group selection would do the opposite, maintaining or even exacerbating conflict.¹⁸⁷ Because the premise is that selection operates at the level of groups, altruistic traits can only spread if altruism helps spread the genes responsible for it at the expense of other genes, and that must occur via intergroup competition or conflict. Therefore, to advocate group selection over individual selection does nothing to reduce predictions regarding human conflict or aggression. If anything, group selection would tend to increase violence, since between-group competition (conflict among strangers) can be more brutal than within-group competition (conflict among kin and fellow group members). In Matt Ridley’s words, to prefer group selection over individual selection “is to prefer genocide over murder.”^{Reference Ridley188} Group selection can promote cooperation and altruism, but only within the group. The dark side of group selection is the unfettered conflict it presupposes and predicts between groups in competition with each other.

Third, it is important to remember that the empirical observation of altruism in nature does not imply or demand group selection. Given group selection’s theoretical constraints, it should be a last-resort explanation (subject to empirical testing), not a first point of call. As we have explained, there are several mechanisms by which altruistic or helping behavior can (and have) evolved because of the benefits of helping others that accrue to oneself—not least, altruistic behavior among kin, reciprocity, and reputation formation. In short, you do not need group selection to explain altruism. In the right contexts, helping others can help oneself. The strategic allocation of resources to others often advances one’s own Darwinian fitness. We do not need to resort to group selection unless altruism cannot be explained by more conventional mechanisms based on individual selection. George C. Williams famously made this point in response to so called naïve group selectionists of the time, and his insight has continued to be reiterated to biology students ever since.^{Reference Williams189}

While biological group selection among humans is unlikely, the selection of cultural traits among groups is possible. The constraints on biological group selection, such as significant differences in a given trait between groups and low migration, are relaxed in the case of cultural traits, since groups actively promote cultural distinctions and have many mechanisms to prevent flows between them.^{Reference Richerson and Boyd190} Therefore, it is not just likely but quite apparent that many cultural traits have evolved out of group-level competition—sometimes referred to as memes, as opposed to genes. (Examples include the spread of Christianity or Islam at the expense of traditional religions over the last 2,000 years.) However, once again, the potential for cultural group selection does not change or challenge our argument. To the extent that cultural group selection extends back into our evolutionary past, cultural traits have not been consistently or powerfully contrary to the evolved traits of egoism, dominance, and ingroup/outgroup bias. Indeed, cultural selection has often reinforced, not reduced, these very behaviors over human history. Moreover, and lastly, cultural differences have themselves represented an additional cause and consequence of conflict. In addition to fighting over resources, we can now fight over ideology as well.

Leaders and offensive realist behavior

A crucial extension of our theory concerns how leaders may differ from the average person in the population. Up to now, our claims have focused on traits that are common to all humans. However, there is, of course, considerable variation in egoism, dominance, and ingroup/outgroup bias. As we have been at pains to explain, much of this variation stems from contextual differences (behavioral ecology)—that is, a given individual’s behavior can change across circumstances. However, another important source of variation is individual differences—that is, specific people exhibit these traits to greater or lesser degrees. Individual differences are important because political leaders may be more likely than the average person to display egoism, dominance, and groupishness.

A state’s elites—its captains of industry and media and its military and political leaders—may be more likely than average to show these traits in abundance for five reasons. First, ambitious leaders self-select themselves into seeking high-profile roles in the first place.^{Reference Ehrenhalt191} Second, strong leaders are selected into power over weak-willed or hesitant candidates.^{Reference Todorov, Mandisodza, Goren and Hall192,Reference Van Vugt and Ahuja193} Third, leaders rise to the top of their respective hierarchies through an intensively competitive process that compels them to be increasingly attentive to self-interest and self-preservation.^{Reference Shenkman194} Fourth, once in power, decision-makers tend to heed hawkish rather than dovish advice.^{Reference Kahneman and Renshon195} Fifth, the experience of power itself is well known to corrupt, precisely because being a leader elevates one’s sense of worth and power.¹⁹⁶ Taking these phenomena together, a skeptic of our argument that humans are generally egoistic, dominance-seeking and groupish may nevertheless concede that the small subset of humans that become political leaders tend to express these traits. Again, the political world mirrors nature: Not everyone can be the alpha male.

Sex differences and offensive realist behavior

Men, more often than women, lead states. If women led them, or were better represented in legislative or executive branches of government, a logical prediction of our theory is that egoistic, dominant, and groupish tendencies—being primarily male traits—would be less likely to influence state behavior. A caveat to this prediction is that women in power may tend to act like men, either because selection effects trump stereotypical sex differences (female leaders may have personalities similar to male leaders), or because egoism and dominance are necessary traits in order to survive in the system of international anarchy (or on Capitol Hill).^{Reference Fukuyama197,Reference Clift and Brazaitis198}

Combining the previous two considerations (leaders and sex) raises another problem: If leaders are especially egoistic and domineering, and if sex is a primary cause, does this not mean that we predict state leaders will undertake actions (consciously or subconsciously) that serve to maximize their own personal reproductive opportunities—perhaps at the expense of state interests? Historically, evidence has often supported this hypothesis.^199,200,201 However, we take the position that, on average, state leaders’ personal interests have significant and genuine overlap with national security interests, not least of which is the survival and prosperity of the state for themselves and their progeny.

Levels of analysis: Individual and state behavior

A recurrent criticism of any theory of international relations based on the role of individuals is why we should expect individual behavior to tell us anything about state behavior. We recognize that many factors may affect the behavior of states, including bureaucracies, types of government, culture, international institutions, or the international system itself, but we also recognize, as traditional theories of international politics have from the time of Thucydides, that humans affect state behavior as well.^{Reference Levy202} Many factors come between an individual leader and the behavior of a state, but that does not mean leaders have no effect at all.

We see several reasons why human behavior is an important predictor of state behavior in the context of this article. First, the preferences of individual citizens are, at least to a degree, represented in those elected to—or tolerated in—office, and those preferences may also be seen in the goals of the state. Second, bureaucracies and organizations are designed, run, and led by human beings, whose own dispositions influence how they function. Third, state leaders are the actors who make important strategic decisions from a set of options, and they are potentially affected by their human dispositions and those of their advisers, even if their actions are tempered by checks and balances. Fourth, group decision-making may actually amplify the influence of human dispositions; it is groups of men that are especially prone to behaviors associated with dominance, aggression, and coalitionary psychology.^{Reference Baumeister, Boden, Geen and Donnerstein203,Reference Janis204,205,206,Reference Wrangham and Wilson207}

Novel solutions to weaknesses in offensive realism

We recognize that offensive realism remains controversial, and we would like to address three fundamental problems with the theory. Although it is not our intention to resolve offensive realism’s theoretical lacunae, an evolutionary account can help to explain them.

First, offensive realism fails to explain why costly wars sometimes occur against the interests of the states that initiate them. As Mearsheimer and others acknowledge, power maximization may not always be a good strategy and, indeed, it has led a number of states into disaster. Conventional offensive realism cannot explain such events well. However, if actors seek dominance at least partly because of evolved behavioral dispositions (of which actors may not even be aware), then we may expect sometimes to observe power-maximizing behavior whether or not it is a good strategy. Although we have stressed that evolved behavior is often contingent on circumstances, this matching is not perfect, especially when human decision-makers are faced with an evolutionarily novel environment—as witnessed today with mass societies, modern technologies, and interactions with distant peoples—for which the human brain was not designed. We should therefore expect instances of “evolutionary mismatch” in which evolved behaviors lead to poor decisions in modern settings. The brain may be responding exactly as it was designed to do, given informational inputs from the environment. However, while the resulting behavior may have been adaptive in our ancestral environment, it may be maladaptive, or even disastrous, today. Evolutionary theory is especially helpful here because it advances our understanding of the proximate (biological) causes of offensive realist behavior and the conditions under which mistakes are more likely to be made (i.e., conditions that exacerbate egoistic, dominating, and groupish behaviors even where such behaviors may not help to achieve strategic goals).

Second, critics of offensive realism point to countering factors such as the “democratic peace” or international institutions. However, an evolutionary perspective raises new doubts about the significance of such evidence. The very existence of these phenomena, not to mention the extreme efforts and expense they continually require to function, only supports the point that international politics needs very special and powerful arrangements to prevent people from acting as offensive realists—predisposed as they are to do so. And, even then, these arrangements often fail to work. When the stakes are high enough, individuals as well as states all too easily revert to egoism, dominance, and fear.

Third, critics point to international cooperation among states as evidence against offensive realism. We realize international cooperation is prevalent, but that does not mean such cooperation is easy to obtain. On the contrary, it is famously hard to initiate and maintain from both a theoretical and empirical perspective, which is why this topic continues to fill huge volumes of scholarly literature in economics and political science.^208,209 As we have emphasized, cooperation is easy to explain where it brings clear mutual benefits to the self-interest of those involved, such as trade or military alliances (in which case offensive realism is as good an explanation of cooperation as any other theory). Collective action to attain public goods, however, is much harder to attain because of the threat of free-riders (as demonstrated, for example, by the slow response to climate change, the reluctance of states to accept Syrian refugees, and Eurozone fiscal responsibility). Where extensive international cooperation does occur, it is often only by virtue of a hegemon willing to sustain it, and cooperation quickly breaks down if core interests and security are put at risk. Furthermore, cooperation is often itself a means to power maximization in the formation of military and security alliances—and thus, cooperation can be a prediction of, not a challenge to, offensive realism. Where a state’s own security is threatened or the state becomes vulnerable to exploitation, alliances offer one means of increasing or preserving power. Similar patterns are evident in nature. Animals do not constantly fight. Natural selection generates contingent behavior because it is more effective than blind aggression. Individuals bide their time, form coalitions and alliances, and cooperate with others, but they also seize power where the opportunity arises. States are much the same. In international politics, the bigger problem may be aspiring hegemons—states that do not need to cooperate to obtain what they want.