A Belgian entity called Unia (The Interfederal Centre for Equal Opportunities and Opposition to Racism and Discrimination) is preparing a criminal case against the philosopher Nathan Cofnas. If successful, a jail term of between six months and three years is possible under a new criminal code that came into force on 1 September 2026. Articles 250 and 251 of that code deal, respectively, with “incitement to discrimination or racial hatred (including denial of genocides)” and “dissemination of racial ideas.”

It is uncommon for a postdoctoral researcher in philosophy to be threatened with prison. What has Cofnas done? Unia’s complaint states:

This American philosopher is spreading a message in our country that would establish a hierarchy amongst human beings on the basis of so-called scientific knowledge, in order to justify discrimination based on origin or skin colour, or to promote forms of segregation. According to Unia, these statements contravene anti-racism legislation and the United Nations International Convention on the Elimination of All Forms of Racial Discrimination (CERD). We reserve the arguments underpinning this action for the courts.

This reservation means that I cannot assess specific allegations or discuss their legal basis. Instead, I will examine the scientific claims at issue and distinguish them from the social conclusions often attached to them. Does the currently opaque case against Cofnas have intelligible moral grounds for such extraordinary possible consequences?

Cofnas has outlined his position on race in academic papers, on his Substack, and in a talk he gave to the Catholic Flemish Students’ Association on 2 April 2026. He states that there are average differences between groups in the distribution of socially relevant characteristics including intelligence. He claims that genetics contributes to causing those average differences between groups. And he suggests that this is a silently held, but reasonably mainstream view among scientists with relevant expertise.

So what is the scientific consensus on Cofnas’s factual claims? We will focus on the socially relevant characteristic, intelligence, here because we care about it. 

First, intelligence is unequally distributed among individuals; the Hungarian-American physicist and mathematician John von Neumann was more intelligent than I am. Second, genetics contributes to the causes of differences in intelligence among individuals; I could no more have come up with the Hilbert-space formulation that flowed so flawlessly from von Neumann’s pen than I could have run a four-minute mile. I could have been boiled alive in the Stygian parenting suffered by poor J.S. Mill, and it would never have propelled me to formulate eigenvectors in a quantum system. From anecdote to data: hundreds of large-scale studies suggest that the proportion of variance accounted for by genetic and non-genetic factors in individual differences in intelligence is around fifty percent each. These estimates may be revised upwards or down and they are specific to a particular population at a moment in time. The point is that there is a very strong scientific consensus that genetics plays a role in individual differences in intelligence.

Criminal Research

European jurists should not seek to arbitrate controversial matters best settled by science.

QuilletteRosalind Arden

What about group differences? Do groups have different distributions of cognitive scores? Yes. There are good enough data to claim that the distribution of intelligence varies among groups. Every group measured (note: the evidence is variable in quality) overlaps, and different means are well established for some groups. For example, because East Asians have an IQ advantage (of several IQ points) over Europeans, some elite educational establishments have resorted to chicanery to restrict the number of East Asians who are awarded places. Yet undergraduate enrolment at Caltech—where they teach the hard stuff and want a really bright intake—is around 46 percent Asian descent (compared with their roughly eight percent share of the US population) according to their Fall 2025–26 statistics. Note that “Asian” in that report denotes a wider category than East Asians. I have written about race and categories elsewhere. Races or groups are partly socially constructed—fuzzy not discrete—yet self-reported race correlates with genetic gradients.

What does the science say about what causes the observed group differences? The causes of intelligence differences between individuals need not match the causes of differences among groups. As a former president of the International Society for Intelligence Research and an attendee at almost every annual conference since before I needed reading glasses, I can say with some authority that the empirical science on the causes of group differences remains unsettled. We do not yet know them. Many scholars think that a mix of factors is likely and would not be startled by evidence of a genetic contribution to group differences. Others intuit that genetics is likely to make no contribution. The opinion that 100 percent of the currently observed gaps among groups is caused by genetic factors is a minority view.

Learning about the causes of group differences in intelligence is hard. And as Glenn Loury said recently when he and John McWhorter interviewed Cofnas, there are many possible causes of group differences. To what kind of data can we appeal? We can examine data from trans-racial adoption studies, but interpretation of these is confounded by early life experience and sample selection. Ancient DNA allows us to explore changes in genetic variation over time and among different groups—but it is a very indirect route to answering questions about current cognitive variation among groups.

What about within-family studies? This is likely the most promising approach. A highly reputable recent study used a clever design to tease apart genetic and non-genetic causes of educational attainment within Mexican families where the siblings differed in their mixture of European and Indigenous American genes. The study showed that having a higher or lower proportion of “Indigenous American genes” or “European genes” made no significant difference to the siblings’ educational attainment. That study did not rule out a genetic contribution to group differences in intelligence (which was unmeasured). As evidence accumulates, we should attend to Carnap’s total evidence rule—rational belief should depend on all the relevant evidence available rather than choosing the subset favoured by our biases. This applies when we have enough high-quality evidence to attend to.

There is no incontrovertible evidence that group differences are caused by genetics. The lack of empirical evidence does not allow us to assume that the evidence will fall on our side when it eventually emerges. But it also cannot be wrong to have an opinion about which way an empirical question is likely to be settled. Dogma on either side of an unanswered question is wrong. Unia’s position that such work is “so-called scientific knowledge” is not just ignorant of the status of domain expertise, but it also risks being a hostage to fortune. Calling all such work pseudoscience traduces those who engage in it, but it also fails to articulate which parts of the claims are being tested.

In 1974, philosophers Ned Block and Gerald Dworkin argued that scientific research into group differences in intelligence should be voluntarily discontinued. Many others in various disciplines agree; they argue that it will be misunderstood, that it will amplify racial abuse and harassment, that it will induce fatalism, that it may sway funding priorities, and that it will deliver no social benefit in any case. Those who defend the importance of this research—including Nathan Cofnas—argue that we should know the causes of things, because if we don’t know the causes of things, we will attribute wrong causes (discrimination, racism, low socio-economic status, for example) to what in the US is called “disparate impact.” Disparate impact occurs when an ostensibly neutral policy disproportionately affects a group. Governments are unlikely to make effective policies if they misunderstand the likely causes of consequential attributes of the people whom they were elected to serve. And recall, a difference in average intelligence does not establish a difference in moral worth.

If the claim that genetics contributes to average differences in test scores is given to the courts to judge, we return ourselves to the burning of Giordano Bruno in 1600. Bruno concluded that the universe was infinite and had no centre; he agreed with Copernicus that the earth moved around the sun; and he was sceptical about the divinity of Christ. It takes a huge imaginative leap now to cast ourselves into a world in which theological or cosmological ideas would convulse societies, shake and shatter people’s core beliefs. But science has been seismic in the past: consider cosmology, evolutionary theory, and plate tectonics. We have come through it all. But Bruno didn’t; he was set on fire for his heretical views, and accused by his contemporaries of being abrasive, quarrelsome, and offensive. This is not to establish Cofnas as a modern-day science martyr. The parallel is that among a vocal crowd, the irrelevance of genetics to group differences is such a passionately held conviction that it causes people to feel justified in hurling accusations and glassware. We must all eventually concede to empirical reality; while we do not yet know where the truth lies, we should not silence those whose views are other than ours.

It is important to correct another misapprehension in Unia’s Cofnas complaint. Even if Cofnas turns out to be right about the causes of group differences, this will not “establish a hierarchy amongst human beings.” A higher group average in a measured trait—even a trait as valuable as intelligence—is one claim; superiority in worth or entitlements is another. Jack may be less bright than Jill. The science cannot say he is inferior or worth less. The way that we value individuals is richer than that, and it does not reduce to a number. That value derives from what we know of a person—all things considered. It can include intelligence but it may not. How would we value a highly intelligent serial rapist against a street-sweeper with low cognitive ability, for example? This is what we generally mean by moral worth. What a person is owed is independent of whether we like or agree with them, and it is forfeited only through due process in response to specific acts of harm.

Nathan Cofnas has inflicted no harm, but the baleful threat of prison is only the latest item on a list of harms that he has experienced for expressing unpalatable views. Cofnas has been accused of racism and even murder. He has received death threats, hundreds of venomous racist emails, and thousands of vituperative social media posts. He has been pelted with yoghurt after a lecture and concussed by a bottle thrown at his head. Students chanted outside his office at Cambridge University. He was suspended from Ghent University (reversed seven days later, apparently helped by the intervention of US ambassador Bill White). Cofnas was banned from campus and teaching for ninety days and he remains under investigation. He has been the subject of 1,633 news articles (within the last year) according to my Factiva search. And now, he faces this criminal case with the possibility of jail. The harms to Cofnas involved different actors, but they all arise from the same mistake: they let feelings about racism determine thoughts about what Cofnas did. They have each mistaken a descriptive claim for a normative claim.

A final word about Jason Arday. Cofnas published convincing evidence that Arday had plagiarised. Arday subsequently died and that was tragic. Arday had been racially abused, which was wrong and great blame attaches to those who did it. But Cofnas did not racially abuse Arday, and the fact that Cofnas made public Arday’s academic wrongdoing does not license the accusation that Cofnas murdered him. Yet he was scapegoated. I was accused of the same crime after an essay of mine appeared on the day of the death, published hours before the news broke. Such accusations are a manifestation of the need to blame someone. And they are, of course, distressing to receive by the hundred.

Jason Arday and Nathan Cofnas: Cambridge’s Double Standard

The Cofnas and Arday controversies show that on matters of race, administrators at Cambridge have failed to conduct deliberate checks, failed in probity, and failed in their tolerance for unsettled science.

QuilletteRosalind Arden

NewsCord stated that 249 newspaper articles were written about Arday. The University of Cambridge claims it provided him with counselling, security measures including panic alarms, and screening of racist posts. They were right to do so. But Cofnas was offered no such care or protection when he became the subject of protests and harassment while he was there. On the contrary, the Ghent University has added to the difficulties experienced by their objectively harassed postdoctoral researcher. An urgent suspension of Cofnas followed Arday’s death. It cited a possible risk to Ghent University itself. Ghent’s public announcement leaves open the impression (without asserting it) that Cofnas was implicated in Arday’s tragic death. The student at Ghent who threw the bottle at Cofnas’s head was positively identified on CCTV. The incident was reported to the university on April 10, this year. Six months later, on 15 October, the university notified Cofnas that the Rector had referred the incident to the Disciplinary Chamber. No health or safety check was provided after the concussion. Basic care is owed to Cofnas. Disliking him or disagreeing with him is not an excuse to withhold what an institution owes to its employee.

Cofnas, too, has a duty of care. His communication sometimes shows a fault of manner. Cultivating a culture of thoughtful speech (and writing), in LSE president Larry Kramer’s words, means that “the focus remains on the idea, as opposed to trying to provoke or score points or grab attention by being intentionally hurtful.” But a failure of tact or sensitivity on Cofnas’s part does not begin to justify the litany of harms he has experienced. Robust rebuttals carry weight. Threatening a scholar with a prison sentence silences the careful along with the clumsy, and it shuts down scientific inquiry. We can reject racism and assert that Nathan Cofnas is being treated shamefully. There is no trade-off. A university and a culture that fail to understand this have lost their way.