sign in sign up
<<
Home , Coevolution, Cultural group selection, Dual inheritance theory, Memetics

± 1

Gene±Culture Coevolution Intermediate article Kevin N Laland, , Cambridge, UK

CONTENTS Introduction Do and culture coevolve? Evidence for transmitted culture Conclusion Types of cultural selection

Evolution in with a dynamic, socially trans- skills were transmitted from one generation to the mitted culture may be different from in next, these simple artifacts represent the earliest other species. Population geneticists have pro- evidence for culture. In fact, comparative evidence posed the gene±culture coevolutionary approach for social learning in a variety of vertebrate species to describe the way in which cultural change may suggests that cultural transmission almost certainly drive a population's biological evolution. preceded Homo habilis by a considerable length of . However, social learning in other animals is INTRODUCTION rarely stable enough to support traditions in which accumulates from one generation to Many researchers have noted analogies between the next. For at least 2 million years our ancestors the processes of biological evolution and cultural have reliably inherited two kinds of information, change. For instance, both genes and culture are one encoded by genes, the other by culture. informational entities that are differentially trans- There is only one evolutionary approach to the mitted from one generation to the next. These simi- study of that takes up the chal- larities have led to the that culture evolves, lenge of understanding genetic and cultural evolu- and prompted the development of mathematical tion simultaneously by focusing directly on their models of . interaction. Gene±culture coevolutionary theory The main scientific approach to the study of (or ), together with evolu- how culture evolves is a branch of theoretical tionary and human behavioral , , known variously as `cultural is one of three principal evolutionary approaches evolution', `gene±culture coevolution', or `dual in- that emerged in the aftermath of the human heritance' theory. This intellectual tradition has debate (Smith, 2000). nothing in common with the nineteenth-century Conceptually, gene±culture coevolution is like a `cultural evolution' schools, which, based on an between and evolutionary psych- erroneous view of evolution as progressive, set ology, although its methods are quite different, out to model stages of societal development. relying as they do on rigorous mathematical Rather, the population genetics approach regards theory. Like memeticists, gene±culture coevolution culture as an evolving pool of , beliefs, values, enthusiasts treat culture as evolving learned know- and knowledge that is learned and socially trans- ledge. Like evolutionary psychologists, these re- mitted between individuals. Researchers focus on a searchers believe that the cultural knowledge single trait, such as a preference for drinking milk, individuals adopt may sometimes ± although cer- or for sons over daughters, and employ a rigorous tainly not always ± depend on their genetic consti- mathematical approach to describe how the cul- tution. Moreover, selection acting on the genetic tural trait changes over time, sometimes coevolving system is commonly generated or modified by the with . Where the cultural entity is spread of cultural information. For gene±culture a discrete package, it has much in common with coevolutionary theorists, the `leash' that ties culture ' idea of the `', defined as a to genes tugs both ways. The advent of culture was cultural analogue of the gene (Dawkins, 1976). a precipitating evolutionary milestone, generating Stone tools appear in the archeological record selection that favored a reorganization of the approximately two and a half million years ago. , and leaving it specialized to acquire, If, as is widely believed, lithic technologies and store, and use cultural information. It was culture, 2 Gene±Culture Coevolution loosely guided by genes, that allowed humans the to the of genes. Gene±culture adaptive flexibility to colonize the world. coevolution exhibits a concern for nonadaptive The quantitative study of gene±culture coevolu- and even maladaptive outcomes of the evolution- tion began in 1976, when two population geneti- ary process. This stance continues both to surprise cists, Luca Cavalli-Sforza and Marc Feldman of and confuse outside observers used to characteriz- Stanford University, published the first simple ing all these evolutionary approaches as `sociobiol- dynamic models with both genetic and cultural ogy'. However, the rigorous theoretical approach inheritance. The fundamental innovation that has led to little experimentation or other forms Cavalli-Sforza and Feldman instigated was that, in of empirical work, and this school remains the addition to modeling the differential transmission prerogative of a comparatively small band of of genes from one generation to the next, they in- workers. corporated cultural information into the analysis, The emerging body of theory has developed in a allowing the evolution of the two systems to be variety of ways. One class of models investigates mutually dependent. However, one curious feature the inheritance of behavioral and personality traits, of the history of gene±culture coevolution is that extending traditional models by incorporating a both archetypal sociobiologists and some of their transmitted cultural component into the analysis. most severe critics almost simultaneously recog- Other models address general questions about the nized the importance of gene±culture interactions, adaptive advantages of learning and culture. More with each starting to develop methods to address recently, these methods have been applied to ad- the problem. By the late 1970s, Charles Lumsden dress specific cases in which there is an interaction and Edward Wilson at Harvard University were between cultural knowledge and genetic variation engaged in a race with Cavalli-Sforza and Feldman that influences its prevalence. These include the to produced the first book on this topic. While evolution of language and of handedness, an an- Lumsden's and Wilson's Genes, and Culture alysis of changes in the genetic sex ratio in the face was published first (Lumsden and Wilson, 1981), it of sex-biased , the spread of was not well regarded (Maynard-Smith and agriculture, the coevolution of hereditary deafness Warren, 1981). In contrast, Cavalli-Sforza's and and sign language, the emergence of incest taboos, Feldman's more cautious tome Cultural Transmis- and an exploration of how cultural niche construc- sion and Evolution (Cavalli-Sforza and Feldman, tion affected (see Feldman and 1981) was much better received. Laland, 1996). Together with many co-workers, Cavalli-Sforza As the rules of cultural transmission are usually and Feldman gradually built up an impressive different from those of genetic transmission, simi- body of mathematical theory exploring the pro- lar selective regimes may result in very different cesses of cultural change and interaction between equilibria. A good example of this is provided by genes and culture. Frequently they took advantage the hypothesis of Boyd and Richerson (1985) that of the parallels between the spread of a gene and can act on cultural variation. The the diffusion of a cultural innovation to borrow or theoretical argument against group selection is adapt established models from population genet- based on models which assume genetic inherit- ics. Drawn by the ongoing sociobiology debate, ance, and the criticisms may not hold for culturally other mathematically minded researchers joined transmitted traits. When individuals adopt the be- the fray, most notably anthropologists Rob Boyd havior of the majority a conformist transmission is and , whose book Culture and the generated (Boyd and Richerson, 1985). As a result Evolutionary Process introduced a variety of novel of its frequency dependence, conformist transmis- theoretical methods and stimulating ideas (Boyd sion can act to amplify differences in the frequency and Richerson, 1985). Gradually a consensus of cultural traits in different subpopulations, but as to the most appropriate methods for tackling reduce variance within groups. Boyd and Richer- gene±culture interactions began to emerge, which son showed that one of the by-products of a con- today forms the basis of modern coevolutionary formist bias is an increase in the strength of the theory. group selection of cultural variation so that it may The technical and explicitly mathematical be a strong force relative to forces acting within of modern gene±culture coevolution is one of sev- groups, such as natural selection. Since selection eral features that distinguishes this perspective between groups may favor beliefs and attitudes from alternatives such as . that benefit the group at the expense of the individ- A second is the incorporation into analyses of a ual, Boyd and Richerson's theory provides a new variety of genetic and cultural processes in addition explanation for human cooperation. Gene±Culture Coevolution 3

EVIDENCE FOR TRANSMITTED different from other aspects of the environment is CULTURE the knowledge passed between individuals. Cul- ture is transmitted and inherited in an endless For most social scientists `culture' is a given. The chain, frequently adapted and modified to produce notion that much of the variation in the behavior of cumulative evolutionary change. This infectious, humans is brought about by their exposed to information-based of transmission is divergent is so widespread and intuitive what allows culture to change rapidly, to propagate that is beyond dispute. While it used to be fashion- a novel behavior through a population, to modify able to define culture as the interwoven complex of the selection pressures acting on genes, and to exert behavior, ideas, and artifacts that characterize a such a powerful influence on our behavioral devel- particular people (e.g. Tylor, 1871), among social opment. scientists this view has been superseded by a more Gene±culture enthusiasts point to countless stud- cognitive perspective that restricts culture to infor- ies that have found that the attitudes of parents and mation stored in the brain. In contrast, biological offspring are similar. They maintain that the most approaches to culture (including those of most obvious explanation for this is that children learn sociobiologists, human behavioral ecologists, and social attitudes in the family. For instance, a study evolutionary psychologists) tend to regard the of Stanford University students revealed that the transmitted elements of culture as either exerting religious and political attitudes were strongly a comparatively trivial influence on human behav- consistent between parents and offspring (Cavalli- ior, or that whatever influence they have is strictly Sforza et al., 1982). The same applies to nonindus- circumscribed by genes. trial societies. For instance among Aka pygmies, an For advocates of gene±culture coevolution these African group of hunter-gatherers, there was evi- biological perspectives underemphasize one crit- dence for parent to child transmission of many ical factor: socially transmitted culture. Too much customs (Hewlett and Cavalli-Sforza, 1986). Such culture changes too quickly to be feasibly explained correlations do not prove cultural transmission by genetic variation, while the fact that different to be prevalent: for instance, there could be herit- behavioral traditions can be found in similar envir- able genetic effects. However, the weight of evi- onments would appear to render environmental dence supports the notion of a transmitted explanations of behavior impotent much of the culture; see Boyd and Richerson (1985) for a time. To give an example, Guglielmino et al. (1995) more extensive collation of evidence for cultural analyzed variation in cultural traits among 277 transmission. contemporary African societies, and found that most traits examined correlated with cultural his- TYPES OF CULTURAL SELECTION tory rather than with ecology. Such findings sug- gest that most human behavioral traits are Researchers in the gene±culture coevolution trad- maintained in populations as distinct cultural trad- ition have described a number of processes that itions, rather than evoked by the natural environ- underpin cultural change. In order to distinguish ment. Genes and environment undoubtedly cultural from biological evolution, Cavalli-Sforza account for some variation in human behavior, and Feldman (1981) defined `cultural selection' as but the socially transmitted component of culture a process by which particular socially learned is also important. beliefs, or pieces of knowledge, increase or de- A capacity for culture is an unusual . crease in frequency owing to their adoption by It allows humans to learn about their world rapidly other individuals at different rates. In contrast, nat- and efficiently. We do not have to scour our envir- ural selection can change the frequency of a cul- onment for sources of and water, devise our tural preference through the differential survival of own means of , or reinvent techno- individuals expressing different types of prefer- logical advances from first principles. Our capacity ence. For instance, in developed countries fertility to acquire valuable skills and information from control (contraception) is at a clear disadvantage in more knowledgeable others, such as parents, natural selection as users typically have fewer off- teachers or friends, as well as indirectly through spring, but has spread by virtue of its advantage in artifacts such as books and computers, furnishes cultural selection since fertility control is a popular us with a short cut to adaptive (and sometimes choice. Working on these two interacting subsys- maladaptive) behavior. Advocates of gene±culture tems simultaneously helps us to understand how coevolution share with the vast majority of social nonadaptive cultural traditions could evolve scientists the view that what makes culture (Cavalli-Sforza and Feldman, 1981). When it has 4 Gene±Culture Coevolution sufficiently high cultural , cultural informa- as frequently seems to be the case, individuals are tion can increase in frequency despite decreasing predisposed to adopt the behavior of the majority, genetic fitness. Cavalli-Sforza and Feldman's this frequency-dependent bias generates conform- framework also considers cases in which cultural ity. People may also use cues about one trait, for selection operates without affecting Darwinian fit- example, wealth, to choose which individuals to ness (e.g. a preference for a particular soft drink). observe in order to acquire information about an- Researchers in gene±culture coevolution are also other trait, such as clothes fashions. Boyd and interested in how information spreads within Richerson call this `indirect bias'. populations. The mode of transmission is the route by which cultural knowledge spread among A Case Study: Coevolution of Dairy individuals (Cavalli-Sforza and Feldman, 1981), Farming and Genes for Processing Milk and different models are required for alternative modes of information transmission. Social trans- The evolution of the ability of adult humans to mission can occur vertically (that is, from parents consume dairy products represents a good to offspring), obliquely (from the parental to the example of gene±culture coevolution. Unlike that offspring generation; for instance, learning from of human infants, virtually all of whom can all teachers or religious elders) or horizontally (that drink milk without problems, the milk digestive is, within-generation transmission such as learning physiology of adult humans varies considerably. from friends or siblings). Genetic inheritance is In fact, if the entire world's population is con- exclusively vertical, and as social transmission fre- sidered, consuming dairy products makes the ma- quently occurs through some combination of these jority of adult humans ill. This is because the modes of information transmission, cultural evolu- activity level of the enzyme lactase in their bodies tion and gene±culture coevolution will exhibit is insufficient to break down the energy-rich sugar quite different properties from those of biological lactose in dairy products, and milk consumption evolution. typically leads to sickness and diarrhea. The ability Boyd and Richerson (1985) extended the tax- of adult humans to digest lactose largely depends onomy of cultural processes by splitting the general on whether they possess the appropriate variants of concept of `cultural selection' into subtypes. One a single gene. A correlation exists between the inci- such subtype is `guided variation', which refers to a dence of the genes for lactose absorption and a process by which individuals acquire information history of dairy farming in populations, with ab- about a behavior culturally, and then modify the sorbers reaching frequencies of over 90% in dairy- behavior on the basis of their personal . farming populations, but typically less than 20% in Here cultural variation is guided by individual populations without dairy traditions (Durham, experience which, as human behavioral ecologists 1991). The correlation is extremely suggestive. envisage, allows behavioral traditions to evolve Milk and milk products have been a component gradually towards the adaptive optimal behavior of the diets of some human populations for over for that environment. 6000 years, roughly 300 generations. Is it conceiv- Another set of processes that Cavalli-Sforza and able that dairy farming might have created the Feldman (1981) and Boyd and Richerson (1985) selective regime under which the allele for absorp- consider is known as `biased cultural transmission'. tion was favored? Biased transmission occurs when, given a choice Feldman and Cavalli-Sforza (1989) used gene± between two alternative behavior patterns, individ- culture coevolutionary models to investigate the uals are more likely to adopt one variant than an- evolution of lactose absorption. They assumed other. Various types of bias exist. In direct bias, that the capacity to absorb lactose was affected by individuals choose which of two or more alterna- alleles (variants) of a single gene, with one particu- tive behavior patterns to adopt. A direct bias might lar allele allowing adults to digest milk. In addition, result from a genetic predisposition to favor certain they modeled the cultural transmission of milk types of information. Stanford University anthro- usage. The analysis suggested that whether or not pologist Bill Durham has argued that the individ- the allele allowing adult lactose absorption and ual choices underpinning these cultural processes milk digestion achieves a high frequency depends are guided, but not determined, by predispositions critically on the probability that the children of and prior knowledge (Durham, 1991). dairy-product users themselves become milk con- In the case of frequency-dependent bias, how- sumers. If this probability is high, then a significant ever, the commonness or rarity of a behavior affects fitness advantage to the genetic capacity for lactose the probability of information transmission. When, absorption will generally result in the selection of Gene±Culture Coevolution 5 the absorption allele to high frequency within 300 unusually fast genetic responses to selection in generations. However, if a significant proportion of humans (Feldman and Laland, 1996). It is thus the offspring of milk users do not exploit dairy entirely feasible that genetic and cultural evolution products, then unrealistically strong selection could operate at similar rates. In fact, the past 2 favoring absorbers would be required for the gene million years of human evolution may even have for absorption to spread. In other words, differ- been dominated by gene±culture coevolution. Cul- ences in the strength of cultural transmission be- ture can, of course, cause rates of environmental tween cultures may account for genetic variability change that really are too fast for human genetic in lactose absorption. The analysis is able to ac- evolution to track, and it is probably doing so in- count for both the spread of lactose absorption, creasingly. In fact, in the last 25 000±40 000 years the and the culturally related variability in its inci- dominant mode of human evolution has probably dence. Moreover, there is a broad range of condi- been exclusively cultural. tions under which the absorption allele does not spread despite a significant fitness advantage, Gene±Culture Coevolution and Niche indicating that traditional genetic models would Construction frequently give the wrong answer. Cultural pro- cesses complicate the selection process to the extent Organisms frequently choose, regulate, construct, that the outcome may differ from that expected and destroy important components of their envir- under purely genetic transmission. onments, such as nests and burrows, in the process The traditional view among the scientific com- changing the selection pressures to which they and munity was that adult lactose tolerance in humans other organisms are exposed. These processes are is an adaptation to reduced exposure to the sun, known as (Odling-Smee, 1988). as both the sun and the enzyme lactase promote Niche-constructing traits are more than just adap- calcium absorption (Durham, 1991). However, tations, because they have the additional role of an analysis by Holden and Mace (1997) of a phyl- modifying natural selection pressures. Like natural ogeny of human cultural groups using sophisti- selection, niche construction can be regarded as an cated statistical techniques found no evidence for evolutionary process potentially capable of gener- the latitudinal theory, but strong support for the ating a complementary match between organism dairy-farming hypothesis. Moreover, their analysis and environment. Organisms may adapt to envir- revealed that dairy farming evolved first, which onments, and environments may be shaped by or- then favored tolerance to lactose, and not the ganisms. In humans, culture has greatly amplified other way around. Holden and Mace's analysis our capacity for niche construction and our ability provides compelling confirmation of the findings to modify selection pressures, and cultural niche of this gene±culture coevolutionary analyses. construction may frequently instigate further bio- logical or cultural change (Laland et al., 2000). DO GENES AND CULTURE Standard gene±culture analyses incorporate COEVOLVE? niche construction implicitly, by assuming that some human cultural activities feed back to modify It is frequently suggested that genetic evolution is some selection pressures in human environments, too slow and cultural change too fast for the latter and thus cultural transmission may affect the fate to drive the former. In fact, selection experiments of some selected human genes. Generally, the rele- and observations of natural selection in the wild vant aspect of human selective environments is reveal that biological evolution may be extremely defined as cultural. For example, the trait that fast, with significant genetic and phenotypic affected human genetic evolution in the lactose change sometimes observed in a small number of tolerance case was milk usage (Durham, 1991). generations. At the same time, observations of Here, gene± is applicable because hominid stone tool technologies reveal that cultural the link between milk usage and its genetic conse- change can be extraordinarily slow. Acheulian and quences are sufficiently simple to allow it to be Oldowan stone tools traditions remained very modeled without bringing in any intermediate similar for hundreds of thousands ± even millions variables (Feldman and Cavalli-Sforza, 1989). ± of years. Even cultural institutions such as labor However, standard gene±culture coevolutionary markets can be extremely persistent, albeit on a models are less appropriate in more complicated shorter time scale. Furthermore, theoretical ana- situations. Take, for example, the case of Kwa- lyses have revealed that cultural transmission speaking yam cultivators in West Africa, who in- may change selection pressures to generate creased the frequency of a gene for sickle cell 6 Gene±Culture Coevolution anemia in their own population as a result of the empirical science. Where gene±culture analyses indirect effects of yam cultivation. These people have been applied to specific case studies they do traditionally cut clearings in the rainforest, creating make a variety of testable predictions; for instance, more standing water and increasing the breeding Soltis et al. (1995) used data on rates of population grounds for malaria-carrying mosquitoes. This, in in New Guinea to test Boyd and Richer- turn, intensified selection for the sickle cell allele, son's group selection hypothesis. Yet no general because of the protection offered by this allele empirical method has been established. The closest against malaria in the heterozygotic condition to a general approach is that advocated by anthro- (Durham, 1991). Here the causal chain is so long pologist Bill Durham, who illustrates with compel- that simply plotting the cultural trait of yam culti- ling examples, each backed by considerable data, vation against the frequency of the sickle cell allele how variability in human behavior and society may would be insufficient to yield a clear relationship be interpreted as resulting from interactions be- between the cultural trait and allele frequencies tween genetic and cultural processes (Durham, (Durham, 1991). The crucial variable is probably 1991). Durham identifies five categories of inter- the amount of standing water in the environment action: caused by the yam cultivation, but standing water . genetic mediation, where genetic differences underlie is an ecological variable, not a cultural variable, and cultural variation, as may be the case for the terms it partly depends on factors (rainfall) that are used by humans to describe color which reflect fea- beyond the control of the population. So here the tures of the human nervous system; simplifying assumption of a direct link between . cultural mediation, where culture drives genetic cultural and genetic inheritance distorts too change, such as with the evolution of adult lactose much to allow their interaction to be modeled in the absorption in populations that consume dairy prod- standard way. This time the two human inherit- ucts; ance systems can interact only via an intermediate, . enhancement, where culture reinforces genetic predis- abiotic, ecological variable subject to niche con- positions, as with the emergence of incest taboos that guard against the deleterious effects of inbreeding; struction, which should be included to complete . neutrality, where are adopted independently the model. of an individual's genotype, as is the case for learning This shortcoming led Laland et al. (2000) to different languages; propose an extended version of gene±culture co- . opposition, where culture leads to maladaptive trad- evolutionary theory, which incorporated niche itions, for instance the cannibalism of the Fore, a New construction as a general evolutionary process. Guinea , which spread the deadly nerve Culturally modified selection pressures are disease kuru. regarded as a part of a more general legacy of modified natural selection pressures bequeathed CONCLUSION by humans to their descendants. Laland et al. (2001) used extended gene±culture coevolutionary Gene±culture coevolutionary theory has rarely models to explore the evolutionary consequences been subject to the same level of criticism as of culturally generated niche construction through human sociobiology or evolutionary psychology. human evolution. The analysis revealed circum- In fact, in the debates over human sociobiology stances under which cultural transmission can and its progeny, it has been almost completely overwhelm natural selection, accelerate the rate at ignored, perhaps because of its technical nature. which a favored gene spreads, initiate novel evolu- However, some social scientists have objected to tionary events, and trigger hominid . Be- the idea that culture can be modeled as if composed cause cultural processes typically operate faster of discrete psychological or behavioral characteris- than natural selection, cultural niche construction tics, while others have questioned the legitimacy of is likely to have more profound consequences than `borrowing' population genetics processes to gene-based niche construction, and is likely to have model culture, or criticized the analyses as promot- played an important role in human evolution. ing a false gene±culture dichotomy. However, for gene±culture researchers these assumptions do not Empirical Studies of Gene±Culture represent ideological stances but are made purely Coevolution for pragmatic reasons. Culture is difficult to ana- lyze unless it is broken down into manageable While gene±culture coevolution has a strong and units. Building on population genetic models is a rigorous theoretical foundation, it is vulnerable to reasonable place to start developing models of cul- the charge that it has not spawned a vigorous tural evolution (providing the differences between Gene±Culture Coevolution 7 biological and cultural processes are accommo- under complex transmission. Theoretical Population dated). Gene±culture methods represent compara- Biology 9: 238±259. tively simple descriptions of the interactions Feldman MW and Cavalli-Sforza LL (1989) On the theory between genetic and cultural processes. However, of evolution under genetic and cultural transmission approaches that focus on a single process (be it with application to the lactose absorption problem. In: Feldman MW (ed.) Mathematical Evolutionary Theory. exclusively cultural or exclusively genetic) have Princeton, NJ: Princeton University Press. made the fundamental and sweeping assumption Feldman MW and Laland KN (1996) Gene±culture that the processes do not interact, or that there is coevolutionary theory. Trends in Ecology and Evolution only one process that matters. 11: 453±457. Gene±culture coevolutionary analyses suggest Guglielmino CR, Viganotti C, Hewlett B and Cavalli- that evolution in species with a dynamic, socially Sforza LL (1995) Cultural variation in Africa: role of transmitted culture may be different from evolu- mechanisms of transmission and adaptation. tion in other species. Culture is a particularly ef- Proceedings of the National Academy of Science of the USA fective means of modifying natural selection 92: 7585±7589. pressures and driving the population's biological Hewlett BS and Cavalli-Sforza LL (1986) Cultural evolution, as was the case for lactose absorption. transmission among Aka pygmies. American Anthropologist 88: 922±934. Culture may generate new evolutionary processes, Holden C and Mace R (1997) Phylogenetic analysis of the for instance . Moreover, evolution of lactose digestion in adults. Human Biology cultural transmission may strongly affect evolu- 5: 605±628. tionary rates, sometimes speeding them up and Laland KN, Odling-Smee FJ and Feldman MW (2000) sometimes slowing them down. Such findings sug- Niche construction, biological evolution and cultural gest that traditional evolutionary approaches to the change. Behavioral and Brain Sciences 23: 131±146. study of human behavior may not always be ad- Laland KN, Odling-Smee FJ and Feldman MW (2001) equate. Cultural niche construction and human evolution. Journal of 14: 22±33. References Lumsden CJ and Wilson EO (1981) Genes, Mind and Culture. Cambridge, MA: Harvard University Press. Boyd R and Richerson PJ (1985) Culture and the Maynard Smith J and Warren N (1982) Models of cultural Evolutionary Process. , IL: University of Chicago and genetic change. Evolution 36: 620±627. Press. Odling-Smee FJ (1988) Niche constructing . Cavalli-Sforza LL and Feldman MW (1981) Cultural In: Plotkin HC (ed.) The Role of Behavior in Evolution. Transmission and Evolution: AQuantitative Approach . Cambridge, MA: MITPress. Princeton, NJ: Princeton University Press. Smith EA (2000) Three styles in the evolutionary analysis Cavalli-Sforza LL, Feldman MW, Chen KH and of human behavior. In: Cronk L, Chagnon N and Irons Dornbusch SM (1982) Theory and observation in W (eds) Adaptation and Human Behavior. New York, NY: cultural transmission. Science 218: 19±27. Aldine de Gruyter. Dawkins R (1976) . Oxford, UK: Oxford Soltis J, Boyd R and Richerson PJ (1995) Can group- University Press. functional behaviors evolve by cultural group Durham WH (1991) Coevolution: Genes, Culture and selection? An empirical test. Current 36: Human Diversity. New York, NY: Stanford University 473±494. Press. Tylor EB (1871) Primitive Culture: Researches into the Feldman MW and Cavalli-Sforza LL (1976) Cultural and Development of Mythology, Philosophy, , Art, and biological evolutionary processes, selection for a trait Custom. London, UK: John Murray.