In his 2003 paper "Heritability and Indirect Causation",\cite{Sesardic_2003} philosopher Neven Sesardic criticizes those critics of behavior genetics who have taken aim at one argument in particular. Specifically, these critics have argued that the position that genes can affect phenotypes indirectly if this effect is mediated by environmental factors, and that such an effect can be validly interpreted as "genetic", is untenable. Sesardic makes it clear that he disagrees with these critics right in the abstract of his paper, in which he writes,This possibility of genetic effects being mediated by environmental influences is often used by scientists and philosophers to argue that heritability is not a very helpful causal or explanatory notion. In this paper it is shown that these criticisms are based on serious misconceptions about methods of behavior genetics.\cite{Sesardic_2003}But are they really? That is what I will try to answer in this paper.First, I will note that Sesardic's paper focuses on an argument used by critics of behavior genetics to argue against the validity of interpreting heritability estimates as reflecting anything "genetic", at least in the sense that almost everyone understands the definition of the word "genetic". Sesardic starts his paper (the body, not the abstract) by focusing on a thought experiment originally described by Christopher Jencks.\cite{christopher1972} Here is how Sesardic summarizes this thought experiment: "Imagine that red‐haired children are for some reason singled out for abuse and are frequently hit on the head by parents and teachers. As a result they get a lower IQ on average than other children." The idea behind this thought experiment is that, if this scenario were true, you could argue that the lower IQs of red-haired children are "genetic". The logic goes like this: the red hair of these children is strongly genetically controlled, and thus the responses of other people to this genetic trait, including hitting people on the head, should also count as "genetic". Because it is hitting children on the head in this thought experiment that causes them to have lower IQs, according to this argument, the lower average IQs of red-haired children can reasonably be said to be genetic. To those making this argument, the lower IQs of said children could be said to be due to environmentally mediated genetic effects, that is, to environmental influences to which the children were exposed only because of the genetic trait of their red hair. Hereafter, I will refer to the argument that, in the context of this thought experiment, the lower average IQs of red-haired children are "genetic" as argument A. There are other formulations of this argument that I will also consider to be examples of argument A: fundamentally, it encompasses any situation in which an individual is treated differently by others on the basis of a genetic trait, which could be not merely hair color but skin color or biological sex. Hence, argument A would also hold that sex discrimination is "genetic" insofar as it is indirectly caused by the genetic trait of having two X chromosomes, and racial discrimination is also "genetic" insofar as it is indirectly caused by the genetic trait of having darker-colored skin. More fundamentally, this can be stated as: both genes -> traits and genes -> X -> traits qualify as examples of a causal genetic effect for a trait, no matter what X is, and in particular if X is some environmental factor.Sesardic notes that identical or very similar versions of this thought experiment are used by a number of other philosophers and social scientists in arguing that heritability estimates of a trait do not, in fact, represent the extent to which variation in that trait is "genetic" in any meaningful sense of the word. The specific way in which critics of behavior genetics will make this argument is by asserting that argument A is invalid, and that because heritability analyses assume that argument A is valid in attributing environmental effects that are responses to genetic traits as themselves genetic, then heritability analyses are also invalid insofar as they purport to shed light on the degree of "genetic" variation in a given trait. This is because even if heritability is higher than zero, it only means that a nonzero proportion of total variance is genetically caused, but this may be the result of a highly indirect process in which one or more environmental factors plays a much more proximate and obvious role. In other words, critics make the following argument: because argument A is untenable, heritability analyses are uninformative about genetic causation. I will refer to this argument as argument B.As an example of argument B, Sesardic quotes Jencks as saying the following: "If, for example, a nation refuses to send children with red hair to school, the genes that cause red hair can be said to lower reading scores…. Attributing redheads' illiteracy to their genes would probably strike most readers as absurd under these circumstances. Yet that is precisely what traditional methods of estimating heritability do. If an individual's genotype affects his environment, for whatever rational or irrational reason, and if this in turn affects his cognitive development, conventional methods of estimating heritability attribute the entire effect to genes and none to environment." Sesardic summarizes the conclusion of argument B as follows: "Heritability estimates, whatever their magnitude, are never by themselves evidence that genes play an important explanatory role simply because even variation that is 100 percent heritable (i.e., ultimately due to genetic differences) may well be 100 percent proximately explained by much more relevant environmental differences." On its face, it seems absurd to claim that a human trait whose average value in a group is wholly due to how that group is treated by other people is "genetic" at all, let alone completely genetic and not environmental at all. In fact, Jencks himself uses the word "absurd" to describe the redhead example he outlined in his 1972 book in which this argument originates. Interestingly, Sesardic agrees with Jencks only with regard to the specific type of situation Jencks outlined in his original thought experiment. Sesardic then tries to draw a distinction between different types of gene-environment correlation (also written as G-E correlation), which is generally when variation in a given gene is statistically related to variation in a given environmental characteristic. This distinction was originally outlined by Plomin et al. (1977), who argued that G-E correlation can be divided into three types: passive, reactive, and active.\cite{Plomin_1977} Plomin et al. (1977, p. 310) define the three types of G-E correlation as follows:Passive G-E correlation "is one in which parents give their children both genes and an environment that are favorable (or unfavorable) for the development of a trait".Reactive (sometimes called "evocative")\cite{Jaffee2008} G-E correlation occurs when "people...react differently to persons of different genotypes" (note also that reactive correlation "differs from the passive type in that it is not limited to environments provided by relatives").Active G-E correlation "occurs as a result of the fact that a child is not merely the passive recipient of his environment. Rather, he contributes to his own environment and may actively seek one related to his genetic propensities."\cite{Plomin_1977} Sesardic then attempts to make the following points:Jencks is correct that argument A is wrong. Specifically, Sesardic identifies argument A as an example of reactive G-E correlation, and argues that behavior geneticists correctly avoid interpreting such correlation as genetic. Sesardic emphasizes the distinction between active and reactive G-E correlation as follows: "In the case of active G‐E correlation, the environments that lead to a phenotypic difference are selected by subjects themselves, whereas in reactive correlation they are imposed by others." It may, in fact, be valid for behavior geneticists to interpret active G-E correlation as genetic when conducting heritability analyses. Furthermore, behavior geneticists do, in fact, sometimes interpret active G-E correlation in this way. As Sesardic writes, "...the influences of those environments that are chosen on the basis of genotype are typically difficult to keep apart from the influence of genotype itself...Therefore, phenotypic effects of such environments are indeed sometimes classified as heritable, on the grounds that they are practically inseparable from direct genetic effects and that they merely represent the self‐realization of genotype." Argument A is not actually made by any behavior geneticists, who apparently all recognize that active G-E correlation can be interpreted as genetically caused, whereas reactive G-E correlation cannot. Specifically, Sesardic writes, "...I am unaware that any serious scholar ever defended the idea that the indirect effects of Jencks‐type scenarios should be treated as heritable." He argues that "the leading authority in the field" of behavior genetics [referring to John L. Fuller] has explicitly argued against Argument A, and criticizes Jencks and other scholars for either failing to give examples of behavior geneticists using argument A or for quoting the examples they do give out of context.He tries to make it clear that he does not mean to say that active G-E correlation should always be considered genetic or that reactive G-E correlation should always be considered environmental: "my claim is that G-E covariance is in the first place a source of variance that is sui generis and, as such, distinct from either the genetic or the environmental component of phenotypic variance (because G-E covariance involves both genetic and environmental influences)."In short, Sesardic argues that to interpret the effects of environmental influences to which an individual is exposed as a result of the actions of other individuals, even if these actions are themselves influenced by a genetic trait of the first individual, as genetic is both incorrect and never actually done in the field of behavior genetics. However, some behavior geneticists do, in fact, make an argument that appears to be, in some cases, identical to this argument. Specifically, a common argument in defense of the most common research design in behavior genetics, the classical twin study, holds that MZ twins experience more similar environments than do DZ twins because of the greater genetic similarity between MZ than DZ twins. This specific argument is used to rebut critics of classical twin studies who point out that such studies depend on the equal environments assumption, which holds that MZ twins' shared environmental variance is equally similar as is that of DZ twins. Hence, any effects resulting from the more similar shared environments of MZ than DZ twins is interpreted as genetic.\cite{jay2013} Though Joseph refers to this argument as "Argument A", I will refer to it as "Argument C" to avoid confusing it with the A or B arguments I have defined earlier in this paper. Argument C holds that this greater genetic similarity leads to a) MZ twins creating or seeking out more similar environments for themselves than do DZ twins and/or b) MZ twins experiencing more similar treatment from others compared to DZ twins. The reason I did not state earlier that behavior geneticists making Argument C are necessarily making the very argument Sesardic claims that they never make (at least to his knowledge) is that a) is clearly equivalent to active G-E correlation as defined by Sesardic. As such, interpreting a) as the result of genetic factors, as those making this argument do when they talk only about MZ twins creating environments based on their own behaviors, is valid, according to Sesardic. But making Argument C including point b) that I defined above to defend counting MZ twins' environmental influences as genetic is not valid, as Sesardic acknowledges. The crucial point I wish to make is that, while Sesardic may not be aware of any behavior geneticists having argued that indirect effects resulting from the actions of others--i.e. from external or imposed factors--should be interpreted as genetic, some behavior geneticists have made this argument. Specifically, some of these researchers have invoked point b) in their specific variations of Argument C in defending twin studies, and it is clear from reading how these arguments have been made that they are making the same point as Argument A. Below I will highlight some examples of behavior geneticists arguing in favor of Argument A in the context of classical twin studies.One relatively blatant example, highlighted by Joseph, comes from Sturgis et al., who wrote the following in a 2010 study: "If MZ twins are treated more similarly because they are biologically more alike, this can hardly be considered a violation of the EEA. For the reason that MZ environments are more similar than DZ environments (if indeed they are) is because of the initial difference in genetic predispositions."\cite{Sturgis_2009} (Emphasis in original) https://www.journals.uchicago.edu/doi/full/10.1086/377384https://www.jstor.org/stable/20446699https://www.sciencedirect.com/science/article/pii/B9780123708779001079
