Imagine that there is a trait observed among people that seems to occur more frequently in some families and not others. One might suspect that the trait is inherited genetically. Imagine researchers looking for the genetic underpinning of this trait and at first, not finding it. What might you conclude? It could be reasonable to conclude that the genetic underpinning of the trait is elusive, perhaps complicated with multiple genes, or that there is a non-genetic component, also not yet identified, that makes finding the genetic component harder. Eventually, you might assume, the gene will be found.
That is probably true sometimes. But we have sequenced the entire human genome, so shouldn’t we know about all the genes? Well, yes and no. We may have a list of genes found in a sample of humans, but “The Human Genome” can consist of a single individual (though it does not) and miss variation between individuals, i.e., it may not be a record of all of the possible alleles (variants) of each gene. Also, beyond the scope of this discussion but worth mentioning, a “gene” is not a simple concept. Whether or not a gene is expressed, where, when, and exactly what product it produces is not entirely encoded in the gene itself, but rather, elsewhere in the genome, or not encoded at all, but rather, dependent on external, non-genetic factors. So that complicates things too. So, if there is a trait that you think must be genetic, but years of research have failed to find it, the existence of a human genome and the prior acquisition of a lot of genetic data does not necessarily mean that the genetic information that determines the trait in question is not there. You can continue to believe that the genetic code for the trait will eventually be found
Except when you can’t.
There are two separate ways in which people sort out which traits are assumed to be genetic from those that are assumed to be not genetic. Both are heuristic, one is valid, and one is not. Let’s start with the one that is valid.
Suppose, as before, there is a trait that is seemingly inherited in families in such a way that a genetic trait would be, in the time tested manner that with respect this trait “offspring resemble their parents” as Darwin noted. The next question you can ask is this: Is it biologically sensible that this trait is inherited genetically, or is there a better, obvious, non-genetic mode of inheritance? If the trait is a physical feature such as eye color, then we have a sensible biological explanation for the trait having to do with developmental process we know something about and a set of metabolic pathways that produce various molecules such as pigments. The idea that this trait is genetic is biologically sensible, so even if you can’t find any, or all, of the genetic determinants of this trait, you can figure they are out there somewhere. Suppose, though, that the trait is a behavioral one that we see people in real life learning. For example, what language a person speaks generally follows the same kind of inheritance pattern many clearly genetic traits follow. With respect to spoken language, most of the time, offspring resemble their parents. But, rather than there being a sensible biological explanation for this trait, there is a sensible cultural explanation for this trait, so we don’t even look for the genetic variants for “French” vs. “Mandarin” vs. “English.” We simply assume this is not genetic.
The second method, the incorrect one, is to work with an article of faith. Broadly speaking, and I oversimplify greatly here, there are two primary articles of faith that often inform people’s thinking, shaping their assumptions, about genetics. Both usually have to do with behavioral traits in humans, but this can apply to physical traits as well. One article of faith asserts that humans are born as a blank slate, and all of their behavioral characteristics, such as their personality, intelligence by one measure or another, and so on, are added by experience. The other is the inheritance assumption, that some or much of an individual’s personality, intelligence, etc is determined by genes. There is not necessarily a consistent logic behind either of these assumptions, though various schools of thinking will include, often, a logical framework. However, this method of coming to a conclusion about the genetics or lack thereof behind various traits relies on one important element regarding genetic systems: Ignorance. If you are a blank slatist, then the absence of a clear pathway from genes to behavior means that your hypothesis can’t be falsified. If you are a genetic determinist, then the lack of such a pathway can be attributed to ongoing ignorance about the genes. The former might then be expected to live in fear that a gene will be found for their favorite learned behavior, and the latter might be expected to to live in a state of hubris, firmly knowing and asserting a truth that is not yet known but someday will be.
My impression is that over time there are fewer and fewer pure genetic determinists out there, and few and fewer blank slatists. I think the reasons for that shift have little to do with increasing knowledge, and more to do with changes in how one plays the academic game of argument, but that is discussion for another time. There is a danger in that shift, though. In the absence of any useful research results, if blank slatists start to admit that there could be some sort of genetics behind behavior, and determinists start to admit that experience and learning can also play a role, then we are converging on an increasingly simplified view of what is really a very complicated process. We should be gaining more complex, nuanced, and better informed views of how behavior arises, not simpler ones. Probably.
Over the last few decades, there have been a few important changes in how we should view human behavior over generational time and variation in those behaviors within and across categories (gender, ethnicity, geography, etc.). In short, certain behavioral traits have shown, synchronically (lacking the perspective of change over time) patterns that look genetic. For example, some families seem to be extra smart. Some have suggested that some “races” are smarter than others (at another time we can discuss why there really are no races, but let’s use “race” here as a potentially valid sampling strategy, which it can be even if the underlying races are fictions). We also see assertions of behavioral differences between the primary sexes (male vs female).
These observations are really statements about variance. Two groups are different, but vary within. There is overlap in the trait (i.e., IQ) but the means vary. We can statistically test the validity of the asserted differences in means by examining the variance in each sample and seeing if the mean of one sample fall within the predicted range of the central tendency of the others. In other words, asserting that there is a statistical difference between two groups is a process that involves understanding the variance of the underlying population(s) and samples. So, the questions can all be reframed in this manner:
Is the variation we see in trait X across certain groups best explained by underlying corresponding variation in the genetic system, or by the variation found in some other cause?
People fight vigorously over the underlying cause of IQ differences between groups. Some say it is primarily genetic, some say it is primarily not genetic, but rather, related somehow to what has become known as “lived experience.” Over the last couple of decades, there have been many attempts to explain observed variation in IQ using socioeconomic status, diet, education, issues having to do with test making or testing procedures. All of these factors have been shown to explain differences between groups to a modest to large degree in several studies. In other words, if you want to explain variation in IQ using non-genetic explanations, you can have some real success.
The genetic explanation of variation in IQ has had success in one main area which is irrelevant. This is the fact that genetically determined developmental differences between people that affect function that are generally classified as disorders predict large IQ differences. But this set of effects is not related to the question being asked.
The strongest evidence for a genetic underpinning of IQ is probably the large scale racial model solidified years ago by J. Philippe Rushton. He demonstrated that there is a grouping of brain sizes by race, with Asians having the largest brains, Caucasians the second larges, and Blacks the smallest (these race terms are his). He then showed that these brain sizes correlated with IQ difference. The modern psychometric literature assumes a racial difference in IQs, and asserts that this difference is real, but does to by citing sources that then site sources that ultimately cite Rushton. Rushtons all the way down, as it were.
The problem with this is that Rushton’s analysis was bogus. The brain sizes were taken from such sources at hat sizes for army conscripts classified by race, with the hat sizes used to estimate brain size. The Black (African) brain got smaller because Rushton subtracted a factor from that estimate of brain size, using an archaic thick skulled African fossil to assume that Africans have very very thick skulls. Correspondingly, the Asians were assumed to have thin skulls, and thus, got larger brains. The IQ data is similarly adulterated. In one part of the study, Rushton needed an “African” (native) IQ value, so he used the results of a test administered by racist anthropologists commissioned by the Apartheid government of South Africa to prove the inferiority of Blacks. And so on. The bottom turtle in this edifice is a fake.
The range of variation across “racial” groups (or other groups) in modern IQ data is very small compared to the change in IQ measured or estimated over decades of time through the 20th century within a single large and diverse population (Americans). If IQ is genetically determined and a stable feature of behavior, then there has been more evolution of these genes over less than 100 years of time in the US than we see across any two groups of modern humans. That is impossible. Again, IQ does not behave nicely as a genetic trait.
The discovery of a gene or set of genes that would underly IQ has not happened. In some recent studies, IQ is assumed to be very complex and the result of many different genes, and there is some statistical evidence for this. But, there is a big problem there too. Any trait can be linked to a set of genetic variants if the set of genes is large enough. That is a statistical effect and it is not really a link. More like a party trick, or a con game. (In fact this method is a con you may have heard of. I send 10,000 people an email predicting that a certain stock will go up, another 10,000 people an email predicting it will go down. One or the other happens. I then send 5,000 of the people who got the “correct” prediction another prediction, and 5,000 of them the opposite prediction. Now, 2,500 people have gotten two correct predictions from me. I keep doing that until I’ve got several dozen people convinced I am a stock market genius, and I take their money.)
Generally speaking, many behavioral traits have been explained, in part and sometimes in large part, by factors that are not genetic, while at the same time, the hunt for the presumed underlying genes have come up empty. There was great optimism up through the 1990s that genetic underpinning of human behavior … genetic variation corresponding to behavioral variation … would be found. But even as early as 1993 this was being questioned. Here is a sidebar, reproduced in full, from a Scientific American article by John Horgan summarizing the work up to that time:
Behavioral Genetics: A lack of progress report (1993)
CRIME: Family, twin and adoption studies have suggested a heritability of 0 to more than 50 percent for predisposition to crime. … In the 1960s researchers reported an association between an extra Y chromosome and vio-lent crime in males. Follow-up studies found that association to be spurious. MANIC DEPRESSION: Twin and family studies indicate heritability of 60 to 80 percent for susceptibility to manic depression. In 1987 two groups reported locating different genes linked to manic depression, one in Amish families and the other in Israeli families. Both reports have been retracted. SCHIZOPHRENIA: Twin studies show heritability of 40 to 90 percent. In 1988 a group reported finding a gene linked to schizophrenia in British and Icelandic families. Other studies documented no linkage, and the initial claim has now been retracted. ALCOHOLISM: Twin and adoption studies suggest heritability ranging from 0 to 60 percent. In 1990 a group claimed to link a gene—one that produces a receptor for the neurotransmitter dopamine—with alcoholism. A recent re-view of the evidence concluded it does not support a link. INTELLIGENCE: Twin and adoption studies show a heritability of performance on intelligence tests of 20 to 80 percent. One group recently unveiled preliminary evidence for genetic markers for high intelligence (an IQ of 130 or higher). The study is unpublished. HOMOSEXUALITY: In 1991 a researcher cited anatomic differences be-tween the brains of heterosexual and homosexual males. Two recent twinstudies have found a heritability of roughly 50 percent for predisposition to male or female homosexuality. These reports have been disputed. Another group claims to have preliminary evidence fo genes linked to male homosexualty. The data have not been published.
This is from a study by Jay Joseph on the “Classical Twin Method in the Social and Behavioral Sciences”
The classical twin method assesses differences in behavioral trait resemblance between reared-together monozygotic and same-sex dizygotic twin pairs. Twin method proponents argue that the greater behavioral trait resemblance of the former supports an important role for genetic factors in causing the trait. Many critics, on the other hand, argue that non-genetic factors plausibly explain these results…. In 2012, a team of researchers in political science using behavioral genetic methods performed a study based on twin data in an attempt to test the critics’ position, and concluded in favor of the validity of the twin method and its underlying monozygotic–dizygotic “equal environment assumption.” The author argues that this conclusion is not supported, because the investigators (1) framed their study in a way that guaranteed validation of the twin method, (2) put forward untenable redefinitions of the equal environment assumption, (3) used inadequate methods to assess twin environmental similarity and political ideology, (4) reached several conclusions that argue against the twin method’s validity, (5) overlooked previous evidence showing that monozygotic twin pairs experience strong levels of identify confusion and attachment, (6) mistakenly counted environmental effects on twins’ behavioral resemblance as genetic effects, and (7) conflated the potential yet differing roles of biological and genetic influences on twin resemblance. The author concludes that the study failed to support the equal environment assumption, and that genetic interpretations of twin method data in political science and the behavioral science fields should be rejected outright.
With respect to psychiatric disorders, from the same author:
The psychiatric genetics ?eld is currently undergoing a crisis due to the decades-long failure to uncover the genes believed to cause the major psychiatric disorders. Since 2009, leading researchers have explained these negative results on the basis of the ‘‘missing heritability’’ argument, which holds that more effective research methods must be developed to uncover presumed missing genes. According to the author, problems with the missing heritability argument include genetic determinist beliefs, a reliance on twin research, the use of heritability estimates, and the failure to seriously consider the possibility that presumed genes do not exist. The author concludes that decades of negative results support a ?nding that genes for the major psychiatric disorders do not appear to exist, and that research attention should be directed away from attempts to uncover ‘‘missing heritability’’ and toward environmental factors and a reassessment of previous genetic interpretations of psychiatric family, twin, and adoption studies.
And from researcher Tim Crow:
A substantial body of research literature, identified by nine out of ten papers on genetics in the recent ISI research front on schizophrenia, claims to have established associations between aspects of the disease and sequence variation in specific candidate genes. These candidatures have proven unreplicated in large sibling pair linkage surveys and a targeted association study. Even if the case for an association be regarded as a lucky guess (assuming one gene in 30 000 was guessed right) the large linkage and association studies provide no evidence of sequence variation relating to psychosis at any of these gene loci. Thus this body of work must be regarded as an indicator of the extent to which the ‘eye of faith’ is able to discern meaning in complex data when none is present.
I could go on. There have been further criticisms of the twin studies, for example. The most interesting, potentially, of these studies was on twins reared apart, more or less separated at birth. Commonalities among such individuals would be strong evidence for a genetic underpinning, because these individuals were raised in completely different environments so there would be no chance of a learned or cultural component other than a general background effect of having been raised n the same planet, or in the same country. Right? Well, no. Twins separated at birth were mostly twins that were not all that separated. After all, where do researchers actually find twins truly and distantly separated at birth, especially in the days when people seeking birth parents had hardly become a thing yet? Many of these twins, probably the vast majority, were separated only in the sense that they were raised by different members of the same family, or separately by divorced parents. Many were raised in the same neighborhood or often, the same house. My brother and I are not twins, but we were “raised apart” by the criteria of the twin studies because my family was distributed among the rooms of a two family residence, so technically he and I had bedrooms at different addresses.
In sum, it is easier to find sociological, cultural, or environmental explanations for variation in human abilities, intelligence, or personality traits. The seeming inheritance by family of some of these traits may well be a combination of something genetic and something experiential or cultural, but when looking for the actual underlying causes, genetics has repeatedly come up wanting while environmental explanations do a good job of addressing a fairly large part of the variation we see. Models of race based differences are so poorly done, and are often highly politically motivated, that they should never be trusted. That scientific ship sailed a long time ago.
Maybe the blank slate theory isn’t so bad after all. It does not imply that just anything can happen when making a human being out of a sperm and an egg. After all, it is a blank slate and not a blank whatever. But it is probably not true that some people’s lived experiences are written on slate, while others on white boards, and still others on smart boards, even if there are some people who I’m sure assume that they were.
Horgan, John. 1992. Eugenics Revisited. Scientific American. June.
Joseph, J. (2011). The Crumbling Pillars of Behavioral Genetics. GeneWatch, 24 (6),4–7. Web page
Joseph, J. (2012). The “Missing Heritability” of Psychiatric Disorders: Elusive Genes or Non-Existent Genes? Applied Developmental Science, 16(2), 65–83. doi:10.1080/10888691.2012.667343
Joseph, J. (2013). The Use of the Classical Twin Method in the Social and Behavioral Sciences : The Fallacy Continues, 34(1), 1–40.
Lewontin, R. Human Diversity. 2000, Scientific American Library.
Marks, J. (2008) Race: Past, Present, and Future. In: Revisiting Race in a Genomic Age, edited by B. Koenig, S. Lee, and S. Richardson. New Brunswick, NJ: Rutgers University Press, pp. 21–38. PDF
Marks, J. (2008) Race across the physical-cultural divide in American anthropology. In: A New History of Anthropology, edited by H. Kuklick. New York: Blackwell, pp. 242–258. PDF
Tizard, B. (1974). IQ and Race. Nature, 247, (5349), 316.
Other posts of interest:
Also of interest: In Search of Sungudogo: A novel of adventure and mystery, which is also an alternative history of the Skeptics Movement.