Bio(un)ethical

#16 Quayshawn Spencer: What is race?

• with Leah Pierson and Sophie Gibert

Share episode

In this episode, we speak with Dr. Quayshawn Spencer, Robert S. Blank Presidential Associate Professor of Philosophy at the University of Pennsylvania, about what race is, why he’s a radical racial pluralist, and what this could imply in science and medicine.

(00:00) Our introduction
(11:23) Interview begins
(20:21) Methodology: What are philosophers of race trying to do?
(32:05) From aspiring physician-scientist to philosopher of race
(41:08) Becoming a realist about race
(44:44) Human continental populations and the OMB racial classification scheme
(1:00:43) The national meaning of “race” in US racial discourse
(1:12:35) Why not be a pure social constructionist?
(1:17:04) Implications: racial inequalities
(1:20:59) Implications: diversifying clinical trials
(1:39:46) Pluralizing race talk

Used or referenced:

Bio(un)ethical is a bioethics podcast written and edited by Leah Pierson and Sophie Gibert, with production support by Audiolift.co. Our music is written by Nina Khoury and performed by Social Skills. We are supported by a grant from Amplify Creative Grants.

Note: All transcripts are automatically generated using Descript and edited with Claude. They likely contain some errors.

Introduction:

Leah: Hi, and welcome to Bio(un)ethical, the podcast where we question existing norms in medicine, science and public health. I'm Leah Pierson, a final year MD-PhD candidate at Harvard Medical School.

Sophie: And I'm Sophie Gibert, a Bersoff Fellow in the NYU Department of Philosophy and soon to be an Assistant Professor at the University of Pennsylvania.

Leah: A few weeks ago, we released an episode with Dr. James Diao, where we discussed the use of race in clinical algorithms. Some of that discussion came down to questions about what race is and what work the concept of race can and should do. For instance, the designers of the new PREVENT equations, which aim to estimate cardiovascular disease risk, decided from the start that the equation should not include race as a variable, because "race is a social construct" and "including race in clinical equations can cause significant harm by implying that it is a biological predictor." In short, people's conceptions about what race is informed their views on when and how to use it in all sorts of real-world situations. So what is race? That's the topic of today's discussion. And without getting too into the weeds yet, we want to frame this discussion by highlighting some important issues about which there is consensus.

Sophie: Most academics agree on a core set of empirical facts with respect to race. We'll highlight four of these.

First, racial categories have shifted substantially over time and included or excluded different people. For instance, at varying times and in varying places in the US, people with Native American ancestry were considered to be Native American if one half, one fourth, or one 16th of their ancestors were Native American. These determinations were in turn usually tied to the conferral of, or more often denial of, rights.

Leah: Second, throughout human history, people have justified racist policies and practices by linking race to genetics. If you can convince people that one group of people is inherently inferior to another, it becomes much easier to dehumanize and mistreat that group. This kind of logic and the policies that have flowed from it led to and perpetuated some of the greatest atrocities in human history, including the Atlantic slave trade, the Holocaust, and the genocide of Native Americans, among other crimes against humanity.

Sophie: Third, to the extent that race has correlated with biological outcomes, this is often the result of social and political factors rather than genetic ones. To use a straightforward example, consider the case of lead poisoning. There is nothing innate about susceptibility to lead poisoning. Lead is something you encounter in your environment. So racial differences in rates of lead poisoning have nothing to do with genetics. Black children in the United States are exposed to much higher levels of lead than white children, which can have profound consequences. There are multiple causes of this, but a key one is that racist policies explicitly prohibited Black Americans from living in neighborhoods with less lead in their housing stock.

Leah: Fourth, as a 2021 New England Journal of Medicine article notes, "Race is also directly associated with genetic ancestry and therefore indirectly related to genetic variance that may affect disease and health outcomes." This issue came up in our recent episode, in our discussion of the fact that patients of Chinese descent tend to get diabetes at lower BMIs, as do patients of Filipino, Korean, Vietnamese, and other Asian descents. For this reason, the WHO Asia Pacific region has set a lower BMI threshold for overweight and obesity for people of Asian descent than the WHO's international BMI cutoffs. Although it's worth noting that risk levels differ across populations and different Asian countries have adopted different cutoffs. People have hypothesized different explanations for the differing relationship between BMI and diabetes in different populations, including ones that have nothing to do with genetics. After all, people in different parts of the world have different diets and habits. But explanations having to do with genetics have also been put forward. For instance, one study found that body fat distribution, which can make people more or less prone to diabetes, varies across ethnic groups and is partly determined by genetics. Importantly, the thing that's directly related to risk here is genetic ancestry. But researchers and clinicians often use race as a proxy for genetic ancestry, an issue we'll discuss more in the episode.

Sophie: So if most scholars are in agreement about these kinds of facts, why do philosophers disagree about what race is? Perhaps the most important answer to this question is that philosophers take different approaches to answering the question "What is race?"

Sally Haslanger, a philosopher and former Bio(un)ethical guest, uses an analogy to make this point. Consider the question "What is water?" There are multiple ways to answer this question. You could provide an explanatory answer, like "water is H₂O, an oxygen atom bound covalently to two hydrogen atoms." Or you could provide a practical answer, like "it's the liquid in lakes and rivers and pools that we swim in, bathe in, and drink." The best answer to the question "What is water?" depends as much on what work the concept of water is meant to do in a given context—for instance, whether you're teaching a chemistry class or telling someone whether it's safe to drink something—as it does on any empirical facts.

Leah: The question "What is race?" is similar to the question "What is water?" in that the best answer might depend on the context. But as Haslanger also notes, the question "What is race?" is also substantially more complicated. First, while we have a stable concept of water because water hasn't changed, we don't have a stable concept of race. Moreover, while people across cultures have a word that tracks the same concept of water, people across different cultures have different beliefs about race, use different racial groups, and racially categorize the same people differently. What Haslanger concludes is that "the idea that there is a single best interpretation of what race is across languages and cultures is not entirely plausible." Because there is no single best interpretation, philosophers of race are often engaging in very different projects. For instance, one philosopher of race might aim to determine what race means in the dominant US context, while another might be focused on determining what the concept of race is in a global context. Philosophical debates about race are thus different from lay debates about race, because much of the discussion comes down to issues related to the philosophy of language and metaphysics: what concepts does and should the word "race" map onto, and do the things they map onto exist? Depending on what projects they're engaged in, modern philosophers reach different conclusions about whether the word "race" maps onto social, political, or biological concepts, and whether the things they refer to are real (like horses) or not real (like unicorns).

Sophie: Today, we're talking to Professor Quayshawn Spencer about how we should understand race. He is the Robert S. Blank Presidential Associate Professor of Philosophy at the University of Pennsylvania, specializing in the philosophy of science, philosophy of biology, and philosophy of race. He has a PhD in Philosophy and Masters of Biology from Stanford.

In his co-authored book "What is Race? Four Philosophical Views," which we'll link in the episode notes, Quayshawn raises a question that's closely related to questions we discussed with Dr. Dao. Specifically, he writes: "We have an interesting and unsettled philosophical question about whether and if so, how race matters in calculating someone's risk for being born with a genetic disorder. Furthermore, answering that question encourages a position on the biological reality of race. If you think that race is not biologically real, then it probably would not make sense for you to include race in a calculation of someone's risk for developing a genetic disorder." The position Quayshawn is discussing here looks a lot like the one espoused by the creators of the PREVENT equations, who opted not to include race in their algorithm because of their view that race is a social construct. But we take it that many doctors and social scientists do not think race is biologically real, and yet do think it makes sense to use race to calculate someone's risk for developing a genetic disorder. So what is going on here?

Quayshawn’s view about the answer to this question has evolved over time. In a 2014 article called "A Radical Solution to the Race Problem," Quayshawn defended the view that the term "race" as it is used in US national racial discourse picks out a biologically real entity, specifically a set of human population groups. The basic idea behind this view—and we'll discuss and question this view in the interview—is that the dominant use of the term "race" in the United States is the one embedded in the US Office of Management and Budget's 1997 categorization scheme, familiar from the census and other official government documents, and that those categories are identical to a certain set of human population groups called the human continental populations.

More recently, Quayshawn has broadened his view, adopting what he calls "radical racial pluralism" or "pluralism" for short. On this current view, the correct theory of what race means, even within the US context, is pluralist. We have different ways of using the term "race," and while one of these refers to a set of human population groups on Quayshawn’s view, others do not.

Quayshawn’s view contrasts with views that listeners might be more familiar with belonging to the social constructionist tradition. To give an example that Quayshawn will mention in the interview, Sally Haslanger is known for having defended a view on which races are "first and foremost, sociopolitical groups marked by biological features that function within a dominance hierarchy. In the case of the current social structure in the United States, the dominance hierarchy is white supremacy, but races are formed within different hierarchies that aren't organized around whiteness." On this view, if we had no dominance hierarchies like white supremacy, there would be no races. Race, then, is a primarily social-political phenomenon, not a biological one.

Today, with Quayshawn, we want to better understand the metaphysics of race debate as well as his views in particular and their implications for real-world scenarios, like the inclusion of race in clinical algorithms and calls to diversify clinical trial populations. As always, you can access anything we reference in the episode notes or on our website, biounethical.com.

Interview:

Leah: Welcome to the podcast. Quayshawn, thanks so much for being here.

Quayshawn: Thanks for inviting me.

Sophie: In your work, you divide the main philosophical views on race according to two axes. First, whether they think race is a real thing or not, and second, whether they think it's a biological thing or not. Could you start by just walking us through these different views?

Quayshawn: Sure. It really came about from the history of the sorts of metaphysical views that philosophers of race have been throwing around since, I would say the mid-1980s, late 1980s, early 1990s. Of course, debates about race go back way farther than that. But in contemporary philosophy, this metaphysical sort of position of the debates kind of dates back to that.

And so you have at that early part, Anthony Appiah arguing with Lucius Outlaw and Naomi Zack. And in this first sort of iteration, you have Appiah arguing that, believe it or not, race has a biological essence. It has biological properties that constitute what makes something a race. And given what those properties are, it doesn't exist.

So that's a view I would call biological anti-realism. And so he's pushing back against a tradition in the social sciences and other parts of the humanities that started with Du Bois. We can call it for now social constructionism, but I'm going to get a little bit more precise with that in a second, where the essence of race is not biological and it is real.

Now in contrast to that, Lucius Outlaw came along and said, well, I'll grant you the biological part. It looks like it's hard to tease apart the biological nature of race, but it also has a social nature. And so he came up with what we can call a bio-social view, but I will put that squarely in the biological camp of views. And given that sort of combination nature of race, it is real. So that would be an example of a biological realism view.

And then you have Naomi Zack, which to probably the surprise of lots of social scientists and humanities professors, comes in and defends the social constructionist view, which I would like to put under a larger umbrella of non-biological realism. So rejecting that race has a biological essence and affirming that race is nonetheless real.

So the three that I think literature supports is biological realism (race is a biological essence and is real), biological anti-realism (race is a biological essence and is not real), and non-biological realism (race has a non-biological essence and it's real). Now, people who are tidy with their logic will notice that there's another possibility here: non-biological anti-realism (race doesn't have a biological essence and it's not real). That hasn't been instantiated in the philosophy literature so far, as far as I can tell. So if there are any ambitious students or budding philosophers listening, that's a potential contribution you can make to the literature.

Sophie: I imagine some of our listeners might find surprising the category that says it's a biological thing and it's not real. Would a good analogy here be something like: you could give an account of what unicorns are, but then conclude that given that that's what they are, there aren't any?

Quayshawn: Um, close. Technically the claim is made about race itself, not just any particular races because you could say race is biological and real, even though right now we don't have any particular biological races. Just like in chemistry—I was a chemistry major in undergrad and a philosopher of science in general, as well as a philosopher of race—we talk about all sorts of elements on a periodic table that currently have no instances.

So Tennessine, one of my favorite chemical elements, element 117, because I was raised in Tennessee, and they discovered that element at Oak Ridge, Tennessee, has exactly 117 protons in its nucleus. That's its sort of essence as a chemical element. It was discovered in the 2010s, and the way that they discovered that that kind existed was that they were able to synthesize three isotopes of that, those just sort of variants of that sort of element.

Now they don't think that nature has ever done that. It's not the sort of atom that pops up in your regular nuclear fission and fusion reactions. So they artificially synthesize these. The fact that we're able to do it gives us adequate evidence that the element exists, but after they synthesize them for a couple of nanoseconds, they were out of existence. It's a very unstable element. And so currently it's plausible to say there aren't any Tennessine atoms floating around in the universe anywhere, but we think that the kind Tennessine still exists.

So a similar move could be made in race theory where you might think that, well, race given its essence is real, but maybe there might not be any particular races that exist right now. It's biologically real, or it's non-biologically real, but there might not be any particular instances. So it's really a claim about race itself, at least with respect to how I divvied up the field.

But you could sort of give a good case for a particular race theory—like if you were a non-biological realist, a standard social constructionist—and say, well, look, I have found that these things that we're calling races, these Black people, Asian people, and so forth, have non-biological essences and are real. So there is this thing race that they're instantiating, this kind of race. So that's a move that you could make to defend that type of race theory, but the categorization itself is about race itself.

Sophie: Okay. So then the analogous biological anti-realist view would be there isn't even any such thing as that element. Not just that it's not instantiated.

Quayshawn: That's right. Absolutely. You can say, well, the term race picks out something with a biological essence, but that thing doesn't exist. So you can think of other examples in the history of biology, like Darwin's "gemmule." Darwin is very well known for his theory of evolution and natural selection, less well known for his many writings on theory of heredity. But he needed a theory of heredity in order to have a plausible reason for thinking that natural selection was true, because you have to inherit the traits that are giving you some kind of fitness advantage.

So he has a couple editions of a book where he lays out what is his own theory of heredity. I won't get into all the details about the theory, but the important thing to note about Darwin's theory is he calls the unit of heredity the "gemmule." And if we don't know anything else in biology, we know the gemmule doesn't exist, given what are supposed to be the central properties of a gemmule. It was actually one of Darwin's cousins that did the important experiments to convince us that there were no such things and there never will be. He was just flat out wrong about what the units of heredity were in animals and plants and so forth.

So that's more of the sort of analogy going there. So now you can kind of see how the biological anti-realist view works. Because it sounded like it's a little bit of a struggle figuring out how one could get a defense of that view off the ground. The defenses in the literature historically have been kind of like what happened to Darwin's gemmule. Because of what's packed in the essence of race, it's biological, but it's a really hard standard to fulfill. The thing doesn't exist.

And you can think about, well, okay, so what's this stuff? You can think about the 1700s to 1800s and the kind of vile biological stuff that they were coming up with with respect to what race was. And this sort of moves that Appiah and some others like today, the biggest defenders, Adam Hochman from Australia, they say, no, no, no, that's what we're talking about. That is what we're talking about. And it just, as a consequence, that thing doesn't exist. So that's kind of the move that the biological anti-realists make.

Leah: Got it. That makes a lot of sense. So one thing that is apparent in your work, and especially in your coauthored book, "What is Race? Four Philosophical Views," is that different philosophers of race understand themselves to be engaged in very different projects. For instance, some take themselves to be analyzing our race discourse, that is, trying to figure out what we mean when we use the term race in different contexts.

Others take themselves to be defining race in such a way that it can play certain functional roles, for example, so it can play a certain role in scientific theories and sociological theories or in ethical theories of domination, oppression, and colonialism. Where do you stand in terms of your methodology? That is, what are you trying to do when you do the philosophy of race?

Quayshawn: Good. So I can give some terms because if philosophers are good for anything else, it's making distinctions. Josh Glasgow calls the one about defining race a kind of normative question. Can you try to do a normative metaphysics of race? Other terms thrown out are ameliorative or ameliorative race theory. Sally Haslanger came up with that language.

And what you're trying to do there is kind of like the project of when scientists are stipulating a definition of a term in order to get certain kind of work done in the field. And this happens all the time in the natural and social sciences, sometimes to the great confusion of ordinary people. So "work" in thermodynamics is not what we mean by work in ordinary discourse. Even terms like "fruit" in botany is quite different from what we are talking about. Green beans are fruit for botanists, for example. It's a very interesting, attractive, important project in philosophy of race. One that, you know, famously Sally Haslanger has been known for kind of jumping off.

But there's a bit of a tradition that goes a little bit longer back to Du Bois. Part of a problem with that question is some folks really see themselves to be doing both a normative metaphysics of race and more of, I guess, what we can call descriptive metaphysics of race project. And Du Bois is kind of like that. He kind of saw himself to be doing both. People like T.K. Jeffers, he sees himself to be doing both.

And the majority of the lion's share of my research, I see myself as doing the descriptive. Now, I have recently taken an interest in the normative for bioethics and medical research, and I'm working with one of my former bioethics students to make a contribution there specifically for the problem of addressing racial health disparities. But the bulk of my research has been on the descriptive question: What is race? Does race exist? If race exists, how does it exist? Is it biologically real? Is it real as a non-biological entity and so forth?

Sophie: Okay. It was striking to me that in the Stanford Encyclopedia of Philosophy, which is a reputable and widely used resource for academic philosophers, there are actually two separate entries about philosophy of race. One called "Race" and one called "The Critical Philosophy of Race." And it seems as though critical philosophers of race sometimes take themselves to be operating within an entirely different subfield of philosophy.

So we were wondering, what are the main similarities and differences between the projects that, say, critical philosophers of race are engaged in, and maybe the people who are doing the more descriptive project or maybe even the descriptive-normative mixed project?

Quayshawn: I've heard that term as well. I'm not the best person to ask about critical philosophy of race. My understanding is, they're a descendant of critical race theorists. I can say a bit about critical race theory. The critical race theorists, you know, they started in law schools and something that's very distinctive about them with respect to what descriptive philosophers of race working in metaphysics of race are doing is they assume at the start that race is real and it has a non-biological essence. So that's a starting point for critical philosophers of race.

And then you answer more interesting questions that you face after making certain stipulations. I mean, from their point of view, it's not a stipulation, they consider that settled fact. And so like investigating how race interacts with other important social categories, like sex and gender identity, things like that. So intersectionality is a big thing for critical race theorists. And then also exploring the depths of racism and how deeply interwoven it is in American society.

So a lot of social scientists are critical race theorists because there's so much work to do to try to untangle all the ways in which race—understood in that way—is intertwined with our well-being and our ability to live good lives.

So that's a project that is very prominent in the humanities outside of philosophy, as well as the social sciences in academia. It's also fairly prominent in philosophy. But if you're talking about metaphysicians of race, like people who I'm arguing with, Sally Haslanger has argued with and so forth, it's surprisingly less prominent.

So I was at a New Waves of Metaphysics of Race conference not too long ago. And if you just looked at the presenters, about half of them were representing the non-biological realist camp, predominantly, not entirely, traditional social constructionists. And so they would be comfortable positioning themselves as critical race theorists, although I don't know for sure that all of them would, but then the other half weren't, and it was about an even split between the anti-realists and the biological realists, who were representing the rest of the conference.

Now that's an unusual conference on race theory outside of philosophy. If you go to a conference on race theory in sociology, psychology, they're all critical race theorists for the most part. And they're all assuming that race is a non-biological existence, typically as some kind of sociopolitical construct, and they're asking other questions, doing other stuff.

So that's an important distinction between critical race theory and philosophy of race because it's not sort of settled in philosophy of race what race is and whether it exists, which put me in a slightly uncomfortable position when Republicans decided to, you know, go on an all-out attack on critical race theorists. Because I was getting all these sympathy messages: "Oh, I'm sorry they're like going after your field." And I'm like, that's not what I do. I attack critical race theorists, too—not in the same way, like in a fair way. But out of solidarity it's like, okay, because they were getting some unfair attacks.

Leah: And how much of the debate in the philosophy of race is driven by disagreement in what metaphysicians of race are trying to do?

Quayshawn: Good. So that's one of the sort of meta issues we debate about—how much metaphysics of race you need to do to do good philosophy of race, period. The metaphysicians of race tend to think you need to have some appreciation of what's going on and take some stance, or some family of stances or something like that. But there are people who disagree, who are doing more standard social, political philosophy, normative ethics, maybe bioethics, something like that. And they're not particularly concerned with what metaphysicists of race are debating about.

Now, as you can imagine, I kind of lean more on the first end, for specific reasons though. So it's going to be really hard not to make some ontological commitments when you're trying to defend some of the major positions in the more normative parts of philosophy of race. Take one of our most popular debates, affirmative action debates or reparations debates. A lot of times they're connected. You will see philosophers making some pretty strong ontological assumptions about race when they're motivating, you know, whether they're for or against affirmative action or reparations or some particular flavor of that.

For example, if you have a diversity rationale for racial preferences in hiring or college admissions, that assumes that you can't just change your race. Because you could say, well, we don't have enough, you know, Black faculty in this department, Joe, you're going to be Black tomorrow. That could be an easy solution if it were metaphysically possible. That's not entirely a silly thing to be concerned about today because we do have a rising phenomenon of—this is a bit of a problematic term, but—transracialism. Philosopher Rebecca Tuvel wrote a paper a couple of years ago, kind of putting some support behind people who self-identify as one race that's different from how they used to.

And in this tradition, a lot of times you're thinking of the self-identification as constituting the racial membership. It's social construction like, well, the way that you're self-identifying is giving your racial membership, in a similar way that, you know, self-identifying might give your status as a man or woman, something like that.

Now, if that was the case, we could have a lot of racial diversity tomorrow in philosophy, chemistry, all kinds of academic departments. But of course, there's been some pushback about maybe some limitations, some metaphysical limitations to our normative options for various moves that we might want to make in those normative debates.

So if it were the case that, say, ancestry was a necessary, a crucial part of your racial membership, that would be a biological constraint that would put some kind of biological property in the essence, so you'd be on the biological side of the race debates. It could be a bio-social view, but you're taking a stance, which would be quite different from a non-biological view, say, constituted by self-identification, or maybe a more popular one, Sally Haslanger's view of your social position.

But hey, we can get diversity with social position. You could trick people into thinking that you should be treated a certain way. They get treated that way. Isn't that what Rachel Dolezal did for a substantial portion of her life? Are you thereby improving the diversity of Black professors by having those Rachel Dolezals in your department before she was discovered as sort of putting on a show?

So these are important but difficult questions that impinge on normative questions. They're non-trivial, especially with that example, because it's such a highly selective enterprise. So when we're giving preferential treatment to someone because of their demographic membership, you might just make one hire. You might just hire one more professor. You might just hire one more computer programmer. And you know, you might want to know if you're going to be giving preferences to that demographic group, what constitutes membership.

Sophie: We want to transition now to talking about your views. And one thing that we were struck by in reading your work was that your views have changed over time. And we have a bunch of questions about that, but first, our understanding is that at one point you wanted to become a physician scientist and somehow ended up becoming a philosopher. And, you know, as an MD-PhD and a philosopher, we're kind of curious about this. So how and why did your intellectual interests evolve the way that they did?

Quayshawn: Right. You know, to the great satisfaction of my mother, who—there were a lot of people waiting for me to be an MD. So it starts for me in high school. I went to a specialist high school for science and engineering, health science and engineering to be exact. And it was in North Nashville, which is historically the African American side of Nashville, where all the HBCUs are. So Fisk was—I could walk to Fisk in a minute, two minutes, not far. TSU was around the corner. Meharry Medical College was around the corner.

I had an interest in biology. I took AP biology. Everybody had to do a science fair project. And I wasn't able to do very in-depth, impressive ones, though, because I didn't have any good connections. A lot of the students, you know, their father was a professor at Vandy or something. They were getting in labs and so forth. That wasn't my situation.

So my bio teacher suggested, hey, you know, I have a friend that's a, I think it was a doctor of veterinary medicine at Meharry. You could do a pretty cool project with him. You could get in his lab. I think he would be receptive. So I contact him, I get in. He was receptive. It was a very fun experience. I loved the life of doing laboratory research.

And I kept moving my way through Meharry. The next year I did a science fair project in a lab of a microbiologist. And my senior year, it was a lab of a biochemist, Samuel Dunyan. And that was—I wanted to be a professor after that experience. He gave me some advice though. He said, if you're going to do it, you probably want to get an MD because they're going to treat you better as an MD. And so that was the plan.

That was the plan strong through junior year, took the MCATs, did internships at Vanderbilt, they were expecting my application. But then I took a philosophy of science course that changed the entire trajectory of my career ambitions. I went to undergrad at Cornell, the professor was Richard Boyd. He was a philosopher of science who was very well known for his contributions to defending a view called scientific realism. But he was expanding his impact by looking at natural kind theories and looking at the interplay of social issues in philosophy of science.

So he offered a course that was particularly interesting. It was basically focused on the limitations of science, based on sort of social biases and social issues that can be hindrances to getting objective knowledge. And they focused on in the course "The Bell Curve."

So that book, probably young people don't know how influential, how much of a boogeyman that book was to many academics, throwing out a hypothesis that tries to give a comprehensive explanation of lots of different test score gaps between lots of different demographic groups. But a lot of the attention focused on racial differences in test scores.

And, you know, this was the issue that educational scientists deal with all the time. It was an issue that was interesting to me because that book came out in 1994 when I was a sophomore in high school, and I kind of saw it up front. Even though this was a school that you went to if and only if you were sort of high performing in science and math, there was still a gap between whites and blacks in the school. And I thought that was weird.

The book came out, gave an explanation. Part of the explanation is environment, but a non-trivial part is genetics. And I said, huh, that's interesting. What does that mean? It was sort of a—I didn't know at the time, but it was a Du Boisian moment because Du Bois starts off his classic paper by motivating why he's doing race theory. And basically the motivation is, well, white supremacy, white folks think that they're superior to us with respect to morality, with respect to intelligence, all kinds of things. And so this is basically supporting white supremacy in the narrow terrain of intelligence. And it was doing it on my turf, biology, so I thought it was interesting.

Didn't really have the philosophical chops to appropriately digest it and criticize it. But here was this course that was just laying it all out, and we're going to get to really kind of go through it. And we looked at the book through many lenses: statistics, psychology, educational science. But what was really fascinating to me was when we got to the philosophers, and particularly papers by people like Ned Block.

And some people were surprised that Ned Block did this, but he has a very influential response to "The Bell Curve" from a philosopher of mind's perspective. And just basically pulling out the rug, like asking really deep foundational questions about how to do good science in the first place.

And I remember that Boyd ended the course by asking what I thought was a tremendously profound and interesting question. So Block asks these questions about what is intelligence really, and how can we really reliably measure it? That was fascinating, but Boyd asked this question like, well, what is race really? And how do we know that races are the sorts of things that can lend themselves to having intelligence? Legitimate genetic differences among them?

So you can classify the world in lots of different ways, but in certain sciences, only some of those ways are going to give rise to non-accidental law-like generalizations. Example, if you were to study the distribution of sickle cell disease in the American population, you could stratify your sample by political party. You could do that, and you will find a statistically significant difference in the rate of sickle cell disease among Democrats and Republicans. In particular, you'll find that a lot of our incidences of sickle cell disease are going to be in the Democratic Party.

Now is that an accidental sort of induction that you've arrived at? Or is that underpinned by something that's more stable and not an accident? It seems like an accident because if you understand what sickle cell disease is, it's a genetic disease. It's caused by a Mendelian sort of underlying genotype. It's very popular among people of African descent for various reasons tied to their environment.

And because of how people got to the United States in general, definitely over 80 percent of sickle cell disease prevalence in the United States is among Africans. Couple that with most of the party registration for African Americans is Democrat, and now you have this connection to sickle cell disease in the Democratic Party. But of course, there's nothing about belonging to the Democratic Party that's causally involved in getting you sickle cell disease. You would be headed down a dead-end scientific research if you were relying upon inductions like that to do your science.

So I thought that was a very interesting insight. It's like, well, you might get statistically significant differences between racial groups in terms of all kinds of test scores and whatnot. Why would that be a sufficient reason for thinking that there's something genetic going on? I thought that question should be really explored well.

The sorts of answers we got were not as in-depth as I wanted. We did get exposed to some of the social structures and so forth. I didn't think that they were careful enough with the genetics, and so I wanted to give it a go. I wanted to see what was there, what really is race. And that will help me answer the other question about testing differences, educational achievement gaps that kind of drew me to this. So that was the experience that changed the entire trajectory of my career ambitions.

Leah: And one of your early stances on what race was biological anti-realism like we talked about earlier. And our impression is that you were in the midst of writing a book defending that view when you changed your mind about this. What made you hold that view initially, and why did you change your mind?

Quayshawn: I got a Ford Foundation fellowship to write a book at Stanford, and the book was called "No Space for Race, The Philosophical Foundations of Biological Racial Antirealism." It wasn't racial anti-realism, it wasn't a flat-out anti-realist view, it was just in biology. So that's kind of a niche position that is compatible with multiple larger positions. That was what I was comfortable defending at the time, because I had done enough study of biology. It's like, look, can't find any defense for this. I'm gonna write this book.

So I go to Stanford. I don't actually interact with the philosophers as much as I had intended to. But I do interact with the biologists a whole bunch. And that was because there was a population geneticist who was interested in social issues, Noah Rosenberg, who had written some very interesting work that's relevant to what I was studying, who was teaching a class on advanced genetics, human population genetics.

And I go in the course and, you know, do all the readings, participate in the discussions. And I flipped my position. Basically what I had realized was the parts of biology I was looking at were not giving a good basis for any kind of legitimate racial classification in biology. And for some background there, I did my master's in biology in systematic biology, which is taxonomy. And traditionally that is a huge place where biologists have tried to do taxonomy since Darwin, actually.

Darwin was the one, in "Descent of Man" in the 1870s, who said, hey, let's start getting precise here. Stop using race in any kind of way. Let's stipulate that races are subspecies and then try to do biological theory of what subspecies are. Since then, there's been a lot of research on race in systematic biology, because people were seeing races as subspecies.

So I was looking there, I was looking at taxonomy and saying, hey, you know, in the way that we're doing taxonomy today, is there any legitimate basis for having a racial distinction in your list of taxa, like your species, order, genre, and so forth, and then race? Should race be there? And I was like, nah, it doesn't look very defensible.

But population genetics is a distinct autonomous subfield of biology, and population analysis have also been interested in the race question since very far back in the early 1900s, when population genetics was getting up and running. One of the sort of pioneers, Theodosius Dobzhansky, has written lots of papers on race, and there was always a question about whether there's some sort of racial distinction among human populations. Is there a level, a racial level of human population structure? So that was kind of how I flipped my mind. I basically saw that I wasn't looking at the whole of biology and there was a space in biology where one might be able to pull this off.

Sophie: So after that experience, it sounds like you started defending a form of biological realism. In 2014, for example, you defended a view according to which race terms pick out biologically real entities, specifically the real entities that population geneticists know as human continental populations. Could you start just by explaining what a human continental population is and how population geneticists come to identify them?

Quayshawn: Oh, yeah. I'm going to do the short way. There's a short and a long way. I have a whole paper on how to sort of precisely get down the identity conditions. But very roughly, population geneticists kind of arose in the 20th century to extend the study of genetics and evolutionary theory.

So in traditional genetics, you're studying organisms and you're trying to find out principles of heredity and so forth from organisms. But what about how genetics works at a population level? You can have genetic phenomena that organisms can't have, like allele frequency. Populations can have allele frequencies, but you need a number of things with alleles to have allele frequencies.

So in evolutionary theory, one project is trying to figure out all the different evolutionary forces and how they impinge on bringing about evolution. But usually the unit that's undergoing evolution is the population or some higher level population, but usually populations. So in evolutionary theory, especially if you're undergoing natural selection, your theory of heredity is going to be connected. So bringing together genetics with evolutionary theory then becomes important. So how does genetics behave at the population level so we can explain how evolution occurs because it occurs at the population level? Typically.

So now that we have a better sense of the aim of population genetics—if you're going to study genetics at the population level, and how evolution works at the population level, you kind of need to know what populations are. There's this fundamental metaphysical question that population geneticists have to answer.

So there's a whole project in population genetics called population structure analysis. And it's a strategy to identify all the levels of populations that you have in a species or a genus, whatever you have. And of course, there are debates about what is a population, but there's at least some convergence on what a population is in some cases.

For sexually producing species, the lowest level is supposed to be in that context a deme or also called a panmictic unit. These are randomly mating groups of organisms, so literally you aren't able to find any more patterns in how mating is happening than at this level.

Now that level actually turns out to be quite high for some human populations. So last I looked, things could be slightly different from the 2010s, but last I looked, the Han Chinese is the largest panmictic population in the human species. Now think about how many people are in the Han. That's about 90, 91 percent of China, all of Taiwan, all of Hong Kong, comprises a huge chunk of the human species. One randomly mating group according to population genetics.

Now, you might also be interested in, in addition to how low you can go, how high can you go? And ultimately, you want to end up at a species. So for that, usually they don't start at the bottom, they start at the top. So you start with the species, see if you can divide that species into two, see if you can divide it into three, see if you can divide it into four. And the lowest you're supposed to be able to go is local populations or panmictic units, whatever you want to call it.

In 2002, we've been doing this with some special software developed at Stanford. And basically what you're doing is, you have a model about how allele frequency differences are being separated over time from evolutionary forces, like drift and so forth. You build that into your model, and you get a whole bunch of genetic data. By genetic, they're just talking about sequences of nucleotides. And you sort it, you sort them into bins of loci, and try to detect genetic patterns that are sort of telling a story about where these originated.

And you might not be able to divide up a species into two or three or four. In fact, there are certain species that don't have any population substructure. I think the cuttlefish might be one. It's just one huge randomly mating group of organisms. That's the species.

So it's an interesting empirical question: what kind of population substructure do you have? So now you can pick a species. In 2002, Rosenberg et al. pick Homo sapiens and unleash this software and was able to detect various levels of population structure. And interestingly, one of those levels looked very similar to a popularly used racial classification in the contemporary United States. And now we're off to the races.

Leah: Can you say more about that?

Quayshawn: Yeah, so we could get pretty reliable breaks at various levels. But at the level that's called K equals 5, it's just breaking up humans into 5 large populations. It looks like it's a pretty continental level.

So here are the groups: Sub-Saharan Africans and their descendants, African Americans, Afro-Caribbeans, and so forth; a group that I'll just use the language in the literature, "Caucasian" is a term that's sometimes used—these are North Africans, West Asians, Europeans, and to some extent South Asians, although they're kind of mixed between these divisions; East Asians, and to a small extent South Asians are in there, but mostly that group is composed of sort of your Northeast and Southeast Asians, folks like Filipinos and Chinese; Native Americans, Indigenous people of the Americas; and Pacific Islanders, or Oceanians sometimes is used because that group is not just the indigenous people of the Pacific Islands like Melanesians, Micronesians, and Polynesians, but also the Aboriginal Australians are in that group.

That's five. And those groups have been, for shorthand, called the continental populations in the human population genetics literature and associated literature like medical genetics and so forth. And population geneticists themselves in the early days, early 2000s, when they found this level of human population structure, said, "Hey, this looks very close, if not the same as a level of racial classification that we are forced to wrestle with because we're taking NIH funding."

So this is the census, or more formally, it traces back to the Office of Management and Budget, a decision they made in 1997. We call them the 1997 races. And the legal names for those groups for the federal government are Asians, whites, blacks, Pacific Islanders, and American Indians.

So it seemed to be a one-to-one correspondence. Might be some debate about the edges. Where do South Asians go? That issue kind of gets settled in virtue of the sorts of populations that population geneticists are talking about. They're not all-or-nothing membership populations. They have graded membership. So the way that they talk about populations in contemporary population genetics is membership isn't limited to you're either in a population or you're not—you can be in a population to a degree of membership.

Which made those groups look even more like what the OMB was talking about because since 1997, they allowed multiracialism. Before that they didn't—you had to pick. I remember those days. I was in high school and the kids with the black parent, white parent, I didn't envy their choices, because in my school, for some reason, we had to actually announce our choice at the beginning of the school year. So the whole class was listening. There was like only a handful of mixed kids in that respect. And we were—you know, the black folks were cheering, "come on, come on, this year, come with us." But the 1990s were interesting times.

But now you could be multiracial for that sort of folk racial classification, and you can have more than one population membership for the modern way of thinking about biological populations. So the more you look at the two, they're like, huh, these two look really, really similar.

And population geneticists, like my colleague here, Sarah Tishkoff, she has a paper where she says these look a lot like the federal government's racial classification. In that paper where this kind of population level was first discovered, they also kind of touched upon this issue, although they never use the term race. They never say race, but it's pretty clear that they're trying to sort of anticipate how some of the results are going to be read and sort of deal with the inevitable issue about whether they discovered a racial classification.

And then there were other population geneticists that made remarks to the effect of, not only do these groups look similar, they're the same groups. Or they're so similar, we shouldn't consider there to be a significant difference, something to that effect. And they were trying to promote using these groups in medicine, which we'll probably get to later. But initially it was the human population geneticists that were making this comparison, and I thought it was an interesting metaphysical question for philosophers.

Leah: And did you agree with that comparison?

Quayshawn: So I thought you had to have some kind of theory as to why there was such a connection. That's not an accident. That's not a statistical accident. That is highly unlikely to occur by accident.

I mean, think about this. Population geneticists didn't just stop with eyeballing and saying, well, these look similar. They did quantitative analyses. There was a joint study between some sociologists and some population geneticists who looked at a nationally representative sample of American college students because they had their self-reported race information.

And then they took that sample, and they used the software to find out how these individuals clustered with respect to the human continental populations. They were able to blind themselves to the racial self-reports and predict the majority membership of the human continental population with over 95 percent accuracy. And they're able to blind themselves to the genetic population, use the self-reports and predict majority ancestry with over 95 percent accuracy. That's not an accident. That's interesting. Somebody needs to explain that.

So I thought there needed to be some theoretical account of why we have this one-to-one mapping between the 1997 OMB races and the continental population level of human population structure. Now, interestingly, there's lots of different theories that philosophers eventually put forth to account for it. My favorite theory was identity—we have discovered what the federal government has been talking about all along. That was my preferred theory, but there were other theories. But at the very least, you can't just say that's just an accident that just fell out of the research.

You can have all kinds of complicated sort of theories for what's going on. I just happened to think that that information was telling us we're talking about the same groups.

Sophie: Could you maybe give us an example of a different theory that doesn't say that the correspondence is identity?

Quayshawn: So for example, Ásta, who is now, I believe, a professor at Duke. She was a student of Sally Haslanger as well. She has developed an interesting alternative view on what that correspondence is. Basically roughly, she thinks that the OMB races themselves, the essence of what they are is a conferred essence. It's a socially constructed conferred essence where those are special kinds of essences that you get to be a certain sort of thing from a status that's conferred onto you by an appropriate authority.

And an interesting trick about social kinds with conferred essences is that they have this relation called tracking. And sometimes it's tracking a biological thing, maybe a physical thing, but there's some kind of target outside of the social world, so to speak, that's being tracked. But it's not constituting any of the essence.

The example she uses in her classic book, where she introduces this stuff, is a strike in baseball. What is a strike? She says people who watch baseball would love it to be a certain physical trajectory, right? That would be great if that were the case. But she says that's not what a strike is. There's no sort of physical trajectory you can give that would constitute all and only strikes. What a strike is, is what the umpire confers on a physical trajectory. Once the umpire says it's a strike in the context of an official baseball game, following all the rules, then it's a strike. And that's just what it is. And that's why we have to eat those bad calls by umpires who are out of pocket.

So with respect to OMB races, she explicitly says in chapter five of that book, we're tracking some kind of biological population in terms of geographic ancestry, something like that. But essentially what it is to be Asian or white or black is that the OMB has conferred that status onto you. And there's an interesting difference between me and her with respect to accounting for some of those correspondences.

So take the correspondence between OMB's American Indian and the human continental population of Native Americans. I would say that's the same group. She says American Indians are tracking that group, but they're different. And how they're different is you don't get to be an American Indian without being a member of a federally recognized tribe. That's what she says.

So there are some people who have as much Native American ancestry as you could possibly have, but not be a member of a federally recognized tribe, and therefore they're not American Indian. Similarly, there are people who are members of federally recognized tribes—maybe they marry into it, or every tribe has their own constitution and rules for membership. Maybe there's some other way of becoming a member.

Actually, there are some tribes that, in virtue of giving reparations to African Americans that they enslaved, or descendants of African Americans that they enslaved (there were some tribes that enslaved African Americans), some of them have granted the descendants membership in the tribe. So those are people that don't have Native American ancestry. So there you have a mismatch.

And she would say, "Aha, see, that's why my theory works." It accounts for the correspondence metaphysically, but the way in which it does it is not an identity theory because there are more expansive, completely social ways of being American Indian that separate themselves from having a certain kind of ancestry.

Sophie: Got it. Okay. So let's jump into more of the details of the view that you came to at that time. The view you favored then, when you wrote your 2014 paper, was, as we understand it, focused on identifying the meaning of race in US racial discourse. You called this the "national meaning" of race terms in the United States. The idea being that just as a country can have a national language, it can have national ways of using terms or national meanings for terms. And you were looking for the US national meanings for race terms.

You were engaged then in what you called earlier the descriptive project in the metaphysics of race, where the question is "What is race?" and the methodology is to consider how ordinary folks use race language. As we understand it, your conclusion was that race is biologically real, and your argument sort of had two parts. First, you argued that the national meaning of race is embodied in the official classification scheme of the OMB—the scheme we've been talking about of black, white, American Indian, Asian, and Pacific Islander. In other words, the term "race" as used in US national racial discourse, on your view at the time, picks out a set of human continental populations. And then second, you argued that that entity, the set of human continental populations, is biologically real.

Assuming that we have this right, we wanted to ask about both sort of parts of that argument. 

 

Leah: So first, regarding your claim that Americans have adopted the OMB categories as their national racial discourse, it seems like one of your reasons for thinking that is that Americans know how to sort themselves and others into the OMB categories, as you mentioned before. But as Sally Haslanger points out in her response to your chapter in the book, this is kind of a long quote, but we liked it. 

She says, quote, "Spencer relies on the broad use of the census and other legal documents to obtain information about race to support the claim that most Americans 'defer to the state and in particular, the OMB to determine what race is.' However, Americans are required by law to fill out forms in which they designate their races for the census, school, employment, et cetera, and are given a mandated set of options to choose from." 

She continues, "I also defer to astrologists in determining what astrological sign I am, and if I am asked to identify my sign, I give the answer based on my birth date, but my ability and willingness to do this says nothing about how I live my life, what matters to me, whether I take the classifications to be warranted. I take astrology to be ridiculous. If I were required by law to complete a census or other legal document that asks me to indicate my astrological sign, not based on my birth date, but on my self-identification, I would give an answer, but my answer would say nothing about what I mean, how I live, or to whom I grant authority. Those of us interested in racial and racializing practices see the OMB as part of the system that creates and enforces race as we know it, not as a source of expertise with respect to what race is." End quote. 

In other words, even if people are used to classifying themselves and others according to the OMB categories, this doesn't mean that they endorse those categories or have really taken them on as their own understanding of race. How would you respond to this kind of critique?

Quayshawn: Good. I'll start by saying, Sally Haslanger was my first professor for race theory. I was a graduate student at Tufts and I took a train to MIT to enroll in her class. It was a fond memory. So being in a position where we're co-authoring a book and she's taking my views as seriously as other experts is kind of a surreal experience for me. So forgive my starstruckness.

But I would not disagree with anything that she said there. Where me and Haslanger are going to disagree is in the relevance of that criticism to the project at hand. And this is going to get back to some meta stuff.

So Haslanger, as you know, straddles a couple of different parts of philosophy of race. Sometimes she's doing the normative stuff, and that has to do with the concerns she had in that criticism with respect to the importance and the significance of some particular race talk to one's life. And so she does have some work where she's trying to explore how we should talk about race, namely for the purposes of getting social justice. Wonderful project. All in support of that project. Not my project—at least certainly not my project in that book.

But there's another, of course, more descriptive side where we're trying to figure out what is race and given what it is, is it real and so forth. Now, there's still a little bit more context that needs to be given here because in that book, we're also not entirely agreeing on the rules of engagement, so to speak.

And I try to be as clear as I can in my chapter about the race talk that I'm interested in. I'm interested in contemporary ordinary American race talk. T.K. Jeffers is not just interested in that race talk in the book. He's Canadian. Why would he just be interested in that race talk? But he is interested in a race talk that's broad enough to incorporate that. He does give American examples. He's talking about Du Bois, who's an American, you know, he's in my territory. I'm not entirely in his.

And in that book, it's not entirely clear to me that Haslanger is just engaging in a descriptive race theory. And to the extent that she's engaging in at least a partially normative race theory, I think we're failing to disagree as opposed to disagreeing.

So I could be completely on board with: I don't particularly have any deep connection or feeling to being black in the sense that the OMB considers me to be black. And this might be surprisingly—people might say, "What about Black Power?" It's like, well, yes, but I actually have more social significance to being African American.

So there's a case to be made that there's lots of reasons to not be particularly deeply embedded emotionally or psychologically in the sort of racial classification that the Office of Management and Budget has forced upon the lion's share of ordinary Americans. Be that as it may, my question is, what is the race talk in the United States in its dominant use? That's an additional sort of modifier that both me and Josh Glasgow are deploying in the book.

I'm deploying it because at the time I wrote that book, I was a pluralist already. In the seventh chapter, I'm sort of kind of easing people into the pluralism, and it's like, well, I used to think there was a single dominant use. That was my 2014 work. And that was the OMB's. Partly because of the sort of special power that the U.S. Government has to kind of control our race talk.

For example, "Hispanic"—they came up with that. That was not something that was kind of grassroots. The U.S. Government came up with Hispanic in, like, the 1970s. So, you know, if you want to get federal funding for your university, you're going to have to report to the Department of Education in the OMB's race talk. So now we have all the race talk spread all throughout the university.

If you care about affirmative action, now you're forced to use that race talk when you're debating about affirmative action. So before that book, I was sort of taken with how widely used that race talk was. Now at the time of the book I had changed my mind though, because I thought there was other race talk that was at least as widely used, namely race talk where we do talk about Hispanics/Latinos as a race. I think that's very widely used, especially in political discourse, when people were debating about how Latinos are going to vote for Trump or not in Florida or whatever.

So that's why I kind of gave up the non-pluralist approach, the monist approach. But I'm still focused on this descriptive project in the book. And so regardless of how psychologically connected people are or how they personally prefer to racially identify, this is at least a dominant race talk in the United States, very widely used in particularly formal contexts of racial discourse: applying for a job, applying for college, applying for food stamps, applying for—it's on FAFSA applications if you want to get federal funding for college and so forth.

That's the target I want to get a sort of tallying of—the dominant ways that we talk about race, develop a theory about that, regardless of whether people like it, whether they prefer to racially identify in that way or not. I just want to find out the facts.

Sophie: So we're curious about what kind of weight you give to the OMB's official statements about what their categories are. Because if you look at their sort of public-facing documentation, they'll say, quote, "The race and ethnicity categories set forth here are socio-political constructs and are not an attempt to define race and ethnicity biologically or genetically." Given that the OMB doesn't think that they're picking out biologically real entities, what are your reasons for thinking that they do?

Quayshawn: Well, first, we need to give some context to that quote. So that's from the 2024 release. So recently this year, the OMB has changed their official racial scheme. And this was a long time coming. The OMB usually does it every 20 years. They started in 1977, then 1997, they changed, and we were expecting something, you know, at least by 2017. But something interesting happened—Trump got elected.

And so the Clinton administration would have made the change earlier because the Obama administration already did studies showing that there were problems with the racial classification, specifically with respect to Hispanic Americans. It kind of fostered less people to answer the question and thereby you get less information in that demographic part of the survey, which is counterproductive.

Lots of Hispanic Americans didn't see themselves as having a correct place in the racial scheme. These are mostly mestizo individuals though, because about 25-30 percent of Hispanic Americans were marking white repeatedly across studies, and to a lesser extent, a small proportion were marking black, like some Dominicans, folks like that.

So they wanted to make the change, and Trump just wanted to do the opposite of whatever Obama had planned. I don't think there was any deep sort of reason there. And so we didn't make the change. And so we held on to the 1997 OMB racial scheme until, I guess, end of March of this year. So that quote is from the new racial-ethnic scheme.

Sophie: Racial-ethnic scheme—they explicitly don't name which groups are races and which groups are ethnicities. That's one way to try to address the Hispanic non-response problem. But if you look at the language used in the 1997 OMB directive, they don't say that. And in fact, they explicitly say that it's both social and biological in terms of ancestry.

Quayshawn: Okay, got it. That's really helpful.

Sophie: You spoke a bit earlier about your reasons for taking on a more pluralist view. And we wanted to ask why you never took on a pure form of social constructivism. I wanted to put this in terms of an exchange between you and Haslanger in the book. You mentioned earlier that you and folks like Sally Haslanger are engaged in different projects that seem compatible, and she writes that way as well.

So she says things like, quote, "It's compatible with my view that there are genetic ancestry groups corresponding to the OMB categories, but it's an empirical question whether using the term 'race' for these categories is, on balance, a politically wise thing to do in the long term or not." So I guess our question is, in your view, what do we lose out on if we reserve the term "race" for referring just to a sociopolitical thing and separately accept the existence of genetic ancestry groups, but without referring to the latter as races?

Quayshawn: That is an interesting question I haven't thought enough about, because that's more of a normative issue. I am starting to get into it a little bit, but in a different way, not so much about the language used, but whether you use the groups specifically in medical research, things like that. With respect to the language, I'm open-minded about, on balance, what's a wiser thing to do.

Actually, there's some scientific research that could be relevant here. So Brian Donovan is an educational sociologist who I actually bumped into in that course I mentioned at Stanford. And he was taking the course because he wanted to figure out ideal ways to design instructional material in life sciences.

And if you look at the textbooks in biology, various biological sciences, you'll find a lot of instances of the use of the term race and examples of groups called races, like in medical textbooks and so forth—population genetics textbooks, genetics textbooks, intro biology textbooks sometimes. And he wanted to know, is this having some kind of undesirable impact on the information that people were taking away, given what the subject matter is in that section?

And so he did some studies. And he found out that—I mean, he's very careful with his statistics, so he wouldn't make too broad of a generalization here. But to the extent that this generalization holds, given his sampling, what he found was: if individuals who were reading a portion of a text that had to do with discussing genetic diseases, and that lesson was being presented in such a way that was just using race terms, not even making any substantive claims about race, any of the nasty stuff we might be worried about, just saying things like, "oh, blacks tend to have higher incidence of sickle cell disease." That's enough with respect to what we're talking about.

That ended up being statistically significantly and non-trivially connected to those individuals developing what we might call more concerning beliefs about race, in particular beliefs about people having racial essences in a way that has been long since rejected. Like, well, I guess if blacks have higher incidence of sickle cell disease, maybe they're better at basketball, you know, like they start making these connections that we would be far more concerned about.

But this happens statistically only for people who didn't have a certain basic level of genetics education. The effect didn't occur at all statistically for participants who did. So there's different ways you can go with that. You could say, well, maybe we shouldn't use race when we talk about genetic diseases. Or maybe we should educate people better about genetics.

So I'm undecided about what I think about the language issue, because of trade-offs like that. It's like, well, I'm an educator. I would think educating people to the point where they wouldn't be making those connections would be ideal. But you know, there could be complexities that makes that not practical or something like that, that then would impinge on my downstream belief about what to do there.

Leah: Okay. Well, we've kind of already gotten into implications, but we want to reserve the remainder of the interview to talk about some of the implications of your views for how we do things in science and medicine. We'll focus specifically on issues related to racial health inequities, and we wanted to start by examining the very broad causes for these inequities before zooming in on the more proximal ones.

So to start very broad, the consensus view in science and medicine in 2024 is that institutional racism or systemic racism is the fundamental driver of racial health inequities. Do you agree with this?

Quayshawn: That's the sort of slogan that philosophers in general don't like. If you ask me, am I pro-life or pro-choice? It's like, well, I want to have a little more subtlety. And here's why I'm a little skeptical about committing myself either way. I don't know how many racial inequities there are. And so I wouldn't be able to give an assessment one way or another about the extent to which getting rid of that is going to be fixed by tweaking with one variable versus another. And this is part of why I've gotten interested in this, because I think one injustice—I think everyone would agree with—is if there's a racial disparity that's unjust that we can't see because we're not able to see it because of the racial classification we're using.

I'll give a political example to make the case, an analogy. One reason that the OMB changed the racial-slash-ethnic now classification was because of pressure that Arab Americans put on them. Arab Americans were unquestionably classified as white in the 1997 classification. Now, because of obligations that various federal agencies have, like the Department of Justice, HUD—you know, Housing Urban Development—one way in which the OMB categories were required to be used was to catch racial discrimination in mortgage lending. There's laws about it. They have to surveil this stuff. And then there's an order that the specific racial classification that should be used on these mortgage lending applications. And if an applicant doesn't put that in, the lender has to put it in.

Now, that's going to do a lot of good in terms of catching anti-black, because of how things are set up, anti-Hispanic or Latino, and so forth, discrimination. Not going to help Arab Americans that much. You can't see it. If people were discriminating against Arab Americans because they're Arab, you're not going to see it because they're marked as white. And when you're looking for disparities, it's not going to pop up. You can't see it.

So one concern that I have about the medical context is that limiting the way that we're racially classifying is literally making us unable to see unjust racial health disparities. And I don't think that there's a single racial classification that can have us see all of them. That's what's pushing me towards a normative pluralism in that context. And so if we're articulating all the ways that we're recognizing race, it's relevant to the health context. Then we're able to actually get a total counting of the unjust racial disparities. So I don't know how many there are. And if we don't know how many there are, we certainly can't say with any sort of a high degree of confidence that there's one or other causal factor that's mostly responsible for them. So that's kind of my approach to answering that question.

Sophie: Okay, so we want to talk a bit about the implications of your views for the growing push to diversify clinical trials. So there are numerous social reasons for calling to diversify clinical trials. For instance, work by Marcella Alsan and colleagues has shown that when racial groups are underrepresented in clinical trials, doctors are less willing to prescribe the treatments that emerge from those trials to members of those groups, and patients are in turn less willing to take them. So ensuring diversity in clinical trials is important for building trust and ensuring uptake of therapies that emerged from research. But we take it that one place where you might disagree with some other theorists is that you might think it's also important to achieve racial diversity in clinical trials for biological reasons. Is that a correct impression? And if so, why do you hold that view?

Quayshawn: Yes, right. That is great research. That was my experience during the COVID-19 pandemic. I couldn't get my family for the life of me to take the vaccine because they didn't have anywhere near enough racially diverse sampling to convince them that they wouldn't be the guinea pig, right? At least for this portion of the human species, for black folks. And different pharmaceutical companies and research groups were better at that than others, but it definitely was leading to vaccine hesitancy in my household and all black friends that I had. But I was looking over at my white colleagues and they were fighting—their families were fighting to get the vaccine, and I was trying to convince them that it was safe and effective.

Yeah, so that's absolutely a concern that drives me. We're all paying taxes. Why don't we all have the same degree of confidence and trust in how our tax money is being used to create pharmaceuticals for us, regardless of race, ethnicity? So that's definitely one motivating interest of mine in that realm.

And so far, absolutely, I'm on board with stratifying your sample in any way that you have a theoretical reason for thinking is relevant for stratification for the causal structures that are at play. And this is not a huge paradigm shifting—I mean, it is to some extent because of how hesitant folks are in the medical world to, you know, go in that direction. But, you know, we do this with sex. I can't tell you how many forms I've filled out in the healthcare context where they ask for your legal sex, your gender identity.

So we're understanding the sex variable in a pluralist way. When it comes to race, one answer. That's it. And what I want to promote is, well, no, if we could do that with, you know, we're getting all this granular data on all these other, hopefully causally relevant variables, why not race? There's more than one way, and this is of course relying on metaphysics a little bit, there's more than one way that we're classifying people by race, why aren't we classifying people more than one way when we're sampling them?

Leah: Okay. That makes sense. So we saw you give a talk where you mentioned the study that was done at St. Jude's where they found that among kids with ALL, a kind of cancer, having greater than 10 percent Native American ancestry was associated with an increased risk of relapsing on the drug. And it seemed like the thing that mattered in that study is whether the patient had the specific alleles or genetic changes that caused them to metabolize the drug differently. And while having some of these alleles may be correlated with race, as we talked about earlier, maybe having sickle cell disease is correlated with your political party affiliation.

Quayshawn: Race isn't the thing that causes you to metabolize the drug differently. The genetic changes. It's what causes you to metabolize the drug differently. So just to kind of like get clear on, on your view is, is the idea basically that having diverse ancestry in clinical trials matters for biological reasons and having racial diversity in clinical trials matters for social reasons, but we don't have to be using these terms like interchangeably here.

You made a turn there. I wasn't expecting to make. But I can answer both what I thought you were going to say and then... So I thought you were getting at the issue of cause in this biomedical race debate stuff. There's a divide about, if we should use some kind of racial classification, how should we use it? Research, treatment, making diagnoses, right, education in medical. And I want to make a very narrow defense about research and even narrower about the type of these observational research.

Whenever you start an enterprise of science, you're basing it on some kind of inductive generalization, some kind of pattern recognition, right? You find some pattern. "Oh, that's interesting. What's the explanation? Is this going to be medically relevant?" Something like that. How you start off your search matters for what you're able to see.

So for example, those researchers in the ALL study, what they did was actually quite interesting. They had a hypothesis that Native American ancestry was relevant. Their sampling was not accidental. They had a hypothesis. In order to get enough statistical power to test the hypothesis, they added a lot of mestizo Hispanics who have a lot of Native American ancestry. They count for the federal government—count as Native Americans, even though various Hispanic folks might be uncomfortable or not, that might not be significant to them. In order to be able to find the difference, if the sampling wasn't high enough in that bin, so to speak, statistically you wouldn't see any difference between relapse rates after doing the standard chemotherapy for that particular cancer. I think everything's fine. You won't see any racial disparity.

They had to establish that a statistically significant racial disparity existed. And so they based that on a hypothesis that, well, you know, I don't know exactly how they came about it. Maybe some of them had some sort of experience in the clinic and had sort of a hunch. Let's beef up the sample. See if we find anything. Wasn't guaranteed. They would—things could have turned out differently, but they didn't.

And so then the question becomes, what do you do now? And notice in that talk that you saw, I didn't say anything about treating people differently based on race after you found out that fact about how the drug is metabolized differently. I would think the next best step would be, well, don't we need to sequence people's relevant genotypes before we recommended a particular type of chemotherapy? You know, that would be the next, you know, practically wise step.

Partly because one thing that they found in a study was there was a lot of people who would look white, would self-identify as white, who had significant unknown Native American ancestry. So they were at risk of not responding well to this, even though they wouldn't dream of identifying as Native American. So, but if you do have as a best practices norm—"Hey, before we do this chemotherapy, let's do a relevant, not a whole genome sequence, a relevant genotyping to see if this might be an issue"—and then we know which direction to go. So that's where I thought you were, because some people do have a concern about, well, does that mean I need to start racially stereotyping when I'm in the clinic? No, that doesn't follow from at least my defense of using biological racial classifications in medicine. Just the research phase and then at the sampling phase when you try to find patterns. Now, now I forget the direction you actually did go.

Leah: Yeah, I mean, the question, the question was something like, you know, research by Alsan and colleagues about trust maybe gives us reason to have like racially diverse clinical trial participation, because we can anticipate that groups that have experienced racism are going to trust the products of research unless they weren't well represented.

Quayshawn: Right.

Leah: And, on the flip side, these different alleles that can be more prevalent or less prevalent in different populations, perhaps gives us reason to ensure that populations in clinical trials have a diverse genetic ancestry, but these two things kind of come apart. Like we can say it's important to have racial diversity in clinical trials for social reasons. It's important to have diverse ancestry in clinical trials for biological reasons, but there's nothing there about it's important to have racially diverse trials for biological reasons.

Quayshawn: Good. So we would have a metaphysical disagreement and it might seem like a sort of small semantic thing, but that's where all the action is. So sometimes people say, "Well, we don't have to call them races. We could just sample East Asians and Sub-Saharan Africans and blah, blah, blah. Make sure we have good samples. We have to call them races." Fine. Don't. Those are still 1997 OMB races, according to my theory. You can call it whatever, you know, leads to best practices, but you're still racially sampling in a way that's distinct from another way of racially sampling, right?

Now, an interesting question that I'm still exploring is, are there other levels of population structure that might be more informative with respect to finding medically significant genetic level differences? That is not a question that we have a good handle on what the answer is, partly because we haven't studied the other levels of population substructure, if there are other levels besides the panmictic level, to anywhere near the rate at which we've studied K equals 5, for obvious reasons. We know K equals 5 exists. We know we subdivide into continental populations. It's not actually been established that beyond continental populations and the local population structure, the panmictic level, that we have any other population substructure. That hasn't been established. I am actually investigating that with some researchers, and hopefully I'll have an answer to that soon.

But in addition to that, even if we do, there's ways in which population geneticist measures sort of genetic significance of population levels. A technique called analysis of molecular variance. It's a way of apportioning how much of the total genetic variation exists at different ways of cutting up a certain. So say that you cut up a level K equals five, you have five groups. Well, you really have three dimensions there. You have the continental group level dimension. So there's variation that's among those groups genetically. You have all the little panmictic local populations within each of those. So you'd ask about what's the variation among all that, and then you have the individual organisms, individual people that are within those panmictic populations. And what we know for K equals five is that most of the genetic variation occurs among the individual people, like over 90 percent. And then the next highest is among the populations—five or so percent. And then a smallest amount is among the continental groups.

However, in the studies where they're doing this sort of analysis, they have looked at other levels. And in those other levels, it's usually the case that K equals 5 has the highest, among sort of that last regional level, genetic variants across all the levels. So K equals 5 has a higher term than K equals 3, K equals 4 and so forth. Now that's in those particular studies. We don't exactly know if those other levels exist because we would have to do a large number of studies, get replication and so forth. So there is some uncertainty about whether we have these other population levels, but if we do, the direction in which things are pointing is K equals five is either the level at which we have the most among region population genetic variants, or at least it's tied.

I think sometimes K equals six or seven comes out pretty close to K equals five. So those are all empirical questions. It's going to be interesting to see how things fall out. But it's not clear at all that we have anything better to go with. If you want to do large granular population sampling at the, you know, any homo sapiens now, I mean, you could do whatever you want. I mean, you could sample by Democrats and Republicans, right? You can always sample however you want, but we're trying to get these non-accidental, right? So if the rationale for using some population structure level is that, well, we have a causal story to support why sampling there would be useful, because over a large number of generations of relative inbreeding, you're going to get built up of a little frequency differences, right? And you also have justification for using it again and again and again, because if there's one difference is probably some others, right?

So it's an interesting question. We'll see how things shake out, but at least I think the continental population level, which I'm going to say is the 1997 OMB racial level, is a legitimate way of sampling if you think that's going to be relevant to what you're looking for.

Leah: Yeah. Okay. And you know more about genetics than I do, but I guess one reaction I have to this is that it really strikes me that these continental populations may just not be that relevant to clinical practice or to biomedical research. So just like one example that immediately springs to mind is like having Ashkenazi ancestry is much more helpful for being able to anticipate what allele someone might have, what diseases they might be at risk for versus saying someone is white. And in fact, if someone is not Ashkenazi, you might predict their level of risk very differently. And I mean, is the view just that like, you're not going to be able to get that fine grained with different categories?

Quayshawn: That's right. I'm absolutely not going to say that because that's a panmictic group. So that you can't get any lower than that. That is a particular randomly mated group of people, as far as population geneticists can tell. It would be fantastic if we were able to, in our clinical studies, have as our population strata, our sampling strata, at the population level, and make the groups now that would be very, very, very difficult for a couple of reasons.

One, that's a whole lot of sampling that you're going to be doing, partly because one of the main theories for how our groups link up with language is that we have about 7,100 plus linguistic groups, some combination of sometimes they're whole languages, sometimes it's kind of like dialects, but they line up nicely as far as we can tell with panmictic groups. So we kind of get an estimate of how many panmictic groups we have by looking at that sort of lowest level of linguistic structure. But if there's over 7,000 of those, so that means every time you do a clinical study, you gotta stratify your sample by 7,000 plus groups. Now, this is not tractable. This is very expensive research. It's just not doable.

So let's go up a level. What level should we go up? We know that K equals 5 would be a level we could go up. It would give us non-accidental reliable inductions in genetics because of what that level is. It's not entirely clear other levels would do that. It's not entirely clear what other levels we have. There you go. Right. So what do we do now? Like, I can give a recommendation to medical scientists right now. Yeah. If you're doing some pharmaceutical sort of research and discovery, and you have a hunch that the way in which your pharmaceutical works could have differences based on the sorts of enzymes people metabolism, and just some hypothesis about that variation might be continentally or some kind of geographically different. Go ahead and put in the human continental levels in your sampling. See what you find. Maybe nothing turns up. But if you sample enough, have enough statistical power, you'll actually be able to see it.

And yes, it would be fantastic if we could just sample every ethnic group. So that would be like Ashkenazi Jews, African Americans descended from U.S. slaves, and just keep going down the list. That's a very expensive project and not really practical.

Leah: Mm-hmm. Okay, so just to summarize, so that would be in theory ideal, but logistically, at least for now, seems hard to actually do in a lot of cases.

Quayshawn: Very close. In theory ideal, they would want to sequence every single human, right? And even with the local populations, you're still dealing with the low frequencies. You could be Ashkenazi Jewish and not have any of this stereotypical alleles tied to increase genetic disease risks that are, you know, stereotype to that group. What matters are your individual genotypes. So what they would really love is the whole sequence of every human that's born and do some big data sort of number crunching by randomly sampling across all the genotypes that exist in the species at the loci that will be relevant for the drug that they're trying to fix, tweak, create something like that and put you in genotype groups. So that would kind of be their idea of that. Now that's a tremendously expensive, probably illegal and, you know, impinging on privacy rights and unjust sort of sampling.

Sophie: Okay. Unfortunately, we have to end the conversation there, but we like to close by asking our guests, what is one rule or norm broadly related to what we've been talking about today that you would change if you could, and why?

Quayshawn: It would be tied to the last thing we were talking about, being a pluralist about race. Like we're pluralists about sex or gender in medical research, I would love the change and I would love to change it because I think it would bring about more racial justice. We'll be able to see more unjust racial health disparities than we're currently able to see. This is particularly an issue for U.S. racial minorities if we're talking about the American context, because we pay taxes too, and it would be fantastic if we didn't have to always question whether the stuff that our taxes are paying for work with 95% efficacy as the Pfizer and Moderna COVID-19. Now, I mean, there weren't any problems with that, it turns out, but there was always this question: can we trust the medical establishment to be working in our favor as well as it's working for white Americans? And I would love to change that.

Sophie: Okay. Well, thank you so much for joining us. We really enjoyed this conversation.

Bio(un)ethical is written and edited by me, Sophie Gibert, and Leah Pearson, with production by audiolift.co. If you want to support the show, please subscribe, rate, and review it wherever you get your podcasts and recommend it to a friend.

You can follow us on Twitter @biounethical (no parentheses) to be notified about new episodes, and you can sign up on our website, biounethical.com to receive emails when new episodes are released.

We promise we won't spam you, but we may reach out to let you know about upcoming guests and give you the opportunity to submit questions. Our music is written by Nina Khoury and performed by the band Social Skills. We are supported by a grant from Amplify Creative Grants. Links to papers that we reference and other helpful resources are available at our website, biounethical.com. You can also submit feedback there or email us at biounethical@gmail.com. Thanks for listening and for your support.