1 The author would like to thank George Benston, Robert Chirinko, Frank Forman, Bruce Johnsen,

Transcription

1 THE STATE OF NATURE AND THE EVOLUTION OF POLITICAL PREFERENCES: Abstract Paul H. Rubin 1 Department of Economics Emory University Atlanta, GA prubin@emory.edu Voice: December 11, 1998 Analysis of political and legal institutions often proceeds from a state of nature argument, where humans are analyzed as if they once existed as solitary individuals with no rules or government. In fact, sociobiology and evolutionary psychology, and study of related species (e.g., chimpanzees), indicate that humans have never existed as humans in such an environment. Rather, hominids (humans and their direct ancestors) have always been a group living species, and rules and preferences about rules would have evolved along with humans. As hominid brains became more complex, rules would also have become more complex, but there was never a time when humans lived with no rules and no time when rules were in any sense created de novo. These evolved rules and preferences for rules have implications for contemporary law and government. Human groups have generally been in conflict, and an important distinction has always been between group members and outsiders. It is useful to distinguish three classes of rules that would have existed in the state of nature. First would have been those defining group membership; these rules seem to be remarkably flexible. Second would have been rules regulating private conduct of individuals within the group. I argue that a libertarian regime, with few such rules, would have been unstable, which may explain why there are few persons with tastes for libertarian governments. Current resistance to libertarianism can be related to rules that would have been useful in evolutionary times, though the issue of the desirability of these rules today is worth further investigation. This analysis can explain why utility functions seem to contain elements of envy. The third class of rules would have been those regulating relationships between individuals within the group essentially, rules of property, contract, and hierarchy. Such rules are well developed and universal among humans. One implication of the argument here is that, while biological forces involving conflict created a value or demand for increased intelligence and group size, supply of energy to support intelligence in the form of, first, meat, and later, agriculture were also needed. The analysis also has implications for the willingness of individuals to contribute to public goods, including voting, and for the emphasis on identified rather than statistical individuals in political discourse. It can also explain the role of religion in economic and legal affairs. 1 The author would like to thank George Benston, Robert Chirinko, Frank Forman, Bruce Johnsen, Margaret Levi, Dean Lueck, Richard Epstein, Jack Hirshleifer, Peter Richerson, Vernon Smith, Som Somanathan, Gordon Tullock, David Sloan Wilson, and participants at seminars at Columiba University, Emory University, George Mason University, Montana State University, and New York University for helpful comments. The usual caveat applies.

2 THE STATE OF NATURE AND THE EVOLUTION OF POLITICAL PREFERENCES Paul H. Rubin Department of Economics Emory University Atlanta, GA Voice: Fax: INTRODUCTION Since at least the work of Hobbes and Locke, the hypothetical state of nature has been an important input into legal, political, and constitutional analysis. The notion is of humans with no law or government choosing a set of rules. The question asked is, What sort of rules would such individuals choose in this situation? A variant, the veil of ignorance is used for normative analysis the constitutional and other rules of society should be those that would be chosen by individuals behind such a veil who do not know their own position in the society to be designed. Major relevant examples include Buchanan and Tullock (1965), Rawls (1971) and Nozick (1974). A related question deals with the primacy of property and of government. It is often argued that that the state is first and that all property rights exist because the state has allowed them; e.g., Sunstein (1993, pp. 4-5). In this view, whatever rights people have are in an important sense a product of law. Obviously, the state of nature is meant as a metaphor, not as a true statement of primitive conditions. It has the advantage or providing a clean starting point for analysis what rules and structure would someone starting with absolutely no law or government choose. Nonetheless, I argue here that, although useful for some purposes, it is a misleading metaphor. That is, I claim that the true original state of humans is so different from the hypothesized state of nature that arguments proceeding from a state of nature origin generally cause confusion rather than enlightenment. Similarly, it is not meaningful to consider property as only defined by law; property and law are co-equal and neither precedes the other. In particular, all of these arguments imply that humans have more freedom in choosing institutions and rules than is actually the case. If real policies

3 The State of Nature Page 2 are based on such false constructions, then real suffering may ensue. Pipes (1990, p. 125) argues that this fallacy, which he traces to Helv tius, a disciple of Locke, was an important input into the Russian Revolution an event that did cause a good deal of misery. The argument that in a state of nature there were no rules is analogous to another fallacious argument common in the social sciences what has been called the Standard Social Science Model (SSSM; Cosmides and Tooby, 1992). In the past (and still, in some circles) it was thought that an individual human was a tabula rasa, a blank slate, and that humans could learn or be trained in almost any way. Students of human evolution, and in particular evolutionary psychologists, now know that this conception was false and in many respects misleading. There is room for wide, but not unlimited, variation in culture. Similarly, the notion that social rules are arbitrary or that such rules can be purely created by reason is false. My analysis is directly based on recent work in primate biology. Ultimately, the argument is grounded in sociobiology and evolutionary psychology as applied to humans. 2 The simple point is this: Social structure, property rights, and rule-like behavior are older than Homo sapiens, so that it is not meaningful to talk about human beings existing solely as individuals in an environment with no political or legal structure. There was never a time when human beings existed with no rules. The simplest and most powerful evidence is from behavior we observe in common chimpanzees, our nearest relative. When two species exhibit similar behavioral or physical characteristics, there are two possible biological explanations. The feature may be analogous derived and evolved separately. This will be the case when the most recent common ancestor of both species did not exhibit the feature. For example, the wings of birds and of bats are analogous, as are the shapes of fish and dolphins. On the other hand, the characteristic may be homologous derived by common descent from a common ancestor. If two closely 2 The original source is Wilson, See also Dawkins, 1976 (1989). For major works applying biological reasoning to human behavior, see Alexander (1987), Barkow Cosmides, and Tooby (1992) and Crawford and Krebs (1997). A pioneer was Robert Ardrey; see, for example, Ardrey, There are numerous applications in economics as well; see for a few examples Hirshleifer, 1977; Rubin and Paul, 1979; Smith, 1992; Robson, 1995; and Siow, Hayek has also written on evolution of preferences; e.g., Hayek, However, his writing was uninformed by more recent advances in evolutionary theory.

4 The State of Nature Page 3 related species exhibit the same characteristic or behavior and if it appears likely that the ancestral species also exhibited this behavior, then it is likely homologous. Moreover, if a feature is homologous in this sense, then all intermediate species from the common ancestor to the current species must also have exhibited the same feature. To the extent that we share traces of political behavior with chimpanzees, then the biological argument is straightforward that the behavior is by common descent. That is, behavior that is common almost certainly is inherited from a common ancestor, and not evolved separately. Just as we can be sure that the common ancestor of humans and chimps (and all ancestral species in between) had five-fingered hands, so we can be sure that this ancestor exhibited certain social behaviors. In particular, for example, chimps, humans, and ancestral and intermediate species lived in groups internally governed by a set of rules and certain hierarchies within those groups. Moreover, all also engaged in intergroup conflict. 3 It is in this sense that we may be sure that no human or close ancestor of any human lived a solitary life. Indeed, two recent books by respected scientists have documented carefully some of the commonalties in social and political behavior between apes and humans. De Waal (1996) has written about desirable features in the behavior of non-human apes, such as respect for property rights; Wrangham and Peterson (1996) have written about less praiseworthy aspects of primate behavior, including murder and something very like warfare. Ridley (1997) discusses in detail some of the relationships between the biological and economic approaches to the study of humans and human society. An important book dealing with the evolution of group and social behavior is Sober and Wilson (1998). Some apes do live isolated lives. Orangutans live alone, with males only visiting females periodically for procreation and females living with offspring only until the offspring are old enough to be independent at about ten years. Apparently, rape is common among these animals (Wrangham and Peterson, ); of all apes, these may be the most purely Hobbesian. Gibbons live in mated pairs, but not in larger groups. It is 3 An exception appears to be the bonobo, the pygmy chimpanzee. This animal seems to have lost much of the aggressiveness of other chimpanzees. However, behavior indicates that the species still has some aggressive characteristics, and that the ancestral species was very similar to common chimpanzees. See Wrangham and Peterson, 1996, Chapter 10.

5 The State of Nature Page 4 interesting that the evolutionary history of the apes indicates that gibbons first and then orangutans split off from the ancestor of the remaining apes, so that humans are more closely related to gorillas and chimpanzees than to these solitary species. Gorillas and chimps both live in groups, although of different sorts. Thus, the oldest common ancestor that may not have lived in groups was the ancestor of all of the apes. It is more likely, however, that the ancestral proto-ape did live in groups since virtually all primates are social (Foley, 1995), and that gibbons and orangutans have reduced or lost the natural sociality common to most primates. Thus, the argument is that when our ancestors first became human, they already had in place a set of rules. There was never any point in time when humans lived without social rules. As our ancestors became more intelligent and more sophisticated, the rules would have become more complex, but a state in which humans lived without rules would not be defined. The state of nature would then describe the rules that existed when humans first came into being as humans. In order to determine what these rules might have looked like, it will be useful to consider what has been called the Environment of Evolutionary Adaptedness (EEA; Barkow, Cosmides and Tooby, 1992; Crawford, 1998). This is the time in which humans evolved from apes or from ancestral species, such as Homo erectus, to humans. Based on these arguments, this paper will provide at least a tentative explanation for on several aspects of behavior that economists find puzzling. Some examples: Why do people contribute to public goods (including voting)? Why is there more cooperation in experimental situations that economic theory would predict? Why do many or most individuals seem to feel envy, and why do social policies often seem to embody this preference? Why do most or all societies interfere with private behavior that creates no obvious externalities, and spend significant resources on regulating such behavior? Why is religious belief, which is untestable, so common among individuals, who are willing to spend real resources in furtherance of these unverifiable beliefs? Why do supporters of particular political policies so often rely on anecdotes and identification of particular individuals who will be influenced by the policy, rather than on objective statistical evidence regarding the policies?

6 The State of Nature Page 5 The argument here is different than Becker s (1996) argument regarding social norms. Becker argues that the upper classes will compensate others for adopting efficient social norms that benefit the upper classes. The norms involved are in some cases similar to the tastes discussed here. However, the argument here is that certain tastes have evolved; Becker does not address this issue. His individuals seem to follow the Standard Social Science Model, discussed above; that is, they have only vaguely specified and highly malleable preferences. In addition, he assumes that people gain utility from attending church, but does not explain why this would be so. In this paper, I assume that tastes, including a taste for belief in religion, have evolved through a biological evolutionary process. In the next section, I discuss the EEA. A key point is that humans lived in groups and distinguished between group members and non-members. I discuss the definition of group membership. I then discuss rules governing private individual behavior within the group and provide an explanation based on gene survival for some otherwise puzzling rules of moral behavior; I show that a libertarian government (a government that provides only those functions that a purely self interested individual would desire) would not have been stable, so that libertarian preferences would not have been selected for. I then consider rules governing relations between group members. The last substantive section discusses the surprising (to an economist) willingness of humans to contribute to provision of public goods, and provides an evolutionary explanation for this willingness. Public goods are divided into two categories, charitable and group goods, with the latter including voting. Explanations are provided for contributions to each type, and testable implications are derived from these evolutionary explanations. The final section is a summary. THE ENVIRONMENT OF EVOLUTIONARY ADAPTEDNESS (EEA) To understand the applicable evolutionary theory, it is necessary to consider the characteristics of the environment in which humans evolved. The relevant time period is the Pleistocene, the period lasting from three million to about 10,000 years before the present, when farming began (Diamond, 1998). It is during this period that humans evolved from a rather ape-like creature to modern humans. There would have been

7 The State of Nature Page 6 selection pressure for behaviors and preferences that would have led to reproductive success in that environment, and modern humans will retain those preferences and behaviors. It is generally believed that the time since farming has been too short to have caused significant evolutionary changes in behavior. There are several features of this environment that are relevant for understanding the state of nature and current political preferences. While the EEA itself would have not been constant, there are some features that probably would not have changed, and these features would have had evolutionary implications. It is these relatively constant features that I discuss. Features that would have remained constant would have provided continuous selection pressure for behaviors adapted to these features. In general, it would be cheaper for natural selection to seek behavior adapted to these constant features, rather than allow individuals each generation to learn how to adapt de novo. The method of adaptation would be to generate tastes for certain behaviors that would have created fitness with respect to these constant features. Thus, these constant features would be relevant for understanding current utility functions. For example, food that tastes sweet always contains sugar, a source of energy. It would have been possible for parents in each generation to teach children to eat sweet foods, but since such foods were always available and always desirable it was cheaper for natural selection to endow humans with a taste for sweetness. This paper deals with similar tastes for various forms of social interaction. Natural Selection There are several additional general points regarding natural selection that are relevant. First, evolution selects for fitness propagation of genes into the future. Economists assume maximization of utility. Utility may be directly related to fitness. We enjoy those things that caused our ancestors to survive and leave more offspring. However, the form of utility functions would be that form that would have led to fitness maximization. If fitness maximization requires unusual or non-standard relations between consumption or wealth and utility, then these relationships would have been selected. For example, there are circumstances where more altruism than economists would expect is fitness maximizing. On the other hand, in some circumstances relative rather than absolute wealth may be the relevant variable. It is also possible to explain risk preferences (Rubin

8 The State of Nature Page 7 and Paul, 1979) and time preference (Rogers, 1994) in terms of evolutionary evolved tastes. Additional examples are provided below; see also Hirshleifer (1977), Frank (1988) and Robson (1995). Second, many of the results discussed below are the outcome of mixed strategies and of density dependent selection processes. Thus, individuals may sometimes randomize their behavior, as discussed by Miller (1997). Also, individuals in a species may follow different strategies, and many selection processes have equilibria with more than one strategy (e.g., Hirshleifer, 1998b.) If some behavior is said to have fitness improving properties, this should not be taken to imply that all individuals will exhibit this behavior, or even that any one individual will consistently exhibit the behavior. These observations are also an explanation for individualism. Humans have different genetic endowments and choose different strategies, which is an argument for different tastes. These differences also explain why individuals want some freedom from social control. Third, evolution has endowed us with certain preferences, those that would have been fitness maximizing in the EEA. However, humans are extremely facultative that is, able to change behavior in response to environmental changes. Humans are the most facultative animals in the world. Indeed, this facultative behavior is the basis of economics, which studies behavioral responses to changes in relative prices. The nature of facultative behaviors as understood by biologists is discussed in Sober and Wilson, 1998, p Thus, while we may have certain preferences based on evolution, we are able to modify behavior in response to changes in environments relative prices, where price is broadly defined. It is these responses that are the subject matter of economics. Addition of biological analysis by no means makes the subject matter of economics redundant. While attention to evolutionary analysis will, I believe, improve economic analysis, I expect that many core areas of economics will be unaffected. Fourth, natural selection is itself a great economizer. It will use whatever mechanisms are easiest and most readily available. If incest is biologically undesirable because children of incestuous relationships are less fit (as is true) and if individuals raised together are generally siblings, then there is no need to evolve a method that siblings can use to recognize each other in order to avoid sexual relations. Rather, humans have

9 The State of Nature Page 8 evolved to find sexual relations with those with whom they were raised undesirable. While this mechanism is not perfect (sometimes unrelated individuals will be raised together, and sometimes siblings will be raised apart), it was relatively easy to evolve, and good enough to work. We will see similar mechanisms in other aspects of behavior. Fifth, natural selection is contingent. While it may be possible ex post to explain the origin of a species, it is impossible ex ante to predict the evolution of that species. Predictive claims can be made regarding the effect of particular parameters on evolution for example, species that evolved in cold climates will, ceteris paribus, be larger in body size than similar species that evolved in warm climates. But in discussing evolution of human intelligence, it is not useful to ask why other species did not also evolve such intelligence. Sixth, the arguments in this paper are about preferences, or utility functions. Human behavior as understood by economists is the maximization of preferences subject to constraints. Thus, actual behavior is only partially determined by preferences. To explain observed behavior, it is necessary to consider prices as well. Finally, this paper ignores the role of culture. But culture will also have an important effect on actual behavior, and will operate through both prices and preferences. Moreover, culture and genes can evolve together, so that some of the evolutionary processes I describe may have co-evolved with culture (Boyd and Richerson are leading advocates of this view; see their papers cited in the references; see also Janicki and Kreps, 1998). Thus, it would be incorrect and counterproductive to view this paper as explaining the actual outcome of behavior. It would also be incorrect to view this paper as implying that behavior is solely determined by the factors identified here. Rather, features considered in this analysis are only one input into the determination of final outcomes. However, they are an input, and understanding these factors will improve our understanding of human behavior. Social Structure Society consisted of relatively small bands (probably, beings) of genetically related individuals. By the late Pleistocene, bands were probably aggregated into socially

10 The State of Nature Page 9 significant tribes of up to a few thousand. 4 Hominids 5 have always been somewhat social creatures. Within these bands, initially the basic family unit was likely a female and her children, and perhaps her mate. As we will see, groups of male kin were also extremely important in understanding human evolution. Over evolutionary time, children became increasingly dependent on long term parental care as intelligence and thus the period of dependency increased. This increase in dependency with increasing intelligence would have been for two reasons. First, increasing brain size at adulthood coupled with a limited size birth canal means that larger brained (more intelligent) offspring must be born with immature and not fully grown brains. Based on other apes, it is estimated that the human gestation period is actually 21 months, with the last 12 months occurring after birth. (Jones et al., 1992, p. 87.) Second, increased intelligence means increased value of learning, and so again a longer dependency period. This increase in dependency in turn caused increased benefits to males from spending resources on raising their children, leading to long term bonding between males and females. This is because males who helped raise dependent children would have had more surviving offspring, so genes for this behavior would have been selected for and would have increased in the population. Based on relative sizes of males and females, it is likely that humans and their ancestors were mildly polygamous. Some successful males would have had perhaps two or three mates. Ancestral hominid bands made their living by hunting and gathering, with the role of hunting probably increasing over time. Indeed, it was probably the additional protein available from meat that enabled humans to bear the high energetic costs of our large brains and thus led to the increase in intelligence of humans relative to all other species (Foley, 1995). Males engaged in hunting, often cooperative hunting, and females in gathering; sexual division of labor among humans is an evolutionarily old characteristic, probably antedating humanness, (Silverman and Phillips, 1998). It is universal among humans (Brown, 1991), although it does not occur among other apes. 4 Peter Richerson, personal communication. 5 Hominids are any member of the human family Hominidae (all species of Australopithecus and Homo). Jones et al. 1992, p. 464.

11 The State of Nature Page 10 Humans and chimpanzees are each others closest living relatives; chimps are closer to humans than to gorillas. Chimps and humans are unique even among apes. They are the only primate species in which male kin groups are important (Foley, 1995; Wrangham and Peterson, 1996). This means that male bonding within the group is more significant than is true for other primates. It also means that conflict between male kin groups is more common. Both of these features are relevant for understanding human evolution and human behavior today. Moreover, humans and chimps are the only primate species in which males engage in coalitional aggression aggression by an organized group of more than two males 6 (Tooby and Cosmides, in press; Wrangham and Peterson, 1996.) Goodall (1986) describes chimpanzee coalitions and aggression in detail. A question that puzzles biologists is the basis for large-scale social organization. Biologists have identified two mechanisms for social cooperation: kin based altruism, and reciprocity. Kin share many genes in common. Kin selection is said to exist when an altruistic behavior (which would include cooperation in a prisoner s dilemma) towards relatives is selected because this increases the representation of the shared gene. For a discussion, see Dawkins, 1976 (1989). To the extent that members of groups are related, then this mechanism comes into play to explain cooperation. Altruism towards kin is dependent on costs and benefits: as the cost to the altruist decreases or the benefit to the recipient decreases, and as relatedness increases, such altruism becomes more likely. An altruistic act from X with a cost of CX providing a benefit to Y of BY where X and Y are related by k 7 is more likely as (BY)(kXY)-CX increases (Shaw and Wong, 1989). However, for large groups, kin based altruism is clearly insufficient. The second mechanism is called reciprocity. In a famous paper, Trivers (1971) discovered contract in a biological setting and provided a biological basis for exchange. However, this mechanism seems to work only for small groups. Hirshleifer (1998b) points out that there are other pathways for the evolution of cooperation, and that biologists err by emphasizing only kinship and reciprocity. Alexander (1987, p. 85; see also Kreps, 1998, p. 347) discusses indirect reciprocity which is a complex multilateral system of reciprocity. 6 Dolphins may also engage in this behavior: Tooby and Cosmides, in press, p. 3.

12 The State of Nature Page 11 Boyd and Richerson (1989) show that this system would be difficult to sustain in a system of networks each containing several individuals. However, this analysis would be greatly simplified if what is called indirect reciprocity were viewed as multilateral exchange within a general equilibrium framework, and that, as economists have known, everyone can benefit from such exchange systems. This is clearly an area of research where the biologists would benefit from inputs from economists. This is an additional economic principle in human evolution, in addition to those identified by Smith (1992). Additional such principles are described below. Richerson and Boyd (1998a) suggest that social cooperation is caused by two separate mechanisms. One is the evolutionarily old cooperative mechanism shared to some extent by all primates. This operates primarily in face-to-face environments and among small sets of individuals. The other is unique to humans and involves much larger networks of exchange and cooperation. Both mechanisms still operate, but in different contexts. For example, criminal organizations use the former mechanism (Richerson and Boyd, 1998, p. 6). Larger mechanisms include cooperation among tribal-sized entities. Rubin (1994) discusses such mechanisms in the context of Russia and other economies lacking legal institutions; see also Landa (1981). In this view, large modern institutions are made up of smaller components consistent with the evolutionary background because there has not been time for mechanisms consistent with this large scale to evolve. However, Richerson and Boyd (1998a) argue that complex societies require command and control mechanisms; they seem to ignore the important role of markets in such coordination because they believe that markets require command and control mechanisms to function. 8 This raises an interesting question: How important are evolutionary adaptations for functioning in large, impersonal markets with either endogenous or outside enforcement mechanisms? The answer to this question (which may well be Not very ) will be important for determining the impact of evolutionary thinking on the core of economics and of law. 7 k is the fraction of shared genes. It is 1 for identical twins;.5 between parents and offspring or between full siblings;.25 between uncles and nephews or between half-siblings; and.125 between first cousins. 8 Peter Richerson, personal communication.

13 The State of Nature Page 12 Conflict The most remarkable feature of humans is our level of intelligence. This is sufficiently high so as to be qualitatively different from any other species. Evolution of this level of intelligence is an extremely improbable event in biology; it has evolved only one time in all the billions of years of evolution of life on earth (Mayr, 1988). A key point for understanding the evolution of human behavior and intelligence is that conflict has long been an important part of ape behavior (Wrangham and Peterson, 1996) and was undoubtedly important in human evolution as well. Chimpanzees and humans, because of their male kin bonding system, place more emphasis on such competition than even other apes (Foley, 1995). This is because it is easier for groups of kin to overcome various free rider problems than for non-related individuals. (For a cost-benefit model demonstrating this formally, see Shaw and Wong, 1989, or Hirshleifer 1998b.) As mentioned above, humans and chimps are the only apes to engage in competition between large coalitions. This is apparently because such coalitional competition requires greater intelligence to monitor behavior and control free riding than is available to other species (Tooby and Cosmides, in press). Current thinking among many biologists is that the only force strong enough to provide sufficient selection pressure to lead to this level of intelligence would have been competition from other hominids. 9 That is, predation and intergroup conflict was so pervasive in proto-human history that it was the selection pressure from this conflict that generated our massive level of intelligence. While data on pre-human hominids is lacking, Keeley (1996) does indicate that warfare in primitive humans societies is a more significant source of death than in advanced societies, even when major wars are included. This pressure for evolution of intelligence driven by competition can be self-enforcing, and can lead to a positive feedback system and runaway competition. In this argument, formation of kin-based coalitions was the force triggering the process. 9 This argument was first made by Humphrey, An important authority who has made this point in numerous publications is Richard Alexander; see, for example, Alexander, Similar arguments are in Bigelow, 1969, Byrne and Whiten, 1988, and Whiten and Byrne, 1997 and, from an economic standpoint, Robson, Dunbar, 1996, also argues for a social origin of intelligence.

14 The State of Nature Page 13 This is a place where economic principles can again be helpful. Coalitional competition can lead to an increased value or demand for intelligence. However, for this demand to actually lead to increased intelligence requires additionally a supply or resource side. Initially, for the growth of the brain, supply was in the form of sufficient sources of energy from meat to support the larger brain, which is a calorically or energetically expensive organ. 10 As we see below, much later agriculture was the supply force leading to increased size of human groups. We may identify three equilibria: For chimps and our early ancestors, there was no meat eating, and the equilibrium was the small band, held together by ancient primate social forces. Meat eating led to a larger equilibrium perhaps at the level of the tribe, where the second level social forces identified by Richerson and Boyd (1998a) would have come into play. Finally, agriculture is associated with much larger groups; indeed, we have not yet observed the equilibrium size group associated with agriculture and modern technology. Here, the command and control forces discussed by Richerson and Boyd (1998a) and market forces are responsible for social coordination. This competition would likely have been both within group competition and between group competition. Alexander (1987) argues that the two would have reinforced each other: as between group competition caused groups to become larger over time, then within group selection pressure would have increased. This within group competition would have also lead to increased intelligence. For example, competition for dominance among males and the corresponding increased procreative success would have placed a greater premium on intelligence as group size increased. Moreover, Alexander (1987) and Crawford (1998) argue that if the major environmental force leading to human adaptation was competition, then the current evolutionarily relevant environment may not be very different from the EEA in many biologically relevant respects; both forms of competition still exist. Rubin and Somanathan (1998) have argued that these sources of intelligence are still relevant for understanding human behavior. 10 Meat also enabled our ancestors to reduce the size of their digestive system, again leading to a saving in energy (Aiello and Wheeler, 1995).

15 The State of Nature Page 14 One key force would be selection to engage in opportunistic cheating, and selection to avoid being the victim of a successful cheater. Cosmides and Tooby (1992) have shown that the human mind has evolved to be well adapted to detect cheating. They have also shown that humans are well adapted to distinguish fair weather from true friends (Tooby and Cosmides, 1996). The other force would be the ability to function as a member of a hierarchical group including both relatives and non-relatives. Such functioning would have been necessary to compete successfully with other groups, where competition would have included actual combat. Tooby and Cosmides (in press) discuss in detail the complex cognitive mechanisms necessary for successful intergroup competition, including the ability to monitor numerous coalition members to assess contributions and measure risk bearing. They argue that only a few species have been sufficiently intelligent for this behavior, even though individuals in other species would benefit from such behavior if they were able to perform it. Note that the key to evolution of intelligence was competition with other humans, not with other species or with the environment. This is shown by the fact that many predecessor hominid species, such as Homo erectus, were quite successful and lived (as a species) for very long periods of time and in many different physical environments. At some time one group of this species evolved into anatomically modern humans, probably with several intermediate species; Homo erectus probably did not change as an entire species into modern humans. (For a contrary view, see Wolpoff and Gaspari, 1997.) Our intelligence enables us to handle physical objects and other non-human elements, but this was not the original source of evolutionary pressure. 11 Of course, as intergroup competition became more important, ability to build and use weapons would also have become more important, and may have played a role in selection as well. Moreover, increased intelligence is associated with better use of the environment ( ecological intelligence ) and greater ability to obtain food. 11 It is possible that climate change and variability in the Pleistocene also had a role in this process, perhaps by leading to larger brains. Consistent with this argument is the fact that brains of many mammalian species increased during this period. However, the sheer magnitude of the increase in human intelligence (both absolutely and relative to all other species) seems to me (and to many others, some cited above) to argue that something more was needed. See Richerson and Boyd, 1998b.

16 The State of Nature Page 15 Social behavior has substantial costs, and a species will be social only if benefits outweigh these costs. For example, conspecifics (members of the same species) are competitors for resources. Also, conspecifics are rivals for mates, generating internal conflict. Social living enables disease to spread more efficiently; the contagious diseases, major killers in historic (and presumably prehistoric) times, cannot exist until population density reaches a certain level, as discussed in Diamond, (1998, Chapter 11). Larger groups must travel farther to find sufficient resources. Thus, living in groups generates large costs of various sorts. Therefore, in many circumstances, it would seem to make individual sense to live a relatively solitary life. One of the major benefits of group living is the ability to minimize predation. There are economies of scale in this activity. Group members can warn each other of the approach of a predator, thus reducing the average time that each individual must spend in defensive observation. They can also unite to fight off a predator that is larger or more powerful than any individual. Most group living species do so for exactly this reason. In the case of humans, there is no non-human predator sufficiently powerful to explain the large size of human groups (even in prehistoric or evolutionary times). But predation from other human or hominid groups is and has been sufficiently powerful to induce humans to live in increasingly large groups. Moreover, this predation would have provided powerful pressures for selection for intelligence. This would have occurred from three sources. First, groups with more intelligent members would have had an advantage in conflict with groups of less intelligent individuals. Second, as group size became larger, it would have required more intelligence simply to keep track of individual relationships with other group members. Dunbar (1996) argues that speech evolved to replace grooming behavior as a method of forming bonds with individuals. He claims that speech became necessary as group size increased beyond the level that grooming could support. Third, competition within the group for resources and mates would have become more intense, and therefore there would have been increasing selective pressure within the group for increased intelligence. Between group pressures would have selected intelligence in males, who are the main fighters. However, within group pressures would have selected intelligence in both

17 The State of Nature Page 16 males and females. Because males invest significant resources in children, females would have sought more intelligent mates. But exactly because males also invest in children, males would also have sought more fit females for long term mating, where fitness would include intelligence, so that children in whom the males invested would have been more fit. For short term liaisons (which males are also selected to seek) fitness of females would have been less important since males would not have invested any resources in offspring of such unions. Females, on the other hand, would seek unions with higher quality or more intelligent males, even if there were not their long-term mates, to improve the quality and fitness of their offspring. Thus, to understand the state of nature we must replace the Hobbesian world of individuals in conflict with a world of groups in conflict. This difference in perspective has profound implications for the nature of the evolution of rules. A world of individuals and conflict between individuals is truly Hobbesian; there would initially be no rules, and the world would be one of each for himself. Any subsequent rules would be created by these individuals to reduce conflict. In a world of groups and conflicts between groups there would be pre-existing, evolved but non-created rules governing behavior within the group. Those groups who utilized the best rules would be the most successful at competing with neighboring groups, and would therefore expand their share of the population. Thus, rules governing social actions of individuals would have come into being along with humans themselves. It is not meaningful to ask about the life of human beings living in isolation with no social structure. Moreover, since rules evolved along with humans, it is not meaningful to ask what rules humans in a totally ruleless state would choose. Such a world has never existed and in principle cannot exist. Types of Rules In examining the nature of the pertinent rules, we may identify three types that would be relevant: those determining membership of a group; those controlling behavior of a single individual within a group; and those controlling behavior between individuals within a group. It is likely that humans have a predisposition to accept certain specific forms of rules. We seem to learn certain types of rules and behaviors, discussed below,

18 The State of Nature Page 17 too easily for there to be no biological mechanism underlying this learning. 12 For example, Israeli kibbutzim attempted to impose a rule that all children would be raised in common with no special attention from biological parents, but humans (even dedicated and committed ones) seemed unable to learn this rule, in that parents continued to want to spend time with their own children (Shepher, 1971.) There is a large literature discussing interactions between genetic and cultural evolution, summarized in Janicki and Kreps, An important lesson from this literature is that cultural and genetic evolution can reinforce each other. For example, if the ability to learn some behavior (such as obedience to a particular rule) leads to increased fitness, then there might be selection pressure for just this ability. Those able to learn the rule most easily would be more successful (have more offspring) and so the ability could become genetic. Shaw and Wong (1989) argue that one thing humans are programmed to learn is ethnocentrism, a preference for members of our own ethnic group, and xenophobia, a dislike or fear of members of other groups. However, they also claim that this bias is relatively weak. The argument of this paper will be based on the assumption that individuals will be selected to desire rules that would have increased their chances of leaving offspring in the EEA. Then rules of behavior may be explained in terms of maximizing the probability that the group using these rules would survive in a conflict with other groups. 13 This argument would apply to all three types of rules mentioned above. The fundamental conflict in groups is between behavior enhancing individual fitness and that favoring the group. One can think of numerous behaviors that would lead to increased probability of survival of the group at the expense of the individual, and conversely; prisoner s dilemmas are everywhere. I discuss some of these conflicts below. I begin by discussing factors determining who is a group member. GROUP MEMBERSHIP A key distinction is between members of the group and outsiders. All humans make this distinction (Brown, 1991, p. 136). Indeed, even infants distinguish between 12 By analogy, the ability to learn language and the deep structure of language is clearly innate; see, for example, Pinker and Bloom (1992). Much of Barkow Cosmides, and Tooby (1992) is devoted to a discussion of innate rules of various sorts.

19 The State of Nature Page 18 group members and non-members. Among infants, males make this distinction more strongly than females. Moreover, although common physical appearance makes identifying group members easier for infants, it is not necessary (Premack and Premack, 1994). All humans also recognize subgroups based on kinship, sex, and age (Brown, 1991, p. 137). Krebs and Denton (1997) discuss some of the features of perceptions of group membership. They indicate that humans quickly classify people as members or nonmembers of the in-group. Members and non-members are treated very differently. All non-members are viewed as being alike and as being enemies, while members are treated as individuals. Group affiliation mechanisms seem quite flexible. For example, in one experiment, when students were randomly assigned to a group they immediately identified with this group, even though they knew that the basis for assignment was random. Mullen et al. (1992) discuss a large number of studies from social psychology on the importance of ingroup bias ; see also Tajfel, 1970, on the power of arbitrary of group loyalties. Rubin and Somanathan (1998) discuss the implications of this flexibility for employee productivity. Kuran (1998) shows that strength of ethnic identification is variable and can shift very quickly. Sober and Wilson (1998, p. 92) define a group in biological or fitness terms: [A] group is defined as a set of individuals that influence each other s fitness with respect to a certain trait but not the fitness of those outside the group. This definition is consistent with individuals belonging simultaneously to more than one group, with different groups associated with different traits. In the EEA, members of the group would generally have been close relatives, so that kin selection would have been important. This may be why humans seem to have evolved to be quite flexible in their group preferences. If everyone nearby were a relative, then there would have been no need for a mechanism distinguishing kin from unrelated neighbors, although ability to measure closeness of kinship would be valuable. This argument is analogous to the argument regarding incest, discussed above. Nonetheless, in a conflict with a neighboring group, all members of one s group could be viewed as kin. Sober and Wilson (1998) argue that there are more general group membership 13 As discussed below, this does not imply group selection, although it is not inconsistent with the possibility of group selection.

20 The State of Nature Page 19 mechanisms. Crawford (1991) suggests that kin recognition mechanisms might have operated in any of several ways: geographic, treating nearby individuals as kin; treating frequent associates as kin; treating individuals physically similar to oneself as kin; and treating individuals with particular genetic markers as kin. Shaw and Wong (1989) suggest that group markers include language, religion, phenotype (that is, body type and appearance), homeland, and myth of common descent. Group identification is stronger as more of these markers are congruent. Shaw and Wong (1989) argue that patriotism can conflict with group or ethnic identity if these markers are more congruent within ethnic groups than within a country. 14 Because Shaw and Wong (1989) are concerned with war, however, they seem to have missed or ignored other non-governmental potentials for group loyalty and identification. Hirshleifer (1998a) also discusses the rule of group identification in war. In addition to the experimental evidence cited above, the types of individuals with whom modern humans feel solidarity show the flexibility of whatever mechanism is involved. Groups include actual kin or ethnic groups, but also: members of the same religious group; citizens of the same country or (with respect at least to sports events, the same city); members of the same occupation; employees of the same division within a company (or department within a university) or the same company; members of the same gender; and numerous other groups. The key distinction is between rules involving individuals in the same group and rules involving outsiders; all humans make this distinction. Religion can serve to define members of the relevant group. All human groups have some religious beliefs (Brown, 1991, p. 139; Burkert, 1996). In one major survey of beliefs, 77% of the respondents indicated that they believed in God, and in no country did fewer than 36% believe in God (Inglehart et al ) I discuss below individual benefits from such belief. It is likely that religions were originally tribal, so that co- 14 They also suggest that in the U.S. identification is strong because, since American citizens are descended from immigrants, there is little ethnic group identification to compete with patriotism, whereas in the (then) U.S.S.R. they indicate that patriotism was generally weak because there was strong ethnic group competition. Subsequent events seem to have proved them correct, at least with respect to the U.S.S.R. 15 This survey, for , was conducted in 43 countries.