Commentaryi'Mealey: The sociobiology of sociopathy Genetic issues in "the sociobiology of sociopathy"

A consideration of the genetics of sociopathy suggests the following. The author's Evolutionary Stable Strategy (ESS) types 2 to 4 are more likely than types 1 and 5 in crimes of violence, and there may not be an ESS for crimes of property or for sociopathy. Correlations between sociopathy and crimes of property are also more likely due to environmental than to genetic variants, and correlations between sociopathy and crimes of property are due more to environmental than genetic variants. Mealey proposes that there are two types of sociopaths, primary and secondary, with differences in their genetics and life histories. Some aspects of the formal, developmental, population, and quantitative genetics of this proposal will be considered and evaluated here. First, I will attempt to describe the formal genetics of the two types of sociopathy. As I understand the author, primary and secondary sociopathy involve the same set of many polymorphic genes (sect. 2.1.3, para. 1 and 2). Primary sociopaths have genotypes at the most extreme end of the binomial distribution of genotypes and are thus most likely homozygous for many of the polymorphic genes. It is not known, however, whether there arc dominance or epistatic effects for these variant genes. Secondary sociopaths have genotypes elsewhere in the distribution, and these do not overlap those of primary sociopaths. Also, if all genotypes in this set were to experience the same environmental conditions, the risk of secondary sociopathy would be a function of the number of sociopathy facilitator alleles. Again, it is not known whether there is dominance or epistatic effects for these variant genes. With regard to developmental genetics, it is incorrect to say that primary sociopathy is more genetically determined than secondary sociopathy and that secondary sociopathy depends more on environmental factors than primary sociopathy (sect. 2.3.1, para. 2 and 3; sect. 2.4.2, para. 3). Rather, each phenotype is a response of each genotype to environmental inputs. It may be, as Mealey suggests, that the cheating strategy or sociopathy of individuals with the genotypes from the most extreme end of the binomial distribution of genotypes is not influenced by advantages or disadvantages in social competition during their life history. But this does not mean that the development of primary sociopathy in individuals with these genotypes is independent of other aspects of environment and life history. In fact, Mealey suggests that such individuals would display less sociopathic behavior if the cost of their detected cheating were increased (sect. 3.2.1). Similarly, it may be, as Mealey suggests that the cheating strategy or sociopathy of individuals with genotypes from the rest of the binomial distribution of genotypes depends on advantages or disadvantages in social competition during their life history. However, Mealey recognizes a genotype-environment interaction for this effect (sect. 2.2.1, para. 4; sect. 2.3; sect. 3.1.2, para. 1). Some genotypes are at a greater risk than others for developing secondary sociopathy at a given level of disadvantage in social competition. With regard to genetic evolution, Mealey suggests that the occurrence of primary sociopathy in a population is due to frequency-dependent selection for ESSs (evolutionarily stable ' strategies) and that this frequency-dependent selection keeps primary sociopathy at low levels in every society. In other words, the fitness (reproductive advantage) of sociopathy depends on its frequency in the population, high fitness and low frequency being associated. However, Maynard Smith (1989) proposes that there are extreme constraints on frequencydependent selection of ESSs in sexually reproductive species. These are (1) a pure or mixed ESS can be specified by a genetic homozygote; these would be represented by Mealey's ESSs of types 2, 3, and 4 rather than 1 and 5 for sociopathy; (2) also, a mixed ESS can be specified when there are only two pure strategies in the ESS and when the genetic system can generate the required phenotypic distribution. The existence of secondary sociopathy with continuous degrees of expression seems inconsistent with the requirement for two pure strategies. In addition, very little appears to be known about the genetic system underlying primary sociopathy, to say whether a distribution with less than 3%-4% male and 1% female primary cheaters is an ESS. With regard to quantitative genetics, the heritability of criminality and sociopathy is relevant to the developmental and ESS aspects of cheating. My comments are based on critical reviews of this material by Carey (1994) and by Raine (1993). First, neither twin nor adoption studies reveal a nonzero heritability for violent crimes. This suggests that there may be no genetic correlation between crimes of violence and either type of sociopathy. However, a single homozygous genotype for our species could, as discussed above, result in an ESS for crimes of violence. This would be Mealey's ESS types 2 to 4 rather than ESS types 1 and 5. Second, both twin and adoption studies show a nonzero heritability for crimes of property. As discussed above, this genotypic polymorphism may not be consistent with there being an ESS for crimes of property. Also, there is a substantial effect of the environment on variation in crimes of property. Thus; correlations between sociopathy and crimes of property are probably due to both genetic and environmental effects. Third, there have been some attempts to assess the heritability of antisocial personality disorder, which may be related more directly to sociopathy than crime is (Baker 1986; Grove et al. 1990). In these studies, the heritability of antisocial personality is low (about: 0.27). To the extent that this is an index of sociopathy, these studies suggest that its heritability is lower than that for crimes of property (about 0.45), and that genotypic correlations for crime and sociopathy will make a modest contribution to their phenotypic correlations. In addition, the genetic polymorphism for antisocial personality may not be consistent with there being an ESS for this trait and perhaps for sociopathy. Diathesis stress model or "Just So" story? Richard M. McFall, James T. Townsend, and