Subject: The Evolutionary Dynamics of a Crude Superorganism hersonManuscript2.htm"

The suffering which comes from [our relations to other men] is perhaps more painful to us than any other. We tend to regard it as a kind of gratuitous addition, although it cannot be any less fatefully inevitable than the suffering which comes from elsewhere.

Sigmund Freud, Civilization and Its Discontents, 1930

Uh-oh!

God, upon dropping a collecting bottle full of humans on Earth and seeing them escaping (according to cartoonist Gary Larson).

Draft 1.2, June1997. Please do not cite or quote without authors’ permission. Comments welcome!

1.0 Introduction

Humans were biologically modern by something like 100,000 years ago and for most of time since people lived in simple relatively simple tribal societies practicing a hunting and gathering mode of subsistence (Klein, 1989). How simple is a matter of controversy (Price and Brown, 1985; Kelly 1995; Arnold, 1993). Since even the simplest hunting and gathering societies known to ethnographers, such as the famous !Kung San and the peoples of Central Australia, lived in multi-band societies linking residential units of a few tens of people into societies of a few hundred to a few thousand people, we can be reasonably certain that such societies were considerably more complex than those of our ape ancestors. Around 10,000 years ago, plant domestication began to raise the human carrying capacity in several regions of the world. Agricultural societies became larger and denser that those of the Pleistocene, to which human social "instincts" are presumably adapted. These societies increased the scale and complexity of social organization compared to the average in hunting and gathering societies. Around 5,000 years ago the first truly complex societies arose. Great innovations in social organization developed to support the unprecedented levels of cooperation, coordination and division of labor. Some of these innovations, especially the development of deep hierarchies, were apparently quite novel and generate enormous conflict. People’s egalitarian impulses frequently rebel at the sharply stratified distributions of rewards in complex societies. Nevertheless, complex societies are generally able to dominate tribal ones, and a ragged but persistent trajectory of social evolution toward ever more complex social systems continues to the present.

The evolution of complex societies is one of the deepest puzzles of the social sciences. Great debates, with roots in the political thought of Plato, Aristotle, and Confucius, have raged over whether the evolution of such societies is voluntaristic or coercive, whether their operations are to be understood in terms of conflicts or functions, and whether the right unit of analysis is the individual or the social institution (e.g. Service, 1975; Carniero, 1970; Kirch, 1984). Sophisticated arguments have been marshaled on both sides of these debates for a century without any consensus being reached. The argument in this paper is that the antagonists in these debates are all right. Using the tools of evolutionary analysis, it is quite possible to construct hypotheses that mix the elements of classical positions. We don’t have to choose between accounts that are built on individual advantage, conflicts between interest groups, or upon societal functions. Each of these positions has something to say for it, on the evidence that many reasonable people still find each attractive! Theory shows that several evolutionary forces can act simultaneously on human populations, each favoring quite different behaviors. Much as Freud imagined, conflicts between evolutionary forces can easily result in conflicts at the population level—between classes, ethnic groups, and like—as well as in painful dilemmas for individuals.

Nor do evolutionary accounts have to choose between genes and culture as the engine behind social evolution in humans. Patently, for all their intelligence and close relationship to humans, chimpanzees cannot learn to behave as even marginally successful members of human society (Whiten and Custance, 1996; Tomasello, 1996). Noam Chomsky’s concept of innate principles and socially learned parameters has to fundamentally correct for all elements of behavior, not just language; children must come equipped to learn human culture, something that other apes cannot do. Humans must have social "instincts" that distinguish us from the other apes. By social "instincts" we mean the genetic and perhaps very conservative culturally coded elements of behavior, more or less universal in all populations. Wilson (1975: 26-7) analyzes two senses of the concept of instinct. In the more restrictive sense, the term means a completely innate behavior pattern which varies little throughout the life of an organism and little from individual to individual. Only limiting cases fit this definition. Far more commonly, innate (genetic) factors are a component of behavior which interacts environment and learning, defining "instinct" in a broader sense. In the human case, we have to contend with socially learned traditions. In the limit of rather conservative transmission from parents to offspring, cultural transmission can mimic the behavior of genes almost perfectly. Since it is not generally practical to distinguish between innate in the genetic sense and very conservative, vertically transmitted cultural universals, we lump conservative culture with genes into the concept of "instincts." We henceforth drop the quotes with the understanding that this definition is unorthodox, but more operational in the case of a culture-bearing organism than the orthodox ones. Those who dislike this terminology may mentally substitute any term they like for the evolutionarily more conservative elements of human behavior. We (Boyd and Richerson, 1985) have used the term "culture capacities." Whatever terms one prefers, we have no intention of privileging social instincts as some Darwinian deus ex machina that holds the dynamic elements of culture in its thrall. Rather, we will argue the opposite is often true; the evolution of more conservative elements may be guided or driven by the less conservative ones.

What is very much more difficult to show on present evidence is the division of labor between, and modes of interaction of, instincts and less conservative cultural elements. There is no consensus even in the well studied case of language (see Pinker and Bloom, 1991, and commentaries thereupon). Humans are complex bundles of skills, beliefs, attitudes, and emotions that arise from some mix of genetic factors, individual learning, social learning, and direct environmental effects. Psychologists have not yet reached any consensus on how the adult human mind is built. Thus, we are agnostic about whether there are many quite separate social instincts, as in the argument of Tooby and Cosmides (1989), or some other architecture. We label only a few instincts here, but proponents of the evolved mental modules idea are free to interpret these as large collections of inter-related domain specific modules. On the evidence of human ecological and social diversity, it seems that our species is an unusually generalized problem solver at the phenotypic and population levels, unrivaled by any other vertebrate. Flexibility at these levels may be supported by many domain specific mental organs deployed opportunistically, rather than by cognitive structures that are themselves generalized. Or the innate cognitive structures may be relatively generalized, with individual experience and social learning furnishing most of the functional details of specific skills, attitudes, and emotions (Bandura, 1986; Vygotsky, 1985; Collier, 1994). These are interesting and important details for the psychologists to clean up, but in the meantime the rest of us have to finesse the fact that we’re still ignorant of them. What are the social instincts (or classes of social instincts, if you prefer) that human populations possessed by the end of the Pleistocene?

In the case at hand, the interesting instincts are those which underpin the learning of culturally variable social behaviors that generate the great diversity of human social systems. What are the instinctive rules which cause us different from our ape ancestors? How did they evolve? We hypothesize (see Richerson and Boyd, in press, for details) that the evolution of human social instincts was substantially driven by cultural evolution, particularly cultural group selection. Cultural group selection can work in situations where genetic group selection would be a vanishingly weak force (Soltis, et al., 1995). In this scenario, cultural group selection set up a social environment that selected for individuals with more prosocial instincts. Because of the advantages of cooperation in economic activities and in inter-group conflict, marginal strengthening of cultural norms for cooperation, ever pressing the envelope of what was possible given the instincts of the moment, gradually selected for instincts that permitted ever more elaborate cultural norms of cooperation to arise. The role of cooperation in inter-group conflict set up an arms race that drove social evolution to ever greater extremes of ingroup cooperation lest the ability to defend against stronger opponents fail. This race apparently reached an equilibrium in the late Pleistocene under harsh glacial environmental conditions, but was renewed with vigor with the food production innovations in the Holocene. It is easy to see the last 10kyr as a race toward ever more complex societies, powered by the ability of large scale complex social organization to produce public goods like defense, and economic security, and intangibles like an interesting life-style.

This hypothesis argues that the existence of an instinctual proximal cause of a behavior is no guarantee at all that the ultimate evolutionary explanation for the instinct not to be found in the dynamics of cultural evolution. There is a careless tendency in debates over the role of genes and culture in human evolution to assume that the showing of an apparently innate component to a behavior means that the evolution of the character was ultimately derived from selection on genes in which culture played no active evolutionary role. It is easy to make models that show how cultural can play a leading role in the evolution of instincts, not a following one (Richerson and Boyd, 1989; Kumm et. al., 1995), and there is at least one good empirical case, adult lactose absorption. High frequencies of adult ability to digest lactose (milk sugar) in adulthood occur only in populations with a long tradition of dairying and fresh milk consumption (Durham, 1991:Ch. 5). Indeed, it would seem that cultural evolution ought generally to play a driving role in the evolution of cognitive capacities supporting any character with a proximal cultural component. This hypothesis is not terribly unorthodox. It is a special case of the long-mooted hypothesis that phenotypic flexibility plays an active evolutionary role (e.g. Wyles’ et al., 1983 and Wilson’s, 1985 case for "behavioral drive"). The mind as a population’s evolving slate is much more nearly blank than the mind of any given individual. Selection can eventually make something only dimly figured in ancestral minds into whole volumes of nearly blank pages (or new specialized instincts).

We also hypothesized that cultural group selection was laid directly over more ancient social instincts favoring cooperation with friends and kin. Human societies include important institutions deriving from kin and reciprocal altruism, but these work best at the smallest scale. It is always easier to trust one’s closest relatives and best friends than distant relatives and passing acquaintances. Thus, there is an inherent conflict build into human social life. Cultural norms, supported by evolved instincts that favor large group altruism, are often at odds with kinship and face-to-face reciprocity. This conflict could only have been resolved by drastically reorganizing human reproductive patterns to create large numbers of close genetic relatives, as in eusocial insects with the reproductive suppression of workers, or some functionally equivalent innovation. Insect style eusociality has arisen many times, including in one mammal, the naked mole rat of Africa. Human evolution might have taken that path to ultra-sociality, but did not (Campbell, 1983). Loyalty to one’s own reproductive interest, to close relatives, and to long-term reciprocators has a strong impact on human decisions, but so do larger loyalties to clan, tribe, class, caste, and nation generated ultimately by cultural group selection. Ideally, people like to do well while doing good, satisfying smaller and larger loyalties with minimal conflict. Arrangements to ensure this happy outcome are seldom perfect: Families are torn apart by civil conflicts, parents send children to war (or not) with painfully mix edemotions, criminal cabals arise to prey upon the production of public goods of larger scale institutions, elites take advantage of key locations in the fabric of society to extract disproportionate private rewards for their work, and so forth. Hypotheses of this type have a long history (Freud, 1930; Campbell, 1975, Maryanski and Turner, 1992), although we suppose (Richerson and Boyd, in press) that conflicts among social imperatives have roots reaching far further back than the dawn of "civilization,." where many commentators mark the beginning of serious social conflict.

By the Upper Paleolithic, human populations came to resemble the hunter-gathering societies of the ethnographic record in most attributes that are archaeologically visible. The first evolutionary question regarding the origins of complex societies is whether or not they could develop by purely cultural evolution using the innate rules generated by selection for simpler societies in the Upper Paleolithic. Simple societies are based upon ingroup cooperation where ingroups of a few hundred to a few thousand people are symbolically marked by language, ritual practices, dress, and the like. Leadership is generally weak and prosocial. People are quite ready to punish others for transgressions of social norms, even when personal interests are not directly at stake. Leaders will coerce followers, but also vice-versa. The evidence is that human psychology is adapted to life in relatively small, relatively egalitarian societies with a modest division of labor. If the answer to the first question is yes, then how exactly did cultural evolution use the productive potential of late Pleistocene humans’ instinctual social "grammar" to accomplish the feat? How did deep hierarchy, strong leadership, inegalitarian social relations, and an extensive division of labor arise?

If our hypothesis is approximately correct, and complex societies largely built on social instincts inherited from the Upper Paleolithic, some of the most telling evidence should be in the adaptations cultural institutions make to accommodate a psychology evolved to support rather simpler societies. We expect that modern complex societies should betray their ancient roots in the form they take. Like the bats’ or birds’ wings, the elementary structures of modern societies should give evidence of the previous function and be constrained by their history. The modifications that permit complex societies should be cultural work-arounds that leave the slow evolving social instincts largely intact. The institutions that support complex societies are to the social instincts what reading and writing are to the language instincts. On the one hand, they are revolutionary changes that permit the development of societies completely outside the range encountered in Pleistocene times. On the other, they must be tolerably adapted to the social instincts that evolved for life in small scale societies. Innovations that simultaneously make larger scale society possible while preserving (or recreating) the sense of living in a small-scale society will tend to spread. People will prefer such arrangements and will adopt them given a choice. Societies that possess such institutions will suffer less conflict between larger and smaller scale units and will tend to be more effective competitors. Where possible, the institutions of complex societies will come to simulate the social life of small egalitarian societies as far as the day-to-day lives of individuals are concerned. Given how unlike life in simple societies life in complex societies is, many of the institutional work-arounds aimed at preserving a sense of living in a small-scale society will work more or less poorly, giving rise to the inevitable conflicts of which Freud spoke. Considerable, if incomplete, evidence suggests that this picture is correct.

This logic works in reverse as well. The structure of complex societies are a very useful natural experiment to infer what the social instincts are. The work-arounds are evidence about the psychology that makes them necessary. What compromises with group-functional outcomes does cultural evolution use to deal with conflicts generated by minds only partly formed by group selection mechanisms and for life in relatively small, politically simple societies? What are the sources of conflict and other imperfections in complex institutions?

In this paper we outline the social instincts for which we think evidence exists, and then try to puzzle out how complex societies of the Holocene were build on the basis of instincts evolved to underpin very much more complex ones. These are large questions, and the following is a mere sketch to establish the plausibility of the hypothesis. We make no strong claims for the details of the story. We have more confidence in the form of the explanation. Somehow cultural evolutionary processes have interacted with evolved psychological dispositions and capacities to create modern complex societies.

2.0 Examples

It will be useful for the ensuing discussion to refer to concrete comparative cases of the evolutionary issues involved in complex societies. We pick two to illustrate the evolutionary issues from opposite ends of the spectrum of complexity, Polynesian chiefdoms and modern armies. The Polynesian example is useful because Polynesian societies span the range from quite small-scale ones that are only marginally if at all more sophisticated than Upper Paleolithic societies to incipient states. The data on Polynesia is unusually rich, and while the scale of social complexity varies over time and space in many other regions, Polynesia was quite isolated and social complexity arose endogenously on remote oceanic islands, not by diffusion from other complex societies. The armies of modern states are interesting because armies are one of the most important organizations of very sophisticated complex societies, because they demand extreme altruism on behalf of very large groups, and because in their degree of hierarchical command driven operation they are extreme even compared to other institutions of complex societies. If Polynesian chiefdoms represent the first steps on the evolutionary path from egalitarian to complex societies, modern armies are about as far removed from the institutions of small egalitarian bands as it is possible to get. The armies of WWII are especially interesting because their combat performance differed sharply for reasons that military sociologists have taken a lot of trouble to dissect, once again giving valuable comparative leverage.

2.1 Polynesian chiefdoms

The evolution of Polynesian political systems illustrates the origins of work-arounds required to generate complex societies. The archaeological record suggests that the Polynesian case is reasonably parallel to more ancient examples of the evolution of complex societies, for example in the early development of monumental architecture and status differences in burial grave goods.

As Sahlins (1963) noted, and Kirch (1984) described in detail, the Polynesian ranked lineage system uses kinship to staff the offices of a hierarchical system based upon the genealogical closeness of lineage leaders to the line of descent from the putative maximal lineage founder. The summit of political power is thus occupied by the nearest male relative, normally the eldest son, of the previous chief, who traces his ancestry back in like fashion to the leader of the founding lineage of the island population. In the basic system, all individuals’ places in society are ranked by their nearness to the senior male line, stretching back to the founding lineage. Kin altruism is thus exploited directly in support of the political system, arranging relatives around the seat of paramount power. In small scale cases, this system is just a particular normative mechanism for generating leadership barely less egalitarian than in the common run of simple societies. For example, in small scale Polynesian societies, chiefs cultivate their own fields, live in houses similar to everyone else, and fulfill their chiefly duties on a part-time basis. However, as a ranked lineage system grows on large islands, ever remoter relatives are incorporated, and kinship becomes a metaphor in an ideology of social order which will often conflict with demands of close kin. Chiefs become full-time political specialists with a retinue of subordinate retainers to transmit and enforce chiefly commands. In the most extreme cases, chiefs become equivalent in power and prestige to the kings of early medieval Europe.

The ancestral Polynesian system developed an ideology of supernatural power to enhance the prestige of chiefs. With inheritance of rank by seniority in the male line, there is obviously no guarantee that a given incumbent will be the most able individual, and personal charisma must be supplemented by a theory of power not tied to individuals. Human minds are apparently adapted to entertain supernatural ideas that can be turned to this purpose. Boyer (1994) gives a detailed argument on the naturalness of religious ideas. If his argument is correct, religious ideas that inspire real subjective senses of veracity and emotional commitment are rooted in our complex, evolved psychology. They can be hijacked for service in lending charisma to leaders who could not, perhaps, sustain it by personal qualities alone.

On Hawaii, the growth of chiefdoms to populations of a few tens of thousands of individuals caused a division of society into commoners, who were dropped from the ranked lineage system, and a nobility that maintained it. The higher ranked lineages became a symbolically marked group prepared to defend an ideology of intrinsic superiority of the noble class. The members of the lower nobility became a class of professional managers and soldiers, operating under chiefly leadership. Chiefly power was backed up by the monopoly of the means of coercion by the chiefly class. Even so, the Polynesian nobility was not free to merely constitute itself as a completely predatory conspiracy to exploit commoners. The Polynesian chiefly system was strongly paternalistic, and even among the Hawaiian nobility paternalistic duties were taken seriously. Chiefs were responsible for organizing the storage of foods in case of natural disasters, regulating fisheries, redistributing surpluses to the unlucky, coordinating a simple division of labor, and maintaining local peace. Failure to fulfill these duties led to popular unrest. Chiefly lineages were in direct, often violent, conflict with one another, and usurpation of chiefs by formally more junior lineages was an ever-present threat. Public opinion could weigh into these conflicts. Chiefs whose behavior was insufficiently paternalistic stood an increased chance of being overthrown.

Thus, by a series of cultural innovations, the simple Polynesian chiefdoms of small islands and early settlement on large islands evolved on large islands into much more complex social systems with a considerable degree of division of labor, coordination, and cooperation organized by the chiefly class. A mixture of paternal, functional attributes is mixed with an emerging class stratification and differentiation of status and wealth associated with the command and control hierarchy. Since the time scale of this change was only on the order of 1,000 years, it is not likely that the social instincts were much changed. What evolved were the cultural institutions necessary to create complex societies in the face of human psychological adaptations, for example egalitarian impulses, that were, on the face of matters, quite hostile to the emerging inegalitarian system.

2.2 WWII armies

Modern militaries, as we’ll see, are most successful when they can recreate many of the elements of the egalitarian tribal ethos notwithstanding the reality of highly formalized hierarchical command and control. The trick is to minimize the psychological difference between desperate drama of modern combat and the similarly desperate acts of offense and defense that were probably common in the intertribal anarchy before 10,000 years ago. (Those taught to believe that tribal life before the origin of states was peaceful must read Keeley, 1996.)

Military traditions in different countries have very different attitudes toward the leadership and formation of bodies of the bodies of troops that actually fight together. Their performance in WWII also differed significantly. Dupuy (1984, 1987) has conducted a quantitative historical analysis of relative combat effectiveness of soldiers that have faced one another in battle in WWII. Controlling for equipment, surprise, and many other factors as well as possible, Dupuy found large differences in the effectiveness of soldiers. In WWII, German per capita effectiveness was highest, followed by Americans and British with about a 20% handicap relative to Germans (all else equal, it would take 120 Americans or British troops to accomplish the same objective as 100 Germans). The German advantage relative to Russians was approximately 2 fold. This difference persisted until the very end of the war despite a steady diet of defeat after 1942 for German units, very heavy casualties, and a poor supply situation. Many indices of German soldier’s military effectiveness accord with Dupuy’s analysis. For example, comparatively few Germans from ordinary divisions surrendered when defeated, most soldiers making their way rearward, even in chaotic defeats that reduced the soldiers of other armies to helpless passivity. Similar factors apply to the relative advantage of Israeli troops over their Arab enemies in 1967 and 1973.

More ethnographic analyses of the performance of the German Army relative to the Western Allies were conducted by Shils and Janowitz (1948), van Creveld (1982) and Fritz (1995). Cockburn’s (1983) account of the Cold War Soviet Army is similarly detailed. Shalit (1988) gives an interesting account of the psychology of combat based on his studies as a member of the Israeli Defense Force. All observers, van Creveld and Fritz most explicitly, agree with Dupuy’s evaluation of the relative superiority of the German Army against all its enemies, and with the relative inefficiency of the Soviet Army. It cannot be objected, these analysts hold, that the German loss of the war is ultimately the test of the combat effectiveness of their army. Lose Germany did, owing to the greater numbers of their enemies, especially the Soviets, and their superiority in material, especially due to American production. Controlling for such factors, German troops were more effective fighters than the Allies. Four aspects of modern armies seem to explain most of the German advantage in soldier’s fighting power, modes of recruitment, training, leadership, and treatment of individuals.

German doctrine placed great emphasis on developing a sense of cohesion and solidarity among the members of small units (squads, platoons, and companies). To begin with, troops were recruited on a territorial basis, so the men served with fellows from the same towns and villages of their birth. Recruits were trained by elements of the same regiment in which they would serve in combat, and sent as a body to the front to make up losses from their regiment. A German soldier always served in the company of comrades, whose bonds of loyalty and fellowship were deliberately designed to cumulatively increase from the day of induction onwards. At the opposite extreme, the American system gave recruits basic training in temporary groups, and sent them on to advanced schools, and eventually to combat, as individuals. Socially isolated neophyte soldiers suffered considerable psychological turmoil during their long, lonely journeys to the front, and were disproportionately likely to become casualties in their first weeks of fighting. Not until actually incorporated into their final destination units could recruits begin to develop a sense of cohesion with any of the comrades with whom they would fight.

German basic training, conducted at the hands of non-commissioned officer drill sergeants, was extremely rigorous physically and mentally. However, this training was explicitly legitimated as preparation for battle, and instructors were typically fair as well as hard. After the beginning of the war, most instructors were veterans with whom the trainees would return to the front. Most drill instructors earned grudging admiration or better from recruits. Training of commissioned officers as well as NCOs emphasized the responsibility of officers for the welfare of their troops, and very often enlisted troops responded warmly to the paternalistic concerns of their field grade officers. At every level, soldiers were trained to seize opportunities and act on individual initiative, rather than await orders.

It is interesting read advice manuals for junior officers in modern Western armies (e.g. Malone, 1983), where leadership by remote command through a chain of command to troops, backed up by extreme coercion for disobedience, ought to be most manifest. Malone advises leaders of small units to display such traits as humility, justice, tact, and selflessness as well as more conventional military virtues as courage decisiveness, dependability, and loyalty. Under several different headings he encourages leaders to conspicuously recognize the contributions of subordinates, downplay their own roles, and defend their subordinates against unfair treatment from outside and from up the chain of command. Ideally at least, modern armies seem to expect leaders to behave much the same way an influential man might in an egalitarian society. More formal studies of bureaucratic leadership, of which military leadership is only a special case, emphasize similar points (Van Fleet and Yukl, 1986; Taylor and Rosenbach, 1992). According to Fritz and van Creveld, the German emphasis on the prosocial traits of NCO and field officer leaders along the lines Malone suggests was a good deal stronger than in the American Army of WWII. No doubt, the chain of command exists, and no doubt it is widely resented. Nothing is more familiar to all of us than complaining about superiors. Without dedicated small-unit leaders that can inspire common action by the same deft force of personality the informal leaders of simple societies use, it is doubtful that soldiers could be inspired to their customary desperate deeds.

American and British training and leadership practices were less meticulous versions of the German system, but Soviet practice was very different according to Cockburn. Recruitment and training were haphazard. In the Soviet Army, the NCO system was very rudimentary; lacking the long-service career NCO cadre that is one of the key components of German, British, and American armies. Commissioned officers were socially remote from recruits, who were informally ranked in terms of length of service (recruits serve 2 or 3 years). The effective face-to-face leaders of Soviet soldiers were typically young, inexperienced draftee junior officers. As a consequence Soviet units not only lacked effective small unit leadership, but the recruit experience engendered divisions between older and younger soldiers within units rather than the strongly felt solidarity of Western European, especially German, small units. Blind obedience to orders coming down from a remote high command was the rule, and local initiative was discouraged. German practice was at the opposite extreme in this regard. German orders to subordinates were drafted to emphasize the mission to be accomplished, with the subordinate expected to devise the means to accomplish the mission, right down to the level of squad and section leaders and individual enlisted men.

According to van Crevelt, the German command system was constructed to seem fair and just to individual soldiers by catering to their psychological needs for strong fellowship and sensitive, if tough, leadership. The German system went to considerable extremes to move the best officers to the front lines, at the expense of leaving rear areas under- and poorly staffed. German procedure greatly simplified reporting by field commanders in the interest of reducing the paperwork burden on fighters. The Americans had a much more manage-by-numbers approach, which tended to keep talent men in rear-area jobs, and which imposed a much larger paperwork burden on front line commanders. The German system for awarding medals was more prompt in its recognition of merit, and more strongly restricted to actual combat accomplishment, than the American system. The development of a very efficient field postal system kept soldiers in touch with their families, and hardship leave (e.g. when the family home was bombed) was common. The Wehrmacht thus went to considerable extremes to demonstrate an interest in a soldier’s personal well-being, minimizing conflicts between soldierly duty and personal interest. Significantly, Fritz describes, a German soldier was very unlikely to face extreme discipline from his normal superiors. In the catastrophic, anarchic retreats of the 1943-5 period, German Field Police units dealt summarily and extremely harshly with stragglers for relative misdemeanors, and the element of raw coercion in the system became very plain. In such circumstances, a soldier’s best option was to rejoin his unit as quickly as possible. Intact German units, even when very badly mauled, continued to be well led and well behaved and didn’t attracted attention from these dreaded detachments. German soldiers felt well cared for despite the fact that service in combat led to the death or maiming of most of them. Russian soldiers, contrariwise, were subject to extreme discipline by their own officers, with attendant lack of intense feelings of fellowship between superiors and subordinates.

There is debate over the role of Nazi ideology in explaining the effectiveness of WWII German soldiers. Shils and Janowitz substantially discount its impact relative to the details of training and leadership. Fritz argues that Nazi ideology underpinned soldier’s attitudes in several respects. First, the Nazi ideology of national solidarity was explicitly built on an analogy with the deeply felt front-line solidarity of small units in the face of a dangerous foreign enemy, a topic on which front-line WWI veteran Hitler could expound with genuine personal familiarity. Army service was endowed with an idealistic demonstration mission for the whole German Volk to a greater degree than in other armies. Fritz elaborates considerably on Hitler’s charismatic appeal to at least some considerable minority of soldiers. Second, the anti-aristocratic element in Nazi ideology served to reinforce attempts to bring officers socially closer to the men they commanded. At the end of WWI, dispirited German soldiers mutinied in large numbers, hastening the end of the war. The deep division between a largely aristocratic officer corps and working class enlisted ranks contributed to the mutinous behavior. No such mutinies occurred even in the last, desperate days of WWII. Van Creveld attributes this difference to a dramatic democratization of the German officer corps, including, much promotion from the NCO ranks, after 1933. Since the American and Soviet societies had ready access to egalitarian ideologies that could have served to reduce the social distance between officers and their troops, perhaps the most plausible explanation is that Hitler’s enlisted man’s experience in WWI gave him an intuitive sympathy for and appeal to ordinary soldiers’ motivations for fighting. In Hitler’s concept of soldierly solidarity, an ideological linchpin of his regime’s claim to legitimacy, happened to work rather well when applied to small units, and his glorification of it sustained soldier’s idealism remarkably well. And, he had the dictatorial power to enforce his ideals of leadership upon the traditional officer class. This power was exercised with sufficient vigor to result in von Stauffenberg’s 1944 attempt on his life, organized by aristocratic officers. In any case, average German performance was better than all but the best units of the American and British armies, whose organization, training, and leadership norms resembled that of the German Army in many respects. It does seem plausible that these differences in training and leadership could give rise to the large differences in per capita effectiveness measured by Dupuy.

It is highly ironic, as van Crevelt, an Israeli not likely to romanticize matters, notes, that a criminal, totalitarian nazi regime managed to find the most successful formula of the period for meeting the conflicting demands of command and control at the nation-state level with the need to provide feelings of egalitarian solidarity and just, prosocial leadership at the psychologically most salient level to ordinary soldiers.

3. New post-Pleistocene social instincts?

Darwin remarked (1874: 237) that "The American aborigines, Negroes, and Europeans are as different from each other in mind as any three races that can be named; yet I was incessantly struck, while living with the Fuegans on board the ‘Beagle,’ with how many little traits of character showing how similar their minds were to ours; and so it was with a full-blooded Negro with whom I happened once to be intimate." Darwin argued in the Descent that the presently civilized societies had once been barbarous, and that a great gulf in turn separated humans living in the simplest societies from the apes. He argued stoutly against the notion that the human races were separate species. All humans belonged to the same species and shared the same social instincts. Modern ethnography and archaeology reinforce the evidence Darwin had at hand. In many places around the world, the climatic amelioration of the Holocene has led to the evolution of complex ultra-social systems based on psychological dispositions that were apparently already present in the late Pleistocene. The pace of this development was complex, uneven, and rapid, as in the case of Polynesian high chiefdoms. Some people, the Highland Scots for example, had maintained rather simple societies right down to the last few centuries, although by Darwin’s time they seemed perfectly civilized. (The Scots-Irish emigrants to America arguably bear the stamp of their rough tribal ancestry to the present day, Nisbett and Cohen, 1996). Thus, the differences between living human populations in social complexity must be a result of what Darwin called "inherited habits" and we call cultural traditions, rather than differences in the organic structures of the brain or cultural universals.

Roger Masters (personal communication) argued to us that we underestimate the power of cultural processes to remodel human psychology, and that perceptible genetic changes may still be occurring in response to cultural imperatives. Years ago, C.D. Darlington (1969) wrote a history of civilization in terms of genetic changes in leadership capacities and the like. Daniel Freedman (1979: Ch.9) reviewed data suggesting large differences in neonatal behavior between major human subpopulations. Most commentators on genetic differences between human populations seem to have accepted that if such differences exist, they are causally prior to cultural differences. Masters (and we) hypothesize that the causal order should be reversed. If minor innate differences exist in the social behavior of different human populations, the differences are most likely due to different cultural environments. Lumsden and Wilson (1981) argued that genetic responses to cultural change could occur on a time scale of 1,000 years "in order to exploit highly efficient new culturgens" (though the specific model from which they drew this conclusion is flawed, Maynard Smith and Warren, 1982).

Some genetic responses to cultural changes do seem to have occurred in the Holocene, such as adult lactose absorption capacity mentioned above. If so, the psychology underpinning human ultra-sociality might have partly arisen in the course of the political evolution of complex societies, or at least been modified by such a changed environment. Perhaps Freedman’s differences between newborn behaviors derive from long-standing differences in child rearing customs, rather than the reverse.

Such speculations notwithstanding, we are sentimentally inclined to the psychic unity hypothesis holding that all contemporary and recent human populations are essentially identical in terms of genetically inherited psychological dispositions. Sentiment aside, it certainly seems that some rather basic psychological capabilities, such as for language, do not differ appreciably between ethnographically and historically known human populations. On Wiessner’s (1983; 1984) data, there is nothing primitive in Kalahari San uses of symbols and style to mark and manage social relations. Their reactions to stylistically alien arrows suggests the same cognitive and affective machinery as is mobilized in Tajfel et al.’s (1971) minimal group paradigm. The recreation in modern armies of small units of tightly bonded compatriots led by a prosocial NCO or junior officer suggests that modern institutions are adapted to Pleistocene instincts, not, yet at least, vice versa.

A bit of psychological evidence in favor of the psychic unity view comes from the cross-cultural work on "cognitive style" (Berry, 1976). The middle class of modern industrial societies have returned to a "frame independent" cognitive style that is also characteristic of living hunter-gathering peoples. People scoring high on frame independence are able to solve problems like adjusting a lighted rod to vertical in a dark room, when the frame around the rod and the adjusting viewer are both tilted. People with "frame dependent" cognitive styles tend to use the tilted frame for reference, rather than separating the image of the frame from their internal sense of the vertical. In agrarian societies and in the lower classes of industrial societies, people tend to be frame dependent. A good case can be made that cognitive style is important to social behavior, and that the present high frame independence of middle class subjects is a recent cultural adaptation, mediated by child rearing practices, to the need for greater independent thinking in professional roles. Werner (1979) argues that hunting and gathering demanded a similar degree of independent thinking, whereas agrarian societies, often with authoritarian political structures, required more blind obedience to convention and discouraged independent thinking. Interestingly, frame dependent children perform better on cooperative tasks than frame independent kids. It is difficult to believe that large changes in cognitive style that presumably occurred in the past two centuries, in a direction in which Hunter-gatherers are already extreme, are mainly a result of genetic rather than cultural changes, especially if it amounts to recovering a style of thinking that disappeared several thousand years ago in the lineage of most human societies.

Ethnographic evidence in favor of psychic unity is also strong. People derived from the same linguistic and presumably genetic stock participated comfortably in quite radically different social systems, as Darwin knew. The Uto-Aztecan speakers of North America included the very simple societies of the Great Basin as well as the Aztec state in the Valley of Mexico. The Celtic and Germanic speakers of the Roman period ranged from fully urbanized citizens of the Empire to the rude tribal societies in the remoter parts of the British Isles and continental Europe north and east of the Rhine. The tribal barbarians that were either incorporated into the Empire by conquest or who crossed for other reasons were substantially Romanized within a few generations (e.g. Germans, Musset, 1965).

It is difficult to rule out small adjustments in human psychology in response to changed socio-cultural environments over the course of the evolution of complex ultra-social societies during the last 10,000 years. There seems to be a fairly large amount of genetic variation for many personality traits and other attributes of human minds (Eaves, et al., 1989), and 400 generations is enough time for microevolutionary responses to selection to occur. In theory, cultural evolution can both evolve work-arounds that obviate the need for a genetic response to a novel environment, or can so change environments as to expose latent genetic variation to strong selection (as must have happen in the case of the evolution of adult lactose absorption). However we think that the weight of evidence, along the lines of the small sketch just given, suggests that societies can rapidly move up and down the scale of social complexity, especially when diffusion political ideas is possible, demonstrating that cultural evolution must play the dominant role. The Upper Paleolithic social "principles" were already sufficiently productive to generate a very great variety of quite complex human societies.

4. Ancient social instincts in complex societies

Since the small-scale elements of human population structure are not drastically remodeled from those of our ape ancestors, the classic social forces of animal sociobiology should still act in human societies. We still live in a family structured social system with a small handful of close kin. We still conduct a retail social life with non-kin, developing trusting relationships with a few other individuals based on extensive personal interactions over a long period of time.

4.1 Kinship

Kin cooperation is the most important social cement most animal societies for the reasons outlined by Hamilton (1964). There is ample evidence, reviewed in Richerson and Boyd (in press) that kinship gives rise important loyalties that more inclusive social norms must deal with in one way or another. In this regard complex human societies have the same problem as simpler ones, but perhaps in exaggerated form, since the relative importance of other ties increases in complex societies. Still, in most societies up to the industrial revolution, the family was the basic unit of production, as well as of procreation. Even in industrial societies, there is apparently no substitute for the family. In the industrial West, family breakdown is associated with a host of pathologies; in the East, attempts to restrict family autonomy, as in Rumania, led to plummeting birth rates. The use of draconian means of indoctrination and coercion to suppress family and friendship in order to build utopian states seems always to lead to severe psychological distress of its victims. More benign social experiments with social forms that conflict with normal family life seem always to fail. The Israeli kibbutz experiments with communal child rearing and the fate of Shaker attempts to create a persistent celibate community through adoption are examples (Brewer, P.J., 1986; Spiro, 1979). Of course, family structure varies enormously in the human species, but kin ties are always important. If there is no human kin altruism instinct, utopian tinkering with family arrangements should be more successful. The existence of nepotistic impulses is something that the institutions of complex societies can use as a building block, but they also constrain their evolution.

Fictive ties of kinship and kinship metaphors are very frequently incorporated into the ideology of complex societies, for example when paramount chiefs, monarchs or dictators are likened to father figures. The Polynesian chiefdoms are good examples. The role of kinship metaphors in Confucian though is an example from a quite complex society (Creel, 1949). We are inclined to think that such metaphors are quite shallow, on the evidence reviewed in Richerson and Boyd (in press) that real kin sentiments are built by the robust mechanism of living in close proximity during a sensitive period. Nevertheless, the success of such ideological constructs in tapping real social instincts bears close investigation.

4.2 Reciprocity

Friendship is a similar matter. Virtually all organizations in complex societies expect face-to-face relations frequently to develop into friendships. Much of people’s ability to carry out their assigned functions in complex organizations would seem to depend upon a network of friendships, others whom one trusts to return favors. At the face to face level, the organizations of complex societies try to re-create many of the features of life in a hunter-gatherers camp. Almost always, there is a face-to-face unit of 5 to 20 adults at the base of a complex organization that is of the same scale as a hunter-gather’s band. The modern infantry squad of 10 or so men is an example. In such groups, it is theoretically plausible that reciprocity can play a strong role (Boyd and Richerson, 1988, 1989). Generally, office parties and similar events are used in an attempt to encourage friendly relations among co-workers. In situations like the successful combat military units of our example where dangerous acts of altruism are routinely expected, very high levels of trust and very deep emotional bonds of comradeship are established by creating a "tribal" ethos that includes bonds individual friendship the primary social cement at the small group level. Formal training and informal mechanisms of socialization grow up to make such bonds routine.

On the negative side, the soap opera of office politics and social games are a costly overhead that organizations must bear to maintain a functional social environment. The larger society is often the victim of predatory cabals and conspiracies organized by closely bonded individuals. Criminal gangs, including those of white collar conspirators, are generally very small organizations in which reciprocity, unaided by any other principle, is the chief cement. Mafias on a larger scale need additional principles of social organization, such as ethnicity to exist at all. During WWII the German Air Force, quite unlike the Army, was riddled with venality, corruption and incompetence that crippled its fighting ability (Irving, 1973).

Human families and friendships are so familiar that we are apt to overlook their significance. Since the conflict between smaller and larger loyalties is an acute problem for virtually all ethnographically known societies, the dependence of ultra-social institutions on these primitive bonds is a strong test of the importance of social instincts in human social life. Utopian reformers, even the Soviet Bloc Marxists with vast resources for persuasion and coercion, were unable to root out such sentiments. It is striking that the rapidly growing religious fundamentalisms in the world stress familial values and face-to-face community in competition with more secular social institutions. It is not clear the extent to which fundamentalist faiths are growing because of the simple demographic results of pronatalism or because of conversions from the secular population or from secularized religious organizations, but growing they are across a large spectrum of societies. In the US, it has been possible to estimate the gain and loss rates to different types of denominations. Conservative Protestant churches are currently growing due to relatively high birth rates, low rates of loss to other denominations and to secularism, and a small intake of converts from other persuasions (Roof and McKinney, 1987). It is plausible that such organizations have found better ways to lessen the conflict between bonds based on reciprocity and kinship and those based on larger loyalties than have more secular ones, explaining their retention of members in the face the powerful attractions of higher status liberal denominations and of secularism. Put the other way around, the low birth rates of the liberal-church and secular middle classes of modern urban societies illustrates how social institutions can arise which conflict with reproductive success (Richerson and Boyd, 1984) and probably the social instincts which subserve reproductive success in most societies (Borgerhoff Mulder, 1988). The relatively high birth rates of fundamentalists will eventually undo the demographic transition in the US if present trends continue.

5. Old social instincts in complex societies

Our hypothesis (Richerson and Boyd, in press) holds that over the course of human divergence from the other apes at least two new social instincts, ingroup cooperation and moralistic punishment of transgressors of social rules, arose under the influence of group selection on cultural variation. Human life in Late Pleistocene and ethnographically observed simple societies involves membership in symbolically marked ingroups. Even when hunting and gathering people live most of the time in small bands or tiny villages, they regularly gather in celebrations and other affairs with far larger units, typically comprised of many hundreds to a few thousand members of the same ethnolinguistic unit. A measure of willingness to cooperate with symbolically marked others can be demonstrated in the psychology laboratory, and ingroup cooperation and outgroup hostility are very commonly observed in simple societies (Keeley, 1996, reviews the archaeological and ethnographic evidence for intergroup warfare in simpler societies). Both cultural group selection (Boyd and Richerson, 1982) and symbolic marking of population boundaries (Boyd and Richerson, 1987) could easily arise for ordinary adaptive reasons, having the propensity to favor group selection as a byproduct. Once the conditions are met for group selection, an arms race will ensue because ingroup cooperation positively affects intergroup competition, whether violent or peaceful. In such a cultural environment, over tens of thousands of years in the late Pleistocene, selection favored people who were adept at recognizing ingroups, and responding to the demands of the ingroup for contributions to public goods, such as defense.

5.1 Moralistic punishment

Physical coercion is frequently used to structure social organization in many animals, witness the common dominance hierarchy. By the late Pleistocene, human dominance evolved into a moralistic system in which both leaders and followers enforced norms collectively recognized as legitimate by most concerned. Rank orders do certainly form in simple human societies, but are substantially based upon face-to-face interactions in which dominant individuals generally display personal characteristics that subordinates find admirable and culturally legitimate. In general, leaders of human groups, as is sometimes the case in animals, receive deference because of their ability to mediate group disputes, provide specialized knowledge, and the like. Seldom, except in small children, is dominance based upon pure physical ability as it tends to be in animals (Eibl-Eibesfeldt, 1989:297-314).

In theory moralistic punishment, in which third parties will coerce rule-breakers, including those who fail to help coerce transgressors, is a very potent mechanism to maintain any sort of norm that becomes established (Boyd and Richerson, 1992). In our models, populations evolve in which there are a few leaders willing to coerce or cajole a larger number of reluctant cooperators into behaving as moralists. These results are highly reminiscent of the political situation in simple societies where relatively weak, prosocial leadership serves to find and implement consensus based upon customary rules and to coordinated the punishment of transgressors.

Great disparities in material condition are seldom tolerated in simple societies. Sometimes extremes of egalitarian behavior are expected from the most successful individuals, as in the even division of meat in the Ache band (Hill and Hawkes, 1983). Most often, leaders of simple societies are coerced by followers when they disobeyed rules or tried to aggrandize themselves, leading to highly egalitarian societies with weak, prosocial leadership (e.g. Jorgensen, 1980: 209ff; Boehm, 1996). Leaders can achieve some influence by force of personality, but could rarely command without the consent of public opinion, in conformance with culturally sanctioned norms. We saw how this principle carried over to constrain Polynesian Chiefs, even in the most complex examples. Salter (1995) describes the overwhelming importance of sanctioned rules in most cases of human dominance in complex societies. Only in situations where such rules are likely to be very weak or absent is dominance mainly a matter of physical coercion (night club bouncers are Salter’s example). Although Salter finds that many dominance situations in complex societies exalt dominant individuals (e.g. the costume and high seat of a law court judge), these devices symbolize sanctioned rules the judge is charged with enforcing, not the judge’s personal power to coerce.

There is some experimental evidence that righteousness and envy are important forces in the evolution of complex societies. The experiments of Insko, et al. (1983) involved setting up two types of laboratory micro-societies, one based on voluntary trade and the other on forced exactions by an elite (their "Carniero" treatment). These experiments were designed to test voluntaristic versus coercive hypotheses for the formation of complex societies. The micro-societies were allowed to evolve as subjects were removed one by one and replaced with new naive subjects. Among many interesting results, the coercive treatment spontaneously generated a lot more "strike, sabotage, and slowdown" behavior on the part of the subordinate groups, much less satisfaction by subjects with the experiment, and a lower rating of the legitimacy of the leadership of the dominant group compared to the voluntaristic treatment. The overall productivity of the coercive treatment groups was also lower; group selection would have favored voluntaristic societies in this microcosm. Dominant human individuals often abuse their official powers or exercise power by direct coercion over those who do not recognize the exercise as legitimate. Insko et al.’s experiments are consistent with the common observation that such dominance will bring only grudging compliance, and that only when avenues of more active resistance are kept costly.

Experiments by Nuttin (1975) and Milgram (1965) show how responsive people can be to moral suasion by weak, apparently legitimate, authority. Recent experiments by Ernst Fehr and colleagues(Fehr and Tyran, in press) show that in a repeated gift exchange game that contributions to the public good decline rapidly from a base rate of 50% contribution without the ability to punish partners. With the ability to award punishment points (at a considerable cost to oneself) to under-average contributors, public good contributions stabilize (strangers) or rise (known partners) from the base rate. It is thus plausible that human social instincts include a sense of righteousness, a tendency to envy social superiors, and a willingness to act as 3^rd party punishers of both equals and superiors who violate (culturally defined) social norms.

It seems a commonplace observation that righteousness and envy are important social forces in complex societies, and that people are more satisfied with collective institutions that they view as legitimate and fair. Some data support this observation. Studies of reported happiness in European societies (Inglehart and Rabier, 1986; Easterlin, 1995) show that people are happiest in countries like Denmark, Holland, and Switzerland, but much lower in countries like Germany, France, and Italy. Happiness in European countries does not vary much as a function of intuitively obvious variables like income per capita. Rather, the best questionnaire correlate of happiness differences between nations is level of interpersonal trust. People dislike living in societies with relatively low interpersonal trust.

Putnam (1993) makes a similar point for regional subsocieties within Italy. There is a big difference between the North and the South, with more functional communal institutions based on interpersonal trust in the North. Fukuyama (1995) argues that the institutions that generate trust vary sharply between nations. In China, for example, the extended family but not large scale voluntary organizations generate trust, whereas in the U.S. and Japan the family is weaker and voluntary organizations much stronger. The consequence of this difference is that Chinese business organizations are largely organized around a founding family, whereas in Japan and the US the corporate form of organization with promotion based on talent rather than family ties is much more common. Put differently, humans appear to prefer to live in societies in which legitimate rules binding both subordinates and dominants guide social life, rules which every individual can invoke to righteously punish transgressors. In complex societies as well as simple ones, living with fellow-citizens or leaders that one doesn’t trust to live according to a common set of such rules is stressful.

The psychology of military service illustrates the power of fear of social approbation and positive effects of trust. The German Army was evidently quite successful at inducing individual soldiers and officers to subscribe to a common concept of correct behavior which kept soldiers fighting effectively without much need for much direct coercion. Leaders trusted subordinates to take proper initiative without detailed orders and subordinates expected to receive fair treatment from superiors. Shalit’s (1988:11) psychological study of Israeli Army personnel includes some interesting figures on what combat veteran troops fear most about battle. Although many officers and men rate fear of death, capture, and serious injury as their worst fears, fear of failing to conform to expectations of peers and social norms is at least as important. While 27.5% of officers feared death, injury, or capture most, 70.7%, most feared behaving badly, with "letting dependents down," 42%, being the most important social fear. Among enlisted men, 50.5% feared death, injury or capture most, but the most common worst fear, 40.4% of respondents, was "letting comrades down."

5.2 Ethnocentrism and allied symbolically marked ingroup favoritism

Theoretical models suggest that the primitive function of symbolic marking is to divide human groups into semi-closed populations that can evolve specialization to particular environments, a sort of cultural pseudo-speciation (Boyd and Richerson, 1985: Ch 7; 1987). The mechanism which favors the erection of symbolic boundaries between groups, which we called indirect bias, involves an episode of cultural transmission in which individuals choose whom to imitate based upon some indicator character, such as dress style or dialect. In the model of a group marker system, the indicator character becomes correlated with an adaptive character, such as the mix of cattle and plant production in an agrarian subsistence system. Empirically, human densities rose substantially and the range of the human species expanded in the late Pleistocene when human symbolic culture became visible in the archaeological record (Klein, 1989; Bettinger, 1992:203-4). Secondarily we believe, symbolically marked groups came under the influence of cultural group selection to make such groups the focus of altruistic institutions, leading to the ingroup favoritism and outgroup hostility characteristic of the ethnocentrism complex (LeVine and Campbell, 1972). Ingroup favoritism likely coevolved as a social instinct in response to this process. Social psychologists (e.g. Tajfel, et al., 1971; Rabbie, 1991) can easily obtain ingroup behavior by symbolic marking of laboratory groups. Ethnographically, sentiments of tribal solidarity often generate quite large scale activities, such as the war parties of egalitarian societies, reaching a few thousand people under ideal circumstances.

Given the common newspaper headline about inter-ethnic violence, the survival of symbolically marked groups as a principle social institution is unmistakable. The main resistance to West European, American, and Soviet imperialism in the 20^th Century was organized around ethnic (and similar) loyalties. The rise of the Solidarity movement in Poland is one of a long list of examples. The epiphany for American social scientists on the durability of ethnicity as a social institution is marked by the publication of Glazer and Moynihan’s (1975) book Ethnicity: Theory and Experience. The expectation that institutions of social organization based upon bonds of common descent and similar traditional forms would be replaced by more modern concepts of class and interest groups turned out to be greatly exaggerated. More recently, the rapidity with which ethnic principles of social organization came to prominence in the wake of the collapse of Communist states in Eastern Europe is testimony to their ability to survive generations of repression. If our hypothesis (Richerson and Boyd, in press) about the function of symbolically marked groups in the Upper Paleolithic is correct, ethnocentrism is a natural mechanism to bond together large societies where cooperation, coordination, and division of labor between strangers and near-strangers is required. If people can be convinced that they are members of a group that has a culture in common, high rates of affiliative behavior can be elicited in large groups. Literal ethnicity is not the only basis for eliciting a sense of group loyalty, but it provides a ready-made set of symbolic identifiers for large groups, and hence is commonly exploited for such purposes. In the past, multi-ethnic empires often ruled through ethnic institutions, granting ethnic organizations formal rights and demanding certain duties of them. Often, but not always, large scale societies acquire nationalist institutions that portray the whole society as an important marked ingroup. Universalistic religions are in principle even more inclusive.

The indirect bias mechanism of choosing whom to imitate is rather analogous to mate choice sexual selection. It has the same potential to support the evolution of exaggerated characters via the runaway coevolution of preferences and indicator traits, or via the handicap principle, in which signaling characters become exaggerated to guarantee their honesty. Displays of conspicuous consumption may arise because of a completely pathological runaway keep-up-with-the-Jones effect. Or it may be that there are individually important reasons to let others know how wealthy, well-educated, etc. one is. In the situation where individuals can gain an advantage by falsely communicating wealth or education, such information can be reliably communicated only by conspicuous consumption or evidence of mastery of otherwise unnecessarily arcane arts like Chinese ideographs or English spelling. Driving a luxury sedan and spelling English akuratly really do require significant investments of money and effort.

The need to exaggerate displays of symbols is plausibly responsible for the creation of vastly elaborated, intricate symbol systems, such as language, art, and religion. According to one recent account, belief in the supernatural is a result of innate structures of human cognition (Boyer, 1994) involving the indirect bias process. Whether or not it is correct to speak of an instinct for the supernatural, there is certainly no doubt that simple human societies maintain complex, sophisticated, and highly variable supernatural traditions (e.g. Barth, 1987). In many cases, anthropologists and sociologists have attributed either individual (e.g. Jensen and Eriksen, 1979) or group level functions (Rappaport, 1979) to religious ideas. We argue (Richerson and Boyd, 1989) that supernatural ideas are a mechanism which can finesse other evolved predispositions. However much supernatural ideas are shaped by innate cognitive constraints, the variety of supernatural beliefs is immense. Even Boyer, in invoking the concept of abductive reasoning to explain religious ideas, allows for them to incorporate otherwise quite improbable beliefs. Abduction is a form of conjectural empirical induction in which it is satisfactory that the explanation is sufficient for the particular case or narrow class of cases. Abductive inferences are not treated as deductive generalizations which must fit a wide class of phenomena in their domain. Circular evidence is counted in favor of abductions. For example, if past examples of disease have been explained as the result of witchcraft, that strengthens the case for using such explanations in the present instance. In simple real-world cases, sufficient evidence accumulates perforce to correct false abductions (a child is disabused of the conjecture that climbing into its highchair will produce food), but it is in the nature of supernatural conjectures that decisive disconfirming evidence hardly ever forces itself on individuals’ minds. To the extent that supernatural beliefs are public and shared, evolutionary forces such as indirect bias and conformity will tend to fix quite arbitrary beliefs in a population. By such means, a culture can come to have deeply held beliefs despite the weakness of evidence considered from cannons of empiricism more familiar to scientists and logicians.

The deep commitment to beliefs which vary dramatically between societies generates between-group variation, on which cultural group selection is likely to act. Even if a cooperative norm originally arises as a rare variant prestige marker, group selection will increase the frequency of such norms in the long run. This picture appears to roughly fit religious systems, where a complex, group-specific set of supernatural beliefs includes and sanctifies a few prosocial beliefs and norms. Religion and other forms of ideology certainly play a conspicuous role in complex societies. The nationalisms of the first half of the 20^th Century and the fundamentalisms of the last half are examples.

Indirect bias also sets up an intensely competitive system, in which individuals able to make a better display of the currently popular symbols may attract disproportionate imitation. The hierarchy of prestige is typically viewed by a society’s members as more legitimate than any inequalities based on luck, physical prowess, or cunning (though elements of these figure in many prestige systems). The extremes of conspicuous personal accomplishment give a few individuals a charismatic appeal to their fellows far above ordinary people. Even in some tribal societies, such as in New Guinea and the North-West Coast of North America, competitive prestige exaggeration overcame the egalitarian impulses typical of small-scale societies, giving rise to societies led by politico-economic "bigmen." Prestige systems and charismatic achievement are among the most important principles ordering social life in complex societies and are one of the central topics of sociology especially. In Weber’s famous Protestant Ethic argument, Calvinist ideology made the vast inequalities of wealth and power spawned by the nascent industrial revolution seem legitimate signs of God’s favor, even as it motivated the wealthy to reinvest rather than spend their fortunes. The secular echo of such beliefs still persist in the US especially, where making a fortune by one’s own hands still invokes a large measure of popular respect as well as the usual envy.

6. Creating Complex Societies: Hierarchy and Stratification

While there is ample evidence that ancient and old social instincts still operate in complex societies, it is also patent that a series of cultural innovations were required (and continue to arise and spread) to create complex societies. The origin of agriculture 10,000 years ago provided the economic potential for large societies with great common benefits to be realized by a complex division of labor, and cooperation and coordination on a large scale. But such advantages could be achieved only by means of social arrangements that are difficult to reconcile with our social instincts. During the Holocene, social evolution has gradually and haltingly lurched upward on a scale of sociopolitical complexity over large parts of the globe. (As we all know, this scale cannot be easily translated into a scale of moral progress, so beloved of our 19^th Century forefolk, as Bloom, 1995, trenchantly observes.)

The most important cultural innovations required to support complex societies are communication, command and control institutions that can systematically organize cooperation, coordination and a division of labor in societies consisting of hundreds of thousands to hundreds of millions of people. Symbolically marked, acephalous tribes can extend cooperation beyond face to face relationships (e.g. Nuer warfare, Kelly, 1985), but only marginally, at least by comparison with complex societies. Complex systems seem to universally develop social stratification in which objective material well-being and culturally defined prestige varies greatly by social role. Those in key positions in the command and control system acquire a disproportionate share of society’s rewards. There is every evidence, as we have seen in the last two sections, that humans are not prepared by their evolutionary experience to tolerate more than the most minimal command and control institutions.

In the face of a human psychology that willing gives only limited deference to leaders and grants only limited ingroup status to strangers, an increase in the scale and sophistication of command and control institutions beyond a very rudimentary level requires major cultural work-arounds. The cultural evolution of social institutions will favor those societies that achieve an efficient division of labor and large scale cooperation with minimal conflict with the ancient and old proclivities that adapted us to live in relatively small, relatively egalitarian communities. Successful cultural innovations will tend to take advantage and use the ancient and old social instincts as building blocks of more complex institutions. Societies that achieve relatively effective arrangements of this type will tend to outcompete those with less successful institutions. Individuals and groups are also likely to attempt to invent and borrow institutional forms that better satisfy oft conflicting impulses regarding existing social arrangements. We suppose that conflicts between individual fitness and ingroup norms are a legacy of the conflict between old and ancient social instincts, subject to new strains as ingroups explode in size and internal complexity. In this section we describe what seem to us to be the main work-around mechanisms, and the conflicts, compromises, and modes of failure that each entails.

6.1 Coercive dominance and exploitation

The cynics’ favorite mechanism for creating complex societies is command backed up by force. The conflict model of state formation has this character (Carniero, 1970). A society successful in war upon a neighboring group, if the defeated cannot flee, as farmers often can’t, can impose itself as a ruling class on the defeated. Conquerors frequently desire to set themselves up as the exploitative elite subsisting on the economic activities of the subjugated. They may also desire to monopolize the reproductive activities of conquered women (Betzig, 1986) and monopolize prestigious roles. Certainly, long stretches of human history (e.g. Grousset, 1970) seem to amount to little more than the repeated conquest of sedentary farming populations be conquering regimes of diverse sorts.

Elements of coercive dominance are no doubt necessary to make complex societies a going concern. Behind law and custom stand ultimate extreme sanctions for transgressors. Social insect colonies contrast sharply with humans in this regard. Eusocial insect individuals are "programmed" to respond to various cues in the nest, and coordinated behavior at the colony level arises spontaneously (e.g. Page, 1995). With minimal propensities to individual selfishness, "trust" is nearly absolute, and coercion and command can be dispensed with entirely. It is not that social insects are generally peaceable; eusocial honeybees raid each others hives and virgin queens engage in sororicide. The quasi- and semi-social insects have sharp dominance contests between potential reproductives (Wilson, 1975: 418ff). If command and control were necessary for routine intra-hive social activities in eusocial species, it would have likely developed by the elaboration and modification of the dominance systems of sub-and semi-social ancestor species, much as it has in humans. In our case, selection at multiple levels makes conflicts between higher and lower level entities inevitable. In complex societies, the moralistic solidarity of tribal societies has to be supplemented with formal police institutions if larger scale public goods are to exist in the face of selfish temptations to expropriate them by individuals, nepotists, and cabals of reciprocators, organized predatory bands, and classes or castes with special access to means of coercion. At the same time the need for organized coercion as an ultimate sanction creates roles, classes, and subcultures with the power to turn coercion to narrow advantage. Ultimately, the police themselves must be policed by institutions that cause them to act in the larger interest to a measurable degree. Such policing is never perfect.

While coercive, exploitative elites are common enough, there are two reasons to suspect that no complex society can be based purely on the coercive principle. The first problem is that coercion of any great mass of subordinates requires that the elite class or caste be itself a complex, cooperative venture. In highly stratified systems, the ruling elite is commonly buttressed by an ideology of ingroup superiority that limits the need for coercive institutions within the elite. In the limit, a ruling elite itself may be egalitarian but quite able to unite to defend its privileged position as exploiters of other people (see e.g. Barth, 1981, on the Pathan feudal system of NW Pakistan and Eastern Afghanistan). Beyond the simple feudal scale, the elite itself must be organized in ways that challenge our egalitarian instincts. Indeed, a common pattern in the aftermath of tribal conquest is that the conquerors fail to achieve any lasting increment of complexity because of violent, Hobbesian quarrels among factions of the conquest band. Thus most Central Asian pastoral conquests achieved little by way of increments to sociopolitical complexity, and the same can be said for Viking conquests in Western Europe and many other similar episodes.

On the other hand, coercive systems with a modicum of internal cohesion can persist for some time. The Hawaiian nobility organized by the ranked lineage system is an example. Aristocracies are generally large complex systems in their own right. An exotic case is the control of Egypt and the Near East by the Mamluke caste of slave soldiers purchased from Asian tribes (Glubb, 1973). Powerful Mamlukes ruled as sultans and lesser officers, but could not transmit Mamluke status to their offspring. Each generation of Mamluke leaders arose by achievement from fresh cadres of steppe boys purchased and rigorously trained by their Mamluke masters. Mamlukes were generally freed by their masters in adulthood, at which time they could begin to climb the hierarchy of the governing Mamluke caste. The Mamluke sultanate persisted from 1250 until 1517 despite much conflict generated by the achievement principle of advancement combined with feudal bonds of loyalty between older Mamluke owners and their younger ex-slaves.

The second problem with raw coercion is that the defeated will seldom accept exploitative subjugation as a permanent state of affairs without costly protest. Insko et al.’s subjects strike, sabotage, and slowdown behavior in their "Carniero" treatment illustrate the underlying logic on a small scale. The frequency of peasant uprisings in early modern states is a larger scale, real-world example. For instance, Highland Peruvian peasants rose up against their Hispanic conquerors in the rebellion led by Tupac Amaru in the late 18^th Century, expelling them from large areas with great loss of Hispanic life. The Hispanic elite reorganized and returned to reconquer lost areas with a brutal punitive slaughter of Quechua and Aymara peasants (Cesar Chavez, 1973). The repeated willingness of subordinate peoples and classes to challenge hated elites at very great risk to the rebels commonly drives the cost of coercive control to high levels. People are moved to desperate acts by deep feelings of injustice generated by manifestly inequitable social arrangements. Kennedy (1987) argues that imperial powers frequently find the costs of dominance insupportable, even those demonstrating a capacity for great ruthlessness, such as the Soviets. The USA was established by rebellion excited by relatively minor inequities imposed by Britain. NcNeill (1980: 25-27) notes that the elites of conquest states generally learn that moderate, regular taxes and policies designed to promote the economy yield more than ruthless exploitation. Subordinate classes benefit from such arrangements, if only to the extent of protection from less disciplined exploiters. Or contrariwise, too-predatory behavior by an established elite may cause subordinate classes to welcome an invading elite offering a better deal, as in the Turkish defeat of Byzantium, whose urban rentiers exploited the peasantry of Anatolia and the Balkans beyond endurance. Christian peasants sometimes welcomed conquest by infidel Turks because they brought lower taxes (McNeill, 1963: 498). We have noted the role commoner Hawaiians could have in tipping the balance in contests between rival chiefs. In general the cost of purely coercive control is to be high and the scope for organizing a complex society by such means alone limited. Durable conquests, such as those leading to the modern European states, Han China, or the Roman Empire leaven raw coercion with more pro-social institutions. The Confucian system in China and Roman law were far more sophisticated and durable institutions than systems of raw coercion. Not infrequently, conquerors are drawn into the institutional framework of the conquered, as in the cases of Chinese Imperial Dynasties like the Yuan (Mongol) and Qing (Manchu) or the German-ancestry emperors of Rome. Thus we believe that other mechanisms are needed to work around human social instincts to generate more than very minimally functional complex societies.

Our Polynesian and modern army examples illustrate how coercion is circumscribed by a common understanding between leaders and followers as to what behavior is legitimate. When such a common understanding fails, the coercive institutions of even the best armies and elites falter. Shalit (1988: Ch. 10) describes how a lack of consensus in Israel and in the Israeli army during the 1982 invasion of Lebanon led to very poor morale and poor performance of officers and men who had fought very well in earlier conflicts.

As Eibl-Eibesfeldt (1989: 314) remarks, even high ranking leaders in modern hierarchies adopt the headman’s deferential approach to leadership. A general may in theory be able to give preemptory orders to a captain, but most of his leadership will be exerted through a staff of colonels, toward whom he generally behaves as only first among equals. Advice on leadership from successful top-ranking US military leaders emphasizes the self-effacing approach (e.g. Marshall and Ridgway in Taylor and Rosenbach, 1982). Marshall reviews the ways in which famous American generals put a very human face on their leadership. Ridgway recounts how as a divisional commander he took care to learn the name of every commissioned officer in his division, fighting in Italy in WWII. In practice brutal sergeants, incompetent colonels, vainglorious generals, and their ilk in other bureaucracies, degrade the effectiveness of social organizations in complex societies. But then, hunters and gatherers also have suffer from the common imperfections of social life. A social system need not be perfect to work tolerably well.

The use of coercion in complex societies offer excellent examples of the imperfections in social arrangements traceable proximally to the coexistence of selfish and prosocial instincts, and ultimately to the widely discrepant behaviors favored by selection acting at larger and smaller scales simultaneously. If human social instincts were entirely prosocial, coerced cooperation would be unnecessary. Given that institutions of coercion are necessary to ensure cooperation yet generally result in special privileges for elites, there is no regular way to prevent three sorts of problems: (1) Intense conflicts among groups with access to the weapons of coercion, resulting in the limit in feudal anarchy. The intense conflict among Hawaiian chiefdoms is an example. Kirch also describes the collapse of the population on Easter Island in the period just before contact, apparently due at least in part to ruinous, violent competition between the two chiefdoms on the island. (2) Moralistic excesses resulting from the problem that the institutionalization of strategies that punish and punish those that don’t punish can stabilize absurd and out-of-date norms as well as group functional ones (Boyd and Richerson, 1992). Conrad and Demarest (1984) argue that the Aztec polity exemplifies this problem. Military expeditions to capture victims for sacrifice were an avenue of upward mobility for successful warriors, but by contact times the depletion of victims within easy campaigning distance of Tenochtitlan and the stiffer resistance of survivors had made the cost of such an institution ruinous. Nevertheless, Montezuma’s attempts at reform failed because the many people materially and ideologically committed to sun sacrifice resisted change successfully, leaving the Aztecs vulnerable to Cortez and his Native American allies. The self-destructiveness of the intensely ideological and popular Nazi regime in Germany (Shirer, 1960) and the militaristic regime of WWII era Japan (Toland, 1970) are modern examples. (3) Elite exploitation or neglect often leads to debilitating divisions in societies. Elite dominance often leads to a pervasive sense of distrust within the exploited and neglected classes. The lack of an effective justice system serving the exploited class can lead to a culture of small-scale coercive violence. The Mafia in Sicily (Gambetta, 1993) and street gangs in American cities (Sanchez Jankowski, 1991) are arguably the result of a combination of elite exploitation and neglect that deprive the victimized communities of the ability to evolve more sophisticated leadership institutions. Even then, the Mafia and street gangs, however imperfectly, are built upon a foundation of prosocial enforcement of community sanctioned norms according to Gambetta and Jankowski. At the other extreme, even in the most enlightened and functional social systems, political power is generally well correlated with income.

The Marxist theoretical economist John Roemer (1988), whose analysis can be viewed in the context here as speaking purely from the prosocial and egalitarian social instincts, provides a rigorous analysis of the various forms of exploitation (morally dubious deviations from equal distribution of society’s rewards) arising in complex societies. Roemer allows that some degree of differential reward may be necessary to motivate hard work and other prosocial acts, but shows that other forms of exploitation can arise in complex societies. In feudal and fascist societies those with direct access to the weapons of coercion reward themselves. In capitalist societies, luck, skill, and inheritance result in concentrations of wealth, which in turn buys political protection. In socialist societies, the power of achieved bureaucratic offices is used to privilege their holders. In each case, Roemer argues that actual inequalities in complex societies seem to greatly exceed those that are socially necessary to motivate prosocial activity.

Frank and Cook (1995) give empirical evidence that the recent growth of income inequality in the USA, especially the very large rewards to star performers in entertainment, sports, and business, is the pathological result of the expansion of what they call winner-take-all competitions. The individual quest for status and income has a quite pathological potential in societies with achieved status. Too many nearly equally qualified people will invest too much competing for the few lavishly rewarded winners slots, starving more equitably rewarded, equally valuable professions of talent. It is notable that there is very wide variation in inequality in contemporary societies (Morris, 1979), and societies with the lowest inequality (Taiwan and the other Asian "tigers;" some NW European societies) are among the most successful. Of course, even in such societies, exploitation may not be reduced to the socially necessary minimum. It is interesting that Insko et al.’s "Carniero" condition produced considerably less total rewards that than their free trade case, thought elite groups made out similarly well in both conditions.

Thus, it also seems empirically reasonable to attribute the largest part of inequality in most if not all complex societies to in-principle socially unnecessary exploitation arising from selfish motives of those with the direct or indirect power to coerce. Excesses of coercion for the benefit of individuals or narrow groups will generally weaken societies in comparison with fairer ones, and will invite internal revolution or reform, defection to alternate elites, and emigration. The rather considerable scale of intertribal violence in Hawaii, the exploitation of the commoner class, and the strategic participation of commoners in elite insubordination, all against a background of a paternalistic ideology (and paternalist economic reality) of chieftainship, illustrate the dilemma of coercive leadership. The German army’s attention to fair treatment of individual soldiers by a competent, respected NCO and officer corps, coupled with the minimization social distance between leaders and men was a sophisticated work-around that allowed very effective command and control with minimal conflict with the egalitarian social instinct.

6.2 Segmentary Hierarchy

Humans probably have a long history of living in egalitarian segmentary societies in which public goods like defense are occasionally provided by the union of small consensually led groups of individuals under some more or less temporary, informal leader, selected by band headmen in consultation with their followers. The segmentary principle can serve the need for of command and control by hardening up lines of authority without disrupting the face-to-face nature of proximal leadership present in egalitarian societies. The Polynesian ranked lineage system illustrates how making political offices formally hereditary according to a kinship formula can help deepen and strengthen a command and control hierarchy. A common method of deepening and strengthening the hierarchy of command and control in complex societies is to construct a formal nested hierarchy of offices, using various mixtures of ascription and achievement principles to staff the offices. The Mongol decimal system of social organization is an example. Each base unit of ten men had a leader, and above ten such men was the next leader in the hierarchy and so on up to the tumen of 10,000. Although its inspiration was military, the Chingizid dynasty imposed the decimal system on society as a whole (Morgan, 1986). The Polynesian ranked lineage system was drafted to perform similar functions as chiefdoms expanded, and modern bureaucracies all use some variant of a pyramidal hierarchy of offices to exert command and control.

In both the kinship based and achieved-office cases, each unit in the segmentary hierarchy consists of a few equal subordinates plus a leader. This is approximately the same structure as a hunting and gathering band with is leading figure, whose headmanship is typically based on some mixture of senior status in a kin network and personal achievement, as argued in the previous section. That is, a typical residential band of hunters and gatherers is on the order of ten families, more or less closely related, centered around an informal weak leader whose role is based on gently encouraging a consensus to emerge. If all is operating properly, the leaders at each step in the hierarchy of a complex society will display the same prosocial leadership and consensual support as the headman of egalitarian band. At each level of the hierarchy the general structure of the band is replicated. As we saw in the last section, even at the highest rank, military leaders are surrounded by a staff of near-equals, and officers with an inability to humanize their relations with subordinates are less successful than those that do. (Shalit, 1988: 131, gives an example of an Israeli general, very successful as a field officer because of his personal courage and audacity, who conspicuously failed as a general officer because his style of leadership was excessively domineering.) The inevitable difficulty is that leaders in complex societies must convey orders downward, not just seek consensus among their comrades. Very careful attention to detail, as in the case of the Wehrmacht in WWII, can go quite far in making leaders responsive to the hierarchy without destroying their subordinates sense that these are the same leaders that would have arisen "naturally," without imposition from above. It is interesting that a very similar move to humanize and legitimate leadership in private business corporations is a current "hot topic" in American business management circles (Petzinger, 1997).

The hierarchical nesting of social units in complex societies gives rise to inefficiencies that would not arise if unquestioned loyalty to high leaders or effective communal solidarity at large scales could be arranged, bypassing the need for hierarchical, segmentary command and control institutions. A byproduct of solidarity in small groups is horizontal intergroup rivalries and factions which degrade the performance of the larger system. Dysfunctional bickering, ranging from bar fights between soldiers of rival units to sniping between cabinet members, is very common. Vertical divisions can lead to marked class divisions in which leader paternalism is limited by upper class arrogance, leading to "strike, sabotage, and slowdown" on the part of the disgruntled lesser classes. The chain of command is often long, devolving substantial leadership responsibility to subleaders far down the chain of command potentially generates great friction if lower-level leaders come to have different objectives than the upper leadership. When leadership at lower levels divergent, exerting more command from the top risks passive or active resistance from subordinates. If coercion from above is made stronger, timid lower level leaders will not take initiative, and the system will be very unresponsive to local contingencies. Successful systems endeavor to inculcate a sense of common mission for the whole organization, provide the necessary minimal and fairly administered coercive command and control, while retaining the flexibility and high individual moral of decentralization. If such a balancing act were easy to perfect, the history of complex societies would be shorter and duller. It is, one hopes, an accident of history that Hitler’s administration of the German army was unusually successful in minimizing the conflict between subordination and cohesion.

6.3 Exploitation of symbolic systems

The increasing density of human populations after agriculture began, the spread of literacy, improvements in passenger transportation, and communications innovations have permitted an ever larger scope for the operation of symbolic systems. The development of monumental architecture to serve mass ritual performances is one of the deepest archaeological markers of the evolution of complex societies, along with the rise in inequality of prestige goods in burials. The use of hypertrophied symbolic systems to structure and organize complex societies is a commonplace. Very nearly as obviously, symbolic systems are a major source of problems for the maintenance of complex social systems. Charismatic innovators regularly launch new prestige systems, which sometimes make radical claims on the allegiance of new members, sometimes make large claims at the expense of existing institutions, and sometimes grow explosively. The ongoing evolution of social systems can be steered in unpredictable, maladaptive directions by symbolic processes. Gibbon (1776-1778) blamed the decline and fall of Rome in equal measure to the growth of barbarian military power (a result of "ethnogenesis" enlarging the scope of barbarian polities) and to the rise of Christianity (a timid and pacifistic ideology unsuited to empire according to Gibbon’s scandalous hypothesis).

6.3.1 Symbolic marking of groups

The symbolic ingrouping instinct is extensively used to delimit a quite diverse array of cultural defined groups in complex societies, within which a good deal of cooperation is routinely achieved. In military organizations, every level except the lowest is generally marked by conspicuous badges of membership. A squad or platoon’s solidarity can rest on personal bonds reinforced by prosocial leadership, but ship’s companies, regiments, and divisions are made real by symbolic marking. Interestingly, ethnic group-like sentiments often seem to be most strongly reinforced at the level of 1,000-10,000 or so men in military organizations (British and German regiments, US divisions). Regiments of around 3,000 soldiers are rather similar in size to the largest military units assembled by tribal scale societies lacking effective hierarchical coercion. Corporations and other civilian bureaucracies frequently evolve distinctive corporate cultures that are symbolically badged (IBM is an example well known around the world). Local regions have distinctive symbolic ingroups in many complex societies. In many historic empires and not a few modern states have distinctive tribal elements within their borders, often fulfilling important social roles (Garthwaite, 1993). The Indian caste system is the most elaborated caseof a complex society built upon the "tribalization" of the division of labor in the form of closed, hereditary castes (Srinivas, 1962; Gadgil, 1983), though specialized economic roles for ethnic or quasi-ethnic groups in complex societies are common and diverse (e.g. Curtin, 1984, on international trade conducted by ethnic diasporas). In modern states, the attempt to achieve a sense of internal unity by constructing ideologies of nationalism to reduce the influence of sub-national tribes and regions is too well known and complex to recount in detail. The American and French Revolutions stimulated the diffusion of the nationalist principle around the world, where it sometimes took vigorous root. Despite universalistic pretensions in some modern societies, powerful nations all seem to be substantially ethnocentric in practice (e.g. Khazanov, 1995, on Russian nationalism in the USSR). In the 20th Century, the success nationalistic appeals is notorious. Perhaps these are the maximal units which mass media propaganda can intuitively plausible to most people. More universalist appeals seem to lack enough plausible primordial realism to be effective, and nationalism is generally confronted with vigorous competition from sub-national institutions.

6.3.2 Charisma

Some gifted, well-trained, or lucky individuals can project a sense of personal competence and prosocial regard for their fellow human beings to a considerably larger number of people than the common run of leaders. They can mimic on a large scale the face to face empathy that elicits an instinctual grant of legitimate authority from anonymous followers. The increases of population density made possible by advanced agrarian techniques made it possible for tens of thousands of individuals to participate in mass rituals where such gifted figures could hold forth. Charismatic leaders can inspire trust, articulate coordination, and motivate spectacular acts by rather large numbers of individuals. Of course, the charismatic individual cannot replicate all the necessary command and control needed to make a complex society actually operate, as Weber’s famous argument that the results of charismatic leadership will persist only if routinized under the leadership of less gifted politicians.

Modern innovations in mass communications amplify the power of personal charisma (and symbolic markers of group membership). Widespread literacy made possible the rise of charismatic nationalist leaders on a scale not possible in the past, and radio and television have expanded the possibility still further. Polls and focus groups attempt to reduce charismatic leadership via mass media to a science. Even so, some leaders can articulate a sense of vision much better than others, for reasons that are still not completely understood (Masters, 1991). To the extent that charisma can reduce the distance between high leaders and followers in a deep hierarchy, it will generate more trust and loyalty, and in principle make larger scale institutions function more effectively. Complex societies generally have ritual systems to elevate the prestige of leaders, who in turn frequently step out of the formally framed prestige role to demonstrate the common touch. Propaganda film of Adolf Hitler typically portrays him either unsmilingly elevated above a vast crowd of cheering Nazis in a stadium draped with dramatic swastikas or cheerfully working a crowd of happy civilians as if he was just another beer hall pol. His emotionally deep father-figure role to so many Germans right up to the end of the war is difficult for us to appreciate today, given the enormity of Nazi crimes (Fritz, 1995).

6.3.3 Exploitation of supernatural and ideological forces

Complex societies make lavish use of supernatural ideologies to buttress social arrangements, including the oft-gross inequalities born of the division of labor and hierarchy. Typically, there is an official, established religion or other ideological system associated with the state, though the relations between institutions of political power and ideology are generally complex. On Cohen’s (1974) classic account, much of the complexity in the relationship between the institutions of power and ideology come from the fact that the organization which controls the symbolic system has a measure of power of its own. Very often, religious organizations and ideological institutions are in a position to compete for power with secular authorities. In cases like the Medieval popes, the Church became itself a significant secular power in addition to standing morally above the many petty secular authorities. Hinduism, as the supernatural justification for the Caste system, similarly stands above and to one side of the historically weak and fragmented polities of India. Nevertheless, typically the secular and sacred or ideological elite are recruited from the same class and generally collaborate in maintaining social order (and social inequality). Functionalist theorists stress the prosocial role of religion and ideology, while conflict theorists stress their role in perpetuating inequality. According to our hypothesis, selection at multiple levels will generally result in a mixture of functions.

6.3.4 Legitimate institutions

At their most functional, symbolic institutions together with effective leadership and smooth articulation of social segments, create a sense of living under a regime of laws and customs that provide public goods efficiently and fairly. Societies like the US and Japan generate a vast array of voluntary associations for economic, charitable, and recreational purposes in which individuals spontaneously associate to produce public goods of various kinds (Fukuyama, 1995). Individuals modern organizations typically feel themselves part of tribal scale groups in which even the remotest individuals are only one step away from face-to-face knowledge, and who are labeled with a shared cultural badge. At the same time, they feel that their "tribes" have an honorable role to play in a larger polity. It seems plausible, as Fukuyama argues, that the trust among members groups of hundreds to a few tens of thousands of individuals (many of whom are strangers or near-strangers) are important principles of social organization in complex societies. In the US and Japan, loyalties to the smaller units generally do not conflict with loyalties to the nation. Quite the contrary, belonging to voluntary organizations is part of citizenship. Khazanov (1995) notes that the near-absence of such organizations (other than the Communist Party, minority ethnic groups, and various mafias) after decades of Communist rule is one of the weaknesses that threatens to derail democratization in the successor states to the USSR. Among the most important small-scale voluntary organizations are those that are felt to offer legitimate means to influence even the highest acts of state. Ordinary people can walk a precinct, lobby a congressional representative, fight city hall, go to court, and join tribal-scale groups that do these things collectively. As long as most individuals feel that existing institutions are reasonably legitimate and that any felt needs for reform are achievable by means of ordinary political activities, there is considerable scope collective social action. On the other hand, individuals who don’t trust the current institutional order’s justness are liable to band together in revolutionary organization, such as the terrorist groups of the contemporary world. More devastating perhaps, those institutions that are not viewed as legitimate will not elicit 3^rd party moralistic aggression toward those who violate them. As illegitimate authorities exert more direct compulsion they risk further alienation. It is a tribute to the Communist leaders of the East Bloc like Gorbachev that most eventually recognized the fatal damage done to their legitimacy and surrendered power with minimal bloodshed.

Boehm (1996) argues that tribal societies often but not always have legitimate, collective institutions by which the society can reach a consensus on actions to take in emergencies, such as the threat of war or famine (see Edgerton, 1992, for many examples of failed social institutions). Even in egalitarian systems without strong leadership, assemblies of adults propose and debate courses of action, gently feeling out courses of action in accord with immediate facts and local custom, and often reach manifestly sensible decisions. Such collective decisions may be self-reinforcing if successful collective decision-making leads to trust in the legitimate process, further use, and even collective decisions to improve the collective decision process. Kirch (1984) describes how Hawaiian chiefs led the development of food production innovations to cope with population growth in the late prehistoric period, activities that no doubt lent so paternalistic legitimacy to an otherwise quite inequitable system. Turner (1995) terms the ability of human societies to deliberately seek solutions to collective problems "Spencerian selection." Complex societies greatly increase the scope for collective decision-making by recruiting specialists to gather and process information about alternative decisions and by providing the means to police conformity to collective decisions on considerable scales.

Note that collective decisions based upon prosocial principles short-circuit the need for group selection to act to improve societies by selection among alternative social systems. More generally, prosocial instincts can substitute for selection by acting as a transmission and innovation biases, favoring successful social institutions by differential invention and imitation, not differential extinction of social groups. It may be that Spencerian selection and individual level biased cultural evolution have largely replaced Darwinian group selection in recent millennia. Keller (1931) considered and stoutly rejected this hypothesis; it strikes us as an open question.

Sabatier and Jenkins-Smith (1993) and colleagues provide fine-grained descriptions of collective decision-making and policy change in the contemporary US. The core units of policy making are "advocacy coalitions," informal tribal scale groups of citizens, legislators, legislative staffers, public servants, and academics that come to share common policy goals. Typical policy arenas have two, sometimes more, advocacy coalitions that struggle in an adversarial framework to influence laws, regulations, and bureaucratic behavior. These debates take place in an institutional arena that coalition participants view as more or less legitimate. They expend effort trying to influence the policy actions of bureaucrats, and legislators through peaceful debate, and other constitutional and customary means, rather than violent or illegal means. Even when defeated, most participants either leave the particular policy arena or continue the fight by legitimate means. Over time scales of a decade or so, measurable change occurs in the opinions of coalition members and in the policies themselves. Unsuccessful coalitions may collapse and disappear due to pervasive frustration. In a modern democracy, an extensively segmented, adversarial-deliberative policy making process is made legitimate. The adversarial element seems Darwinian and the deliberative Spencerian in Turner’s taxonomy, and both are intricately intertwined in Sabatier’s and Jenkins-Smith’s picture of the process of policy evolution.

Most complex societies have had much less open and decentralized decision-making systems. Authoritarian systems at least formally concentrate power in the hands of a maximum leader. Even such systems are often granted a fairly strong writ of legitimacy by their subjects, as in monarchies. Sophisticated authoritarian systems allow for considerable political participation, although the "franchise" is generally sharply limited by class. For example, it is notorious that the ancien regime monarchies of early modern Europe accreted a huge and inefficient burden of special privileges won by wealthy aristocrats, but also by towns, guilds, religious orders, and other more middling groups (McNeill, 1963:655). Empires are often responsive to the suasion of tribal representatives, as in the relationship between Ottoman sultans and the Greek Orthodox Church, whose hierarchy mediated relations between the Sultan and his Greek subjects. The persistence of such authoritarian systems, and the rising cost of social control when large masses of people feel subject to illegitimate rule, suggest that authoritarian systems generally have reasonably effective institutions for generating senses of legitimacy. The Chinese Confucian authoritarian tradition was a moral order with an elaborate theory of legitimacy of emperors captured by the concept of the Mandate of Heaven, which dynasties had to earn by correct behavior (Fairbank, 1992). The mandarin elite exerted much influence over the Mandate, but the costs of imperial failure also included peasant uprisings and external invasion.

6.3.5 Disfunctions of symbolic systems

A major difficulty with loyalties induced by appeals to shared symbolic culture is the very language-like productivity possible with this system. Language itself is a classic badge of an ethnic group, along with other symbols. Dialect markers of social subgroups emerge rapidly along social fault-lines (Labov, 1972). The rhetorical power of language, together with "body language" are the main instruments of charismatic figures. Religion and ideology are very diverse and their evolution is notoriously unconstrained. The extensive division of labor in complex societies, the flows of immigrants from diverse ethnic backgrounds, and the elaboration of formalized political and religious ideologies mean that individuals have considerable scope for choosing which ties to emphasize in particular circumstances (Barth 1969). Elites, stigmatized groups, or any semi-closed unit may evolve away from the mass of society, creating great ideological gaps where once narrow ones existed. The number of cross-cutting symbolically marked groups is much larger in the dense, literate populations of the last few thousand years compared to the tribal past when practical choices of individuals were limited to choosing the level of a segmentary system on which to focus loyalty, or perhaps sometimes to assimilate to neighboring group. Ongoing cultural evolution can produce unpredictable shifts in the sense of which groups should be primary. Thus, in the 18^th and 19^th Centuries, German speakers in Western Europe gradually developed a sense of nationhood that eventually replaced the local loyalties that had typically kept German polities small throughout their long history.

Charismatic leaders frequently drum up small, new groups by their personal magnetism, as in the commonplace cults and revitalization movements of the daily news. Usually, any such cults persisting past the life of the founder have quite extreme beliefs, as the theory predicts (the smaller a marked group is, the more extreme markers are required to maintain separation, Boyd and Richerson, 1987). Occasionally, such cults grow into major organizations, as in the case of the Mormon Church in the US (Brooke, 1994). As Max Weber appreciated, and the case of Hitler and many other charismatic leaders illustrates, charisma is difficult to harness to reason or tradition. Charisma is a wild card in the affairs of complex societies. Occasionally, charismatic leaders can cause a sudden shift in the loyalties of large populations, leading societies from outworn traditions to a new and more functional system. Occasionally, towering political figures rise up to rescue and endangered society from the inevitable problems that beset it. However, given the intricacy of the problems of complex societies, pure charismatic leadership has no guarantee of faithfully serving either tradition or reason. Some social traditions seem to routinely take reasonable advantage of the charisma of their maximum leaders when the possess this quality. The liberal democracies of the modern world have been reasonably successful at productively capturing the charisma of their Nehrus, Reagans, and Churchills, whereas other types of societies have had disastrous luck with Napoleons, Hitlers, and Perons.

Since one form of cooperation is participation in the coercion of outgroup members (and ingroup members perceived as defectors), it seems quite likely that shifts in loyalties should have "outbreak" dynamics that cause sudden rather than gradual shifts in mass adherence to particular ingroups. As a new ingroup or ideology grows slowly in the face of coercion by the established institutions, individuals will at a certain point judge that balance of coercive power is close to equal. Those inclined anyway to affiliate with the new group may begin to shift loyalties, precipitating a cascade of loyalty shifting, perhaps with an assisting shove from a charismatic figure. Perhaps for this reason, ruling elites often go to very considerable lengths to maintain control over the whole symbol marking process by insisting that all groups have official sanction and conform to official ideology. Historical examples abound. Modern totalitarian states are examples, as is the collaboration of the Catholic Church with Medieval and early Modern states to maintain religious conformity in Western Europe. Other solutions include the multi-ethnic empire with a leading elite tolerating but controlling others (Ottoman Turkey), and caste systems linked by an overarching ideology (Hindu India). The tolerant modern democratic state united by a mild nationalism is perhaps the currently most successful system but it is a historical rarity, and is still far from the commonest sort of society in the contemporary world. The elites, at least, of less tolerant societies are prepared to make the case that tolerance of ideological diversity is an invitation to schism and violent conflict. Other critics of liberal systems deny that they are truly tolerant, as in Marcuse’s (1964) "repressive desublimation" critique of consumer capitalism. More empirical critiques (Easterlin, 1995; Frank and Cook, 1995) of the cult of consumerism, and any number of environmentalist broadsides, make cases at least some institutions of the modern liberal society have dangerous ideology-driven flaws from which there is little effective dissent.

In the contemporary world, many modern states, liberal and illiberal, are beset by religious fundamentalisms, and in a few fundamentalism has come to rule (Roof and McKinney, 1987; Marty and Appleby, 1991; Finke and Stark, 1992). Typically, fundamentalisms grow out of folk religious beliefs, whose quaint character provides a badge of membership in a group foreign to the secular elite of the state. American Protestant fundamentalism (Ammerman, 1991) should be seen against a historical background religious diversity, including a history of separatist sects, such as the Mormons. Nevertheless, for most of American history, the major Protestant churches were an important institutional pillar of the segmentary hierarchy of the nation. The local churches provided primary groups and a "tribal" structure that nested comfortably within the national society. Americans were expected to belong a set of common Protestant denominations, and Catholics, Jews, and members of many smaller denominations suffered discrimination. In the late 19^th Century, critical biblical scholarship and kindred natural science advances swept up the intellectual leaders of the mainline denominations, casting fatal doubts on ancient articles of faith (Bartley, 1984). For thoughtful moderns, the Bible fell to the status of ancient myth, fascinating in its own way, but not literally truthful. Many followers of traditional Protestant churches were not willing to follow their elite in such a large concession to secular ideas or belonged to churches and sects in which modern intellectual ideas never achieved much currency. Mainline churches monopolized many sources of power and cultural influence, including radio and television network broadcast time, and waged a concerted campaign, led by the liberal National Council of Churches to marginalize fundamentalist churches and sects (Finke and Stark, 1992). They were completely unsuccessful. New leaders rapidly emerged (without direct government support as established churches, American Protestant leaders had always to have a streak of entreprenuership). For example, as more and more old-denomination seminaries fell under the sway of critical scholarship, fundamentalist bible schools were founded to perpetuate inerrant interpretations of the Bible and to train fundamentalist pastors. Conservative churches have grown rapidly at the expense of old denominations, who have lost most of their members to the secular community. The current trajectory is toward an increasingly polarized society with most people either secular or members of quite conservative churches.

No less than Shi’ite fundamentalism in Iran, American Protestant fundamentalisms illustrate the potential for independent evolution of the ideology of symbolically marked subgroups. American Protestantism consists of dozens of churches and sects, with new variants always arising, which engage in a rather Darwinian struggle for hearts and minds of the faithful. We say Darwinian advisedly, as the data in Roof and McKinney (1987: 148-185) suggest that differential birth rates are not the least of the advantages of conservative Protestants. There is a certain irony in the fact that the stoutest enemies of Darwin are benefiting from natural selection! While large-scale societies appear to require intermediate institutions to "tribalize" the world at a level that attracts spontaneous loyalties of individuals, the evolution of such groups is, for better or worse, unpredictable, uncontrollable, and often subversive. All attempts to create monolithic, hierarchically ordered symbolic systems seem to have succeeded only at considerable ongoing costs to suppress heresy. Even the most successful—Confucianism, Roman Catholicism—have been epic struggles, with many battles lost. The sense of legitimacy can be compromised as well as supported by symbolic institutions.

7.0 Conclusions

The complex human societies of the past few thousand years are a tightly wound bundle of contradictions. On the one hand, they sometimes produce public goods on a massive scale, generate a sense of solidarity among millions of citizens, and construct symbolic worlds of great sophistication, beauty, and attractiveness. On the other, inegalitarian distributions of rewards, inefficient management of large scale organizations, ruthless repression of harmless deviants, and participation in bouts of large scale violence are common. Sometimes complex societies seem incredibly powerful and robust against the severest insults. Other times they shatter spontaneously for obscure reasons. Sometimes state institutions engage in feats of strategic decision-making that are far outside the bounds of small-scale societies. Other times, small-scale interests hamstring elective democracies, monarchies, tyrannies, and every other variant system of government tried.

Our argument departs from an infamous remark of E. O. Wilson’s (1975: 562) in Sociobiology: "Human beings are absurdly easy to indoctrinate—the seek it." Like many other innately primed human capacities like language, the social instincts do certainly lead us to acquire many elements of our social behavior with remarkable speed and ease. We are primed to live in symbolically marked groups, respond to coercion by leaders, cooperate with ingroup members, subscribe to objectively absurd supernatural beliefs, and discriminate in favor of kin and friends. We are tolerably adept at managing the substantial conflicts that arise from multiple demands upon our loyalties. If our argument is correct, the derived, species specific psychological mechanisms that support our peculiar form of ultra-sociality arose by coevolution with culture. As with language, the social instincts are not comprehensible apart from the cultural traditions the provide the specific recipes for social action. Absent cultural specification of ingroup and outgroup, the nature of legitimate authority, the nature of solutions to coordination games, the beliefs which we are supposed to hold sacred, and the like, why would we have this combination of strong predispositions and an absurd susceptibility to indoctrination? The existence of social instincts, such as a real regard for the success of the groups to which we belong, in turn underpin social institutions like legitimate authority that can speed the rate of discovery and diffusion of cultural innovations. On the other hand, these same instincts under other circumstances lead to crazy cracks in the face of a once solid social system, breaking a complex society back down to a anarchic, feudal, dark age remnants.

We believe that this apparently paradoxical state of affairs is readily explicable by the evolutionary dynamics that are inherent in human populations. Human reproductive biology is similar to our ape ancestors, leading to selection for propensities to cooperate only on the smallest scales of close kin and intimate companions. Humans’ massive use of culture led to the potential for group selection to act on cultural variation. Ultimately, the coevolution of psychological predispositions with group selected cultural institutions led to the evolution of social instincts adapting us to living in cooperative ingroups with (weakly) coercive, prosocial leadership, and moralistic ethical systems. These simple societies were no conflict free paradise. Intra-psychic and social conflict is bound to be a common byproduct of having selection operating on genes for cooperation in very small groups, but selection acting on culture to favor cooperation on larger scales. No human can really trust anyone but kin and close compatriots to take a heartfelt interest in her reproductive success. On the other hand, a human without a tribe misses the provision of important goods, not least defense from the organized predation of other tribes. The evolution of institutions that minimize such conflicts will always be favored by human decision-makers and ceteras paribus by selection, but solutions are difficult to achieve and rather fragile.

Food production systems increased human densities many-fold, favoring social systems that used command and control institutions to enlarge the scope of cooperation, coordination, and division of labor. Such systems had to evolve work-around institutions to minimize the impact of scale and hierarchy on people whose social instincts are adapted tribal societies in which leadership and privilege are sharply circumscribed and in which the most socially distant members were just barely strangers.

None of these increments of social scale did away with the older, smaller scale institutions. Individuals still have to look out for their personal reproductive interest, blood is still thicker than water, and tribe, town, faith, interest group, and fraternity commonly attract allegiance. Human societies are segmentary in structure, and there is considerable evolutionary pressure exerted at all levels, both currently and by the operation of the social instincts. The social instincts themselves, we argue reflect a history of conflicting selection pressures from which derive the ancient and old insincts. There is a general arms-race pressure favoring the largest scale units, but dynamics at smaller scales are not easy to predict and even less easy to reliably control. The instincts are powerful pieces of psychology that quite generally cause dissatisfaction with the ubiquitous institutions of command and control and the elite privilege that always have attached to them. Nations and empires are fragile constructs as a consequence. The best of them develop a tolerable working partnership of institutions at different levels, so that individuals feel a minimum conflict between loyalty to family, friends, co-"tribals," and nation, and a minimum of dissatisfaction with elite behavior. Often, organizations that are quite admirable in certain respects, like the Wehrmacht of WWII and conservative Protestant sects, are quite unattractive in others. In the worst cases, incompetent, repressive national elites demand absolute obedience on the pain of torture and execution, or small scale factions square off in uncontrolled civil strife. On the bright side, we can aspire to competent, legitimate social systems that accord for the most part with our social instincts, produce public goods in large quantities, and are robust in the face of the inevitable internal and external shocks.

There is an apocryphal story of two explorers who come upon some tribesmen deep in a tropical forest who have constructed a crude internal combustion engine of wood and rope, fueled by the juice of a jungle vine. The question in the ensuing conversation is whether to admire them for their isolated ingenuity, or pity the crudity of the result. If honeybees could talk, they might discuss our societies in the same terms. What amazing social feats we accomplish with such cumbersome primate raw material! What pitiful messes compared to the much more smoothly functioning superorganisms of the eusocial insects!

8.0 References