Skip to main content

A biological realist about race holds that races are real divisions of the human species, discoverable by science, whatever people believe about them. The view has an ugly past and a more careful present, and it is important to keep the two apart.

The old racialism

Nineteenth century race science claimed far more than visible difference. Writers such as the French aristocrat Arthur de Gobineau, whose Essay on the Inequality of the Human Races appeared from 1853 to 1855, held that races had fixed essences, that intelligence and character came packaged with skin and skull, and that the races could be ranked. The philosopher Michael Hardimon calls this bundle the "racialist" concept of race, and nearly everyone today, including biological realists, agrees it is false. After the Second World War, a 1950 UNESCO statement drafted by the anthropologist Ashley Montagu went so far as to suggest replacing the word "race" with "ethnic group."

A minimal concept

Hardimon argues that rejecting racialism does not mean rejecting race. His "minimalist" race is a group marked by visible physical features, linked by common ancestry, and originating in a distinct geographic region. That concept makes no claims about intelligence, morals or rank. Hardimon thinks minimalist races exist: there really are groups of people with characteristic appearances who descend from ancestors in, say, sub-Saharan Africa or East Asia. The ordinary concept, stripped of its bad add-ons, survives.

Genetic clusters

Newer evidence comes from population genetics. In 2002 a team led by Noah Rosenberg analyzed DNA from 1,056 people in 52 populations using a program that groups individuals by similarity without being told where anyone came from. When asked for five groups, the program produced clusters that matched major geographic regions: Africa, a broad Eurasian region, East Asia, Oceania and the Americas. Many realists took this as evidence that continental ancestry leaves a measurable trace in the genome.

The statistician A. W. F. Edwards added a point about method. Critics had stressed that any single gene varies more within groups than between them. Edwards argued in 2003 that differences across many genes, taken together, can sort individuals into groups reliably, even when each gene alone tells you little. A single word rarely reveals which language a text is in; a paragraph almost always does.

Philosophical versions

Philosophers have built several versions of realism on these findings. Robin Andreasen proposed in 1998 that races could be defined cladistically, as branches of the human family tree, defined by descent rather than by appearance. Quayshawn Spencer argued in 2014 that the race categories used by the United States government pick out real genetic population groups, roughly those five clusters. Spencer stresses that this kind of reality is thin. It says nothing about which group is smarter or better, and it does not make race important for most purposes. A category can be real, he says, without being the kind of thing that ought to organize a society.

Race in medicine

The strongest practical case for realism comes from medicine. Some genetic conditions and drug responses vary by ancestry, and researchers such as Neil Risch argued in 2002 that self-identified race is a useful proxy in biomedical research. In 2005 the drug BiDil became the first medicine approved in the United States for a single racial group, after a trial in self-identified Black patients with heart failure. Defenders say that ignoring ancestry in medicine could cost lives, especially among groups underrepresented in research.

Fair criticism

Skeptics reply that the clusters depend on researchers' choices about which populations to sample and how many groups to ask for. Ask the program for three groups or twelve and it will oblige. They also note that self-identified race is a rough proxy for the genes that matter, and that medicine might do better by testing for those genes directly. Realists answer that every classification involves choices, and that a fuzzy boundary does not make a category unreal, any more than dusk makes day and night unreal.

Worked example: the kidney equation

For two decades many American labs estimated kidney function with formulas that multiplied the result by a fixed amount for patients recorded as Black, based on studies suggesting higher average muscle mass. The adjustment made Black patients' kidneys look healthier on paper, which could delay referral for specialist care or a transplant. In 2021 a task force of two leading kidney societies recommended dropping race from the equation. How should a philosopher read this? A realist can say the original problem was crude measurement: race stood in for something better measured directly, such as a blood marker. A skeptic can say the case shows how a social label, applied to a hugely varied group, slips into science and passes as biology. Both agree on the remedy, which is to measure the relevant biological fact instead of guessing it from a census box. Notice that the debate about whether races exist did not need to be settled to fix the equation. What mattered was whether race was the best available variable for that purpose.

Quiz

5 questions · Practice: scored, not saved

1. What does Hardimon call the old bundle of fixed essences and rankings?
2. Which features define Hardimon's minimalist race?
3. What did Rosenberg's program produce when asked for five groups?
4. What did Edwards argue in 2003?
5. How did Robin Andreasen propose to define races?
6. What did Quayshawn Spencer argue in 2014?
7. What was notable about BiDil's approval in 2005?
8. According to skeptics, what do Rosenberg-style clusters depend on?

CivilSimian.com created by AxiomaticPanic, CivilSimian, Kalokagathia