Evolutionary Selection Research Paper

In more recent times the social sciences have become dominated by a more reductionistic perspective, sometimes referred to as ‘methodological individualism,’ in which ‘groups and social organizations have no ontological reality—that where used, references to organizations, etc., are but convenient summaries of individual behavior’ (Campbell 1994, p. 23). In addition, evolutionary biologists reached a consensus during the 1960s that groups are highly unlikely to have evolved into adaptive units comparable to individual organisms (Williams 1966). These two developments, which occurred largely independently of each other, appeared to relegate the concept of groups as organisms to the antique shop of history.

Developments in evolutionary biology since the 1960s have led to a new consensus that is represented by the term ‘multilevel selection’ (e.g., Frank 1998, Michod 1999, Sober and Wilson 1998). Natural selection acts on a hierarchy of biological units, from genes within individuals to groups within a meta-population (a population of groups). Individual organisms are an upper level of this biological hierarchy, social groups that have become so functionally integrated that the whole is more conspicuous than the parts. If individuals are themselves groups, the concept of groups as individuals can no longer be regarded as theoretically improbable. In fact, an especially strong case can be made for human groups as units of natural selection (Boehm 1997). The group-level perspective is back in evolutionary biology, with profound implications for the human social sciences. However, groups can never be assumed axiomatically to function as adaptive units. Special conditions are required and opposing forces always exist, including subversion from within. The modern multilevel view can draw upon the originators of the social sciences for insight, but is not a return to the grandiose and often naıve groupism of the past.

1. Evolution And The Fundamental Problem Of Social Life

Prior to Darwin, supernatural agents provided the most ready explanation of purpose and/order at all levels, from celestial bodies, to human society, to the actions of individual people and other creatures. Darwin provided the first successful scientific theory of adaptations, which remains today the only theory in an ultimate sense. Human artifacts have a purpose and did not evolve by natural selection in a proximate sense, but the ability of humans to make artifacts with a purpose itself presumably evolved by natural selection. All intentional processes are probably rooted in a process of blind variation and selective retention (Campbell 1960).

Evolution explains adaptations but it is not nearly as permissive as supernatural explanations. In particular, group-level adaptations are not as easy to explain as individual-level adaptations. To see why, consider the evolution of a nonsocial adaptation, such as cryptic coloration. Imagine a population of insects that varies in the degree to which they match their background. Every generation, the most conspicuous individuals are detected and eaten by predators while the most cryptic individuals survive and reproduce. If offsprings resemble their parents, then the average individual will become more cryptic with every generation. Anyone who has beheld an insect that looks exactly like a leaf, right down to the veins and simulated herbivore damage, cannot fail to be impressed by the power of natural selection to evolve breathtaking adaptations at the individual level.

Now consider the same process for a social adaptation, such as members of a group warning one another about approaching predators. Imagine a flock of birds that vary in their tendency to scan the horizon for predators and to utter a call when one is spotted. It is not obvious that the most vigilant individuals will survive and reproduce better than the least vigilant. If scanning the horizon detracts from feeding, the most vigilant birds will gather less food than their more oblivious neighbors. If uttering a cry attracts the attention of the predator, the sentinel places itself at risk by warning others. If these suppositions are correct, then birds that do not scan the horizon and that remain silent when they do see a predator will survive and reproduce better than their more vigilant neighbors.

These two examples show that the evolutionary concept of adaptation does not always conform to the intuitive concept, especially at the group level. It is easy to imagine a bird flock as an adaptive unit and to predict its properties. We would expect members of the flock to adopt the creed ‘all for one and one for all.’ We might expect sentries to be posted at all times to detect predators at the earliest possible moment and to relay the information to feeding members of the flock. Unfortunately, individuals who possess these behaviors do not necessarily survive and reproduce better than individuals who enjoy the benefits and do not share the costs. Since Darwin’s theory explains adaptations only on the basis of differential survival and reproduction, it appears unable to explain groups as adaptive units. This can be called the fundamental problem of social life. Groups function best when their members produce benefits for one another, but it is difficult to translate this kind of social organization into the currency of biological fitness. Evolutionary theory has difficulty explaining any kind of group as an adaptive unit, including those that might be found in our own species.

2. Darwin’s Solution To The Fundamental Problem

Darwin was aware of the fundamental problem of social life and proposed a solution. Suppose there is not just one flock of birds but many flocks. Furthermore, suppose that the flocks vary in their proportion of vigilant callers. It is true that vigilant callers do not have a fitness advantage within a single flock, but groups of vigilant callers will be more successful than groups whose members do not look out for one another. Darwin used this reasoning to explain examples of apparent altruism in nonhuman species, such as the bee’s suicidal sting, and also human moral virtues that appear designed to promote group welfare. In his own words: ‘At all times throughout the world tribes have supplanted other tribes; and as morality is one important element of their success, the standard of morality and the number of well-endowed men will thus everywhere tend to rise and increase’ (Darwin 1871, p. 166).

Darwin’s solution is elegant and perhaps even obvious in retrospect. If adaptations evolve by differential survival and reproduction, it makes sense that the evolution of adaptive groups requires the differential survival and reproduction of groups. Adaptation at each level of the biological hierarchy requires a corresponding process of natural selection at that level. However, Darwin’s solution only shows that adaptive groups can evolve in principle. Conditions are required that may or may not exist in the real world. In addition, the adaptations favored by group selection can be opposed at other levels. In the case of our birds, group selection favors vigilance but selection within groups favors nonvigilance. If we wish to explain bird flocks as adaptive units, not only must we demonstrate a process of between-group selection, but we also must show that it overpowers within-group selection.

Darwin’s solution has another disturbing implication, which demonstrates that the biological concept of adaptation remains limited, even in its multilevel form. We have already seen that adaptations at one level of the biological hierarchy can be highly maladaptive at other levels. A male lion that displaces another male and systematically kills his rival’s offspring so that he can proceed to mate with the females is behaving adaptively, in a narrow biological sense of the word. However, this adaptation is good only for the individual male; not for the vanquished male, the females, the group, the species, or the ecosystem. It is possible to imagine a process of group selection that favors noninfanticidal males. After all, groups with noninfanticidal males might grow larger by producing more cubs, enabling them to outcompete groups with infanticidal males. However, this would not eliminate conflict so much as elevate it up the biological hierarchy, to the level of between-group rather than within-group competition. The most that group selection can do is evolve groups that are like organisms in the harmony and coordination of their parts. We already know what individual organisms do to one another and nothing different can be expected of groups. Harmony among groups would require a process of natural selection operating at a yet higher level.

3. Evolutionary Theory’s Wrong Turn

Group selection has been portrayed as a process that can occur but which also must contend with other forces that pull in other directions. As previously mentioned, a consensus emerged among evolutionary biologists in the 1960s that group selection is such a weak force that it can be ignored for most purposes. Even though it is theoretically possible for groups to evolve into adaptive units, it almost never happens in the real world. Of course, group selection could not be rejected without alternative theories that explain apparent group-level adaptations in more individualistic terms. The theories that became the foundation for the evolutionary study of social behavior were kin selection (Hamilton 1964, Maynard Smith 1964), in which individuals evolve to benefit copies of their genes present in genetic relatives, and reciprocal altruism (Trivers 1971, Maynard Smith 1982, Axelrod and Hamilton 1981), in which individuals indirectly benefit themselves by helping others. With these theories, ‘biologists seemed to have little further need for the metaphor of society as an organism’ (Konner 1999, p. 31).

The rejection of group selection was hailed by evolutionary biologists as a great accomplishment but subsequent developments have shown it to be a massive wrong turn from which the field is only starting to recover (see Sober and Wilson 1998 for a detailed historical and conceptual analysis of this period). We can see why by returning to our bird example. Suppose that uttering a cry does not increase the risk of being attacked by an approaching predator. On the contrary, it tells the approaching predator that it has been spotted by the crier and that a less vigilant member of the group should be targeted. If these are the facts of the matter, then the evolution of so-called warning cries could be explained entirely by within group selection. Individuals who call survive and reproduce better than individuals in the same group who do not call. The function of the adaptation is not to warn other members of the group but to communicate with the predator in a way that actually endangers other members of the group. We would be right to reject group selection in this case. But now suppose that our original story was correct; crying is selectively disadvantageous within groups and evolves only because groups of criers fare better than groups of noncriers. It is still the case that the average crier in the metapopulation is more fit than the average noncrier—that is just another way of saying that the crying behavior evolves—but the evolutionary force that causes the average crier to be more fit is group selection, not selection among individuals within groups.

It follows that distinguishing levels of selection requires several steps. First we must compare the fitnesses of individuals within groups. Then we must compare the fitnesses of groups in the metapopulation. Finally we must combine these effects to determine the bottom line of what evolves. It turns out that many evolutionary models do not follow this procedure. Instead, whatever evolves in the model is vacuously defined as the product of ‘individual (or gene) selection,’ despite the presence of groups and counteracting selection pressures within and among groups. A good example is n-person evolutionary game theory, which has become the framework for studying reciprocal altruism (Dugatkin and Reeve 1998). The term ‘nperson’ signifies that a large population is broken into groups of size n within which social interactions occur. The fitness of an individual is given by a payoff matrix and is determined by its own behavior and the behavior of other members of its group. Our bird flock example could easily be modeled with n-person game theory, in which the fitness of an individual depends on whether it is a caller and on the number of other callers in its group. Despite the presence of groups and fitness differences within and among groups that are plain to see in the payoff matrix, the standard procedure in evolutionary game theory is to average the fitness of the alternative behavioral types across groups and to say that the type with the highest average fitness evolves by individual-level selection. Another example is kin selection theory, which implicitly assumes that a large population is broken into groups of genetic relatives, within which social interactions occur.

Elliott Sober and I call the vacuous practice of defining anything that evolves the product of individual or gene-level selection ‘the averaging fallacy’ (Sober and Wilson 1998). Avoiding the averaging fallacy has two major implications for evolutionary theories of social behavior. First, kin selection and reciprocal altruism become examples of multilevel selection theory rather than alternative theories. When the groups are identified in these theories, it becomes clear that the prosocial behaviors being modeled (e.g., cooperation and altruism) are selectively disadvantageous within groups and evolve only because groups with prosocial members outperform other kinds of groups. Second, not only does multilevel selection theory include kin selection and reciprocal altruism as special cases, but it goes further to explain other cases that are beyond the imagination of these theoretical frameworks. This is fortunate, because kin selection and reciprocal altruism have always seemed poorly suited to explain the kinds of prosocial behaviors that take place in human groups.

4. Ancestral Human Groups As Adaptive Units

Our species evolved in small groups that are roughly approximated by the hunter-gatherer societies of today. Prior to the resurgence of multilevel selection theory, evolutionary biologists found it difficult to envision human groups as adaptive units. Genetic relatedness among members of ancestral human groups was only moderate and the principle of reciprocal altruism has always been difficult to extend beyond pairwise relationships. However, modern multilevel selection theory enables human groups to be seen as potent units of selection.

Extant hunter-gatherer societies around the world are remarkably egalitarian. It is not that hunter-gatherers lack selfish impulses, but rather that selfish impulses are effectively controlled by other members of the group. This form of guarded egalitarianism has been called ‘reverse dominance’ by anthropologist Chris Boehm (1993). In many animal groups, the strongest individuals are usually able to dominate their rivals, taking a disproportionate share of the resources. This is within-group selection, pure and simple. In human groups, an individual who attempts to benefit himself or herself at the expense of others is likely to encounter the combined retaliation of the rest of the group. In most cases even the strongest individual is no match for the rest of the group, so self-serving acts are effectively curtailed. Boehm’s survey of hunter-gatherer societies includes many examples of reverse domination, ranging from ridicule to ostracism, to assassination, of those who attempt to impose their will upon others. Hunter-gatherer societies are first and foremost moral communities with a strong sense of right and wrong that organizes the practices of the group.

The concept of a human group as a moral community shows how much has been missed by the concept of genetic relatedness. In a kin selection model, behavioral uniformity within groups can be achieved only by genetic uniformity. In a moral community, behavioral uniformity can be achieved (or at least approached) by a shared belief in what constitutes right behavior and an ability to discourage wrong behavior effectively. Social norms can create a degree of behavioral uniformity within groups and differences among groups that could never be predicted from their genetic structure and which radically shifts the balance between levels of selection in favor of group selection.

The concept of a human group as a moral community also shows how much has been missed by the concept of reciprocal altruism. Consider a moral community whose members believe that it is right to help others in proportion to need rather than the likelihood of return benefit. Individuals who abide by the norm are rewarded, those who violate the norm are punished, and the group (let us say) prospers compared to groups whose members restrict helping to those who will return the favor. It is true that giving according to need appears less altruistic when enforced by social controls, but social control itself emerges as a group-level adaptation when analyzed in multilevel terms (Boyd and Richerson 1992, Sober and Wilson 1998). Causing others to provide a public good is itself a public good, as economists have long realized (Heckathorn 1993).

Moral communities have both a cultural and an innate psychological component. The specific content of social norms can be modified on the basis of experience and socially transmitted from one person to another. On the other hand, the emotional reactions associated with rightdoing and wrongdoing are almost certainly innate and evolved by genetic multilevel selection. Finally, it should be obvious that wrongdoing remains as much a part of human nature as rightdoing, ready to surface in at least some individuals whenever the maintenance of social norms becomes lax. For this reason alone, human groups can never be axiomatically assumed to function as adaptive units.

Many of these principles that emerge from multilevel selection theory are supported by social psychology experiments that simulate social dilemmas (e.g., Ostrom et al. 1994). In a typical experiment, members of a newly created ten-person group are each given 10 dollars, which they can keep or anonymously put into a central pot where it will be doubled and redistributed to all members of the group. If everyone cooperates they can double their money. However, a single individual who defects keeps his or her 10 dollars and gets 18 dollars back from the central pot, providing a powerful incentive to cheat. If everyone cheats the benefits of group life are lost and no one increases their money. If a single individual cooperates in a group of cheaters, he or she gets a paltry two dollars for his or her investment of 10 dollars, providing a powerful incentive not to cooperate. This is the fundamental problem of social life in a nutshell.

Most people in this experiment initially display a moderate degree of generosity, putting about half their money into the pot, but there is variation around the average. As soon as they realize that others are cheating, the more generous members of the group become both angry and stingy. Of course, there is no outlet for anger within the framework of the experiment so the level of giving quickly drops to zero. In a modified version of the experiment, the members are allowed to contribute their money, not only to the first pot, but also to a second pot that is used to punish the cheaters. Those who contribute to the second pot are second-order public good providers. They enforce contributing to the first pot at their own expense. Nevertheless, when this outlet for righteous indignation is provided within the framework of the experiment, the second pot overflows with contributions!

Generalizing from these experiments, we can provisionally envision a human nature that is reluctant to give in ways that can easily be exploited but is willing and even eager to contribute to building and maintaining a social system that realizes the advantages of social life while preventing exploitation from within. Of course, a well-protected social system is required because at least some people, and perhaps all people some of the time, are inclined to exploit the generosity of others. Furthermore, the whole purpose of a group unified by a well-protected social system may be to exploit other groups. This complicated mix of right-doing and wrongdoing, and behaviors that change their moral status depending upon the level at which they are viewed, may be the legacy of thousands of generations of multilevel selection operating on human populations that were subdivided into small foraging groups.

5. Modern Human Groups As Adaptive Units

A theory that explains small-scale human societies as (largely) adaptive units is an enormous advance over previous theories that seemed restricted to groups of genetic relatives and narrow reciprocators. However, more work is required to explain the nature of human societies at a scale above a few hundred or at most a few thousand individuals. Large-scale societies did not even exist before the advent of agriculture, which is a short (although not negligible) period of time as far as genetic evolution is concerned (see Durham 1991, Wilson 1994 for discussions of genetic adaptations to modern environments). Nevertheless, this has not prevented cultural processes from creating a parade of new social forms that succeed or fail on the basis on their properties.

Superficially, large-scale human societies appear much less egalitarian than hunter–gatherer groups but the apparent inequalities can be interpreted in two very different ways. On the one hand, social control mechanisms are probably strongest in small groups in which everyone knows and depends on everyone else. Many inequalities that exist in large-scale societies are therefore exactly what they seem: some individuals profiting at the expense of others within the society. These inequities should not be interpreted as grouplevel adaptations but rather as individual-level adaptations with consequences that are often dysfunctional at the society level. On the other hand, purely from the group-level functional standpoint, societies must become differentiated and hierarchical as they increase in size. Thirty persons can sit around the campfire and arrive at a consensual decision. Thirty million persons cannot. It is therefore an open question whether extreme status differences and other seeming inequalities in large-scale societies represent pure and simple domination or design features that enable the society to function at a large scale, especially in competition with other societies. There can be little doubt that the scale of a society itself is a group-level adaptation. Larger societies tend to replace smaller societies unless their larger size is offset by problems of coordination and internal conflicts of interest. It is possible to imagine human history as a process of multilevel cultural evolution in which smaller groups coalesce into larger groups, just as for long-term biological evolution.

6. Conclusion

Despite its turbulent history, the concept of human groups as organismic units can be partially justified by multilevel selection theory. The implications extend the length and breadth of the social sciences; from cognition as a group-level process (Wilson 1997), to individual differences in prosocial and antisocial behaviors (Wilson et al. 1996), to historical and cultural studies (MacDonald 1994), to modern economic policy (Bowles and Gintis 1998). Social scientists have every reason to question the dominance of methodological individualism and to revisit the perspective of the originators of their own disciplines in the light of modern evolutionary theory.

Bibliography:

1. Axelrod R, Hamilton W D 1981 The evolution of cooperation. Science 211: 1390–6
2. Boehm C 1993 Egalitarian society and reverse dominance hierarchy. Current Anthropology. 34: 227–54
3. Boehm C 1997 Impact of the human egalitarian syndrome on Darwinian selection mechanics. American Naturalist 150: S100–21
4. Bowles S, Gintis H 1998 Is equality passe? Boston Review 23(6): 4–26
5. Boyd R, Richerson P J 1992 Punishment allows the evolution of cooperation (or anything else) in sizable groups. Ethology and Sociobiology 13: 171–95
6. Campbell T D 1960 Blind variation and selective retention in creative thought and other knowledge processes. Psychological Review 67: 380–400
7. Campbell D T 1994 How individual and face-to-face-group selection undermine firm selection in organizational evolution. In: Baum J A C, Singh J V (eds.) Evolutionary Dynamics of Organizations. Oxford University Press, New York, pp. 23–38
8. Darwin C 1871 The Descent of Man, and Selection in Relation to Sex. Murray, London
9. Dugatkin L A, Reeve H K (eds.) 1998 Game Theory and Animal Behavior. Oxford University Press, Oxford, UK
10. Durham W H 1991 Coevolution: Genes, Culture and Human Diversity. Stanford University Press, Stanford, CA
11. Frank S A 1998 Foundations of Social Evolution. Princeton University Press, Princeton, NJ
12. Hamilton W D 1964 The genetical evolution of social behavior, I and II. Journal of Theoretical Biology 7: 1–52
13. Heckathorn D D 1993 Collective action and group heterogeneity: Voluntary provision vs. selective incentives. American Sociological Review 58: 329–50
14. Konner M 1999 Darwin’s truth, Jefferson’s vision. American Prospect 45: 30–8
15. MacDonald K B 1994 A People that Shall Dwell Alone: Judaism as a Group Evolutionary Strategy. Praeger, Westport, CT
16. Maynard Smith J 1964 Group selection and kin selection. Nature 201: 1145–6
17. Maynard Smith J 1982 Evolution and the Theory of Games. Cambridge University Press, Cambridge, UK
18. Michod R E 1999 Darwinian Dynamics. Princeton University Press, Princeton, NJ
19. Ostrom E, Gardner R, Walker J M 1994 Rules, Games, and Common-Pool Resources. University of Michigan Press, Ann Arbor, MI
20. Sober E, Wilson D S 1998 Unto Others: The Evolution and Psychology of Unselfish Behavior. Harvard University Press, Cambridge, MA
21. Trivers R L 1971 The evolution of reciprocal altruism. Quarterly Review of Biology 46: 35–57
22. Wegner D M 1986 Transactive memory: A contemporary analysis of the group mind. In: Mullen B, Goethals G R (eds.) Theories of Group Behavior. Springer-Verlag, New York, pp. 185–208
23. Williams G C 1966 Adaptation and Natural Selection: A critique of some current evolutionary thought. Princeton University Press, Princeton, NJ
24. Wilson D S 1994 Adaptive genetic variation and human evolutionary psychology. Ethology and Sociobiology 15: 219–35
25. Wilson D S 1997 Incorporating group selection into the adaptationist program: A case study involving human decision making. In: Simpson J, Kendrick D (eds.) Evolutionary Social Psychology. Erlbaum, Mahwah, NJ, pp. 345–86
26. Wilson D S, Near D, Miller R R 1996 Machiavellianism: A synthesis of the evolutionary and psychological literatures. Psychological Bulletin 119: 285–99