Genopolitics and genosociology: From genetics to human social and political behavior
A significant number of papers have tried to link polymorphisms with three genes encoding monoamine oxidase-A (MAOA),the serotonin transporter (5-HTT),and the dopamine D4 receptor with a wide array of behavioral patterns and traits; partially these three genes encode proteins that regulate brain functions. Here, I shall examine merely two of the three genes.
In a widely publicized paper that built upon an earlier paper (Fowler et al. 2008); Fowler and Dawes (2008) boldly claimed “two genes (i.e., 5-HTT and MAOA) predict voter turnout.” In a series of papers, Guang Guo and his colleagues have claimed to find strong association between these genes and delinquency behaviors and the number of sexual partners (e.g., Guo, Roettger, and Shih 2007; Guo, Tong et al. 2007; Guo, Ou et al. 2008, Guo, Roettger, and Cai 2008; Guo,Tong, and Cai 2008; see also American Journal of Sociology, vol. 114, SI, 2008).1
Unfortunately, these authors of genopolitics have only superficial understanding of the relationship between genotype and social behaviors (e.g., voting behavior and delinquency). Most critically, the so-called “behaviors” in their studies, such as voting behavior, delinquency, and the number of sexual partners, are immensely complex social outcomes (see the following). As a result, they are even more reductionist than sociobiology' and EP (Charney 2008a, 2008b), and all my criticism against sociobiology and EP can be applied to these new so-called geno-social sciences.
Methodologically, all of these works rely on statistics. But “genetics is not a subdiscipline of statistics” (Charney and English 2012, 2013). Indeed, even their statistics is not that great. In short, it is almost inevitable to find robust association between a single gene (and its SNPs) with a social or political trait, especially when the threshold level of significance for rejecting the null hypothesis of random association has been set to as low as p<0.05 to 0.01 (Guo et al. 2007; Fowler and Dawes 2008) to p<0.0001 at the highest (Guo, Tong et al. 2007). Unfortunately, such a level has been shown to be far below the threshold level of significance for rejecting the null hypothesis of non-association in GWAS studies (Benjamin et al. 2012; see also Charbris et al. 2013; Richardson 2017a, 2017b).
Genoeconomics: The long reach of human genetics in economic development
Some genoeconomists have attempted to link genetics with macro social outcomes such as the Industrial Revolution and long-run economic development.
Focusing on the coming of the Industrial Revolution in England, Gregory Clark (2007, 231) asserted a sweepingly simplistic thesis: “The Industrial Revolution ... was the product of the gradual progress of settled agrarian societies toward a more rational, economically oriented mindset.” For Clark, England was one of those unique places that were “becoming increasingly middle class in their orientations. Thrift, prudence, negotiation, and hard work were becoming values for communities that previously had been spendthrift, impulsive, violent, and leisure loving” (ibid, 166). Although Clark did not state it explicitly, his point was clear: “A plausible source of this apparent evolution of human preferences is the survival of the richest that is evident in preindustrial England” (ibid, 167; italics added).Thus, by the 18th century, England had come to possess “middle-class values,” and presumably it had been these dynamics that explained the coming of Industrial Revolution in England.
Unfortunately, Clark has provided skinny evidence for his grand narrative. As Allen (2008) has detailed, there was not even a single table of regression results in Clark’s whole book. Indeed, Clark did not even specify the scope of“middle-class values”: Was it among the peasants (the poor), a portion of the whole population, or the whole population?
For our discussion here, the key question is: how did “middle-class values” get transmitted from one generation to another without much depreciation (or erosion)? Clark had no viable answers or mechanisms for this question other than the remote and ludicrous possibility that middle-class values must have been carried by our genes or genomes.2
There is no doubt that “natural” selection might have played a role in selecting individuals who were better off and those who survived made a key contribution to the coming of the Industrial Revolution. Likewise, natural selection certainly has played a key role in bestowing us with the initial physical and intellectual capacity that enables us to innovate (and survive). Yet, after human beings as a species had gained these capacities, however, artificial selection came to be a more important force in shaping our evolution. As such, any theory that relies exclusively on natural selection can only provide limited mileage for understanding the evolution of economic man and the pattern of economic development (van den Bergh and Stagl 2003, 293-297).’
More recently, Ashraf and Galor (2013) went further and argued that deeply rooted genetic diversity within a population (with a country today) has a robust hump-shaped relationship with economic development of countries. More concretely, both too much genetic diversity (in today’s African countries) and too little genetic diversity (in today’s Latin American countries) are bad for economic development.Too little genetic diversity discouraged technological innovation, presumably via producing more geniuses or brain power genetically. In contrast, too much genetic diversity contributed to their high level of ethnolinguistic diversity, which has been consistently shown to be detrimental to economic growth. Both groups had not fared well. Only the groups with a more-or-less optimal level of genetic diversity came to have economic success in history. Unsurprisingly, these countries are Asian-European countries and their offshoots (e.g., America).
Ashraf and Galor (2013) has been mistaken in identifying an optimal genetic diversity. Although too little genetic diversity (within a mostly isolated small population) is inimical to the production of genius, with a population size above 50,000, genetic diversity is larger enough. Certainly, after the coming of the Neolithic Revolution around 12,000—10,000 BP, it has been social and political development, rather than genetic diversity, that has been the main driver of economic development.
Moreover, several studies have found the statistical results of Ashraf and Galor (2013) wanting:The supposedly robust significant statistical results of genetic diversity shown in Ashraf and Galor (2013) were not robust at all.4 Indeed I have shown that by controlling for a single Eurasia dummy, the statistical significance of genetic diversity is gone entirely (Tang 2016b). These results are consistent with Jared Diamond’s (1997) Eurasia advantage thesis rather than with Ashraf and Galor’s thesis of genetic diversity.
GW4S as the third wave of geno-social sciences
At first glance, GWAS seems to show more promise in identifying SNPs and perhaps groups of genes for some important developmental outcomes such as intelligence and social outcomes such as educational attainments, partly because educational attainment as a social outcome is more closely determined by intelligence. Beyond these traits, however, it is unclear whether GWAS can do any better than classical CGAS based on twin studies. Indeed, even two highly sympathetic reviews have admitted that few of the positive results reporting association between SNPs and social behaviors and outcomes can be replicated (Beahchamp et al. 2011; Benjamin et al. 2012). Worse, these studies have little out-of-sample predictive power.
In light of the difficulty of finding specific genes, some behavioral geneticists now proposed a fourth law: “A typical human behavioral trait is associated with very many genetic variants, each of which accounts for a very small percentage of the behavioral variability” (Charbris et al. 2015). Unfortunately, the fourth law is also seriously flawed because it cannot possibly defend or define what a human behavioral trait is within GWAS. Unless you define a human behavioral trait very broadly, all those associations amount to almost all meaningless correlations.
More devastatingly, Richardson (2017a, 2017b) has convincingly shown that with GWAS, some kind of statistical associations (or correlations) between SNPs and some so-called social “behavioral traits” are inevitable but also meaningless (see also Turkheimer 2012) because social stratifications have made social classes and ethnic groups, and endogamy (in the broader sense, say, within class and group) has sustained stratifications. Moreover, many of the key “traits” or social outcomes such as intelligence and educational attainment have been maintained and shaped by social stratifications. Witness the recent case of bribing into top universities by the rich and powerful.’ Such widespread corruption of the elite universities has gone undetected for years, and this fact alone should give us pause about the genetic foundation of educational attainment (and by implication, professional attainment).
