The hypothesis is examined that genes bias the development of complex social behavior in one direction over alternatives. Studies of altruism and political attitudes in twins estimate that =50% of the variance is associated with direct genetic inheritance, virtually 0% with the twin's common family environment, and the remainder with each twin's specific environment. Studies of human marriages show that spouses choose each other on the basis of similarity, assorting on the most genetically influenced of a set of homogeneous attributes. These data imply a genetic canalization of social influences such that, within the constraints allowed by the total spectrum of cultural alternatives, people create environments maximally compatible with their genotypes. Most if not all of the components of cognitive development examined to date have supported the gene-culture model of transmission (1-7). Corresponding data on personality and social development have been attended to far less. Here we review research in support of the hypothesis that epigenetic rules bias the development of complex social behavior, choosing as illustrative examples, altruism and mate choice. In so doing, we contrast Lumsden and coworkers' (1-7) approach with models of cultural transmission hypothesized by Feldman,. While the latter authors incorporated the phrase "gene-culture coevolution" in a study of the transmission of altruism (10), they have generally failed to adopt the concomitant prescription that epigenetic rules bias individual development. As a result, much of their perspective is discordant with knowledge about social learning in families. By contrast, gene-culture coevolutionary formulations, in the sense originally intended by Lumsden and Wilson (1), appear to be highly compatible with these same facts.
Epigenetic Rules in Social DevelopmentExtensive theorizing in both the evolutionary and social sciences errs in not taking into account that social learning is dependent upon the innate capacities and biases of the learner (e.g., refs. 8, 9, and 11). For example, most models of cultural transmission within the family (i.e., vertical, from parent to child, and horizontal, from sibling to sibling) imply that siblings will resemble each other, over and above shared genes, as a result of a common family environment. Geneculture theory, in contrast, leads to the expectation that siblings will differ from each other in part because their nonshared genes incline them to acquire patterns of behavior best fitting their particular genotype (gene-culture transmission). While it may seem intuitively correct to assume that common family environment shapes individual development, consideration of data reveals quite a different set of relationships.Behavior 16,17). In a twin study, the raw data are the between-and within-pairs variances and covariances. The between-pairs mean squares reflect both pair resemblances and pair differences, and the within-pairs mean squares, pair differences. The genetic models are fitted to these mean squares....