4 September, 2026

In 2020, Nathan Cofnas, then a doctoral student in philosophy at Oxford, published a paper in Philosophical Psychology arguing that closing off academic debate about the sources of “group differences in intelligence” carried under-appreciated moral risks. In particular, Cofnas argued that it was a mistake to dismiss hereditarianism, according to which there are genetic or biological differences between races that explain measurable differences in traits or in social circumstances, as unworthy of investigation. Nine academics, from philosophy and other disciplines, challenged Cofnas’s paper, and the journal’s decision to publish it, in a response piece. The central claim of this response was that the hereditarian theory of group differences, which Cofnas himself advocates, relies on the notion that races are real. This notion, they wrote, is undermined by contemporary science. Over the past few weeks, too, many critics have taken the view that hereditarians like Cofnas err by attributing some sort of reality to race.

This is a mistake; the question in metaphysics or philosophy of science about whether race is a “real” category has little to do with the explanatory claims these hereditarians try to make, and it also has little to do with what ends up offending people about such claims. So I will argue in this essay. First, I will explain how you can be a metaphysical realist about race, even as a biological category, without being a hereditarian. Then I will explain how you can be a hereditarian without being a metaphysical realist about race. Finally, I will explain how neither hereditarianism nor metaphysical realism about race exhaust the political offensiveness of claims about race. Some further points will emerge as I explore these topics: explanations given in biological terms can answer sociological questions without thereby biologizing social categories; broad appeals to the arbitrariness of categorization or taxonomy should not be satisfying to opponents of hereditarianism; and so on.

Metaphysics of race without hereditarianism. Among philosophers this will be the least controversial section. While the view that race isn’t “real” was very popular in the 1990s, it has been superseded as the conventional wisdom by the view that race is real but a “social construct”. The view that race was not “real” came – fairly or unfairly – to be associated with colorblindness, which – again, fairly or unfairly – came to be viewed as a conservative, even racist-friendly attitude. Supreme Court Justice Sonia Sotomayor captured the new progressive way of speaking well in her dissent in Schuette v. Coalition to Defend Affirmative Action, writing:

Race matters. Race matters in part because of the long history of racial minorities’ being denied access to the political process. . . . Race also matters because of persistent racial inequality in society—inequality that cannot be ignored and that has produced stark socioeconomic disparities . . . And race matters for reasons that really are only skin deep, that cannot be discussed any other way, and that cannot be wished away . . . Th[e] refusal to accept the stark reality that race matters is regrettable.

However, it is not only social construct theorists who can accept that race is real without being hereditarians. There are also non-hereditarian biological theories of race. Penn professor Quayshawn Spencer is probably the most notable academic proponent of this sort of view. He advocates, among other things, that the actually-existing races in fact include the “lay races” asked about on, say, a census questionnaire, and that these races are biologically classifiable by correlations in genes that do not code for important traits. Some very large percentage of human DNA is “non-coding” like this.

Overall orthogonality. It is perhaps less obvious that hereditarianism does not entail that race is a biologically significant category—although, since the hereditarian position is not taken very seriously by many philosophers, it is also unclear just what sorts of reactions this lack of an entailment would inspire. In conversation, one philosopher suggested to me that, without any metaphysical baggage, the hereditarian position would no longer be interesting. I doubt that!

Painting with a broad brush, hereditarian explanations tend to have the following structure. There is some social pattern with regard to race, some variance in circumstances or outcomes or measures. This social pattern is then given a cognitive explanation, in terms of some variance in underlying cognitive or psychological traits, which is also racially patterned. For later reference, we will call this step trait-ism. Then the cognitive pattern is in turn given a genetic explanation. The idea that the hereditarian position makes metaphysical assumptions about race seems to be based in the idea that this final step of giving a genetic explanation requires that there be genetic racial groups of the sort that, say, a geneticist with training in biology might be interested in investigating.

However, that is not the case. For instance, suppose that Dayton high schools kept beating Akron high schools at basketball. One might explain this fact in terms of variance in a biological trait, like height. Then the ultimate explanation of the on-average height difference between Dayton students and Akron students might be that there is, perhaps for reasons of sheer coincidence, some difference between Dayton and Akron when it comes to the distribution of the genes that determine height. None of this requires thinking that “Dayton students” and “Akron students” are biologically or genetically interesting categories. One wouldn’t write a paper in a biology journal about Dayton students or Akron students. But that’s fine; the question as posed isn’t a question in biology. More generally, explaining patterns in one area of inquiry with facts from another area of inquiry does not require that the categories in which the patterns were framed are robust in the second area of inquiry. The patterns are part of the explanandum (what is to be explained), not part of the explanans (what is to do the explaining).

This also gives a more general answer to those who claim that once we understand that race is ultimately a historically contingent, politically motivated, or socially constructed category, we will stop seeking explanations that advert to biology for facts about race. The more general answer is that the kind of category something is – whatever that means to begin with – doesn’t determine what kinds of explanations are relevant to it. In fact, different explanations, at different levels, can be given for all sorts of facts. Biology can influence culture, which can turn back to influence biology, and so on. There’s no reason to think that the right explanatory story will respect our silly categories of academic investigation.

Against categories altogether? Some opponents of hereditarianism have resorted to more extreme attacks on biology or taxonomy. But these are too general to have much bite. On Twitter, Josh Duncan writes: “Pretty fucking bleak that the Cofnas affair has revealed just how many people believe that race is some sort of objective, biologically determined category. The spirit of scientific racism never died it just bubbled away. As if any biological category can be objective and free of the social anyway.” But if no biological category can be objective and free of the social, it doesn’t do much to say that race is not objective and free of the social. If we can attack on the same grounds, for instance, the distinction between fish and birds, the problem is a much more general one.

Similarly, LSE sociologist Alonso Gurmendi started an enormous, broad discourse on Twitter about the purported arbitrariness of categorization by posting a picture of three clusters of dots, in which a few of the dots were not easily classifiable into any one cluster. But if all categories – or at any rate all categories with borderline cases – are arbitrary, it is hard to see how racial categories will be specifically undermined. After all, we are perfectly happy to use many categories, clearly delineated or not, in both our casual and our scientific investigations. So the theorist of category arbitrariness must say something much more specific about race if they want the critique to make a dent.

These sorts of critiques actually undermine the notion that there is some sort of general principle of within-category explanation that researchers tend to respect. Scholars in the humanities and social sciences do not in general avoid explaining scientific findings in terms of their own disciplines. Whole swaths of research findings might be explained from the perspective of economics, psychology, sociology, history, or whatever else. There is no a priori reason to think that our best explanations will be within-category explanations in this way.

Specifications and subpopulations. Critiques of a connection between biology and race suggest that it obscures more than it clarifies. So Kenan Malik wrote on Twitter:

Ironically, sickle cell anaemia is a good example of the problems of reducing issues to racial categories. Because African Americans and Britons of Caribbean origin suffer disproportionately from the trait there is a widespread perception that sickle cell anaemia is linked to race, and to being black. Sickle cell is actually a disease of populations originating in areas with high incidence of malaria. Some of these populations are black, some are not. . . . The majority of people in Africa do not suffer from sickle cell disease. . . . In America, the presence of the sickle cell trait can help distinguish between those with African origins and European origins. In South Africa it can’t, because black South Africans have a very low incidence of the disease. . . . It is the social imagination, not the biological reality, of race that turns sickle cell into a ‘black’ disease.

Malik’s discussion is very illustrative. It seems to suggest that, for race to have some biological reality, or for race to be “linked” with biology, there would need to be questions for which race is always and everywhere an important biological consideration. If that is really a requirement, then of course race has no biological reality. Whatever correlation there might be between social groupings and biology is always subject to change. For instance, if malaria were suddenly to spread around certain parts of the world and not others, correlations could suddenly emerge between social groupings and sickle cell genetics. Extreme examples can bring this point home. If for several generations only tall people reproduced in Dayton and only short people reproduced in Akron, then soon there might be correlations between the genes for height and urban residency. But hereditarians need not deny that. The facts they think they can explain are facts from the present day or other specific moments in history. They don’t claim that there are timeless racial essences which are unaffected by broader historical circumstances – the sort of racial thinking characteristic of the horrors of the nineteenth and twentieth centuries.

Another type of response has it that there are natural, non-arbitrary biological groupings of humans, but that these are not races but are instead “subpopulations” – for instance, the types of groupings a 23andMe test might provide. Again, few hereditarians deny this. Indeed, the structure of their explanations might rely on it – say if 2/3 of the Dayton subpopulations had certain height-linked genes and only 1/3 of the Akron subpopulations had those genes.

What actually offends people? The point about subpopulations makes for a point about offense. On Twitter the historian Moshik Temkin made a similar observation, and was asked by an anonymous commenter: “If Cofnas said ‘ethnic groups differ in intelligence due in large part to genetic differences’ instead of ‘race’ I don’t think anyone would be less outraged.” Temkin responded: “They would be equally outraged because it’s equally false and offensive.” Temkin is surely right that people would be equally outraged. Moreover, members of one subpopulation can discriminate against another and commit all sorts of ethnic crimes. Recourse to subpopulations to undermine a purported metaphysical assumption thus cannot suffice for the anti-hereditarian because the anti-hereditarian believes that their view, and the moral caution or offense associated with it, covers groups that they might think of as far less arbitrary.

Indeed, it is not clear that it is hereditarianism at all that causes offense. Here is the second paragraph of a piece from the British newspaper The Observer:

Cofnas is a self-described “race realist” whose academic research focuses on the philosophy of biology and ethics. In February 2024, he published a post on Substack in which he claimed that race is linked to intelligence and that without DEI initiatives there would be no black academics at Harvard and that “under a colour-blind system” Black people “would disappear from almost all high-profile positions outside of sport and entertainment”. He called for a “revolution” against the “DEI devil” and for people to “accept that talent is not distributed equally within or across groups”.

This is the most commonly-cited offensive passage from Cofnas’s writings. (He has written more offensive things, but that’s a separate point.) Strikingly, however, it does not require a commitment to hereditarianism at all, let alone to some view on the metaphysics of race. Perhaps the word “talent” has some innatist connotations, but in general one can believe in an uneven distribution of qualifications without being a hereditarian. The old view, now considered racist in its own right, was that members of different racial groups were often also members of different cultures, and that culture was a significant determinant of qualification – that some cultures simply produced less qualified people than others.

What is the view that doesn’t cause offense? Roughly speaking, it is that bias and discrimination themselves cause different outcomes. However, the middle step in the hereditarian explanation, what I called trait-ism, is skipped in these kinds of explanations. You virtually never hear someone say that racism affects people’s outcomes by affecting the actual traits which might make them susceptible to certain outcomes. (This is a rather modern shift. Historically important arguments against slavery, for instance, often held that slavery made both masters and slaves not just worse off but in some sense worse people, somehow morally corrupted or degraded: the former by failing to uphold human dignity, the latter by foreclosing the sort of character development that comes from agency and responsibility.)

It therefore seems to be trait-ism itself that most upsets people. The suggestion that different groups have different distributions of important, valuable, and meritorious traits is a hard one to bear. (We find it pretty hard to handle even on the individual level, in fact.) One softening approach, taken by writers like Freddie deBoer and Paige Harden, as well as some hereditarian figures like Charles Murray, is to deny that the traits in question are actually important, valuable, or meritorious in a deep sense. Here hereditarianism might actually be on safer ground than other kinds of explanations, because it makes it easier to adopt a kind of group luck egalitarianism. Briefly stated, the luck egalitarian perspective would be something like: we are not responsible for our genes, so we are not responsible for the traits they code for. Since what really matters in evaluating a person depends on what they’re responsible for, these traits aren’t what really matters. A challenge is to see how anything can really matter in that case.

But of course someone might deny trait-ism altogether. They might do this by denying the validity of the measured differences which differences in traits were meant to explain: for instance, denying the validity of IQ tests as a measure of intelligence. Alternatively, they might deny the coherence of the posited explanatory trait: for instance, denying that there is such a thing as general intelligence or g. They also might give a direct explanation of the measured difference in terms of bias or discrimination, or something similar: for instance, holding that differences in results on standardized tests and intelligence tests are due to stereotype threat. These moves are in effect arguments against trait-ism. Hereditarian explanations assume trait-ism, and so they are arguments against hereditarianism as well.

When something controversial happens in the world, like the Arday–Cofnas scandal, academics often look for ways for their areas of expertise to be relevant. Thus, specialists in metaphysics might think that arguments about metaphysics are what the intellectual world needs to be able to challenge Cofnas’s views. Historians might think that what we need is knowledge of history, social theorists that it is an understanding of the dynamics of social construction, linguists and philosophers of language that it is an account of slurring and propaganda, psychologists that it is an account of biased cognition, and so on. This can look opportunistic, but we are all trying to explain the world and its events in the best ways we know how. One explanation need not exclude another, even – perhaps especially – if they come from different areas of inquiry.

Oliver Traldi is an assistant professor in philosophy at the University of Toledo’s Institute of Constitutional Thought and Leadership and author of Political Beliefs: A Philosophical Introduction.