Jump to ContentJump to Main Navigation
What Is Race?Four Philosophical Views$

Joshua Glasgow, Sally Haslanger, Chike Jeffers, and Quayshawn Spencer

Print publication date: 2019

Print ISBN-13: 9780190610173

Published to Oxford Scholarship Online: June 2019

DOI: 10.1093/oso/9780190610173.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (oxford.universitypressscholarship.com). (c) Copyright Oxford University Press, 2020. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use. date: 22 October 2020

Is Race an Illusion or a (Very) Basic Reality?

Is Race an Illusion or a (Very) Basic Reality?

(p.111) 4 Is Race an Illusion or a (Very) Basic Reality?
What Is Race?

Joshua Glasgow

Oxford University Press

Abstract and Keywords

Joshua Glasgow argues in this chapter that ‘race’ in the ordinary sense is defined such that races are supposed to be large groups of humans organized according to certain visible traits, like skin color. Biology cannot validate the existence of such groups, even if it can identify other human populations. So Glasgow argues that race is not biologically real. And because the relevant traits do not change when social facts change, race is not a social construction. This suggests that race is not socially real, either. Consequently, it seems that race is not real. However, Glasgow also considers another possibility: that race is real, neither biologically nor socially, but in a more basic sense. According to this view, races are real by virtue of facts that find no home in any of the sciences, biological or social. Their only significance is that which people choose to give them. In the course of making his arguments, Glasgow explores how we should identify the meanings of our terms and how to proceed when scientific and ordinary meanings diverge. He concludes by leaving it open whether race is simply not real, or whether it is real in that basic sense.

Keywords:   race, biology, identity, social construction, basic realism, anti-realism

It was 1915, and Vaishno Das Bagai had several thousand dollars. He was educated. His high school headmaster recommended him as “a high-caste Hindu [from] a respectable family” and an individual of “very good character.” But these advantages were not enough. He still needed to escape India’s domination. His wife Kala would later remember him saying, “I don’t want to stay in this slave country,” in reference to Britain’s colonial rule. “I want to go to America where there is no slavery.”

After the long journey, it must have been difficult for Vaishno, Kala, and their three children to get stuck at the immigration center on Angel Island. But the officials there wanted to know how the family would provide for themselves in a new country. A few days in, the Bagais made their wealth known, which seems to be all that was needed to gain entry to San Francisco.1 The local paper featured a piece about Kala, headlined “Nose Diamond Latest Fad; Arrives Here from India.” It reported that she was “the first Hindu woman to enter this city in ten years.”

In the late 1800s and early 1900s, the United States was engulfed in a flood of anti-Asian racism. Seven years before the Bagais arrived, a white mob in Marysville, California, robbed and ran out of town seventy “Hindus,” telling them not to return. Whites repeatedly engaged in race riots that targeted Chinese and Filipino lives and property. Following an 1882 ban on Chinese immigrants, (p.112) the Immigration Act of 1917 banned immigration from India, Siam, Arabia, and elsewhere.

Nevertheless, Vaishno became a naturalized citizen six years after entering the United States. Because California’s Alien Land Act of 1913 said that only citizens could own land, this new membership card translated into the right to own property. Of course, legal rights are not everything. His citizenship did not prevent racist neighbors from locking the Bagais out of the house they bought in Berkeley, compelling them to return to San Francisco, where they lived above their store on Fillmore Street. Still, Vaishno’s citizenship was in hand, at least for the time being.

Three years after the US Constitution was adopted, the Naturalization Act of 1790 established that only white people could become naturalized citizens, and while this policy would take different forms over time, naturalization was racially restricted until the 1940s, when the Nazi specter shamed any nation that still stained itself with racialized population policies. In the meantime, the Naturalization, Immigration, and Alien Land Acts were just a few of the many bludgeons used for racial domination and exclusion. But enforcing this system required classifying people into races, and the law buckled under this task.

A singular piece of conceptual acrobatics came from the Supreme Court of California in People v. Hall (1854). A white man, George Hall, had been convicted and sentenced to death for killing another man, Ling Sing. But Hall’s conviction was based on the eyewitness testimony of three Chinese people, and at the time California law put racial restrictions on eyewitness testimony. Specifically, neither black people nor Native Americans were legally eligible to testify against white people. What, then, was the legal status of the eyewitnesses in Hall’s case? On appeal the Court declared both that Chinese people were black, because they were not white and ‘black’ just meant non-white; and also that Chinese people were American Indians, because indigenous Americans’ ancestors originally arrived via the Bering crossing from China. Legally speaking, then, the witnesses were now both black and Native American. Racist law and a contorted judgment about racial identity set free the murderer Hall.

This was just one moment in America’s long legal confrontation with racial categorization. Proving your race could be the key to proving which school your child was eligible to attend, to proving that you could marry your beloved in a country infested with anti-miscegenation laws, or to proving that your parents’ marriage was legitimate under those same laws, entitling you to your inheritance. Prior to the abolition of slavery, plaintiffs sought freedom (p.113) by arguing that their race made their enslavement illegal. And of course lives, relationships, treasure, and freedom were not the only goods that hung in the balance. If you wanted the protections and privileges of citizenship, you had to authenticate your whiteness here, too. The proving ground was the courthouse.

Generations earlier, witch-hunters in Salem identified their prey by looking for a hidden “witch’s mark.” When it came time to identify people’s races, things were hardly better. Admissible evidence of one’s race included one’s behavior, one’s reputation, even the shape of one’s feet. The rules about how to classify people of mixed racial ancestry were particularly notable for their sheer variety from state to state. And slotting individuals into races was not the only challenge. By the early 1900s, when events surrounding World War I sent new waves of immigrants seeking safe harbor around the world, America’s courts were tasked with sorting out the racial status of entire peoples. They had to decide: legally speaking, were Arabs white? How about Syrians or Armenians? Inconsistent rulings using varying standards piled up.

Enter the Supreme Court. Takao Ozawa, a Japanese immigrant, petitioned to be legally recognized as white for the purpose of becoming a citizen. In making his case, he pointed to his light skin color. He noted that he was assimilated in education, religion, language, and style. He emphasized his love of America. Late in 1922, the Court ruled. Justice George Sutherland—himself a naturalized citizen but apparently not one who felt liberated to expand the club—wrote that from a legal perspective, ‘white’ meant Caucasian. Sutherland recognized that there was some controversy about who exactly is and is not Caucasian. But he held that Ozawa “is clearly of a race which is not Caucasian and therefore belongs entirely outside the zone on the negative side.” The Supreme Court was unanimous: Ozawa was not Caucasian, so he was not white. Consequently, he was ineligible for citizenship.

By making being Caucasian the legal key to whiteness and therefore citizenship, the Court appeared to open a door for Bhagat Singh Thind. Citing scholarly authorities such as Johann Friedrich Blumenbach, Thind marshalled linguistic and other evidence to prove that Punjabis like himself were part of the Caucasian race. In fact, like Bagai, Thind had official certification that he was of “Aryan origin.” Thus from a scientific perspective, it looked like Thind was Caucasian, and since it was now legally established that ‘white’ meant Caucasian, citizenship appeared to be within reach.

Just three months after the Ozawa ruling, Justice Sutherland again codified whiteness from the bench: for legal purposes, Thind’s evidence was beside (p.114) the point. In this context, the words ‘white’ and ‘Caucasian’ were to be understood as “words of common speech and not of scientific origin.” Technical biological classification did not matter. Ancestral connection did not matter. Cultural assimilation and language did not matter. What mattered in this context was that common sense said that South Asians were not white. And under the Naturalization Act, that meant they were barred from citizenship. Thind lost his case.

Twelve years later, Congress allowed World War I veterans to become citizens, and due to his service Thind was naturalized. And by 1946, Congress and President Truman would create new law that Indians (and Filipinos) could become naturalized US citizens. But in 1923, the Thind decision stripped sixty-five Indians of their naturalized citizenship. Among them was Vaishno Das Bagai.

No longer being citizens, Bagai and others were forced to sell their real estate under California law. (Some would transfer their holdings to their children who, being born in the United States, had birthright citizenship; the Bagai kids were born in India.) And that was not the only setback. Bagai had renounced his British citizenship when he became an American citizen. Now there was no home to give him a passport. He had no way to travel to India again.

Bagai had sacrificed to tear off the cuffs of colonial oppression. Now his sanctuary had turned against him, rendering him dispossessed, legally bottlenecked, stateless, and geographically stuck. He had escaped the suffocating scope of British imperialism only to meet the relentlessness of American racism. What did it mean to trade chains for a box?

By this point he was desperate and heartbroken. Fed up, he wrote the following in a letter to the San Francisco Examiner:

But they now come to me and say, I am no longer an American citizen. They will not permit me to buy my home and, lo, they even shall not issue me a passport to go back to India. Now what am I? What have I made of myself and my children? We cannot exercise our rights, we cannot leave this country. Humility and insults, who is responsible for all this? Myself and American government.

I do not choose to live the life of an interned person; yes, I am in a free country and can move about where and when I wish inside the country. Is life worth living in a gilded cage? Obstacles this way, blockades that way, and the bridges burnt behind.

(p.115) Bagai had settled on a plan to escape his gilded cage: he would protest, by committing suicide. To spare his loved ones, he traveled to San Jose and rented a room, paying the whole month’s $35 in advance. He wrote a separate note to his family. Then he closed the windows, locked the door, and released the room’s illuminating gas. On March 16, 1928, Vaishno Das Bagai was dead from self-inflicted gas poisoning.2

And what was this thing behind it all, race? This beam of support for the Naturalization Act? This puzzle George Sutherland tried to solve to the exclusion of Japanese and Indian immigrants? This quality that racist Berkeleans saw as some sort of reason to lock the Bagais out of their own house? This line that lay underneath twenty white Californians running seventy innocent souls out of Marysville? This thing that has been behind white people regularly forcing non-white people out of towns, and much, much worse? What kind of thing is this?

It is not obvious how to answer that question. For starters, we could easily be confused about our concept of race. After all, when people thought that the word ‘whale’ referred to a kind of fish, this did not mean that whales actually were fish. What we care about, ultimately, is what the word ‘race’ actually means, not what we think it means. We want what Sally Haslanger (2012) calls the operative meaning of the term ‘race’—the meaning that governs our use of the term, even when we are unaware of it.3

Further complicating matters is that there are very likely several meanings of ‘race.’ What I am focused on here is how the term ‘race’ is used by ordinary, linguistically competent people in the contemporary United States, to describe groups of humans. Now this group of speakers brings with them a tremendously diverse set of perspectives. Still, I believe that people from these many positions have been having a collective, long-lasting, and broad conversation about race. I aim to focus on this conversation.

Although this conversation is inclusive and broad, it does exclude some communications that use the word ‘race.’ In particular, if some specialists or (p.116) experts (genomists, anthropologists, psychologists, etc.) somewhere use ‘race’ in a way that deviates from the meaning implied by the way ordinary non-specialists use ‘race,’ then that usage won’t be relevant to the following discussion (Glasgow 2010). They simply have their own definition, and so their own conversation. There is not necessarily anything wrong with this, it’s just that it is a conversation that would not be relevant to the conversation that I am addressing.

The same goes for non-expert speakers who use the language of race in a way that semantically deviates from that broad and inclusive conversation. If I said, “A couch is a race,” you would say that I mean something different by ‘race’ than what you mean when you use that word, assuming we mean the same thing by ‘a couch.’ Less unusually, some people may (some people do, I think) use ‘race’ in novel or revolutionary ways that depart from the broad conversation I’m focused on. Even narrowing the scope of our investigation to the context of ordinary usage, there is still plausibly more than one meaning of ‘race.’ I want to focus on whichever concept of race has the most currency—call this, for simplicity, the ordinary concept of race.

Again, this concept is no doubt thought of in many different ways and understood from many different perspectives. Nonetheless, to the extent that people from those different standpoints are at least sometimes having a coherent conversation about race, and at least sometimes participate in a shared system of racial identification and practice with each other, underlying that common conversation and practice is what I’m calling the ordinary concept of race. As I mean it, this is the concept that has shaped our approach to race for so many years, that has given definition to our identities and experiences, and that has been complicit in so much oppression and, for some, a key to liberation from that very oppression. In this way, by referring to the discourse surrounding the “ordinary” concept of race or “our” concept of race, I mean to refer to the same discourse that Sally Haslanger and Chike Jeffers are diagnosing. (The degree to which this is also the same discourse that Quayshawn Spencer is analyzing will be explored in subsequent chapters.)

By this idea of a ‘coherent conversation,’ I just mean that there must be some shared concept of race for us to even talk—even just disagree—about race. We can meaningfully disagree and converse about something only if we use our words in such a way that they have a shared meaning. Otherwise, we’ll talk past each other, using the same words to talk about different things. If you invite me to “cut the deck of cards” and I slice each of them in half with my scissors, we’re having a communication breakdown, not a genuine disagreement. In this way, shared meaning is (p.117) required for communication.4 I’m interested in whatever meaning of ‘race’ is presupposed by our frequent communications about race and by our use of racial categories, even when we disagree.5

I believe that the operative, ordinary meaning of ‘race’ is something like the following. This is a rough way of putting it, and it only captures part of the definition, but it’s a good starting point: Races, by definition, are relatively large groups of people who are distinguished from other groups of people by having certain visible biological traits (such as skin colors) to a disproportionate extent.6

With that working definition in hand, the next question is whether there is anything in the world that fits that description. Does ‘race,’ so defined, refer to anything real? Do races exist?

My coauthors have argued in the first three chapters of this book that race is biological, sociocultural, and sociopolitical. I think there are substantial merits to all three of these views, and I feel the pull of the arguments that have brought us to this point. Nevertheless, next I argue for a different view. I believe that the overall balance of considerations pressures us to conclude that races are neither biologically nor socially real. And from this a pretty compelling argument by elimination follows: if races are neither biologically nor socially real, then race is an illusion. This is racial anti-realism, and it is a stark claim. It maintains that Vaishno Das Bagai had no race. Neither do you, neither do I, neither does anyone else. Slavery, the Naturalization Act, Jim Crow, global domination and exploitation by Europe’s colonial powers, the exclusion of Ozawa and Thind, the racial biases and discrimination that shape our world today, and all the rest have been premised on a terrible confusion. Race does not exist.

Now, like any argument by elimination, it is possible that this argument for anti-realism is missing a viable but hidden option, beyond the options that race is biologically real, socially real, and not real at all. In particular, maybe (p.118) race is real in some more basic way, one that is irrelevant to the biological and social sciences. Toward the end of this chapter I will also explore this view—basic racial realism. So ultimately we’ll have to choose between the idea that race is an illusion and the idea that race is real in a basic, scientifically irrelevant sense. This will turn out to be a difficult choice.

But before confronting those alternatives, return to the step that starts us down this path. It is hard not to gravitate toward the common-sense idea that race is a biological fact, and really a fairly obvious one. So we start there. Can’t the biologists deliver race?

4.1. Mermaids, Werewolves, and Other Fantastical Things: On the Gap between the Biological Facts and the Concept of Race

4.1.1. The Spectrum

Races would be biologically real if—and only if—the concept of race played some legitimate role in good biological science. So if something is part of bad science or non-science, such as astrology, or if it’s useful for a non-biological science like sociology, then that thing is not biologically real.7 With this and the working definition of ‘race’ given earlier, our question then becomes whether it is useful for the biological sciences to divide humanity into large groups based on the relevant visible traits, like skin color. The answer is: no.

It has long been noticed that if you line up all of humanity by skin color, from darkest to lightest, you’ll see a spectrum of shades. Joseph Graves puts it this way: “If we were to only look at people in the tropics and people in Norway, we’d come to the conclusion that there’s a group of people who have light skin and there’s a group of people who have dark skin. But if we were to walk from the tropics to Norway, what we would see is a continuous change in skin tone. And at no point along that trip would we be able to say, ‘Oh, this is the place in which we go from the dark race to the light race’ ” (Herbes-Sommers 2003). The same goes for shapes and sizes of facial features. And for hair textures. And for any other visible trait that we might associate with race. Each variation transitions gradually to the next.

Call this fact The Spectrum. The Spectrum means that there is no principled biological reason to put one racial boundary here and another racial (p.119) boundary there based on visible traits. From the perspective of (good) biological science, there is no use in carving humanity into visible trait groups one way rather than another. And so, if races just are visible-trait groupings, races are not biologically real.

Now obviously we can establish some boundaries between groups based on visible traits. Put the lines of demarcation wherever you want! All the same, biologists still will have no scientific reason to recognize your lines—or any others based on visible traits. The boundaries within The Spectrum could be put here, they could be put there. From the perspective of biology, each of these boundaries is arbitrary. So we can draw lines around groups of humans, which we then call ‘Asian,’ ‘white,’ or ‘Latinx.’ But the lines that separate those bounded categories are imposed by us onto a blurred image of humanity. This is the sense in which race is not biologically real. Skin colors are biological traits. And we can divide ourselves up according to those traits. But our lines of racial demarcation are not discovered in the biology. Which means that racial groups themselves are not in the biology. It is akin to dividing ourselves into the height categories of ‘short’ and ‘tall’—individually, our properties (our particular skin colors and heights) are biological, but we project the categories (white people, short people) onto the world (Relethford 2017).

In addition, attempts to ground race on certain bio-medical conditions, like heart disease or sickle-cell trait or Tay-Sachs disease, run into the roadblock that the populations with these conditions do not correspond to ordinary races. Moreover, to whatever extent there appears to be a correlation between races and medical conditions, this correlation is ultimately best explained by social, rather than biological, causes (Diamond 1994; Herbes-Sommers 2003; Kaplan 2010; Sullivan 2013). For example, black Americans suffer heart disease disproportionally, so you might think that something biological about being black gives rise to this disease susceptibility. But black people who are recent immigrants to the United States do not suffer disproportionate heart disease, suggesting that the spike in heart disease is due not to being black itself, but rather to the experience of being black in America (plausibly, to the stress that comes from an unrelenting, intergenerational subjection to racism). To that extent, self-identified race may be worth taking seriously in a medical context, but that is because it can indicate risks due to social environment, not long-standing genetic risks. And to be sure, it is very hard to find medical conditions that correspond to ordinary races. Sickle-cell trait is an adaptation to fight malaria, so it is found in people with ancestral connections to malarial regions, including parts of Africa and southern India but excluding parts of South Africa. A coalition of some Indians and some (p.120) Africans is not a race on the common-sense understanding of race. Likewise, Tay-Sachs disease is disproportionally found among those with Northern European Jewish ancestry, but that smaller group of people is not its own race, either, on the common-sense concept of race.

This will be a recurring theme: biologically real groups of humans do not seem to line up with races, as ordinarily defined. This is the Mismatch Objection (Mallon 2006), and its source is that the ordinary concept of race is pegged to visible traits, but groups based on visible traits are not useful for biology.8

4.1.2. “Words of Common Speech”

More sophisticated biological theories of race, including Quayshawn Spencer’s contribution to this book, may seem to sidestep the Mismatch Objection. To give them a broad characterization, genealogical theories say that races are populations produced by the breeding patterns of our ancestors. This might get cashed out in terms of the ancestral lineages themselves, or in terms of genetic populations that result today, or in terms of some other genealogical element.9 What is crucial is that the genealogical element in question is the only thing distinguishing each race from the next. These are not the racist biological theories of yesteryear that held that important qualities like intelligence, virtue, and beauty were part of some inherited racial package. Race, on these accounts, is merely a matter of ancestral relations or that microscopic residue of ancestry, our genes.

I believe that genealogical theory faces its own Mismatch Objection: scientifically provable genealogical populations are not races.10 The source of the mismatch is that genealogical populations are ultimately determined by (p.121) reproductive patterns, while races, again, are supposed to be determined by the way we look. The two can come apart.

For instance, Roberta Millstein (2015) points out that individuals can migrate between populations, at least on one plausible articulation of what a population is. If you descend from population A but start reproducing with people from population B, then population-wise you will have stopped being an A and started being a B, she argues. Race obviously does not work like that. When “Hindustani” men in the early twentieth century migrated to California’s Imperial Valley and married and reproduced with women racialized as “Mexican” (Hart 1998), that wrinkle in mating behavior may have changed their population, but it did not amount to a literal race change, on the ordinary concept of race. So migratability is one difference between populations and races.

In addition, two genealogical populations could, in principle, be visibly identical, whereas two races cannot. One hundred years and four days after the Bagais arrived in San Francisco, the prosecutor’s office in Cuyahoga County, Ohio, received a letter from lawyers representing Samaria Rice. The letter asked the prosecutor to pursue an aggravated murder charge against police officer Timothy Loehman for killing Samaria’s 12-year-old son, Tamir. Loehman had gunned down Tamir as he played in a park while in possession of a toy weapon. With no apparent justification, Loehman jumped out of the police car before it had stopped and only let two seconds elapse before he started firing on Tamir. Two seconds to say anything to Tamir. Less than two seconds for Tamir to potentially comply with any orders. Now imagine that some activists conclude that it is time to do something radical about the police persistently brutalizing and murdering black people in this way. The activists decide that the best solution is just to make us all look the same. So they develop a chemical agent that changes the genetic makeup of anyone who ingests it such that they end up looking exactly, and permanently, like the Dalai Lama. They then infuse the global water supply with this agent, and sure enough, within a few weeks, every human being on earth looks like the Dalai Lama.

Now also imagine that at least for a few generations we keep the ancestral populations we had prior to the change: geographical and cultural forces (such as language, dress, and popular places to find potential mates) impact reproductive choices, preserving genealogical lines. But everyone looks like the Dalai Lama. In that world, ancestral populations have not (yet) faded away. But race has disappeared in that world, because we look the same. There aren’t any black people in a world of only Dalai Lamas. But there are still (p.122) people with recent ancestry that is entirely from sub-Saharan Africa. Here again, we have a mismatch: races must, by definition, be visibly distinct, but populations need not be.

Migratability and the irrelevance of visible traits are features of genealogical populations that diverge from what we are trying to talk about when we talk about race. In addition, genealogical theory also has a different groups problem: the particular racial groups we ordinarily recognize are not the same as the genealogical populations recognized by science. One part of the problem is that some individuals get classified by science in a way that they themselves, and others, may reject—a phenomenon that has been highlighted with Americans who are racialized as black but would not fit the genealogical category of ‘African’ (McPherson and Shelby 2004). A second part of the problem is that genealogy theory fails to recognize some common-sense racial groups altogether. In particular, categories like Arab or Latinx find no home in the flow charts of population genetics, but they are routinely racialized in the United States. Importantly, the people in these groups routinely racialize themselves in this way (Alcoff 2006, Ch. 10–11; Haney López 2005). Until recently, it appeared that the 2020 US Census might add Hispanic and Middle Eastern or North African to its list of races, partly because the people who fall under these labels have found no place for themselves on previous Censuses (Krogstad and Cohn 2014). That proposal has been shelved, at least for now. In any event, it looks like certain identity commitments do not answer to population science.

Now one reply to the Mismatch Objection is that it doesn’t matter if science and common sense fail to map onto one another, since science can have its own concept of race (Andreasen 2005). And, indeed, some might like the fact that genealogical theory recognizes different races than common-sense racial thinking recognizes. Thind certainly wanted to let science dictate our legal racial categories.

There’s a kernel of truth in this reply: science can have whatever concepts it wants. So can the rest of us. It is up to us how to use our words. If you want, you could use the term ‘race’ to refer to window, in which case—as long as we believe in real windows—we would want to think that race is real. And if genealogical theorists want, they can use ‘race’ to refer to genealogical populations.

But for our project, what ‘race’ refers to is constrained. Our goal is to see if the term ‘race,’ as it operates in ordinary talk, maps onto anything real. And for that purpose, we have to stick with the operative, ordinary concept of race. We can’t just use whatever concept we want, and in particular we can’t use a (p.123) technical, scientific concept that is dislodged from the ordinary concept of race. If we are more committed to ordinary racial categories than we are to conforming our racial discourse to mismatching scientific categories, then a scientific meaning that deviates from the ordinary meaning is not relevant to our project (Glasgow 2010; cf. McPherson and Shelby 2004, 187–188).11

In short, it looks like some of our racial identities and some features conceptually bound up with race do not map onto the populations recognized by science. The ordinary term ‘race’ purports to refer to something other than the genealogical lines validated by biology.

4.1.3. Surprising Referents, Take 1: Whales Are Not Fish

But wait. We used to think that ‘atom’ was defined as the smallest particle and therefore indivisible. Then we split the atom, and it turned out that our attempt at defining the word ‘atom’ was wrong. We also used to think that whales were by definition fish. Then we discovered that whales are warm-blooded and breathe with lungs. Our early attempts at defining ‘whale’ had failed, too. Sometimes our concepts are opaque. We can make mistakes in how we define terms. So why not say something similar about race, that ‘race,’ even in the ordinary sense, refers to something unexpected and contrary to common thought? Actually, haven’t we already said similar things about race? For example, some argue that Italians went from being classified as non-white to being classified as white.12 Is that really so different from what genealogical theory is requiring us to do with the category Latinx? Sometimes our words refer to things in ways that are surprising to us. So why not allow that ‘race’ is (p.124) one of those words—why not say that the term ‘race,’ like ‘atom’ and ‘whale,’ has a surprising referent? In particular, why not accept that the surprising referent for ‘races’ is the set of populations recognized by the most plausible version of genealogy theory?

Distinguish the surprising-referent view from a view that recommends that we should change what ‘race’ refers to. On the latter view, you might concede that the ordinary term ‘race’ is currently supposed to refer to something that is incompatible with genealogy theory; simultaneously, you might also maintain that we should improve the meaning of ‘race’ so that going forward it can refer to certain ancestral relations. To give this kind of view a name, let’s call theories that recommend a change in a term’s referent revolutionary. Revolutionary theory, however appealing it may be, is not applicable to our present project (though it will reappear later in our discussion). It answers the questions, what should ‘race’ refer to, and what should its definition be? Our present question is what the current, operative concept of race is. The surprising-referent proposal does answer this question. It says that there really are biological races in the way that whales really were mammals all along: these facts are just hidden from us. To paraphrase Ron Mallon (2017), we might call these covert kinds of thing. This view does not recommend that we change what racial terms actually refer to, as revolutionary theory says. Instead, we correct what we mistakenly think racial terms refer to. On this approach, ‘race’ already means that there is no Latinx race and that all races could look exactly alike; we just haven’t realized it yet, since the true nature of race has been covert to date. As with ‘whale,’ the surprising referent of ‘race’ just needs to be revealed to us. Call this view revelation theory. On the genealogical version of revelation theory, it is surprising to learn that individuals can change races by reproducing with someone of another race, that there is no Latinx race, and so on. But it was also unexpected to learn that atoms could be divided and that whales were not fish, so why not accept some unanticipated aspects of race, too?

Revelation theory presents the big challenge for the race debate (cf. Glasgow 2009, 126–132). It puts us at a crossroads. On the one hand, we are inclined to preserve the phenomena we believe in, including race. This existence commitment pushes us toward genealogical theory and whatever surprises come along with it, if that is the most likely way to vindicate our race-talk. On the other hand, ‘race’ having a genealogical referent is incompatible with other commitments, as we have seen. We are committed to racial groups being organized by visible traits. We are committed to the idea that you are born with your race and cannot change it simply by having sex (p.125) with someone of a different race. Many are committed to certain racial identities, including Latinx. Call these, collectively, our features-and-identities commitments. The big challenge is what to do when our existence and features-and-identities commitments conflict. If our most fundamental, stickiest conceptual commitments are the features-and-identities commitments, then we have to accept that races cannot be the ancestral populations recognized by biology. But if we’re willing to negotiate on features and identities in order to preserve existence, then races could be genealogical populations, surprisingly. It all depends on which commitments are so firmly held that they get embedded into the very meaning of the word ‘race.’

There is no general principle that dictates when one commitment has conceptual priority over another (Appiah 1991). Sometimes, as with ‘whale’ and ‘atom,’ we accept surprising referents. Other times, we do not. Apparently one source of the werewolf legend was that people with rabies were thought to be possessed by wolves. So we could have said that werewolves are real: ‘werewolf’ surprisingly refers to person with rabies. But we don’t say that; ‘werewolf’ is defined by the features commitment that werewolves have to be human by day and wolf by night. To preserve this, we’re willing to give up the existence commitment and accept that werewolves are not real. Similarly, sailors supposedly mistook manatees for mermaids. So we could have said that ‘mermaid’ surprisingly refers to manatee. But instead we stuck with the features commitment that mermaids, by definition, must be half-people/half-sea creatures, meaning that there are no mermaids. We could have said that phlogiston was really oxygen in disguise; but we stuck with the commitment that phlogiston is supposed to be a certain kind of substance that just does not exist.

So sometimes, as with ‘atom’ and ‘whale,’ we are willing to negotiate on our features commitments to preserve the existence commitment. Other times, as with ‘werewolf,’ ‘mermaid,’ and ‘phlogiston,’ we give up existence and favor our features-based definitions. The challenge for us is to identify which set of commitments is stronger when it comes to ‘race.’ What is non-negotiable for us in our shared racial discourse?

Our predicament is that there is no obvious way to figure out which commitments are stickiest. We may never satisfactorily meet the challenge. I am a fan of experimental approaches to questions about our concepts, where we poll people on what they think about race, ask them to classify test cases, and so on, as a way of exploring the definition of ‘race’ (Glasgow 2008; Glasgow, Shulman, and Covarrubias 2009; Shulman and Glasgow 2010). The results of such studies provide evidence about where our commitments (p.126) lie. But we have to be careful here: this is just evidence; it is not a conclusion. Some say that each folk theory of race reveals a separate operative concept of race (e.g., Ludwig 2015, 255). Another possibility is that some single experiment can tell us what the content of the concept of race is (Pierce 2015, Ch. 4). I believe something else is closer to the truth: we have multiple commitments of varying strengths; our most strongly held commitments ultimately define ‘race’ and limit what it can refer to; and experimental studies only hint at what our commitments are and what their degree of conceptual entrenchment is.

The operative concept of race consists of the commitments people would stick to when they are forced by consistency to abandon part of a set of incompatible commitments (Carnap 1955; Chalmers 2012; Glasgow forthcoming a), and this is hard to demonstrate experimentally. Potential confounds abound, whether we poll many people or just reflect internally on our own thinking about race. Are participants really keeping all relevant cases in mind as they answer a prompt? Do they recognize the conflicts? How do you test the potential conflicting commitments about features and identities in a way that can rank their conceptual priority? Are there other commitments that are left out of the study? Are the experimental prompts distorting—or changing—participants’ answers? So my view is that while experimental studies can be suggestive, no experiment could decisively prove which complex set we would be willing to negotiate on and which we would insist on keeping when they are thrown together into multiple inconsistent sets. (That said, I hope an experimenter proves me wrong on that one.)

In the empirical studies we do have, people regularly point to biological features like skin color as definitive of race, but though this dovetails with my argument, it really only suggests that race is defined in terms of biological traits and skin color in particular. Even if everyone prioritized biology, it still would not be decisive. After all, if a representative poll found ancient Europeans univocally saying that whales are fish, this would not decisively prove that ‘whale’ by their definition referred to a kind of fish. The question is what those language users would (and ultimately did) say when forced to confront evidence that the objects they call ‘whales’ are not actually fish. (Again, this is what Haslanger calls the operative meaning of the word ‘whale.’) Existing research does not force respondents to choose between vast and complex sets of propositions about race—and potentially cannot do this, given the confounds that are built into such choices. The most we can do is identify our commitments, map their inconsistencies, speculate as to which ones would survive attempts to reconcile them into consistent sets of beliefs, (p.127) and try to partially defend our speculations with the best experiments and arguments we can find.

For now, our ongoing dialogue about race is arguably the best trial we have. This natural experiment can indicate which commitments we keep and which ones we jettison. In effect, books like this one—or the most perfect version of them, anyway—are the tests, and reader reactions are the results. But this is an inherently flawed sort of test, because it will always be possible that reader reactions constitute an online redefinition of ‘race’ rather than revealing our hidden-all-along definition. Revolution might overtake revelation.

So the best we can do is to interpret the data we have. The view I have tried to argue for here is that the features-and-identities commitments are stickier than the existence commitment. In essence, to say this is to predict that, if we could generate an ideal experiment, the ordinary concept of race would be shown to be inconsistent with any view that denied that Latinx identity is a possible racial identity (not least because of the tremendous stakes at play in such a denial). It will also be inconsistent with the idea that one could change races just by having children with someone from another race. It will also be inconsistent with the claim that all races could look exactly alike. As with mermaids, werewolves, and other fantastical things, we are more willing to stop believing in the existence of biological race than to abandon the idea that these core features and identities are central to race, at least when we mean ‘race’ in the sense in which it operates in our day-to-day lives.

Those who agree will find that biological theories of race face insurmountable problems. The most stringent version of the mismatch standard says that to demonstrate that race in the ordinary sense is biologically real, we’d have to show that it is useful for biology to recognize a Latinx race but not, say, a biological race of musicians; that there are biologically principled groups organized according to the relevant visible traits, including skin color; and that these groups cannot be entered and exited by reproducing with someone in a different group. Given what we know about biology, it seems unlikely to deliver race in that sense. Some may accept a weaker standard than this. Perhaps, for instance, you’re willing to accept that Latinx does not need to be shown to be a biological race—you can accept that Latinx identity may not turn out to be a racial identity—but you also maintain that the ordinary concept of race requires that a person’s race cannot change simply by virtue of her mating behavior. (Or perhaps, vice versa.) Even then, biology seems unlikely to deliver race in that sense.

(p.128) 4.2. Amnesia, Babies, and Equals: On the Relevance and Irrelevance of Social Facts

4.2.1. “But They Now Come to Me and Say . . .”

If the racial boundaries we think we see are not biological, then it seems we must have projected them onto ourselves. How we draw these borders has been a matter of choice: different societies, and the same societies in different eras, have recognized different racial lines. This behavior provides the raw materials for an alternate theory: Don’t we create race by these acts of dividing humanity? We treat one another differently based on how we perceive each other’s visible traits; perhaps that treatment, or perhaps the results of this treatment, are the very thing that generates races. Race, on this constructionist view, is a social rather than biological kind of thing.

The most rudimentary version of constructionism says simply that when we classify ourselves into races, we create real races, just as classifying ourselves into students and teachers creates actual students and teachers. We create the real groups of nurse, ballplayer, engineer, and dancer, so why not say we create race, too?

I do not think that this view can work, for it faces its own Mismatch Objection, in the form of a different features problem.13 Imagine that we all forgot about race for a few minutes, due to some physical force—imagine the activists infused the global water supply with a different drug. The racial amnesia sets in quickly. We no longer recognize any lines of racial demarcation. We just see a spectrum of individualized traits, not groups based on them. Then after a few minutes our amnesia passes, and we resume classifying ourselves and others racially.

Rudimentary constructionism says that, because we no longer recognize racial divisions, we literally lose our races during our bout of amnesia. Jonathan Franzen goes from white to non-white to white again, in a matter of minutes. Oprah Winfrey stops being black. Sonia Sotomayor stops being Latina. Then they change back. Notice how strong the implication is. It is not merely that we stop classifying or perceiving people in these ways—that is true by hypothesis. Rudimentary constructionism implies that when we stop classifying people racially, something more substantive changes, too: people actually stop being Native American, Asian, black, Latinx, and white. In the way that paper money stops having monetary value if we stop classifying that paper as money, rudimentary constructionism says that if we stop recognizing race, then race itself actually disappears.

(p.129) But race does not work that way. On the operative ordinary concept of race, you cannot lose your race simply because we happen to forget what your race is. Franzen doesn’t really stop being white just because of what we think or fail to think. It therefore looks like socially recognized groups have different features than races must have. Socially recognized groups disappear when the recognition disappears, but races are supposed to persist throughout different forms of social recognition.

Before complicating this argument, there is an important truth in constructionism that we need to capture. Let’s call groups that society recognizes as races racialized groups (Blum 2002; Shelby 2005). Racialization, as distinguished from race, is socially real. Because differential treatment follows racialization, we bend ourselves into groups that come with their own possibilities, experiences, and identities. This racialized interaction has tremendously significant consequences for people’s lives, and we have urgent responsibilities to address the massive moral failures that have historically marched in lockstep with it. In slogan form, we must recognize racialization. On that much we should all agree. But at the same time, constructionism posits, further, that racialized groups just are races. Against this, the amnesia thought experiment suggests that races and racialized groups have different features, because racialized groups, unlike races, disappear with changes in social recognition.

There are students and teachers, readers and authors, citizens and government officials. There are the powerful and the powerless, oppressors and oppressed. Chefs and diners, stoics and whiners, hardscrabble coal miners. And there are racialized groups. Not one of those categories is biological. The people who fall under these labels are brought together by the fact that they inhabit certain social roles, and I thoroughly agree with constructionists that such labels name realities that are explainable only from a social, rather than biological, perspective. I am just arguing that race is not one of those realities, for race is supposed to stick around even when the social facts change.

4.2.2. Sophisticated Constructionism

Most constructionist theories are more sophisticated than the rudimentary form I’ve just examined. You’ve seen two such views in the chapters by Haslanger and Jeffers.14 Because constructionists disagree with each other about which social facts make a group of people into a race, race could be created by social facts that are more durable than the mere fact of engaging (p.130) in racial classification. And because these more durable facts might not disappear during a bout of collective racial amnesia, a more sophisticated constructionism could avoid the amnesia counterexample. Consider the way that Ta-Nehisi Coates (2015) puts it in his penetrating reflection on blackness in America:

the power of domination and exclusion is central to the belief in being white, and without it, “white people” would cease to exist for want of reasons. There will surely always be people with straight hair and blue eyes, as there have been for all of history. But some of these straight-haired people with blue eyes have been “black,” and this points to the great difference between their world and ours. We did not choose our fences. They were imposed on us by Virginia planters obsessed with enslaving as many Americans as possible. Now I saw that we had made something down here, in slavery, in Jim Crow, in ghettoes. At The Mecca I saw how we had taken their one-drop rule and flipped it. They made us into a race. We made ourselves into a people.

Coates here positions race as essentially a matter of creating categories based on visible traits and attaching them to different positions in a power hierarchy, updating W. E. B. Du Bois’s much-discussed claim that “the black man is the person who must ride Jim Crow in Georgia.” Skin color for Du Bois is a “badge” that indicates which race one has been given by society, but the essence of race ultimately lies in who has power over whom. On these accounts, a white supremacist social system, be it in the form of Jim Crow laws or a one-drop rule designed to maximize the economic gains of slavery, is what imposes race on human bodies. According to Coates, whiteness in particular loses its raison d’être, if white supremacy disappears. By placing power at the center of race, Coates, Du Bois, and Haslanger (2012) could likely agree with this constructionist analysis of race from Hazel Rose Markus and Paula Moya (2010, 21):

Race is a doing—a dynamic set of historically derived and institutionalized ideas and practices that . . . associates differential value, power, and privilege with these characteristics; establishes a hierarchy among the different groups; and confers opportunity accordingly.

Not all constructionists place power in the very definition of ‘race.’ Ian Haney López (1996, 14) defines race as “the historically contingent social (p.131) systems of meaning that attach to elements of morphology and ancestry.” This is a constructionist view, but it allows for races that do not necessarily involve a power difference. But whether power is thought to be essential or merely very common to racialization, racialization on all of these accounts is more complex than simply engaging in racial classification. The idea of race has been thoroughly woven into our actions and institutions, from slavery and George Sutherland’s writings through the fact that a police officer wantonly killed Tamir Rice without proportionate and just consequences. Racialization shapes our family lives through inherited wealth and poverty, along with privileges and discriminations that are renewed daily. It is in a misunderstood glance and in a hiring decision that is unconsciously guided by unknown biases. It is manifested in massive inequalities and reflected in our most hallowed and influential chambers. It is in how we live and where we live. It is in our laws, and it is in our kitchens. It is incorporated into our bodies (Alcoff 2006; Haslanger 2012, Ch. 9).

Because racialized practice as a whole is more complicated and enduring than racial classification, the amnesia thought experiment is not sufficient to undermine sophisticated constructionism. Even if we collectively forget about race for a few minutes, that doesn’t change the fact that Tamir Rice’s perceived blackness was likely a factor in Timothy Loehman murdering him. It doesn’t change the fact that Vaishno Das Bagai was ruled ineligible for citizenship. It doesn’t change the way that power has been wielded and resources distributed. And so if race is created by a set of social forces that outlast mere classification, then they can explain how race can outlast a brief period of racial amnesia. That counterexample can’t work against all forms of constructionism.

4.2.3. The Persistence of Race

I have learned a tremendous amount from constructionists about how we filter our lives through the lens of race. Still, I also believe that all constructionist accounts of race are exposed to counterexamples that are structurally similar to the amnesia case. For all constructionists, if the relevant social facts, the ones that they think create race, are not in place, then race would disappear. This exposes all versions of constructionism to the different features problem: on the ordinary concept of race, race persists even when the social facts change.

Consider the constructionist view that what makes race is inequality. It could be inequality in power, wealth, health outcomes, educational (p.132) opportunities, or anything else. Regardless of the details, the problem is that these views make racial equality impossible. For power-based theories and all other inequality-based theories of race, inequality is the very essence of race. So if we gain equality, we lose race, on these accounts. The most that these views can deliver on the equality front is post-racial equality. Racial equality for these theories is literally a contradiction in terms. But in that way, inequality-based constructionism fails to capture our ordinary concept of race (Glasgow 2009, 120). We, and therefore our operative ordinary concept of race, are committed to the idea that racial equality is possible. It might be a long way off. It might be hard to imagine. If you are more pessimistic about the future than I am, you might think racial equality is even more unrealistic than drugs that make us look like the Dalai Lama. But racial equality is not incoherent. Yet inequality theories make it literally incoherent. They turn racial equality into the squared circle of social life—something that is impossible even in our dreams. And that flies against the ordinary concept of race. The ordinary concept of race—the operative one, the one we are committed to in the practices and discourses that constructionists rightly demand that we focus on—allows that racial equality is a possibility, a legitimate goal, an ideal that is conceptually sound. Inequality constructionism, including sociopolitical constructionism, invalidates that goal by turning it into a contradiction, and in so doing it violates the concept of race. Consequently, inequality theories of race are not actually theories of race. At best they are theories of racialized groups, or even more plausibly, unequal racialized groups.

Against this, Haslanger (2012, 255) has argued that there would be no point to society “recognizing” or “addressing” color-based categories if we get rid of racial inequality. This seems to be built on the notion that the reason to recognize race is to promote justice (Haslanger 2012, 250, 258–260). But why can’t ultimate justice be the achievement of racial equality, rather than the elimination of racialization altogether (Outlaw 1996)? Moreover, what justice requires is different from what is in the world. There are people with hitchhiker’s thumb—a thumb that can curve significantly away from one’s fingers. Presumably there is no justice-based reason to recognize people with hitchhiker’s thumbs. But even if they enjoy equality with those who lack hitchhiker’s thumbs, they exist all the same. Similarly, inequality-based analyses of racialized practice are crucial to addressing injustice, but that is different from the task of identifying what exists.

All forms of constructionism say that race disappears if and when the relevant social facts, the ones that the constructionist thinks are race-making (p.133) facts, disappear, be they facts of classification, inequality, or anything else. That is why all forms of constructionism are inconsistent with the ordinary concept of race: it is coherent to imagine race persisting through any social change, and constructionism rejects this possibility—as not just unlikely but downright incoherent.

In arguing against constructionism, I want to emphatically underscore the significance in the racialized facets of our social reality. We must attend to their impacts, in all their complex detail and history. Racialized behavior has affected lives in ways that we neglect at our moral peril, ways that arguably should draw the bulk of our attention. At the same time, no matter which social facts we attend to, we can always imagine them disappearing while race stays. And if race is conceptually able to persist across all social practices, then by definition it is not a social phenomenon.

One thought experiment challenges all constructionist views: imagine a world of only babies. Everyone else has died off. A new technology keeps the newborns alive and cares for them until they can care for themselves. Before the adults perished, they acted to prevent the terror wrought by centuries of unjust racist behavior. Wanting their children to avoid the same racial struggles with which humanity had plagued itself, the parents decided to wipe any trace of racialization. They destroyed any records that refer to our racially fraught history. In fact, just to be safe, they erased all history and culture other than what was needed to provide the babies with enough science to maximize their well-being. All babies are given equal resources. A variety of therapies become available to allow them the equal chance for equal health outcomes. And so on. Any other information is eradicated in an attempt to present the Reboot Generation with a social blank slate.

Because every racial practice, along with every result of our racialized past, dies off with the adults, constructionism is forced to say that a (racially) Asian baby stops being (racially) Asian when the last adult dies. According to constructionism, that baby was firmly Asian, but her Asian-ness somehow instantaneously vanishes at the age of four months and six days, even though the only thing that changes is that some adult, some stranger on the other side of the world, passed away. That is not how race purports to work. Surely the babies would still have their races after the adult perishes, if they have any races to begin with. This is how constructionism fails to capture race in the ordinary sense of the term.

Of course, committed constructionists will not (and do not!) share my reactions to these thought experiments. They will say that we can lose our races in the amnesia case, that racial equality is impossible (Haslanger 2012, (p.134) 9), that races wholly sharing culture is impossible, and that the Asian baby truly stops being Asian.

The constructionist and I disagree on these points. You, reader, must judge for yourself. Interestingly, our judgments as to what would and would not make race disappear themselves determine how we should react. Our judgments about how to properly deploy the term ‘race’ give definition to that word, and in so doing they set boundaries for which judgments are correct. This conceptual bootstrapping may seem odd, bur reactions to these cases—our patterns of applying or refusing to apply the relevant term, in this case ‘race’—dictate the content of the concept. If I have correctly anticipated the commitments shared by our linguistic community, then we have found a new part of the definition of ‘race.’ This part of the definition was not explicit in the working definition given earlier. If we are to use the language of race—which is an open question here—we have to say that on the currently operative ordinary concept of race, the British were not spontaneously made white in the early 1490s, that an Asian baby is Asian regardless of whether some random adult dies, and that people of all racialized groups could, in principle, enjoy equality and harmony. It’s not just that races by definition must be organized by visible traits. In addition, races by definition must be non-social.

4.2.4. Surprising Referents, Take 2: Gender

Recall that sometimes our words can refer to objects in highly surprising ways. Whales are not fish, and atoms are not indivisible, despite what people thought. Such revelations are possible with social objects, too. In particular, sometimes we think we are talking about a biological thing, when really we are talking about a social thing (Haslanger 2012, Ch. 4). Consider the behaviors associated with gender. It was once common to think that biological hardwiring as males and females dictated who was most fit for taking on child-care responsibilities, changing lightbulbs, earning income, doing dishes, wearing pink. We now know differently. We invented the assignment of these behaviors to sex, just as we constructed what it means to be a ‘foodie.’ Boys used to wear pink and girls did not; then it was switched. We treat one another differently, which creates almost all of the gender role differences we see. For lack of a better word, we genderize.

Racial constructionists ask us to see race in the same, revelatory way. It may be true, on this revelatory theory, that race appears to be biological. It may also be true that race appears to persist across any social practice, as in the cases I have presented. But appearances can be deceiving. While it used (p.135) to appear that women’s biology rendered them unsuited for being medical doctors, we now know that to be false. Race is similar, on this version of revelation theory: it is revealed to be different from what we thought it was. Race is, on this account, a covert kind.

Again, I take revelation theory to bring the race debate’s big challenge into relief. We could protect our commitment that race appears to be real by allowing that, to our great surprise, an Asian baby can actually lose its Asian-ness if a random adult dies on the other side of the planet. In that case, race could be socially real. Or we can stick with our features commitment, preserving the apparent semantic fact that race is supposed to be simply about our bodies and thus persist even when our social practices change, which would mean that constructionism is false. Which commitment is stickier?

Haslanger (2012, Ch. 10) holds that socially racialized groups act as a “reference magnet” for the term ‘race.’ Once we give up our belief in visible-trait-based biological races, we should search for the next best thing for ‘race’ to refer to—the ‘magnet.’ And, she claims, the next best thing, the magnet, is a set of social groups. Now we should be careful here: Why should we think that ‘race’ is drawn to a social magnet rather than some other magnet? If we need something for ‘race’ to refer to, why pick apparently mismatched socially racialized groups rather than, say, the biological populations on which genealogy theory focuses? What counts as the next best thing for ‘race’ to refer to? That’s one gap in the constructionist reference-magnet argument. But there is also a more basic question: Why should we think that the term ‘race’ successfully refers to any reality at all? Why think it has any magnet?

Some people think that every term has a real-world referent to which it is magnetized, in which case racial terms “refer to whatever is the best candidate for reference given how we use those terms” (Pierce 2015, 60). But words do not always refer to the real object closest to their perceived reference. ‘Werewolf’ does not refer to a person with rabies. When we discovered that the people we called ‘witches’ did not have supernatural powers, we could have said witches are, roughly, people who are treated in socially witchized ways, in the manner that the racial constructionist thinks that white people are, roughly, people who are socially racialized as white. That would be the “best candidate” referent for ‘witch’ in the real world. But we remained semantically committed to a feature of being a witch, namely, they had to have the supernatural abilities to cast spells and commune with the devil. Since nothing in the world has those abilities, there just aren’t any witches. (At least, this is part of the meaning of ‘witch’ when we say that there are no witches. Obviously, terms are ambiguous, and people sometimes (p.136) use ‘witch’ to refer to practitioners of Wicca, who do exist. There are witches in one sense of the word but not in another.) In this way, not every word is magnetized to an object in the world. Our features commitments sometimes block that option.

So it is not enough to say that ‘race’ could have a surprising referent. The big challenge is to figure out whether ‘race’ is magnetically drawn to a surprising referent or instead purports to refer to some sort of entity that just doesn’t exist. And as with genealogical revelation theory, the answer lies in which of our competing conceptual commitments are more entrenched, that is, which principles are most deeply rooted in our concept of race. If we find ourselves unable to accept that an Asian baby could stop being Asian when the last adult disappears, then race is not socially real.

To revisit a point made earlier, an answer to this question is out there, but it is elusive. I believe that the cases I presented earlier speak against constructionism in a way that is more powerful than the considerations constructionists call on. That said, though race is not socially real, racialization is. There are facts about inequality. There are facts about patterns of discrimination. There are facts about who is exercising power over whom. There are facts about violence, about privilege, about pain, about culture. We build the fences, and lives are consistently and unjustly taken and made worse off. Redress is required. We interact with one another on the assumption that race is real, and in that we create new realities and new moral obligations. But beyond that, constructionism adds that when we racialize, we create real races. If my semantic claims are accurate, then the concept of race does not work like that. Race, as opposed to racialization, is something that by definition must persist beyond the social, if it is to exist at all.

4.3. A Confusion

The arguments I have presented suggest that the biological sciences do not, and likely will not, deliver race in the relevant sense, though they might vindicate genealogical populations. The social sciences, similarly, can deliver racialized groups, but not races. What this adds up to is that the sciences cannot find race. From there it is easy to conclude that the world doesn’t seem to contain anything that fits the operative, ordinary concept of race. If race is not a biological reality, and if it is not a social reality, then it looks like race just is not real. This is racial anti-realism: race does not exist. We have no races. Vaishno Das Bagai had no race, nor did Tamir Rice. Their lives were cut down early by terrible injustices multiplied by an exceptional confusion.

(p.137) Some worry that anti-realism won’t allow us to make sense of and deal with the lived reality of race.15 They argue that racial anti-realism compromises people’s identities, invalidates their experiences, or forces us to ignore racial injustices and neglect our voluminous racial ills by stripping us of the words needed to address them. If race isn’t real, then we shouldn’t talk about race—but if we can’t talk about race, then how can we fight racism?

These are serious charges, and it would be a devastating problem if racial anti-realism could not answer them. So it is important to see that anti-realism does not require us to compromise identity, invalidate experience, or undermine progress in these ways. Anti-realism about race can capture every moral phenomenon that racial realism can. The difference is in how they frame the phenomena. In particular, Lawrence Blum (2002), Tommie Shelby (2005), and others have drawn attention to this solution:

We can talk about real racialized groups even if there are no races.

We can say that anti-black discrimination is discrimination that targets people who have been racialized as black. We can say that white privilege is the unequal benefit that accrues to people racialized as white. We can say that identifying as racially Asian is better understood as identifying as someone who is racialized as Asian. We can direct restorative and reparative programs toward victims of racist action. Anti-realism about race can in this way make sense of every consequence of every action that is based on the belief in race: these realities are facts of racialization, rather than race.

That said, if our belief in race is a mistake, we will have to confront the fact that lot is built on this mistake. Identities, communities, and more have been based on the premise that race is real. So the risk in anti-realism is not that we’ll be unable to attend to injustice, but that once we accept our mistake, we might have to dismantle certain entrenched parts of our social lives, and not necessarily in a healthy way (Outlaw 1996). In particular, if we’re not going to believe in race in the currently operative, ordinary sense, what will happen to our identities?

Early waves of racial anti-realism recommended simply eliminating the concept of race (e.g., Appiah 1985), and others have recommended replacing race with neighboring categories like socioancestry (McPherson 2015) or (p.138) genogroup (Montagu 1964, 23). I have argued that it would be better to repair the mistake and preserve something very close to our present racial identities and communities, but in a new and improved way (Glasgow 2009, Ch. 7). To do this, we’d have to attend to not only the massive pain, abuse, and injustice that has been inflicted in the name of race, but also to the mistake upon which that name rests. Two pieces of the concept of race, namely that ‘race’ is defined non-socially and in terms of visible traits, have conspired to make our belief in race a mistake. So to fix the mistake, we must change the definition of ‘race’ accordingly. In particular, ‘a race’ should be redefined to mean something like a socially racialized group. This is one kind of revolutionary theory: we now ask, not what our concept of race is, but what we should want it to be. Given that our most pressing needs are to address that pain, abuse, and injustice, my view has been that we should redefine ‘race’ to directly and explicitly capture our social practices. I call this version of revolutionary theory ‘reconstructionism.’

To emphasize: ‘a race’ does not currently mean a socially racialized group. We have seen that races are not racialized groups, because race—as presently defined—is supposed to persist even when social racialization expires. But we can always stage a semantic revolution and change our terms’ meanings. If we eliminate the non-social element in the definition of ‘race,’ then race could be as the constructionists think it is now. On the reconstructionist picture, race-now is an illusion, and in particular it is not a social reality; but race-future could be socially real, if we redefine our words. Constructionism would become true, even though it is false presently.

Such a change in our definition of ‘race’ would require a significant shift in how we understand race, a shift so profound that it destabilizes the conceptual core of racial discourse. We would have to start accepting that a baby could start out Asian, and then literally stop being Asian if some adult on the other side of the world dies. We would have to accept that race would no longer persist beyond the social. But accepting tomorrow what is perplexing today is the price of revolution.

4.4. Could Race Be a More Basic Kind of Thing?

I have found this picture of anti-realism and reconstructionism compelling for more than a decade. I remain confident that if race is not real, then we ought to fix our mistaken belief in race by embracing conceptual revolution. But another theory of race has made me question whether anti-realism really is true.

(p.139) Several years ago, a then-student of mine, Jonathan Woodward, asked this question: Granting that race is not a biological or social thing, why does that mean it couldn’t exist? We eventually called this idea basic racial realism: rather than race being biologically real, socially real, or illusory, this view says that race is real in a way that is more ‘basic’ than what science aspires to (Glasgow and Woodward 2015).

Science, be it social or natural science, is in the business of finding out how things work and identifying what must exist in order to explain how things work. The sciences tell us that there are quarks and genes and genders, but no mermaids or phlogiston or witches. But, arguably, science is not in the business of finding everything that exists. For example, everyone agrees that science does not and should not focus on what we call sundogs—things that are either suns or dogs. No science cares about sundogs. The sciences care about suns, and they care about dogs. But there is no good reason for science to recognize sundogs as such, since that category has no role in its theories. For science, whether something is {a sun or a dog} is irrelevant. Barring a radical change in the universe, you’ll never see sundogs show up in a biology or sociology textbook. Nevertheless, it sure looks like there are things that are either suns or dogs. Fido is a sundog, because Fido is a dog. So it is plausible to say that sundogs are real, even if they are scientifically irrelevant.

Now obviously sundog is a disunified category—dogs and suns share little in common besides being material objects. There are other categories that are more unified but still scientifically irrelevant. Take the category, stuff around trees. All stuff around trees has another thing in common beyond being stuff, namely being near a tree. But still, science does not care about stuff around trees any more than it cares about sundogs. The category is closer to being scientifically relevant than sundog, particularly because it significantly overlaps with categories that are scientifically relevant, such as fire fuel. But given that some stuff around trees, such as ash from a previous fire, is not fire fuel, the two categories are not the same. Stuff around trees itself is scientifically irrelevant. It plays no role in scientific theories.

Call kinds of things that are unified by some similarity but irrelevant to the sciences basic kinds of things. Because they don’t rise to the level of scientific relevance, basic kinds, including stuff around trees, are not biological kinds or social kinds. Nevertheless, Woodward and I argue that basic kinds are real. After all, they exist. There really is stuff around trees.

You can see where this is going: If there really are basic kinds of things, like stuff around trees, why not say that there are races in this more basic sense, too? The argument for anti-realism held that, if race is neither biologically (p.140) real nor socially real, nor real in any other science, then since there is no other way that it could be real, race is an illusion. But if a kind of thing can be real without being backed by any science, then that argument fails. Race might still be a real, basic kind of thing.

The other realist views are undermined by the conceptual fact that races are supposed to be—on the operative, ordinary definition—certain non-social groups distinguished by their visible traits. Neither biological nor social theories of race can deliver objects that fit that definition. But basic racial realism can avoid this Mismatch Objection: just make the unifying trait of race visible traits, and don’t add anything social. What makes each race a race, then, will be that it has a distinctive visible profile. (What we will identify as races are groups that we perceive to have this profile.) Basic realism can deliver groups that have the exact features that are embedded in the definition of race.16

These visible traits are biologically real traits. Nonetheless, basic races are not biologically real, for the boundaries dividing us into races are still biologically arbitrary. And you might hesitate there. Why not just go ahead and call them biological groups, if they are based on biological traits?

It is worth keeping in mind that calling races biologically real is potentially dangerous. One study showed that prompting people to think about race in certain biological ways leads them to more broadly accept inequality between the races and to reduce interracial interaction (Williams and Eberhardt 2008); another found that biological race-thinking correlates with racist attitudes (Glasgow, Shulman and Covarrubias 2009). But not all studies reveal such dangers (Shulman and Glasgow 2010), and just because a fact is dangerous, that does not make it any less of a fact. So a more direct consideration is simply that when calling something ‘real’ for a science, it should somehow be useful to that science (Spencer 2012). Though basic races are unified by traits that considered independently of race might be useful for biology, the ways of demarcating within those traits, and the groups themselves—races—are not useful for biology. So there are no biological races, even if the traits that give character to basic races are biological.

(p.141) To be sure, sometimes basic kinds can step in and act as imperfect proxies for scientifically relevant kinds of things. A firefighter might shout, “Clear that stuff away from the tree!” during a fire. She would be putting her point in terms of stuff around trees, but that’s just paraphrasing. Really she would be talking about fire fuel. (If ash or a pail of water was near the tree, she wouldn’t care if it was removed.) Similarly, it might be easier to gather data on people in a basic racial group than it is to find data on a significantly overlapping disease population; so to simplify, the research community might settle for using basic race as an imperfect proxy for the disease population. But this is just a proxy relationship. The basic racial group is not itself scientifically relevant. It just is easier to find than a scientifically relevant kind that it overlaps with to a significant degree.

4.4.1. Choosing Fences

The racial classifications that we ended up with sprung from a stew of ignorance and faulty human cognition, seasoned with the corrupting motivations of power, wealth, and status. In particular, as Coates emphasizes, Americans generated the one-drop rule, according to which having any black ancestors makes you black. Barack Obama, who has one black and one white parent and whose visible traits are basically right in the middle of The Spectrum, regularly ends up classified as black. This classification is the result of our perceptual capacities being trained to represent people in a way that adheres to the one-drop rule (Alcoff 2006).

We could have settled on other potential boundaries for social regulation; the one-drop rule was just the one that stuck in this one community. We could have adopted a reverse one-drop rule, as Haiti reportedly did (Mukhopadhyay et al. 2013, 170; cf. Mills 1998: 46–47). If we did—if our perception had been trained differently, with all the implications for social practice that go with that—then we would perceive Obama to be as obviously white as he is obviously black to us now. We took one way of carving humanity and made it socially relevant. We now seem more willing than we have in a long while to recognize mixed-race or multiple-race identity. The next time someone like Barack Obama is elected, the New York Times might announce that we elected a mixed-race president, not a black president (Nagourney 2008).17

(p.142) In this way, basic racial realism says that President Obama belongs to at least three races, which have uneven amounts of social salience. He belongs to one basic race—one collection of people organized according to one set of scientifically irrelevant defining visible traits—called ‘black,’ which comes into focus when we use the one-drop rule. He simultaneously belongs to another, called ‘white,’ which would be relevant if we used the reverse one-drop rule. And he simultaneously belongs to a third, ‘mixed-race,’ which is socially operational in contexts where we recognize being mixed-race. The first identity is the one that the United States has prioritized historically, but all three are equally real in the basic sense. That is, while the social adoption of the one-drop rule has had tremendous implications for the way lives go in the United States, from the perspective of basic racial realism, it does not change the fact that Obama is equally a member of all three races, and many more. There are multiple, cross-cutting ways of dividing us up by visible traits. None is better or worse from the perspective of what is real, though some better capture our social realities. It is therefore very likely that you, reader, are also part of multiple basic races. The one(s) you identify with is (are) just the one(s) that we have vested with social relevance.18

Basic races are out there in the world. We do not invent them. They are not social constructions. We have our basic races regardless of social relevance. Their lines are already drawn: we each look a certain way, and because of that we resemble some people more than others. However, we do get to choose which boundaries, which basic races, we care about. Faced with The Spectrum, we chose how to divide humanity. Many possibilities are given by the world; we decide where to put society’s fences, and we could move them yet again if we like. All the way through, we’d be tracking real but basic, non-biological, unconstructed kinds of thing. We’d just change which ones we care to track.

Because basic races exist independently of any of our social practices, basic racial realism more closely hews to the operative ordinary concept of race than (p.143) constructionism does. As common sense says, if we achieve perfect equality of power, (basic) race continues to exist, because we still have our skin colors and other visible features. Likewise, if we achieve economic equality, (basic) race continues to exist. If all the adults die off and take all their racial practices with them, a baby born Asian remains Asian. And before we ever got our idea of race, way back when on the savannah and in the caves, in our ancestors’ little bands where they had no idea of the amount of human variation or of the pain and division that would eventually accompany it, they had races in the basic sense, too.

So says basic racial realism, anyway.

4.5. Is Race Real?

So in the end, what is race? What is this division in which we invest so much that we need the Supreme Court to tell us where the fence is? This feature that meant Vaishno Das Bagai unwittingly traded British oppression for American oppression? This concept that says that our first mixed-race president is our first black president? This ghostly idea that hovers over Tamir Rice’s two-second murder?

It is tempting to think that race might be a biological, sociopolitical, or sociocultural reality. My coauthors have demonstrated that there are real strengths to all three of these views. But I believe that the balance of arguments pushes us to conclude that race is neither social nor biological. That said, to know what race is not is not to know what race is. Is it an imagined piece of biology, or is it a real but scientifically irrelevant, basic kind of thing? The answer lies buried in the meaning of ‘race,’ somewhere in a partially concealed web of intentions about how to use racial terminology.

We stand again at the crossroads: Which is our deepest commitment? If, on the one hand, our racial discourse is most committed to the idea that racial groups are biological, non-social entities organized by visible traits, then race is an illusion. On the other hand, if the essential idea is just that racial groups are based on (biological) traits like skin color, then races can be real, basic kinds of things—otherwise random, intrinsically unimportant, and unscientific collections of individual people that we spin up into socially momentous entities. We need hypothetical thought experiments no longer, for the actual world has given us the test on this one: there are no visible-trait-based biological racial groups, but there are traits that could be the basis for real, basic races. Is that enough to have race?

(p.144) On this question, I’m afraid that I am at a loss. All I have are weak and wavering leanings about which of these commitments is entrenched in the meaning of the word ‘race.’ It may be that we have some conversations in which we deploy one meaning of ‘race’ and other conversations where we deploy the other, allowing basic racial realism to be true for some conversations while racial anti-realism is true for others. It might be that we have not taken a stand either way, in any conversation, in which case ‘race’ is semantically indeterminate on this question, meaning that there simply is no fact of the matter whether basic realism or anti-realism better fits what we mean by ‘race.’ Or it may be, instead, that there is a determinate, decisive answer in one direction that I am not seeing. Perhaps you can do better at navigating through this particularly heavy fog.


Bibliography references:

Alcoff, Linda Martín. 2006. Visible Identities: Race, Gender, and the Self. Oxford: Oxford University Press.

Andreasen, Robin O. 1998. “A New Perspective on the Race Debate.” British Journal for the Philosophy of Science 49: 199–225.

Andreasen, Robin O. 2000. “Race: Biological Reality or Social Construct?” Philosophy of Science 67(Supplement): S653–S666.

Andreasen, Robin O -. 2004. “The Cladistic Race Concept: A Defense.” Biology and Philosophy 19: 425–442.

Andreasen, Robin O. 2005. “The Meaning of ‘Race’: Folk Conceptions and the New Biology of Race.” The Journal of Philosophy 102: 94–106.

“An Indian Merchant’s Suicide: Patriotic Protest against Racial Discrimination.” 1928, May 22. The Hindustan Times, p. 5. Reproduced at: https://www.saada.org/item/20130709-2976. Last accessed August 7, 2018.

Appiah, K. Anthony. 1985. “The Uncompleted Argument: Du Bois and the Illusion of Race.” Critical Inquiry 12: 21–37.

Appiah, K. Anthony. 1991. “Social Forces, ‘Natural Kinds.’” In Abebe Zegeye, Leonard Harris, and Julia Maxted (eds.), Exploitation and Exclusion: Race and Class in Contemporary US Society. London: Hanz Zell, pp. 1–13.

Appiah, K. Anthony. 1996. “Race, Culture, Identity: Misunderstood Connections.” In Appiah and Gutmann, Color Conscious: The Political Morality of Race. Princeton, NJ: Princeton University Press, pp. 30–105.

Arthur, John. 2007. Race, Equality, and the Burdens of History. Cambridge: Cambridge University Press.

Atkin, Albert. 2017. “Race, Definition, and Science.” In Naomi Zack (ed.), The Oxford Handbook of Philosophy and Race. Oxford: Oxford University Press, pp. 140–149.

(p.145) Bagai, Rani. n.d. “‘Bridges Burnt Behind’: The Story of Vaishno Das Bagai.” http://www.aiisf.org/stories-by-author/876-bridges-burnt-behind-the-story-of-vaishno-das-bagai. Accessed January 5, 2016.

Blum, Lawrence. 2002. “I’m Not a Racist, but . . .”: The Moral Quandary of Race. Ithaca, NY: Cornell University Press.

Burchard, Esteban Gonzalez, Elad Ziv, Natasha Coyle, Scarlett Lin Gomez, Hua Tang, Andrew J. Karter, Joanna L. Mountain, Eliseo J. Perez-Stable, Dean Sheppard, and Neil Risch. 2003. “The Importance of Race and Ethnic Background in Biomedical Research and Clinical Practice.” The New England Journal of Medicine 348: 1170–1175.

Carnap, Rudolf. 1955. “Meaning and Synonymy in Natural Languages.” Philosophical Studies 6: 33–47.

Chalmers, David J. 2011. “Verbal Disputes.” Philosophical Review 120: 515–566.

Chalmers, David J. 2012. Constructing the World. Oxford: Oxford University Press.

Cillizza, Chris. 2014a. “Is Barack Obama ‘Black’? A Majority of Americans Say No.” The Washington Post, April 14. https://www.washingtonpost.com/news/the-fix/wp/2014/04/14/is-barack-obama-black/

Cillizza, Chris. 2014b. “The Next America Poses Challenges, Opportunities for Politicians.” The Washington Post, April 13. https://www.washingtonpost.com/politics/the-next-america-presents-challenges-opportunities-for-politicians/2014/04/13/66d72e3e-c311-11e3-b574-f8748871856a_story.html

Chesnutt, Charles W. 1889. “What Is a White Man?” The Independent 41 (May 30): 5–6. Reprinted at the Charles Chesnutt Archive: http://www.chesnuttarchive.org/works/Essays/whiteman.html. Accessed October 28, 2015.

Coates, Ta-Nehisi. 2015. “Letter to My Son.” The Atlantic, July 4. http://www.theatlantic.com/politics/archive/2015/07/tanehisi-coates-between-the-world-and-me/39D7619/

Condit, Celeste M. 2005. “‘Race’ Is Not a Scientific Concept: Alternative Directions.” L’Observatoire de la génétique 24.

Diamond, Jared. 1994. “Race without Color.” Discover 15(11): 82–89.

Dobzhansky, T. 1941. “The Race Concept in Biology.” Scientific Monthly 52: 161–165.

Du Bois, W. E. B. 1987 [1897]. “The Conservation of Races.” Reprinted in David W. Blight and Robert Gooding-Williams (eds.), The Souls of Black Folk. Boston: Bedford, pp. 228–238.

Feldman, M. W., R. C. Lewontin, and M. King. 2003. “Race: A Genetic Melting Pot.” Nature 424: 374.

Franklin, Ruth. 2015. “Trial and Error: Three Centuries of American Witch Hunts.” Harper’s 331(1986): 89–94.

Glasgow, Joshua. 2003. “On the New Biology of Race.” The Journal of Philosophy 100: 456–474.

Glasgow, Joshua. 2008. “On the Methodology of the Race Debate: Conceptual Analysis and Racial Discourse.” Philosophy and Phenomenological Research 76: 333–358.

(p.146) Glasgow, Joshua. 2009. A Theory of Race. New York: Routledge.

Glasgow, Joshua. 2010. “Another Look at the Reality of Race, by Which I Mean Racef.” In Allan Hazlett (ed.), New Waves in Metaphysics. New York: Palgrave Macmillan, pp. 54–71.

Glasgow, Joshua. Forthcoming a. “Conceptual Revolution.” In Teresa Marques and Åsa Wikforss (eds.), Shifting Concepts: The Philosophy and Psychology of Conceptual Variability. Oxford: Oxford University Press.

Glasgow, Joshua. Forthcoming b. “‘Race’ and Description.” In Quayshawn Spencer (ed.), The Race Debates. Oxford: Oxford University Press.

Glasgow, J., J. Shulman, and E. Covarrubias. 2009. “The Ordinary Conception of Race in the United States and Its Relation to Racial Attitudes: A New Approach.” Journal of Cognition and Culture 9: 15–38.

Glasgow, Joshua, and Jonathan Woodward. 2015. “Basic Racial Realism.” Journal of the American Philosophical Association 1: 449–466.

Gracia, Jorge J. E. 2005. Surviving Race, Ethnicity, and Nationality: A Challenge for the Twenty-First Century. Lanham, MD: Rowman & Littlefield.

Gross, Ariela J. 2008. What Blood Won’t Tell: A History of Race on Trial in America. Cambridge, MA: Harvard University Press.

Haney López, Ian F. 1996. White by Law: The Legal Construction of Race. New York: New York University Press.

Haney López, Ian F. 2005. “Race on the 2010 Census: Hispanics and the Shrinking White Majority.” Dædalus 134: 42–52.

Hardimon, Michael O. 2003. “The Ordinary Concept of Race.” The Journal of Philosophy 100: 437–455.

Hardimon, Michael O. 2017. Rethinking Race: The Case for Deflationary Realism. Cambridge, MA: Harvard University Press.

Haslanger, Sally. 2012. Resisting Reality: Social Construction and Social Critique. Oxford: Oxford University Press.

Haslanger, Sally. 2016. “Theorizing with a Purpose.” In Catherine Kendig (ed.), Natural Kinds and Classification in Scientific Practice. London: Routledge, pp. 129–144.

Hart, Jayasri. 1998. Roots in the Sand. Hart Films.

Herbes-Sommers, Christine. 2003. Race: The Power of an Illusion. Episode One: The Difference between Us. California Newsreel.

Hirschfeld, Lawrence A. 1996. Race in the Making: Cognition, Culture, and the Child’s Construction of Human Kinds. Cambridge, MA: MIT Press.

Hull, David L. 1998. “Species, Subspecies, and Races.” Social Research 65: 351–367.

Jorde, Lynn B., and WStephen P. Wooding. 2004. “Genetic Variation, Classification, and ‘Race.’” Nature Genetics 36(Supplement): S28–S33.

Kaplan, Jonathan Michael. 2010. “When Socially Determined Categories Make Biological Realities: Understanding Black/White Health Disparities in the U.S.” The Monist 93: 281–297.

(p.147) Keita, S. O. Y., and Rick A. Kittles. 1997. “The Persistence of Racial Thinking and the Myth of Racial Divergence.” American Anthropologist 99: 534–544.

Keita, S. O. Y., R. A. Kittles, C. D. M. Royal, G. E. Bonney, P. Furbert-Harris, G. M. Dunston, and C. N. Rotimi. 2004. “Conceptualizing Human Variation.” Nature Genetics 36(Supplement): S17–S20.

Kitcher, Philip. 1999. “Race, Ethnicity, Biology, Culture.” In L. Harris (ed.), Racism. Amherst, NY: Humanity Books, pp. 87–117.

Krogstad, Jens Manuel, and D’Vera Cohn. 2014. “U.S. Census Looking at Big Changes in How It Asks about Race and Ethnicity.” The Pew Research Center, March 14. http://www.pewresearch.org/fact-tank/2014/03/14/u-s-census-looking-at-big-changes-in-how-it-asks-about-race-and-ethnicity/

Lambah, Mool Chand. 1920. “Caste Certificate for Vaishno Das Bagai.” https://www.saada.org/item/20130701-2900. Last accessed January 5, 2016.

Ludwig, David. 2015. “Against the New Metaphysics of Race.” Philosophy of Science 82: 244–265.

Mallon, Ron. 2006. “‘Race’: Normative, Not Metaphysical or Semantic.” Ethics 116: 525–551.

Mallon, Ron. 2017. “Social Construction and Achieving Reference.” Noûs 51: 113–131.

Markus, Hazel Rose, and Paula M. L. Moya. 2010. Doing Race. New York: W. W. Norton.

Mayr, Ernst. 2002. “The Biology of Race and the Concept of Equality.” Dædalus 131: 89–94.

McNish, Emily. 2013. “Rani Bagai on Vaishno Das Bagai.” https://www.saada.org/item/20130821-3099. Last accessed January 5, 2016.

McPherson, Lionel K. 2015. “Deflating ‘Race.’” Journal of the American Philosophical Association 1: 674–693.

McPherson, Lionel K., and Tommie Shelby. 2004. “Blackness and Blood: Interpreting African American Identity.” Philosophy and Public Affairs 32: 171–192.

Mills, Charles. 1998. Blackness Visible: Essays on Philosophy and Race. Ithaca, NY: Cornell University Press.

Millstein, Roberta L. 2015. “Thinking about Populations and Races in Time.” Studies in History and Philosophy of Biological and Biomedical Sciences 52: 5–11.

Montagu, Ashley M. 1964. The Concept of Race. New York: The Free Press.

Mukhopadhyay, Carol C., Rosemary Henze, and Yolanda T. Moses. 2013. How Real Is Race? A Sourcebook on Race, Culture, and Biology, 2nd edition. Lanham, MD: Rowman & Littlefield.

Nagourney, Adam. 2008. “Obama Elected President as Racial Barrier Falls.” The New York Times, November 5. http://www.nytimes.com/2008/11/05/us/politics/05elect.html?pagewanted=all

“Nose Diamond Latest Fad; Arrives Here from India.” September 1915. The San Francisco Call & Post. Reproduced at: https://www.saada.org/item/20130508-2734. Last accessed August 7, 2018.

(p.148) Outlaw, Lucius. 1996. On Race and Philosophy. New York: Routledge.

Pierce, Jeremy. 2015. A Realist Metaphysics of Race: A Context-Sensitive, Short-Term Retentionist, Long-Term Revisionist Approach. Lanham, MD: Lexington Books.

Pigliucci, Massimo. 2013. “What Are We to Make of the Concept of Race? Thoughts of a Philosopher-Scientist.” Studies in History and Philosophy of Biological and Biomedical Sciences 44: 272–277.

Pigliucci, Massimo, and Jonathan Kaplan. 2003. “On the Concept of Biological Race and Its Applicability to Humans.” Philosophy of Science 70: 1161–1172.

Piper, Adrian. 1992. “Passing for White, Passing for Black.” Transition 58: 5–32.

Relethford, John H. 2017. “Biological Anthropology, Population Genetics, and Race.” In Naomi Zack (ed.), The Oxford Handbook of Philosophy and Race. Oxford: Oxford University Press, pp. 160–169.

Risch, Neil, Esteban Burchard, Elad Ziv, and Hua Tang. 2002. “Categorization of Humans in Biomedical Research: Genes, Race, and Disease.” Genome Biology 3: 1–12.

Root, Michael. 2000. “How We Divide the World.” Philosophy of Science 67: S628–S639.

Sainsbury, Mark. 2014. “Fishy Business.” Analysis 74: 3–5.

Sarich, Vincent, and Frank Miele. 2004. Race: The Reality of Human Differences. Boulder, CO: Westview Press.

Shelby, Tommie. 2005. We Who Are Dark: The Philosophical Foundations of Black Solidarity. Cambridge, MA: Harvard University Press.

Shulman, Julie L., and Joshua M. Glasgow. 2010. “Is Ordinary Race-Thinking Biological or Social, and Does It Matter for Racism?” Journal of Social Philosophy 41: 244–259.

Sidelle, Alan. 2007. “The Method of Verbal Dispute.” Philosophical Topics 35: 83–113.

Spencer, Quayshawn. 2012. “What ‘Biological Racial Realism’ Should Mean.” Philosophical Studies 159: 181–204.

Spencer, Quayshawn. 2014. “A Radical Solution to the Race Problem.” Philosophy of Science 81: 1025–1038.

Stalnaker, Robert. 2002. “Common Ground.” Linguistics and Philosophy 25: 701–721.

Sullivan, Shannon. 2013. “Inheriting Racist Disparities in Health: Epigenetics and the Transgenerational Effects of White Racism.” Critical Philosophy of Race 1: 190–218.

Sundstrom, Ronald R. 2002a. “Race as a Human Kind.” Philosophy and Social Criticism 28: 91–115.

Sundstrom, Ronald R. 2002b. “‘Racial’ Nominalism.” Journal of Social Philosophy 33: 193–210.

Taylor, Paul. 2004. Race: A Philosophical Introduction. Cambridge: Polity Press.

Templeton, A. R. 1998. “Human Races: A Genetic and Evolutionary Perspective.” American Anthropologist 100: 632–650.

Williams, Melissa W., and Jennifer L. Eberhard. 2008. “Biological Conceptions of Race and the Motivation to Cross Racial Boundaries.” Journal of Personality and Social Psychology 94: 1033–1047.

(p.149) Witherspoon, D. J., S. Wooding, A. R. Rogers, E. E. Marchani, W. S. Watkins, M. A. Batzer, and L. B. Jorde. 2007. “Genetic Similarities within and between Human Populations.” Genetics 176: 351–359.

Yudell, Michael, Dorothy Roberts, Rob DeSalle, and Sarah Tishkoff. 2016. “Taking Race out of Human Genetics.” Science 351: 564–565.

Zack, Naomi. 2002. Philosophy of Science and Race. New York: Routledge.


(1.) One report from 1915 had the family holding $17,000 in cash; another from 1928 had them carrying $25,000 in gold, the equivalent of over half a million dollars today.

(2.) 260 U.S. 178 (1922)—Takao Ozawa v. United States; 261 U.S. 204 (1923)—United States v. Bhagat Singh Thind; “An Indian Merchant’s Suicide: Patriotic Protest against Racial Discrimination”; Bagai (n.d.); Chesnutt (1889); Franklin (2015); Gross (2008); Haney López (1996); Lambah (1920); McNish (2013); “Nose Diamond Latest Fad; Arrives Here from India.” Details were also provided via personal communication with Rani Bagai.

(3.) That said, even mistaken beliefs about what a word means can have real consequences. An 1818 jury ruled that for the purposes of the New York City fish tax, whales are fish on the grounds that a fish, by definition, is any sea creature (Sainsbury 2014).

(4.) This is misleading, strictly speaking. Shared meaning is required for conversation when we “take it for granted that the contents of [our] speech acts are true or false” (Stalnaker 2002, 702). But in those conversations where we merely presuppose something false, we can have conversation without shared meaning (see Stalnaker 2002 for more). My arguments here implicate the former diagnosis of our racial conversation, not the latter.

(5.) Lionel McPherson (2015) is skeptical about our prospects for landing on one overarching meaning of ‘race’ to stably ground conversations about race. What follows here will suggest otherwise. To emphasize the present point, though, this does not imply that there is only one meaning of ‘race.’

(6.) This definition is, of course, unoriginal. A very wide range of other commentators, from A(lcoff 2006) to Z(ack 2002), also identify visible traits as essential to race.

(7.) I’m happy to adopt Spencer’s (2012) way of understanding useful to good biological science, namely as being valid in well-ordered scientific research programs. My argument aims to be neutral on that question.

(10.) The Mismatch Objection is not the only reason to be concerned about genealogical theory. For example, some worry that the data used to back it up are not taken from globally representative samples (Witherspoon et al. 2007).

(11.) David Ludwig (2015) argues that because each side in the race debate can adopt its own semantics of race, each side can be correct within the definition of ‘race’ it privileges (cf. McPherson 2015), and it has long been noted both that there are different specifications of what race is supposed to be and that each specification has its own consequences for the status of race (e.g., Appiah 1996; Glasgow 2010; Mallon 2006). But when we decide to figure out if the term ‘race’ in ordinary discourse refers to anything real in the world, then not all definitions are equal: we need the meaning operative in ordinary discourse (cf. Chalmers 2011; Sidelle 2007). And this decision matters: the ordinary meaning of ‘race,’ whether race in that sense is real, and if so what its nature is, are questions that are both interesting in themselves and consequential insofar as the answers impact policies, actions, and lives (Glasgow 2009, Ch. 3; Haslanger 2012, Ch. 10). However, as we will revisit below, none of this means that we are necessarily right when we try to analyze the ordinary concept of race—the contents of our concepts are often hidden to us (and see Glasgow 2010; Glasgow forthcoming a, forthcoming b; Haslanger 2012).

(12.) See Gross (2008) for a penetrating discussion of how, though treated differently, Italians and other groups with contested whiteness were always white legally.

(13.) For elaboration, see Glasgow (2009, esp. Ch. 6).

(14.) Other constructionist theories abound. Just in philosophy this camp includes Alcoff (2006); Gracia (2005); Mills (1998); Outlaw (1996); Piper (1992); Root (2000); Sundstrom (2002a); and Taylor (2004), among others.

(15.) See, for example, Du Bois (1897); Gracia (2005, 93, 97–99, 144); Hardimon (2003); Haslanger (2012, 199); Outlaw (1996); Sundstrom (2002b); Taylor (2004, 126).

(16.) You might wonder: if race is based on visible traits, how can people who have atypical traits for their group still be members of those groups? The short answer is that ancestry or some other factor could determine the racial membership of each individual even if ancestry or that other factor does not make a group of people into a race (Glasgow 2009, 79; Hardimon 2017, sec. 2.3; cf. Mills 1998).

(17.) The numbers are moving toward recognizing Obama as being mixed race (Cillizza 2014a). More generally, the one-drop rule may be losing its sway (Glasgow, Shulman, and Covarrubias 2009; Cillizza 2014b).

(18.) Pierce (2015) independently developed a view similar to basic racial realism that he calls “biological constructionism.” The main difference is that, while basic racial realism says we have our basic races even if they are not socially relevant, Pierce’s biological constructionism says that without the social relevance, those groups are not races (Pierce 2015, 85, 90). (We also disagree on terminology: he thinks that to call a view ‘constructionist,’ it is enough if it says that races are made socially significant (74, 89); I think constructionists must say that society makes races exist.) Haslanger (2012, 302, 306) very briefly considers a view like basic racial realism as well (and also see Haslanger 2016), but then she reiterates her constructionist view that the defining features of race are not visible traits but social facts.