Burt does an outstanding job of clearly describing, at just the right level of necessary detail, how modern genomic techniques work, and clearly articulates the many limitations associated with drawing strong inferences about the role played by genes in determining any particular behavioral or social outcome. She also acknowledges that most researchers who use these techniques (even the “enthusiasts”) are aware of these limitations, and typically do what they can to mitigate them. The crux of her argument is that these attempts are inadequate (maybe even unavoidably so), and so continued efforts in that direction are a waste of limited resources, and potentially dangerous, because they might be “obscuring social–structural and physical environmental influences and promoting the individualization of social problems” (target article, sect. 8, para. 4). These perceived potential dangers reveal a laudable (but perhaps misguided) political motivation for discouraging continued research into genetic influences, but Burt is explicit about the fact that she is not arguing from a political perspective, but a scientific one. From a scientific perspective, deliberately ignoring genetic influences on even the most complex and apparently purely “social” phenomena is a straightforward mistake.
One reason for rejecting Burt's proposal on scientific grounds is that it rules out the possibility of understanding the evolutionary origin of complex social phenomena, which, of course, depends on explicating the role genes might play in such effects. The reason this is important is because the evolutionary selected consequences of a gene's actions are its effects. Burt uses Jenks' (1972) classic thought experiment about discrimination based on hair color to bolster her argument for purely social factors creating what look like genetic influences on a social outcome. The thought experiment asks to imagine a world in which red-haired children are barred from school. Under such conditions, there would be a strong relationship between genes (those responsible for pheomelanin production; Valverde, Healy, Jackson, Rees, & Thody, Reference Valverde, Healy, Jackson, Rees and Thody1995) and educational attainment, but Burt's argument is that this relationship would be spurious, because the genes themselves do not “code” for anything directly related to an individual-level propensity to acquire an education (e.g., intelligence, or perseverance, or carefulness). Although it is intended to illustrate the opposite point, this thought experiment can be used to exemplify the fundamental premise of evolution; that genes only have selective consequences in environments – this interaction underpins all selection. This example makes it obvious that the selective consequences of the gene need not be direct. Indeed, as genes only make proteins (or regulate the genes that make proteins), it can never be direct. From an evolutionary perspective, if school-level education increased reproductive success (increased “fitness” in evolutionary biology), for whatever reason, then the genes for red hair would be selected against, because they reduced educational opportunities. This is one of the gene's consequences in that environment, and so there is a real sense in which this is one of its effects, despite being indirect. The only reason this looks like a spurious association is because we know that the hypothetical educational barrier is artificially constructed. We rarely have that insight for most genes and their real effects, but understanding all evolution depends on understanding the effects of genes in environments. For example, imagine a world in which rather than red-heads being barred from school, they were considered more attractive, and so more easily attracted mating partners. Again, whether this preference was because of a sexually selected “genetic” preference (Endler & Basolo, Reference Endler and Basolo1998) or an entirely arbitrary socially determined preference (if such a thing could exist), the possession of the genes for making red hair would have evolutionary consequences, and those consequences would be one of its effects.
A major theme of Burt's critique is that modern genomic techniques are guilty of neglecting the kind of developmental interactionism described above, and have a tendency to ignore the fact that such findings are therefore context- and population-specific. Although this criticism may be true of the most ardent enthusiasts, the vast majority of researchers in this field clearly demonstrate awareness of these issues by trying to factor them in (or out) in their analyses. The irony of this criticism is that Burt cites a host of social factors that are presented as causes of social outcomes without any direct evidence of their casual efficacy and without any acknowledgement that those factors can only have their effects via interacting with genes. It is true, as Burt points out, that there is no meaningful way we can identify social-context-independent genetic “potential” for a particular social outcome (like educational attainment), but it is equally true that there is no meaningful way to identify gene-independent social “potential” for any social outcome. No matter what socially determined education-relevant advantages (or disadvantages) a person has, that can only possibly translate into actual educational attainment via a long, complicated interaction between their genes and those social factors. Part of the problem here is that “social” variables are discussed as causes and outcomes, without a complete fleshing out of the individual-level causal mechanisms involved in linking them. It is individuals who attain certain levels of education, or have same-sex sex, or who play golf, to use some of Burt's examples, and so we need to understand how genes and environments dynamically interact to understand why individuals behave in certain ways in certain environments. Two obvious individual-level factors that influence many of these outcomes are personality and intelligence, and both of those are traits that are a consequence of complex gene–environment interactions. For any outcome we might be interested in understanding, all of these factors (social, individual, and genetic) complexly interact, and so there are logistical barriers to identifying “the role” played by any one factor, and logical reasons to avoid any such approach. We need good data on all of the factors involved, and so discouraging any approach is counterproductive.
You have
Access
Burt does an outstanding job of clearly describing, at just the right level of necessary detail, how modern genomic techniques work, and clearly articulates the many limitations associated with drawing strong inferences about the role played by genes in determining any particular behavioral or social outcome. She also acknowledges that most researchers who use these techniques (even the “enthusiasts”) are aware of these limitations, and typically do what they can to mitigate them. The crux of her argument is that these attempts are inadequate (maybe even unavoidably so), and so continued efforts in that direction are a waste of limited resources, and potentially dangerous, because they might be “obscuring social–structural and physical environmental influences and promoting the individualization of social problems” (target article, sect. 8, para. 4). These perceived potential dangers reveal a laudable (but perhaps misguided) political motivation for discouraging continued research into genetic influences, but Burt is explicit about the fact that she is not arguing from a political perspective, but a scientific one. From a scientific perspective, deliberately ignoring genetic influences on even the most complex and apparently purely “social” phenomena is a straightforward mistake.
One reason for rejecting Burt's proposal on scientific grounds is that it rules out the possibility of understanding the evolutionary origin of complex social phenomena, which, of course, depends on explicating the role genes might play in such effects. The reason this is important is because the evolutionary selected consequences of a gene's actions are its effects. Burt uses Jenks' (1972) classic thought experiment about discrimination based on hair color to bolster her argument for purely social factors creating what look like genetic influences on a social outcome. The thought experiment asks to imagine a world in which red-haired children are barred from school. Under such conditions, there would be a strong relationship between genes (those responsible for pheomelanin production; Valverde, Healy, Jackson, Rees, & Thody, Reference Valverde, Healy, Jackson, Rees and Thody1995) and educational attainment, but Burt's argument is that this relationship would be spurious, because the genes themselves do not “code” for anything directly related to an individual-level propensity to acquire an education (e.g., intelligence, or perseverance, or carefulness). Although it is intended to illustrate the opposite point, this thought experiment can be used to exemplify the fundamental premise of evolution; that genes only have selective consequences in environments – this interaction underpins all selection. This example makes it obvious that the selective consequences of the gene need not be direct. Indeed, as genes only make proteins (or regulate the genes that make proteins), it can never be direct. From an evolutionary perspective, if school-level education increased reproductive success (increased “fitness” in evolutionary biology), for whatever reason, then the genes for red hair would be selected against, because they reduced educational opportunities. This is one of the gene's consequences in that environment, and so there is a real sense in which this is one of its effects, despite being indirect. The only reason this looks like a spurious association is because we know that the hypothetical educational barrier is artificially constructed. We rarely have that insight for most genes and their real effects, but understanding all evolution depends on understanding the effects of genes in environments. For example, imagine a world in which rather than red-heads being barred from school, they were considered more attractive, and so more easily attracted mating partners. Again, whether this preference was because of a sexually selected “genetic” preference (Endler & Basolo, Reference Endler and Basolo1998) or an entirely arbitrary socially determined preference (if such a thing could exist), the possession of the genes for making red hair would have evolutionary consequences, and those consequences would be one of its effects.
A major theme of Burt's critique is that modern genomic techniques are guilty of neglecting the kind of developmental interactionism described above, and have a tendency to ignore the fact that such findings are therefore context- and population-specific. Although this criticism may be true of the most ardent enthusiasts, the vast majority of researchers in this field clearly demonstrate awareness of these issues by trying to factor them in (or out) in their analyses. The irony of this criticism is that Burt cites a host of social factors that are presented as causes of social outcomes without any direct evidence of their casual efficacy and without any acknowledgement that those factors can only have their effects via interacting with genes. It is true, as Burt points out, that there is no meaningful way we can identify social-context-independent genetic “potential” for a particular social outcome (like educational attainment), but it is equally true that there is no meaningful way to identify gene-independent social “potential” for any social outcome. No matter what socially determined education-relevant advantages (or disadvantages) a person has, that can only possibly translate into actual educational attainment via a long, complicated interaction between their genes and those social factors. Part of the problem here is that “social” variables are discussed as causes and outcomes, without a complete fleshing out of the individual-level causal mechanisms involved in linking them. It is individuals who attain certain levels of education, or have same-sex sex, or who play golf, to use some of Burt's examples, and so we need to understand how genes and environments dynamically interact to understand why individuals behave in certain ways in certain environments. Two obvious individual-level factors that influence many of these outcomes are personality and intelligence, and both of those are traits that are a consequence of complex gene–environment interactions. For any outcome we might be interested in understanding, all of these factors (social, individual, and genetic) complexly interact, and so there are logistical barriers to identifying “the role” played by any one factor, and logical reasons to avoid any such approach. We need good data on all of the factors involved, and so discouraging any approach is counterproductive.
Competing interest
None.