Glasgow Et Al - What Is Race Four Philosophical Views (2019)

i
W H AT I S R A C E ?
ii
iii
WHAT IS RACE?
Four Philosophical Views
Joshua Glasgow
Sally Haslanger
Chike Jeffers
Quayshawn Spencer
1
iv
1
Oxford University Press is a department of the University of Oxford. It furthers
the University’s objective of excellence in research, scholarship, and education
by publishing worldwide. Oxford is a registered trade mark of Oxford University
Press in the UK and certain other countries.
Published in the United States of America by Oxford University Press

198 Madison Avenue, New York, NY 10016, United States of America.
© Oxford University Press 2019
All rights reserved. No part of this publication may be reproduced, stored in

a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by license, or under terms agreed with the appropriate reproduction
rights organization. Inquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above.
You must not circulate this work in any other form

and you must impose this same condition on any acquirer.
Library of Congress Cataloging-in-Publication Data

Names: Glasgow, Joshua, author. | Haslanger, Sally Anne, author. |
Jeffers, Chike, 1982– author.
Title: What is race? : four philosophical views / Joshua Glasgow, Sally Haslanger,
Chike Jeffers, Quayshawn Spencer/
Description: New York : Oxford University Press, 2019. |
Includes bibliographical references and index.
Identifiers: LCCN 2018053834 (print) | LCCN 2019007102 (ebook) |
ISBN 9780190610210 (Online content) | ISBN 9780190610197 (updf ) |
ISBN 9780190610203 (epub) | ISBN 9780190610173 (hardback) |
ISBN 9780190610180 (paperback)
Subjects: LCSH: Race--Philosophy. | Race relations--Philosophy. |
BISAC: PHILOSOPHY / Political. | PHILOSOPHY / Metaphysics. |
PHILOSOPHY / Movements / General.
Classification: LCC HT1523 (ebook) | LCC HT1523 .G643 2019 (print) |
DDC 305.8001—dc23
LC record available at https://lccn.loc.gov/2018053834
9 8 7 6 5 4 3 2 1
Paperback printed by Sheridan Books, Inc., United States of America

Hardback printed by Bridgeport National Bindery, Inc., United States of America
v
To our children,
Samantha Rose Glasgow-Shulman, Aminata Lilla Jeffers, Ayo Jelani
Jeffers, Aza Katherine Ida Jeffers, Julian Buo-Hon Spencer, Quentin
Buo-Yi Spencer, Isaac Amazu Haslanger Yablo, and Zina Siyasa
Haslanger Yablo,
in the hope that their generations will find a more just world.
vi
vi
CONTENTS
Acknowledgments ix
Introduction 1
1. Tracing the Sociopolitical Reality of Race—S ally Haslanger 4
2. Cultural Constructionism—C hike Jeffers 38
3. How to Be a Biological Racial Realist—Q uayshawn Spencer 73
4. Is Race an Illusion or a (Very) Basic Reality?—J oshua Glasgow 111
5. Haslanger’s Reply to Glasgow, Jeffers, and Spencer 150
6. Jeffers’s Reply to Glasgow, Haslanger, and Spencer 176
7. Spencer’s Reply to Glasgow, Haslanger, and Jeffers 203
8. Glasgow’s Reply to Haslanger, Jeffers, and Spencer 245
Index 275
vi
ix
ACKNOWLEDGMENTS
The theories presented in this book are derived from articles and books
that we have previously published. Numerous people helped us as we devel-
oped those earlier publications. We would like to reaffirm our gratitude for
their help.
With respect to the present work, we gratefully acknowledge the
following:
Sally Haslanger would like especially to thank the other authors of the
book, Joshua Glasgow, Chike Jeffers, and Quayshawn Spencer, for their on-
going work, helpful conversations, and friendly collaboration in putting this
book together. In addition, she would like to thank Lawrence Blum, Jorge
Garcia, Adam Hosein, Karen Jones, Lionel McPherson, Megan Mitchell,
José Jorge Mendoza, Deborah Mühlebach, David Plunkett, Laura Schroeter,
Francois Schroeter, Greg Restall, Tommie Shelby, Isaac Yablo, Stephen Yablo,
and Zina Yablo for ongoing conversations on the topics discussed.
Chike Jeffers would like to thank Linda Martín Alcoff, Veromi Arsiradam,
Tiffany Gordon, Tyler Hildebrand, David Ludwig, Tina Roberts-Jeffers,
Katie Stockdale, and his coauthors Sally Haslanger, Quayshawn Spencer, and
Joshua Glasgow for their feedback on drafts of his chapters in this book. He
also presented versions of Chapter 2 at the University of Memphis and at
Northwestern University. He is thankful to all who were present and espe-
cially to those who provided feedback on those occasions.
Quayshawn Spencer is thankful for feedback from Mariana Achugar,
Linda Alcoff, Elizabeth Anderson, Luvell Anderson, Robert Brandon, Liam
Bright, Mazviita Chirimuuta, Haixin Dang, Michael Devitt, Brian Epstein,
Patrick Forber, Justin Garson, Joshua Glasgow, Michael Hardimon, Sally
Haslanger, Jay Garfield, Chike Jeffers, Saul Kripke, Meena Krishnamurthy,
Edouard Machery, Lionel McPherson, Charles Mills, Sandra Mitchell,
Jennifer Morton, Albert Mosley, Wayne Norman, Sumeet Patwardhan,
George Smith, and Daniel Wodak. He would also like to thank the faculty
x
x • Acknowledgments
and students in the philosophy departments at the following universities for

allowing him to receive feedback on drafts of his chapters at a colloquium
talk: Carnegie Mellon University, College of the Holy Cross, CUNY Graduate
Center, Drexel University, Duke University, the University of Michigan at
Ann Arbor, Smith College, Temple University, Tufts University, University
of Maryland in Baltimore County, and Virginia Polytechnic Institute. Also,
Spencer would like to thank the participants of the University of Pittsburgh’s
Summer Program in Philosophy of Science for Underrepresented Groups in
2017 for all of their useful feedback on Chapter 3, and he’d like to thank the
Center for Philosophy of Science at the University of Pittsburgh for providing
two opportunities to present and receive feedback on Chapter 3. Finally, he’d
like to thank the participants at the 2015 Metaphysics of Social Categories
Workshop at the University of Colorado at Boulder for offering helpful feed-
back on work that would become part of Chapter 3, and the participants at
the ninth annual Ida B. Wells conference at the University of Memphis for
offering helpful feedback on work that would become part of Chapter 7.
Josh Glasgow is thankful for the help of Shani Long Abdallah, Paul
Bloomfield, Matt Ingram, Robert Ingram, Esa Díaz León, Dan López de Sa,
and audiences at the University of Barcelona and University of Geneva. He
is also grateful to Sally Haslanger, Chike Jeffers, and Quayshawn Spencer for
both their collaboration on this project and their ongoing work. Finally, he’d
like to thank Rani Bagai for conversation about Vaishno Das Bagai, Kala
Bagai Chandra, and events surrounding the Thind decision.
1
INTRODUCTION
Historically, efforts to make sense of human diversity led to classi-

fying humans by where they live, by culture or custom (including
language or religion), by family or ancestry, by appearance, by per-
sonality type or temperament, by physical and intellectual capac-
ities, and, of course, by race. But what, if anything, is distinctive
of racial groups? A race is not just a group of people linked by a
common ancestry. While this may account for the idea of the
human race, since all living humans share a common ancestor (so-
called Mitochondrial Eve, who lived about 200,000 years ago), it
also would make any group of people who are biological siblings a
race, which is not how anyone talks about race. Nor is a race just
a group of people linked by a common geographic origin. If that
were true, then African Americans and Afro-Caribbeans would be
American, instead of Black, because both of these groups of people
originated in the Americas, not Africa. And it is difficult to unpack
race in terms of shared skin color, since people of different races
can share the same skin color and people of the same race can have
different skin colors.
So, how should we understand what a race is supposed to be?
This is the question we address in the chapters that follow.
One might think that the best way to answer the question is
to consult current biology. However, whether race is essentially a
biological thing is controversial, as this book demonstrates. For
that matter, there is not even agreement among biologists. In one
recent survey, 45.5% of American biology professors believed that
“[r]‌ace is biological,” 50% believed that “[r]ace is not biological,”
and the remaining 4.5% were uncommitted (Morning 2011, 111).
Moreover, the term ‘race’ has been used in multiple ways over the
centuries, and in some uses, it clearly does not function as a biolog-
ical term (Herzog 1998, Ch. 7; Stevens 1999, Ch. 4). Don Herzog
(1998) characterizes the problem:
2
2 • Introduction
Take Coleridge: “Men of genius are rarely much annoyed by the com-
pany of vulgar people; because they have a power of looking at such
persons as objects of amusement, of another race altogether.” A race of
vulgar or stupid people: is that a genuine or merely metaphoric race?
Does a concept as nefarious and slippery as that of race admit of a gen-
uine or innocent or primary usage and then parasitic and metaphoric
usages? Or are there no clear core cases to which it applies? If there is
a clear core, does it have to do with something revolving around blood
or genetics? Or, as Coleridge seems to suggest, around hierarchy and
contempt? (290)
Each of the four authors included in this book proposes a different account of
race. Quayshawn Spencer (Chapter 3) is a naturalist about race and believes
that there are racial groups distinguished by biological features (e.g., shared
genomic ancestry), but that the biological features in question do not jus-
tify any social hierarchy among the races. Chike Jeffers (Chapter 2) and
Sally Haslanger (Chapter 1) argue in different ways that races are socially
constructed groups. Jeffers holds that the cultural aspect of racial difference—
that is, the ways in which races are groups differentiated by distinctive ways
of life—is underappreciated in its importance to the past and present and is
the centrally important feature when thinking about how we might continue
to construct races in the future, past the end of racism. Haslanger, by con-
trast, argues that races are first and foremost sociopolitical groups, marked
by bodily features, that function within a dominance hierarchy. In the case
of the current social structure in the United States, the dominance hierarchy
is White supremacy, but races are formed within different hierarchies that
aren’t organized around Whiteness. Joshua Glasgow (Chapter 4) maintains
that the concept of race includes, as a necessary condition, that there be a set
of visible features that are disproportionately held by each race. If this condi-
tion is strongly interpreted to require that visible-trait groups are supposed
to be legitimized by biology, then, Glasgow argues, there are no races (ra-
cial anti-realism); but if, instead, this condition is weakly interpreted to mean
simply that there are differences between humans, then race is arguably real in
a non-biological and non-social way (basic racial realism).
After laying out the fundamentals of our views in the first four chapters,
we each reply to our coauthors in Chapters 5 (Haslanger), 6 ( Jeffers), 7
(Spencer), and 8 (Glasgow).
We don’t consider our four views to exhaust all of the plausible metaphys-
ical views one can have about what race is and whether it’s real, nor do we think
3
Introduction • 3
that our particular ways of defending the metaphysical views represented are
the only ways to defend these views. For instance, the arguments over whether
race is biological cover a wider terrain than we survey here. And it’s possible
that race is, essentially, a social grouping of people without necessarily being
a cultural grouping of people (à la Jeffers) or a political grouping of people (à
la Haslanger). For example, Ron Mallon and Daniel Kelly (2012) have argued
that race is a socio-psychological grouping of people.
These limitations notwithstanding, we aim for this book to highlight
some central themes and compelling arguments in a difficult and contested
debate. The history of sorting out the nature of race has been fraught, impli-
cated both in oppression and in efforts to liberate from that very oppression.
Our hope is that by focusing on some promising lines of argument, analysis,
and inquiry, we can facilitate understanding and progress.
References
Herzog, Don. 1998. Poisoning the Minds of the Lower Orders. Princeton, NJ: Princeton
University Press.
Mallon, R., and D. Kelly. 2012. “Making Race Out of Nothing.” In Harold Kincaid
(ed.), The Oxford Handbook of Philosophy of Social Science. New York: Oxford
University Press, pp. 507–532.
Morning, Ann. 2011. The Nature of Race. Berkeley: University of California Press.
Stevens, Jacqueline. 1999. Reproducing the State. Princeton, NJ: Princeton
University Press.
4
1 TRACING THE SOCIOPOLITICAL REALITY

OF RACE
Sally Haslanger
1.1. Methodological Preliminaries
The question before us is: What is race? When we pose questions of
the form “What is X?” there are a variety of ways we might go about
answering them. For example, if, pointing to a small wiggly thing in
the corner, I ask, “What is that?” I will probably want someone to
help me figure out the species of insect it belongs to, as determined by
entomology. If you tell me it is a silverfish, I might also pose a question
about the kind of thing (e.g., “What is a silverfish?”). Plausibly I am
asking how silverfish, as a group, are classified: what features some-
thing must have to count as a silverfish, what to expect of silverfish,
and how they are related to other sorts of creepy crawly things.
These sorts of questions seem to presuppose that we have a
well-developed science that will provide us with empirically based
answers. However, sometimes our “What is it?” questions take us
beyond what science has figured out. For example, if in the sev-
enteenth century someone pointed at a burning log and asked,
“What is that?” a straightforward answer would be “Fire.” But if
the speaker already knew that and proceeded, “But what is fire?”
the question is probably attempting to probe features of fire that
aren’t apparent from our ordinary familiarity with it; and it would
(and did) take substantial empirical research and future scientific
theory to reach any answers.
In the sorts of contexts just considered, it would be, at the very
least, odd to answer the questions by consulting our linguistic
intuitions.1 Our judgments about when to use the term ‘silverfish’
1. Some parts of this section are repeated and developed more fully in Haslanger
(forthcoming).
5
Tracing the Sociopolitical Reality of Race • 5
don’t tell us what a silverfish is. However, there are a variety of “What is X?”
questions that many philosophers seem to think can be answered by discov-
ering the meaning of the term(s) substituted for X, as determined by our
disposition to apply the term(s) in question (e.g., ‘knowledge,’ ‘moral worth,’
‘justice,’ ‘a person,’ ‘causation’). In some of these cases, one might think that
this a priori methodology is warranted because the boundaries of these kinds
depend in some way on us and our practices. Perhaps moral worth, justice,
personhood, and the like, don’t exist independently of our judgments of what
counts as moral worth, justice, and personhood. So, of course, we should at
least begin by investigating our judgments and putting them in order. (This is
more plausible in some cases than in others; e.g., the answer would have to be
more complicated in cases such as ‘causation’ or ‘intrinsic property.’)
But the idea that (some) philosophical kinds somehow “depend on us” is
not entirely clear; nor is it clear why our a priori (linguistic) reflections should
be sufficient to provide an adequate theory of them—for example, “What is a
sheriff ?” Even if you are a competent user of the term ‘sheriff,’ you may not be
able to tell me what a sheriff is. A full answer would presumably require infor-
mation about the jurisdiction of sheriffs, what their responsibilities are, how
they are chosen, etc. as determined by law. We might need to consult experts
in civics to get answers (and the answers will depend on what country we are
in). We can’t just depend on common sense or linguistic intuitions. But surely
what counts as a sheriff depends on us—there are no sheriffs outside of a hu-
manly constructed system of government.
In the case of ‘sheriff,’ there will be a well-defined role specified by statute,
and someone who knows the relevant statutes will know the answers to
our questions. But there are also social phenomena that in some sense “de-
pend on us” but are not stipulated or planned by us. Such social phenomena
range from macro-scale economic depressions, globalization, urbanization,
and gentrification, to more local social practices and relations (e.g., within a
town, religious congregation, or family). These phenomena call for explana-
tion, and the social sciences (broadly construed) endeavor to provide theories
that enable us to understand them, usually identifying kinds of institutions,
economic relations, cultural traditions, social meanings, and psychological
predispositions, to do so. The kinds in question are social kinds, in the sense
that they are kinds of things that exist in the social world (and so, in some
sense, depend on us). But we discover these kinds through empirical inquiry,
just as we discover chemical kinds through empirical inquiry.
For example, accounts of gentrification often make reference to the “urban
pioneer,” sometimes characterized as artists and “bohemians” who take
6
6 • W h a t I s  R a c e ?
advantage of low rents in poor neighborhoods. When single people who share
rent enter a neighborhood, businesses (such as cafés and pubs) take interest,
and landlords see opportunities to raise rents, which drives out the locals.
Urban pioneers are a functional kind that identifies a particular role in an
evolving real estate market. The term ‘pioneer’ is chosen due to the perceived
parallel with pioneers who “settled” the western United States, displacing the
local population. If someone were to object to the term ‘pioneer’—perhaps
thinking that it carried an overly positive connotation—this would not un-
dermine the explanatory claims.2 The adequacy of explaining gentrification
by reference to singles moving into an urban neighborhood does not depend
on our linguistic intuitions about applying the term ‘pioneer’ to them. The
choice of terminology was intended to illuminate a parallel; if the termino-
logical choice doesn’t work, then another term could be used as a substitute.
However, insofar as philosophical kinds such as justice and personhood
“depend on us,” it is not in the sense that we stipulate what they are (like
sheriff), or in the sense that they serve in explanations of social phenomena
(like urban pioneers). Rather, it is something along these lines: the adequacy
of our theory is not to be judged simply by reference to “the facts,” but also
by its responsiveness to our prior understandings. In the case of sheriff, you
might think that there aren’t any independent facts we’re trying to accommo-
date. Oversimplifying, we simply create sheriffs and then talk about them. In
the case of urban pioneer, the prior understandings of ‘pioneer’ are not crucial
to the explanation provided by the theory. But in the case of justice, there is
something we are aiming to understand that is not simply constituted by what
we say, but at the same time, our conclusions cannot float completely free of
the discursive tradition in which we are aiming to understand it.
How might we explain this? Note that in the philosophical cases, we are
not situated as anthropologists trying to understand the social life of the
“natives.” Nor are we legislators specifying new practices. We are seeking an
understanding of practices in which we are currently engaged as participants.
The practices are not fully understood, however. And they are open-ended,
revisable, possibly self-defeating. In making sense of them, we are making
judgments about how to better understand what we are doing, and how then
to go on. This is not primarily a linguistic exercise: we aren’t just deciding how
2. Metaphors and analogies can play an important and even ineliminable role in theorizing and
can aid in explanation. My claim here is only that the choice of terminology for the functional
kinds in the proposed mechanisms of gentrification (specifically the influx of singles) is not es-
sential to the success of the model for some purposes (though it may be for others).
7
to use existing terminology, but how to collectively orient ourselves toward

the world and toward each other. Language provides tools to achieve this.
But language is a practice within practices and is itself a proper target of phil-
osophical inquiry: meanings are not simply constituted by what we believe,
yet we are situated within a tradition of linguistic practices that have already
shaped our meanings and our world; so ignoring those practices would be a
mistake. We are situated inquirers, and the question is how we should go on,
given where we have been, where we are now, and where we are trying to go
(Lear 1986).
1.2. What Is the Question?

The question arises, then, what sort of question is at issue when we ask, “What
is race?” Is it an empirical question that we should answer using the methods
of biology? Or should we use the methods of empirical sociology or history?
Is it a question about what ‘race’ means? And how might one determine the
meaning of ‘race’? Do we get to stipulate the meaning? Are we seeking a phil-
osophical tool for explanatory purposes? Or is the question best understood
as arising for us as participants in racializing practices?
I don’t think there is one right way to pose the question, “What is race?”
In fact, I think it is useful to ask different versions of the question in order
to understand the phenomenon, and different forms of the question, raised
in different contexts, will call for different answers.3 In my own work (e.g.,
Haslanger 2012, Ch. 7), I have explored the question as a critical theorist.
There are multiple ways of characterizing critical theory, but for our purposes
here, I will draw on Tommie Shelby’s characterization of a social critic:
There is also the discourse of the social critic, which is identical with
neither everyday discourse nor scientific discourse. Social critics don’t
merely systematize common sense or popularize scientific findings.
Social critics seek to inform, and possibly shape, public opinion with
clear and careful thinking, well-established facts, and moral insight.
They will of course draw on and engage both common sense and
3. I embrace an “eretetic” approach to explanation that takes explanations, and theories more
generally, to be answers to questions. So the first task of any theoretical project is to clarify
the question being asked. Apparent disagreements can sometimes be resolved by noting that
the parties to the disagreement are answering different questions (see Garfinkel 1981; Risjord
2000; Anderson 1995).
8
8 • W h a t I s  R a c e ?
scientific thought, but they do so without taking a slavish attitude to-

ward either . . . [In the context of debates over race and racism,] the
principal role of the philosophical social critic, as here conceived, is to
shed light on the most fundamental conceptual and normative issues
that race-related questions raise. (Shelby 2014, 63)
Plausibly, all inquiry is situated. Inquiry begins with questions, and all
questions have presuppositions. And any serious effort to answer a question
relies on a method that is taken to have at least some epistemic credentials. I’ve
been suggesting that certain forms of philosophical inquiry are situated in an
additional sense; that is, the project is not simply a descriptive or explanatory
project, but aims to shape or guide our thinking and acting. Social critics take
this even a step further: we are situated as critics of ordinary social practices and
offer tools and understandings that are designed to improve them (Fraser 1989;
Marx 1843). The social critic embraces the normative dimension of philosoph-
ical theorizing, and also relies crucially on empirical research. The idea of race is
already embedded in our customs, practices, and institutions, and facts about
its role in our lives are crucial to the critical project. Such empirical information
and normative concerns are also important, on some accounts, for adjudicating
linguistic meaning, and so, in particular, for understanding what ‘race’ means.
1.3. The Semantic Strategy

Quine (1953) has taught us that if we are engaged in an ontological debate
about the existence of some kind of thing, say, races, we should semantically
ascend. In other words, instead of asking directly whether races exist, we
should ask whether the term ‘race’ picks anything out in the world, and if so,
what. This is an especially helpful move if parties to the debate don’t agree on
what the term ‘race’ means, for if, say, a racial realist and a racial anti-realist
have different understandings about what ‘race’ means, then the conflict be-
tween them may be only apparent.4 It may be true, for example, both that
biological races don’t exist and social races do exist.
4. I assume for the purposes of this discussion that a racial realist believes that at least some
statements involving the term ‘race’ are both truth-apt and true. Anti-realists disagree. Anti-
realists may hold that all statements involving the term ‘race’ are not truth-apt (they are “non-
cognitivists” about race talk), or they may hold that race talk is truth-apt, but false (they are
“error theorists” about race). In this book, Spencer, Jeffers, and I are all realists (though we
disagree about what makes race talk true); Glasgow is either an anti-realist error theorist or
a “basic” realist that allows ‘race’ to refer, but to uninteresting groups. (Other anti-realists in-
clude Appiah [1996] and Zack [2002]; Blum [2002]; Hochman [2017].)
9
Ron Mallon (2006, 527) considers this “semantic strategy,” widespread

in the race debate, and argues, however, that it is not helpful. The problem
is that different parties to the debate seem to adhere to different theories of
meaning. So the controversy is just pushed up a level. For example, the realist
and anti-realist, it would seem, don’t agree on how we might determine the
meaning of the term ‘race.’ When the realist claims that ‘race’ refers to a so-
cial kind, and the anti-realist says that ‘race’ does not refer to anything, they
haven’t established a basis for debate because they are committed to different
theories of meaning or reference. As a result, there is a risk that the race debate
just collapses into a debate in the philosophy of language. Given the unlike-
lihood that we will be able to settle on a theory of reference any time soon, it
looks like the race debate is left hanging.
However, Mallon suggests that there are important questions that should
be asked and whose answers shouldn’t depend on a metaphysical or semantic
theory. His proposal is that we take up a normative approach to race. The im-
portant question isn’t the metaphysical one (i.e., whether races exist or not)
or the semantic one (i.e., what ‘race’ means, if anything). Rather, the question
is normative: how we should think and talk when it comes to matters of race.
While there is (or should be) a wide basis of metaphysical agreement on

the expanded ontological consensus, there is profound disagreement
over the practical and moral import of ‘race’ talk. Resolving this dis-
agreement requires a complex assessment of many factors, including,
the epistemic value of ‘race’ talk in various domains, the benefits and
costs of racial identification and of the social enforcement of such
identification, the value of racialized identities and communities fos-
tered by ‘race’ talk, the role of ‘race’ talk in promoting or undermining
racism, the benefits or costs of ‘race’ talk in a process of rectification for
past injustice, the cognitive or aesthetic value of ‘race’ talk, and the de-
gree of entrenchment of ‘race’ talk in everyday discourse. The point is
that it is on the basis of these and similar considerations that the issue
of what to do with ‘race’ talk will be decided, not putative metaphys-
ical or actual semantic disagreements. (Mallon 2006, 550)
I am sympathetic with Mallon’s (2006) suggestion that resolving the issue

depends on normative and empirical considerations.5 However, as he frames
5. For a parallel argument concerning gender terms such as ‘woman,’ and ‘man,’ see Saul (2012)
and Kapusta (2016).
10
10 • W h a t I s  R a c e ?
it, the questions we must answer still concern “race talk.” But recall that, by
hypothesis, the different parties to the race debate disagree about what ‘race’
means because they embrace different accounts of meaning. What counts,
then, as “race talk?” It can’t be identified by talk about race, for we don’t
agree on what race is, or even whether there is such a thing as a race. And we
can’t just consider talk that includes linguistic items pronounced as English
speakers pronounce ‘race,’ for we use that sound also for boat races, running
races, and the like. Although Mallon is right that we need to ask a wide range
of epistemic and moral questions of the sort he lists, his characterization of
the task retains too much of the semantic strategy. What’s at issue isn’t just our
talk and thought, but racial structures and practices of all sorts—linguistic,
cultural, medical, political, juridical. We begin our theorizing already situated
in these practices. What we are trying to do is understand how they work,
what is salvageable (if anything), and how to go on.
Consider a comparison with the notion of moral worth. We begin with
our practices of distinguishing the worth of an action from its consequences.
We seem to be prepared, at least sometimes, to commend an action as good,
even if it has unfortunate consequences, and to condemn an action as bad
even if it has good consequences; yet we don’t have a clear idea of what moral
worth is. For example, if I bring my new neighbor a bouquet of flowers, not
knowing that she has severe allergies, and she suffers as a result, my action
was nevertheless kind and thoughtful, and seems to have moral worth, even
though it had bad consequences. When we ask, “What is moral worth?” we
consider a full range of cases, the presuppositions and effects of this practice
of attributing moral worth, and what function it has. The point is not to look
at “moral-worth-talk,” since the language of ‘moral worth’ is rather rare in
common parlance. We are attempting to capture a set of practices of moral
evaluation. After careful scrutiny, we may find that the feature that seems to
distinguish worthy actions isn’t as valuable as we thought, or the worthy fea-
ture is more rarely present; and this justifies a revision to the practice. If we
are consequentialists, we may find that the practice isn’t justified at all and we
may recommend discontinuation.
As mentioned before, ideas of race are “woven into” many of our everyday
practices (i.e., racial distinctions seem to play a role in so much of what we do,
where we go, with whom we associate, in what resources are available to us
and what is required to access them). This is not to say that race is explicitly
and intentionally functioning in these practices. But our lives are shaped by
a racial geography. As in the example of moral worth, we begin by collecting
a full range of apparent examples, consider their presuppositions and effects,
1
and consider what function they have. What is it, if anything, about these
practices that makes them “racial”? In the contemporary “post-racial” climate,
some will no doubt argue that there is nothing specifically “racial” about them
(e.g., they are to be understood in terms of class). But there is also plenty of
evidence that racial distinctions, racial assumptions, and racial identities con-
tinue to structure our lives together (and apart).
1.4. Representational Traditions: ‘Water’ as an Example

Laura and François Schroeter (2015) offer an account of meaning that not
only seems to be compatible with the spirit of Mallon’s suggestion, but also
situates our linguistic activities within our broader social practices. They focus
on the example of ‘water,’ and suggest that to determine what ‘water’ means,
we should undertake an inquiry into what water is. But how do we do this?
We cannot assume from the start that this is a task for the chemist, for when
the chemist says that water is H2O, she may be using the term in a technical
sense, in which case it would not provide an account of what the ordinary
person means by ‘water.’ (Note that the same might be said of the biologist’s
use of ‘race.’) But neither can we just undertake reflection on linguistic usage
or common sense.
Before you explicitly reflect on the question of what water is, your own
assumptions about the topic are bound to be heterogeneous, incom-
plete, and partially contradictory—and this heterogeneity is only exac-
erbated when you take your whole community’s views into account.
Thus justifying an answer to a ‘what is x?’ question is nothing like slot-
ting some missing values into an implicitly grasped formula. Your goal
in rational deliberation is to find some principled way of prioritizing
and systematizing your own and your community’s commitments
about water, so as to identify the appropriate normative standards
for evaluating the truth and acceptability of beliefs about the topic.
(2015, 430)
The broad idea is this: when we deliberate about what X [water, race, free-
will, moral worth . . .] is, we have to start with something. In the sorts of cases
we are considering, we can take ourselves to be situated within a broad rep-
resentational and practical tradition concerned with X. We are not starting
from scratch and stipulating the meaning of theoretical terms. And we may
assume that the tradition has a certain epistemic ambition, so we may “take
12
12 • W h a t I s  R a c e ?
our words and thoughts to represent genuinely interesting and important

features of the world—not just whatever happens to satisfy our current cri-
teria” (Schroeter and Schroeter 2015, 436). So scientific inquiry, although not
definitive, is relevant, since it discloses some parts of the world that are impor-
tant for many of our purposes. But where do we begin? The Schroeters (2015,
426) give a sample of inputs to deliberation in the case of water (the examples
are theirs):
• Particular instances: there’s water in this bottle, in Port Phillip Bay, Lake
Michigan, etc.;
• Perceptual gestalts: the characteristic look, taste, odor, tactile resistance,
and heaviness of water;
• Physical roles: water’s rough boiling point, its transformation into steam,
its role as a solvent, the fact that it expands when it freezes, etc.;
• Biological roles: water’s necessity for the survival of plants and animals;
how it’s ingested; the effects of water deprivation; etc.;
• Practical roles: the roles water plays in agriculture, transport, washing,
cooking, surfing, etc.;
• Symbolic roles: water is strongly associated with cleanliness and purity, it
plays an important role in many religious rituals, etc.;
• Explanatory roles: water has a non-obvious explanatory structure, which
explains many of its characteristic roles; water is composed of H20;
• Epistemology: water is easy to spot but hard to define; our beliefs about
water may be mistaken or incomplete; observation of instances of water
grounds induction to unobserved cases.
Our aim is to answer to the “What is X?” question. The project is not se-
mantic but meta-semantic; that is, we are not trying to find what the X-term
means. Rather, we are trying to determine what the kind X is. The inputs just
considered help us narrow down the kind so we can investigate it further.
As we proceed, we may find that some of our background beliefs are false
and our theoretical efforts misguided. It is only the result of our investigation
that gives us the meaning of the term. But what do we do with these inputs?
How do we balance various considerations? Schroeter and Schroeter (2015)
propose that
. . . ideal epistemic methods for answering ‘what is x?’ questions hinge

on rationalizing interpretation of one’s representational traditions.
You need to diagnose the most important representational interests at
13
stake in a representational tradition with ‘x’, and you should identify

the correct verdict about the nature of x as the one that makes best
sense of those interests. (2015, 430)
A rationalizing interpretation, on their view, is not determined by reports of

beliefs and intentions of participants in the tradition, nor is it a causal expla-
nation of the tradition:
From the deliberative perspective of a rational epistemic agent, the

interests that are relevant to adjudicating ‘what is x?’ questions are
those that help justify or rationalize that tradition. Ideal methods for
adjudicating ‘what is x?’ questions don’t simply construe representa-
tional practices as meeting psychologically or causally fixed represen-
tational interests. Our interpretive methods construe them as meeting
representational interests that help make sense of our practices—that
help construe them as having a point or rationale. (2015, 435)
In the case of ‘water,’ there are at least two candidates. One set of interests
served by our attitudes toward and talk of water are explanatory, another
set is practical. These two interests may come apart; for example, our prac-
tical interests do not require that we identify water with H2O, for liquids
that are mostly H2O but contain other ingredients (harmless trace chemi-
cals, fluoride) are fine for most purposes (drinking, bathing, swimming, etc.).
However, scientific inquiry enables us to explain the properties of water—
and how it can actually serve our practical interests—by reference to its chem-
ical structure. This divergence of possible interpretations of what’s at stake in
the tradition leaves us with two candidate answers to “What is water?” and so
two candidates for the meaning of ‘water.’ Water is H2O, or water is the wa-
tery stuff found in lakes and rivers (etc.). It might appear that this leaves the
term ‘water’ as ambiguous, or perhaps with no determinate meaning.
On the Schroeters’ (2015) view, there is a best interpretation of the repre-
sentational tradition, where the scope of that tradition is determined by com-
mitment to de jure sameness of reference and shared linguistic and epistemic
practices (428). (We all take ourselves to be referring to the same thing in our
thought and talk and are engaged in talking and thinking together.) What
I mean is not just a function of what I think water is, or any old interpretation
of our representational tradition: I can get the meaning wrong if I don’t do ad-
equate justice to the interpretive task. For example, if I decide that, given our
interests and collective uses of the term, water is the alcoholic beverage also
14
14 • W h a t I s  R a c e ?
known as ‘beer,’ I would be wrong. I would have failed to capture a reasonable

interpretation of our representational tradition. But I could also be wrong if
I miss what is worth talking about:
As rational epistemic agents, we normally take our words and

thoughts to represent genuinely interesting and important features of
the world—not just whatever happens to satisfy our current criteria.
When asking about the nature of water (or free will, color, etc.), we
don’t assume that we (or our community as a whole) already implicitly
know the right answer. (2015, 436)
We postulate ambiguity or opt for an error theory only as a last resort.

The Schroeters’ view seems to me to provide a kind of middle ground
between adopting the “semantic strategy” and moving entirely to norma-
tive considerations. Recall the previous suggestion that in undertaking at
least certain kinds of philosophical inquiry, what we are doing (roughly) is
interpreting an indeterminate tradition and deciding “how to go on” with our
practices. In doing philosophy, we are both interpreting and recommending.
However, they seem to suggest that there is one “best” way to rationalize the
representational (and practical) tradition, and so just one way to go on. I find
this implausible and unnecessary.6 Different communities may highlight
different parts of the tradition because of what is important to them, what
practices they are committed to, what questions they ask, and how the world
around them pushes back (e.g., what else they come to know).
1.5. Representational Traditions: ‘Race’

Does the Schroeters’ model give us resources to make progress in under-
standing what race is? The case of race is clearly more complicated: there
are substantive disagreements about the different roles the idea plays in our
representational tradition, and the tradition has clearly changed over time.
This is something we should take into account. Moreover, there are signif-
icant differences in how the idea of race functions in different cultures; for
6. Botchkina and Hodges (2016) defend a view similar to theirs, but that allows for multiple
reasonable interpretations. Moreover, the Schroeters’ view is more individualistic than my own.
On their view, a primary normative constraint is to provide a rationalizing self-interpretation,
i.e., to make sense of one’s own beliefs and practices (linguistic and otherwise). I see this as a
more collective project. See also Haslanger (forthcoming) for an elaboration of the idea that
conceptual amelioration through such reflection is possible.
15
example, which racial group one belongs to may differ depending on the
country one is in, and the background beliefs about races may differ. In the
case of water, there is a basic human interest in being able to refer to the stuff
in question, and at least most languages will have some way of talking about
it. This is much less obvious in the case of race. So the idea that there is a single
best interpretation of what race is—across languages and cultures—is not en-
tirely plausible.
For example, the United States has relied—sometimes implicitly and
sometimes explicitly—on a rule of hypodescent (i.e., assuming a racial hier-
archy, the child of individuals of different races is assigned the “lower” race
of the two parents).7 However, social scientists have found a variety of other
rules for assigning race in the case of “mixed” offspring. In some societies
(such as Hawaii, at least before statehood), “mixed” offspring are fully in-
cluded as members of both races (Davis 1995, 116). In other societies, chil-
dren of differently raced parents constitute a separate group that, depending
on the case, may be considered inferior to, superior to, and or between the
racial groups of the parents. In parts of Latin America, the race of “mixed”
offspring does not depend simply on ancestry, but also on “economic and ed-
ucational achievements”:
Whites are at the top of these class structures and unmixed blacks and
Indians on the bottom. Blacks are defined as only those of unmixed
African descent. Although the many rungs on the long status ladder
are indicated by terms that describe the highly variable physical ap-
pearance of mulatto and mestizo individuals, this racial terminology
can be quite misleading. These are actually class systems in which life-
style is much more important than racial ancestry or physical traits.
“Money whitens” as the phrase goes, and a person who rises in educa-
tional and economic status is identified by whiter racial designations.
(Davis 1995, 119)
And finally, in some cases, there is a possibility of assimilation, so that after

some number of generations, “mixed” offspring can become members of
the superior or dominant race. Thus, at least currently in the United States,
individuals who are, say, one-eighth Asian and seven-eighths White may
count as White with an “Asian” heritage (Davis 1995, 120). The practices of
racial identification are also evolving.
7. https://www.encyclopediavirginia.org/Racial_Integrity_Laws_of_the_1920s
16
16 • W h a t I s  R a c e ?
So when I suggest we consider “our representational tradition,” what do

I have in mind? For the purposes of our discussion I will be focusing on
what race is in the United States, keeping in mind that the goal is to pro-
vide an interpretation of what has plausibly been at issue (though not al-
ways clearly at issue) “all along,” as evidenced not only by what we say, but
what we do, such as the practices we engage in, the laws we pass, and social
scientific explanations of these. Given the history of the United States, the
representational tradition draws upon some historical uses of the term and
practices in Europe as well. In the following I’ve made a start on the relevant
inputs to deliberation about what race is within this tradition. Inputs in-
clude both ideas that I take to be broadly shared in the United States, ideas
from both natural and social sciences, and normatively relevant ideas, as the
model recommends, though some will be controversial. I will use the term
‘raceus’ to designate what I take to be outputs of deliberation about this rep-
resentational tradition. I hope that in interpreting our own tradition we will
gain insight into related ones.8 I also use an initial upper-case letter for the
names of purported races.
• Particular instances: When we say that Martin Luther King, Jr., is Black,
Hillary Clinton is White, Che Guevara is Latino, Sacagawea is Native
American, and Aung San Suu Kyi is Asian, we are classifying each as
belonging to a different race. Everyone belongs to at least one race, pos-
sibly more than one. The criteria for racial membership varies depending
on context and is not consistent: the US government relies primarily on
self-
identification; epidemiologists and demographers sometimes rely
on self-reports, but also on birth certificates, mother’s birth certificate,
death certificates, doctor’s (or other’s) attribution of race (Root 2001,
2003, 2009). Generally, however, one’s racial designation is confirmed or
disconfirmed by facts about whether one’s ancestry derives from a partic-
ular geographical region or regions. It is possible for someone to belong to
a race without knowing that they do, e.g., an illiterate Kayin peasant from
Myanmar is racially Asian, even though she may know nothing about Asia
8. I have found it challenging to judge which instances of the word ‘race’ should include the
subscript ‘us’. My goal has been to leave the subscript off when we are considering candidate
inputs to the deliberation—allowing that they may or may not be aptly considered a core part
of the phenomenon and may be simply associations or related phenomena—but adding ‘us’
when drawing conclusions about the US phenomenon we’re aiming to track. I’m not sure I’ve
been wholly consistent in this because it isn’t obvious, to me at least, what occurs as part of de-
liberation and what occurs as a result. My apologies for any confusion this may cause.
17
as a continent or US racial practices. There is disagreement about whether

Latino(a)s are a race (Gracia 2007).
• Perceptual gestalts: Members of different races can usually be distinguished
by physical features such as skin color, hair texture and color, eye shape; it
is sometimes difficult to identify the racial makeup of mixed-race people,
so perceptual gestalts are fallible, and some individuals do not have the
distinctive features associated with their race (and so may intentionally or
unintentionally “pass” as a member of a different race).
• Biological roles: People inherit their race (though the criteria for inher-
itance have been contested and variable over time, and seem to differ,
depending on the race at issue). Race is correlated with differences in life
expectancy, various diseases, etc. Historically, it was thought (and some,
but not all, people continue to believe) that one’s race is part of one’s na-
ture, and at least in the case of some races, it is passed along to biolog-
ical offspring (though Whiteness, apparently, is not always passed on,
though Blackness is, according to the system of hypodescent!). Scientific
research suggests that there isn’t a meaningful biological basis for racial
distinctions, sufficient to postulate racial “natures” or essences, though it
is currently a matter of controversy whether there are minor biological
differences among groups roughly corresponding to the most commonly
assumed racial groups (Black, White, Asian, Native American, Pacific
Islander) (see Spencer’s Chapter 3 in this volume; also Andreasen 2000,
2004; Kitcher 2007).
• Historical roles: Attributions of race have played a major role in world his-
tory. For example, the trans-Atlantic slave trade and European coloniza-
tion across the world were justified on the basis of beliefs about race. The
most common racial divisions have been based (roughly) on appearances
that differ between continents, but there has not been unanimity on what
races there are (Bernasconi and Lott 2000; Herzog 1998, 288ff ), and argu-
ably we are in a historical moment when those of (apparent) Arab descent,
having been White, are being re-racialized as non-White ( Jamal and Naber
2008). US federal and state law has restricted the civil rights of members
of non-White races, and there have also been attempts at legal remedies,
e.g., affirmative action. Race continues to be a matter of heated social, po-
litical, and legal debate. The history of science reveals ongoing scientific
attempts to justify claims about racial differences, especially in intelligence
and character.
• Practical roles: Race is a significant factor in the organization of social
life, e.g., in patterns of association, housing, religion, employment, crime,
18
18 • W h a t I s  R a c e ?
athletics. It is also an important part of many people’s identity and sense of

solidarity with others, and contributes to shaping their life plans and polit-
ical views.9
• Symbolic roles: Race is strongly associated with cultural norms, artistic
traditions, and forms of life; historically it has been associated with char-
acter traits and degrees of moral worth.
• Explanatory roles: Race is used to explain a broad range of differences be-
tween social groups, including educational attainment, patterns of arrest
and incarceration, health outcomes, social history, etc. It is also used to ex-
plain different interests, cultural and artistic tendencies, and political affili-
ation. These explanations vary in their form. Some purport to be biological
explanations, others sociological explanations.
• Epistemology: A person’s race is usually taken to be evident based on widely
accepted perceptual gestalts; however, race has been hard to define, and
many assumptions about race have been undermined by scientific inquiry;
our beliefs about race may be mistaken or incomplete; nevertheless, obser-
vation of racial regularities grounds induction to unobserved cases.
Given these inputs (I don’t mean for these to be exhaustive—this is just a

sample), the task is to provide an interpretation of our representational tra-
dition.10 What are we doing when we divide humans into different races?
What interests are being served? Is there an interpretation that rationalizes or
justifies the tradition? If not, then should we reject the idea of race completely?
One might argue that, as in the case of ‘water,’ there are several different
reasonable ways to go here. A first option is to note that the representational
tradition concerning race includes a history of drawing distinctions between
groups of people on the basis of certain bodily features (skin color, hair tex-
ture, eye shape, and the like) and postulating racial “natures” underlying these
observable differences to explain further cultural and behavioral differences
9. There is a broad literature on the content of racial identities, and African American or Black
identity in particular. A helpful philosophical snapshot may be found in Gooding-Williams
(1998); Appiah (2002); Shelby and McPherson (2004); Shelby (2005); Gracia (2007);
Kendig (2011).
10. As I read Quayshawn Spencer’s (2014) argument for a modest racial naturalism, he could
agree with the Schroeters’ approach that has us trying to make sense of a representational/
practical tradition; he chooses to place great weight on the federal discourse around race as
regimented by the census. I don’t think his choice of emphasis takes sufficient account of the
many functions of race in our ordinary discourse and places too much weight on the role of the
state; but as I make clear in my reply, our priorities are different.
19
and to justify unjust treatment of non-Whites. The tradition was in the ma-
terial and cultural interests of Whites and continues to play a role in many
Americans’ thinking about race. The Schroeters are clear, however, that the
best interpretation of the representational tradition must capture what we
have been thinking and talking about “all along.”
The method we have sketched precisely aims at determining what’s

interesting and important relative to the subject’s own past represen-
tational tradition. So from the point of view of a rational epistemic
agent, these pragmatic meta-cognitive methods are ideally suited to
getting us closer to the truth about the interesting and important
topics that we were thinking and talking about all along. (2015, 436)
It is reasonable to claim that our linguistic forebears were thinking and talking
about races distinguished by racial natures or essences. Yet at this point we
know that there are no racial “natures,” (i.e., a set of properties that a member
of a race has necessarily, by virtue of which they are a member of the race, and
that explains their characteristic behavior and abilities).11 If the point or pur-
pose of the tradition was to attribute racial natures to humans, and there are
no such racial natures, the representational tradition has failed and we should
give it up. In short, we should be anti-realists about raceus, more specifically,
error theorists.
One need not think that our representational tradition is invested in ra-
cial natures, however, in order to account for the inputs to deliberation about
race. Michael Hardimon (2017; also 2003) has argued for a minimalist ac-
count of race according to which:
A race is a group of human beings
(C1) that, as a group, is distinguished from other groups of human

beings by patterns of visible physical features,
(C2) whose members are linked by common ancestry peculiar to
members of the group, and
(C3) that originates from a distinctive geographic location. (2017, 31)
11. I prefer the term ‘nature’ to ‘essence’ in this context because of the complexities in the his-
torical and contemporary use of ‘essence,’ though ‘essence’ is the more commonly used for this
postulation. Think of something like the nature of a tiger—each tiger has a set of properties
necessarily by virtue of which it counts as a tiger (tiger is its kind), and this set of properties
explains its behavior and abilities, e.g., it is by nature a feline, a carnivore, etc.
20
20 • W h a t I s  R a c e ?
Hardimon argues that there are groups that satisfy the minimalist race con-
cept, and so racesus exist. Given the simplicity and plausibility of Hardimon’s
conditions, it would seem that his account is an excellent candidate for an
interpretation of the inputs regarding race. Should we take this view to be
sufficient and the task to be complete?
Joshua Glasgow (2009) rejects a minimalist view such as Hardimon’s.
He maintains that according to the “ordinary concept of race” (i.e., the one
that has the most currency in the contemporary United States), “while an
individual’s particular race might depend on social factors, each racial group
is, as a conceptual matter, defined only in terms of its visible, biological pro-
file” (2009, 123) (see also Alcoff 2005). The condition Glasgow isolates, how-
ever, is not satisfied: human appearance falls on a broad spectrum, and the
supposed visibly notable and biologically relevant clumping that would be re-
quired by the condition does not occur.12 Glasgow concludes that “since these
groups’ putative distinctiveness is not, as a point of fact, legitimated by the bi-
ology, there are no races” (2009, 123). So the term is vacuous, and statements
employing the term are false.
I agree with Glasgow (2009) that there are no existing human groups
that meet the condition that there are inherited visible features that demar-
cate the races. Note, however, that an error theory about raceus has substan-
tial costs, given that we are attempting to give an interpretation of the inputs
described earlier. Not only would we have to claim that our attributions of
race to individuals are false, but that the historical, symbolic, explanatory,
practical, and epistemic roles of race are all founded on illusion. We would
need to give up the idea that race explains certain group differences (from ar-
tistic traditions to health outcomes); we would need to give up the idea that
race provides reasons for certain practical, historical, and symbolic choices.
We could potentially replace these claims with the suggestion that false racial
beliefs about racial groups explain the broad range of racial formations. But
this is to take a substantive stand on difficult explanatory questions in the his-
tory and sociology of racial practices, racial institutions, and the like.
Although false beliefs about racial groups may be the best explanation of
early forms of racial hierarchy (though I find even that questionable, given
the economic and other forces at work), it is implausible that such beliefs are
the best explanation of ongoing racial injustice, including the perpetuation
12. An extension of this argument is also relevant to discussion of Quayshawn Spencer’s (2014)
minimal naturalism (see Chapter 3 for details).
21
of economic and political injustice, social segregation, and cultural stigma

(Haslanger 2016, 2017a). For example, the waning of racial essentialism is not
sufficient to undermine the legacy of economic deprivation because belief in
racial essentialism—or racial naturalism more generally—is not what runs
the economy; nor does correcting false beliefs about race correct the legacy
of centuries of legal and political wrongs. So an error theory about race needs
to be supplemented with alternative explanations of apparently racial phe-
nomena, including our abilities to deliberate about race, perform induction
on racial regularities, and find meaning in racial identification.
Consider an analogy. For centuries, philosophers have attempted to pro-
vide necessary and sufficient conditions for being a person. Candidate neces-
sary conditions have included the following:
• X is a person only if X has a soul.

• X is a person only if the stages of X are psychologically continuous.
• X is a person only if X occupies a continuous living human body.
There are compelling examples suggesting that accepting any of these

conditions fails to capture what we mean by person due to a mismatch be-
tween what the condition requires and the cases we judge to be persons.
Should we conclude that there are no persons and be error theorists about
person-talk? Surely not! For ordinary cases, we do fine in judging whether an
individual is a person or not. There are hard cases and our practices in these
cases are contested, but we usually have rules or laws for settling them, at least
for the time being. The question remains whether these rules or laws for the
controversial cases are appropriate or justified, and they are open to revision
based on actual cases that need to be settled. Our practices are evolving: new
technologies and medical discoveries have forced us to answer questions that
never had to be faced before. Such open-endedness and revisability do not
show that there are no persons, that person-talk is vacuous, or that we should
discontinue practices that rely on a background idea of persons.
In discussing the example of person, I am relying on the methodology
I have recommended: given the function of language to guide and coordi-
nate our ongoing practices, we should investigate what something is by pro-
viding an interpretation of the representational (and practical) tradition that
helps us make sense of the new and challenging cases. Admittedly, there may
be circumstances in which the tradition cannot or should not be sustained.
But holding fixed one condition on the application of a concept that actu-
ally serves multiple purposes is not a sufficient reason to reject or eliminate
2
22 • W h a t I s  R a c e ?
the concept. Communication in a context does not require a rigid represen-

tational tradition that anticipates every empirical discovery and every tech-
nological change. We make do with rough overlapping understandings of
phenomena that concern our shared practices, and update as life goes on.
A third option for interpreting the inputs sketched in the preceding
would be to turn to a social constructionist account of raceus. When the rep-
resentational tradition was historically postulating racial natures, the point
was to provide an explanation of the striking observable differences in human
appearance, behavior, and culture found through voyages of exploration and
conquest. The explanations that were ready to hand at the time were based
on biblical interpretation or neo-Aristotelian biology (Stocking 1994). These
explanations of human differences, we have found, are faulty. But there were
and continue to be differences among the groups that were then designated
as races that call for interpretation and explanation. A better approach looks
to social formations.
A social constructionist account (e.g., sociopolitical account or cultural
account, to be discussed more fully in the next section) proposes that the
conditions for being a member of a racial group are to be given in social terms,
rather than in physical, biological, or other non-social terms.13 Consider, for
example, slaves. Aristotle seems to have thought that there were natural slaves
(Politics, Bk. 1): natural slaves are individuals who are incapable of sufficient
practical reason to lead an autonomous life. Natural slaves, on his view, were
justifiably owned by others and were better off as a result. But the idea of a
natural slave is badly mistaken. Slaves are a social category, that is, to be a slave
is to be owned by someone according to the laws or customs of one’s social
milieu. A social constructionist account of slaves is an improvement on the
naturalistic account that defines slaves in terms of their cognitive capacities.
There are many other cases in which social constructionists have challenged
naturalistic accounts, for example, of sex, gender, sexual desire, disability, par-
enthood, family, and race.
How does attention to social relations and social structures help us un-
derstand the observable differences among human groups? There is over-
whelming evidence that differences between racial groups in educational
attainment, health outcomes, incarceration rates, and the like are due to
the looping effects of social structures that impose a racial hierarchy. Many
13. Social constructionism about race takes many forms. For other examples, see Omi and
Winant (1994); Mills (1997, 1998); Gooding-Williams (1998); Sundstrom (2002); Mallon
(2003); Taylor (2004); Alcoff (2005).
23
of the great achievements and cultural traditions of different races are also a
product of living within such structures (Taylor 2016). This social structural
hierarchy is partly a product of a history of false beliefs about races and ra-
cial natures, but such beliefs are systematically linked to cultural and material
factors that are equally important in accounting for the systematic nature of
racial differentiation and racial injustice; false beliefs are a small part, maybe
even an eliminable part, of what sustains the system. For example, it is in-
sufficient to explain racial differences in educational achievement simply in
terms of false beliefs about the abilities or “natures” of Whites and people of
color. Additional factors include the racial patterns of wealth and poverty,
patterns of housing segregation, the dependence of school funding on prop
erty taxes, the expense of university education, hiring discrimination, and the
social meaning of intelligence and education.14 Such social phenomena do
not depend entirely on the psychological states of individuals (Epstein 2015;
Haslanger 2017b).
This second, social constructionst, approach gains further support from
the parallels with other scientific advances. Early explanations of many natural
phenomena have been rejected over time and have been supplanted by better
explanations without disrupting our representational traditions. Hippocrates
was aware of and treated cancer, though he thought it was caused by an excess
of black bile (thought to be one of the four humors) in the body; it is plau-
sible that Hippocrates is part of our representational tradition concerning
cancer (he is credited with the origin of the word karkinoma), in spite of the
fact that some of his core assumptions about cancer have been thoroughly
rejected. The idea that empirical hypotheses about the nature of a kind are an-
alytically entailed by our use of a term would make scientific inquiry difficult
(this is an old point made by Quine, Putnam, Kripke, and many others). If we
could not substantially revise our understanding of kinds, then as we develop
new hypotheses about a phenomenon, we would not be improving our un-
derstanding of a poorly understood kind, but investigating a new kind, thus
obscuring the dynamics of inquiry.
One might argue, however, that shifting from a “natural” to a “social” kind
is more than meaning can bear. But shifts across different categories of expla-
nation are not uncommon. For example, medical conditions that were once
14. There is a huge social science literature on racial health gaps, educational achievement
gaps, etc. For a glance at the numbers in 2014, see Irwin et al. (2014). An important approach
to explaining this is offered in Mills (2017, Ch. 7). See also Anderson (2010) and Haslanger
(2014).
24
24 • W h a t I s  R a c e ?
thought to have been the result of God’s punishment, or evil thoughts, or anx-
iety, have been shown to have straightforward physical causes. Various caste,
class, and ethnic divisions have been thought to be established by divine law
or nature, but are now understood in terms of the workings of social systems.
For example, monarchs were once thought to gain their political legitimacy
from God. To be a monarch is to have sacred power, invested in the family
lineage. This explanation of a monarch’s legitimacy was eventually rejected,
and yet we did not give up the idea of a monarch. Instead, alternative social
accounts of monarchy were supplied.
A social constructionist account of raceus will also face challenges in ac-
commodating some of the inputs to deliberation listed in the preceding. For
example, it is commonly thought that race is inherited. But social position is
only inherited metaphorically; one usually occupies a similar social position
to one’s parents, not by virtue of “blood” but by virtue of social conditions
and pressures. I mentioned earlier, however, that there are multiple ways of
“tracing” race through ancestry, hypodescent being only one of them, even
in the United States. This suggests that a commitment to the idea that race
is inherited is not a fixed point. Moreover, the use of ancestry to track race is
a phenomenon that an error theorist will also need to explain. Any account
of raceus—whether realist or anti-realist, naturalist or constructionist—will
need to include details that make sense of or explain away the complexities of
representational and practical tradition. So there is much work to do.
1.6. What Is Race?
Even if the representational tradition concerning race allows for a social scien-
tific analysis of the explanatory interests being served, two questions remain:
(i) How exactly should a social constructionist capture what race is?
(ii) Are our current interests served by continuing with the representational
tradition concerning the term ‘race,’ or should we replace race with an-
other term, e.g., ‘racialized group’?
I will consider (i) in this section and (ii) in the next.

Two forms of social constructionism about race have been proposed in the
literature. One is the sociopolitical account, the other is the cultural account.
The two accounts agree on many points, for example, that the current domi-
nant racesus emerged in a particular historical context of White racial domi-
nation; that members of races are “marked” as having a particular appearance;
25
that the “marking” is taken to be evidence of where, geographically, the group

mostly lived at a key moment in time; that racial groups function differently
within the contemporary sociopolitical structure, and are positioned on a hi-
erarchy. The primary differences between the two accounts are (a) the cultural
account requires that races, as a group, share a culture,15 whereas the sociopo-
litical account does not, and (b) the sociopolitical account takes the sociopo-
litical hierarchy to be a defining feature of race, whereas the cultural account
does not. I defend the sociopolitical account. Chike Jeffers (2013) elaborates
and defends the cultural account.
In my earlier work (2012, Ch. 7), I argued that critical theorists should
adopt the following core account of race, and use this to explicate other ra-
cial phenomena, such as racial identities, racial norms and traditions, racial
narratives, racial oppression, racial justice, and the like.16
Social/Political Race (SPR): A group G is racialized relative to context C
iffdf members of G are (all and only) those
(i) who are observed or imagined to have certain bodily features presumed
in C to be evidence of ancestral links to a certain geographical region
(or regions)—call this “color”;
(ii) whose having (or being imagined to have) these features marks them
within the context of the background ideology in C as appropriately
occupying certain kinds of social position that are in fact either subor-
dinate or privileged (and so motivates and justifies their occupying such
a position); and
15. A cultural constructionist view does not require that every member of the group participates
fully in the culture; rather, a group does not count as a racialus group unless it represents a
particular form of life. DuBois is often taken as offering a paradigm of the cultural account,
suggesting that a race is “a vast family of human beings, generally of common blood and lan-
guage, always of common history, traditions and impulses, who are both voluntarily and in-
voluntarily striving together for the accomplishment of certain more or less vividly conceived
ideals of life” (DuBois 1991[1987], 75–76), also quoted in Jeffers (2013, 405). In other words,
the set of conditions that make a group a racialus group may include reference to a form of life,
but the conditions for being a member of a racialus group may not include this condition; e.g.,
the condition could simply be that one’s parents are a member of the group. So, for example,
the Jewish people have a particular form of life, but not all Jews are observant. Nevertheless,
one is a member of the Jewish people by virtue of being born of a Jewish mother (or in some
forms of Judaism, a Jewish mother or father, or by conversion), not by virtue of observing the
practices of Judaism.
16. Note that the term ‘iffdf’ is sometimes used to indicate that the biconditional is offering a
definition of a word or a concept, I intend it here to indicate that I’m answering a “What is X?”
question or “What is it to be X?” question, i.e., to give what is sometime called a “real defini-
tion.” See, e.g., Rosen (2015).
26
26 • W h a t I s  R a c e ?
(iii) whose satisfying (i) and (ii) plays (or would play) a role in their system-
atic subordination or privilege in C, that is, who are along some dimen-
sion systematically subordinated or privileged when in C, and satisfying
(i) and (ii) plays (or would play) a role in that dimension of privilege or
subordination.17
The idea is that racesus are racialized groups, that is, those groups demarcated
by the geographical associations accompanying perceived body type, when
those associations take on evaluative significance (or social meaning) con-
cerning how members of the group should be viewed and treated, and the
treatment situates the groups on a social hierarchy.
Thus, to say that Martin Luther King, Jr., is Blackus is to say that he is a
member of a group that meets these conditions, and in particular, that he is
marked in the United States as having relatively recent ancestry from Africa,
and this situates him as subordinate in the social hierarchy of the United
States. Moreover, to say that Whitesus have higher educational achievement
than Latinxus is to say that a group that is marked as having recent ancestry
17. There are several aspects of this definition that need further elaboration or qualification.
First, the definition does not accommodate contexts such as Brazil in which membership in
“racial” groups is partly a function of education and class. This is because my project here is to
capture what race is in the contemporary United States, i.e., raceus. However, a related racial
phenomenon can be found in other representational/practical traditions and another version
on which appropriate “color” is relevant but not necessary might be captured by modifying the
second condition:
(ii*) having (or being imagined to have) these features—in combination with factors
such as economic and educational status—marks them within the context of C’s cul-
tural ideology as appropriately occupying the kinds of social position that are in fact
either subordinate or privileged (and so motivates and justifies their occupying such
a position).
The first condition already allows that the group’s members may have supposed origins in
more than one region (originally necessary to accommodate the racialization of “mixed-race”
groups); modifying the second condition allows that racialized groups may include people
of different “colors” and may depend on a variety of factors. Second, I want the definition to
capture the idea that members of racial groups may be scattered across social contexts and may
not all actually be (immediately) affected by local structures of privilege and subordination. So,
for example, Black Africans and African Americans are together members of a group currently
racialized in the US, even if a certain ideological interpretation of their “color” has not played
a role in the subordination of all Black Africans; there are parallel phenomena in the case of
other races. So I suggest that members of a group racialized in C are those who are or would be
marked and correspondingly subordinated or privileged when in C. Those who think (plau-
sibly) that all Blacks worldwide have been affected by the structures and ideology of White
supremacy do not need this added clause; and those who want a potentially more fine-grained
basis for racial membership can drop it.
27
in Europe and that is situated as privileged, as a result, has higher educational

achievement than those marked as having recent ancestry in Latin America
and who are disadvantaged as a result. This claim reveals a correlation be-
tween certain forms of social subordination/privilege and outcomes. It does
not itself make a causal claim. A relevant causal claim might be this: those
who are marked and privileged as Whiteus have higher educational achieve-
ment because of their racially marked privilege. This is not a tautology, nor
is it a vacuous explanation: a group with racial privilege could have educa-
tional success due to other causes. However, the explanation is far from being
complete, for we would want to know how the privilege is more specifically
related to the achievement.
The proposed SPR account also helps us explain certain aspects of racial
meanings, artistic traditions, and cultural norms. On my view, races are dis-
tinct from ethnicities. An ethnicity is a cultural grouping—involving shared
language, customs, social meanings, cultural formations—that typically (but
not always) relies for its existence and coherence on geographical and gene-
alogical connections, and sometimes carries (defeasible) presumptions about
appearance. So Germans, Italians, Basques, Armenians, Berbers, Croats, Fula,
Hausa, Gujarati, Icelanders, Kurds, Luo, Manchu, Mongols, etc., are ethnicities.
Ethnicities are often positioned hierarchically within a society or broader soci-
opolitical formation. On my account, the hierarchical positioning of an ethnic
group within a broader society (or broader political formation) is a process
of racializing the group. The ethnicity may predate the racialization, and will
(hopefully) continue after racialization has ended. Moreover, multiple ethnic
groups may be racialized together as a single race; this may result in what Yen
Le Espiritu (1992) calls pan-ethnicities. So, Asians are considered a racial group,
but include many different ethnicities (e.g., Bamars, Bengalis, Gujarati, Han
Chinese, Hindustani, Hmong, Hui, Japanese, Kashmiri, Khmer, Konkani,
Korean, Manchu, Marathi, Mongols, Napali, Sinhalese, Tais, Telebu, Tibetens,
Uyughur, Vietnamese, Zhuang, to name a few). Such ethnic groups do not
share a form of life, and may have long-standing conflicts over land, religion,
and politics (see also Alcoff 2000 on the different ethnic groups considered
Hispanic or Latinx). The cultural differences between the ethnic groups does
not prevent them from forming a race, however, because racialization is not
in the first instance a matter of identity or shared culture, but of an imposed
(ascribed) position in a sociopolitical formation.
The Africans who were forcibly brought to the United States came not
as “blacks” or “Africans” but as members of distinct and various ethnic
28
28 • W h a t I s  R a c e ?
populations. As a result of slavery, “the ‘Negro race’ emerged from the

heterogeneity of African ethnicity” (Blauner 1972, 13). . . . Diverse Native
American tribes also have had to assume the pan-Indian label in order to
conform to the perceptions of the American State. . . . Similarly, diverse
Latino populations have been treated by the larger society as a unitary
group with common characteristics and common problems. . . . And the
term ‘Asian-American’ arose out of the racist discourse that constructs
Asians as a homogeneous group. Excessive categorization is fundamental
to racism because it permits “whites to order a universe of unfamiliar peo-
ples without confronting their diversity and individuality (Blauner 1972,
113).” (Espiritu 1992, 6)
The development of a pan-ethnicity may emerge, for example, among

Asian immigrants in the United States, but is an accomplishment, not a given.
And such pan-ethnic identities do not necessarily extend to communities of
origin; “group formation is not only circumstantially determined, but takes
place as an interaction between assignment and assertion. . . . In other words,
panethnic boundaries are shaped and reshaped in the continuing interaction
between both external and internal forces” (Espiritu 1992, 7). Thus, Asian
American may be a pan-ethnicity because Asian immigrants to the United
States, and their descendants, form a sense of shared Asian American cul-
ture. This suggests that there are three relevant types of groups: ethnicities,
pan-ethnicities, and races. Ethnicities have distinctive cultures. Races typi-
cally consist of people from multiple cultures. Pan-ethnicities emerge when
multiple groups are racialized and treated as one group, and form an iden-
tity and way of life as a result.18 So Hmong, Japanese, Khmer, and Korean
are ethnicities. They are all treated as Asian in the United States, and Asian
Americans form a pan-ethnicity. Some individuals living in Asia may come
to see themselves as Asian in response to the racialization of Asians in the
United States (and elsewhere), so there may be a group larger than just Asian
Americans who are members of the pan-ethnicity; we might call these the
pan-ethnic Asians. But this does not make the group of people living, or with
recent ancestry, in Asia who are “marked” as Asian, a pan-ethnicity. The large
heterogeneous group does not have a shared culture. Asian is not recognized
as an identity by those living outside of a process of Asian racialization; none-
theless, it is, or has been historically considered, a race. Plausibly also, Black is
18. See also Alcoff (2000) on ethno-race, and Gooding-Williams (1998).

29
not a pan-ethnicity, even if African American, or Diasporic African, is. This is

a central difference between the SPR account I support from a cultural con-
structionist account that takes shared culture to be a defining feature of race
( Jeffers 2013).
I agree with Jeffers that often identities and cultural practices associated
with races (e.g., “African American” or “Asian American” or “Latinx”) offer
creative (and protective!) resources for those who have been racialized (2013,
422) that go well beyond a response to oppression; and a pan-ethnicity such
as “White” offers other creative opportunities, in addition to resources for
domination—or even more often, escape from subordination. Pan-ethnic
groups share at least some minimal culture. How this works will vary from
context to context. At one time the ideology invoked racial essences to jus-
tify the differential treatment of different “colored” groups, and conceptions
of the essence or spirit of a people was a basis for identity; ideology has also
linked racially marked people with cultural traditions, histories, and talents.
People so marked have shared experiences, and some have bonded together
in celebration and resistance. This has resulted in racially identified artistic
movements, cultural norms, and forms of association. I do not claim that ra-
cially inflected culture is all about the position of the group in a hierarchy.
Culture is dynamic and relatively autonomous from, and so not determined
by, economic, political, or historical factors with which it is always manifested
(Sewell 2005). But such pan-ethnicities are not races, or so I would argue.
An individual ethnic Hmong living in China or Laos is, I would maintain,
Asianus, even if there is nothing distinctively “Asian” about Hmong culture,
and she does not identify as Asian (and maybe has not even heard of the des-
ignation). That she counts as Asianus is clear, however, by how Hmong are
viewed and treated within the United States (Fadiman 2012), and how she
would be viewed and treated if she came here.
Although I believe that SPR is a reasonable interpretation of the rep-
resentational tradition concerning race, there are also reasons to resist it.
This is to be expected, given that I embrace the idea that we can reason-
ably draw different conclusions about what is crucial to our representa-
tional and practical traditions, depending on the questions we ask and the
purposes we bring to the inquiry. For example, some definitely take their
racial identity to be an important part of who they are, and it is offensive
to them to regard it as a response to racial subordination or privilege. It is
important to note, however, that it does not follow from SPR that a racial
identity must focus on facts of subordination or privilege, and nothing
I have said entails that it is wrong or illegitimate to embrace an identity or
30
30 • W h a t I s  R a c e ?
the distinctive way of life that has emerged with the pan-ethnicity. In fact,
I believe that many forms of racial identity are important, valuable, and in
some cases even inevitable responses to racial hierarchy. As I see it, a racial
identity is a kind of know-how for navigating one’s position in racialized
social space (Haslanger 2012, Ch. 9). The apt content for a racial identity,
then, may be positive, affirming, and empowering, even if the racialized
social position one occupies is oppressive.
There is a key normative difference, I think, between the sociopolitical ac-
count of race and the cultural account that becomes clear when one asks why
hierarchy is built into race according to the SPR. Why not say that races are
groups who are “marked” by reference to ancestry and geography, where this
marking has implications for the group’s social position, without claiming
that the social positions in question need be arranged hierarchically? If I drop
the hierarchy condition, then the account comes much closer to the cultural
account, on the assumption that those who occupy the same social position
are likely to share some non-trivial practices that would amount to at least a
thin “way of life.” Jeffers argues that we should adopt an account of race that
does not have the result that race is eliminated once racial hierarchy is elimi-
nated. He suggests:
From the cultural perspective, though, a situation in which racial

groups persist but in a state of equality rather than socioeconomic and
Eurocentric cultural hierarchy, respecting and mutually influencing
each other while remaining relatively distinct, is a coherent and admi-
rable goal. ( Jeffers 2013, 421)
I worry, however, about the extent to which we should embrace cultural groups
marked by ancestry and appearance in the long run (of course in the short run,
they are necessary to achieve justice). Currently, ethnic groups carry a pre-
sumption of shared ancestry, appearance, and geography, but this is merely
a presumption. At least many cultural groups (understood as groups sharing
a way of life, a language, a religion, a set of common practices) have porous
boundaries: one can marry into them, convert, immigrate, look very different
from other members, not originate where other members originated. Jeffers
emphasizes the benefits of racial cultural unity, but not the costs of racial seg-
regation. As I see them, the costs include tendencies to cultural norming and
authenticity tests of those with a “marked” racial appearance (this results in
the arguably slurring racial terminology of ‘oreo,’ ‘banana,’ ‘twinkie,’ ‘apple,’
‘coconut,’ and ‘egg’). It also suggests that those without the right physical and
31
ancestral credentials don’t belong in the culture, shouldn’t participate in the

way of life, and are suspect when they build strong alliances with and take up
the practices of those who satisfy the racial conditions. Living, myself, in a
mixed-race (Black-White) and cross-cultural ( Jewish-Christian) family, I may
be overly influenced by the huge contemporary challenges posed by racial (an-
cestry and appearance-based) membership criteria in cultural practices and
cultural communities. These challenges could—and I think should—subside
under conditions of justice. I find problematic the idea that a just world is
one in which cultural groups can restrict their membership on racial grounds.
I embrace, instead, a model of multiple coexisting cultures that are mutable,
flexible, and creatively tolerant around issues of ancestry and appearance.
Clearly there is more to be said about the ways in which SPR does or does
not make sense of our representational tradition concerning race. I believe,
however, that it is an excellent candidate, given the Schroeter-style method,
for determining at least one thing race is, and so at least one thing we can
mean by ‘race.’
1.7. Going On: The Normative Dimension

of Racial Classification
I think it is unquestionable that SPR captures an important set of social
groups. They are those groups that have been racialized. Drawing on the
Schroeters’ methodology, I have also argued that there is a good case to be
made that SPR is a reasonable interpretation of our ongoing representational
tradition and social practices with respect to the idea of race. However, I do
not want to be committed to there being a single best interpretation of that
tradition, nor do I think that how we should go on with our representational
practices depends entirely on what our past practices have endeavored to
identify as an important matter of shared concern. Even if the best interpre-
tation of the tradition shows that it is semantically permissible to use the term
‘race’ along the lines that SPR suggests, that does not settle how or whether
we should continue to use the term. In other words, even if we can isolate a
set of social groups that are reasonably considered races, we could still decide
not to use the term anymore, or to use a new term.
So the question remains whether our current interests are served by con-
tinuing with tradition of using the term ‘race.’ For example, some theorists
have chosen to reject the term ‘race’ because of its problematic history in justi-
fying racial injustice, and have opted instead for terminology that echoes but
does not maintain the term (e.g., ‘race’ is replaced by ‘racialized group’; Blum
32
32 • W h a t I s  R a c e ?
2002). As mentioned at the start, I enter this debate as a social critic, and be-
lieve we can criticize our past practices and recommend changes to them. This
includes changes to our linguistic practices.
On my view, this is a practical and political issue that is best answered by
well-informed activists at a specific historical moment. As Mallon suggested,
there are empirical and normative considerations that matter, for example,
“the epistemic value of ‘race’ talk in various domains, the benefits and costs
of racial identification and of the social enforcement of such identification,
the value of racialized identities and communities fostered by ‘race’ talk, the
role of ‘race’ talk in promoting or undermining racism, the benefits or costs
of ‘race’ talk in a process of rectification for past injustice, the cognitive or
aesthetic value of ‘race’ talk, and the degree of entrenchment of ‘race’ talk
in everyday discourse” (2006, 550, also quoted earlier). How we go on also
depends on the sources of solidarity that unify and empower a movement,
and the importance of consistent demographic information across time and
domain. These are clearly not questions that can be addressed a priori, and
depend enormously upon context and moment (Shelby 2005).
To say that the issue is best addressed by well-informed activists, however,
is not to relinquish philosophical input. Suppose we find reasons to think
that the racialization of groups is a bad thing and that society would be better
if we were to acknowledge and respect ethno-cultural differences but cease
to think and act in racial terms. (I think there are compelling reasons of this
sort, and briefly discussed this in the previous section.) It would be unreal-
istic, I think, to suggest that we can achieve such a society simply by ceasing to
use racial terminology, by becoming “color blind,” or by denying that races are
real. This is because racialization has caused tremendous social and economic
harms, and reparative justice is required. But how can we go on, if on the one
hand, it would be wrong to continue our current racial practices, and on the
other, it would also be wrong to ignore the legacy of what’s been done?
One strategy mentioned earlier is to employ a new term for the groups that
have been racialized. But there are two risks here. First, most neologisms don’t
catch on. Second, racial identity has a deep and pervasive grip on Americans.
It is very difficult to cast off an identity without offering another in its place,
for identities shape our relations to others, the practices we engage in, and the
possibilities we imagine. A second strategy is to offer a debunking account
of race. Debunking accounts aim to shift our understanding to reveal how
our prior thinking is false or misguided. The point is to disrupt our ways of
thinking, to motivate a new relationship to our practices. This is the sort of
account I think SPR provides. The hope is that if we can see that what we are
3
tracking with our racial classifications is something captured by SPR, then we

will begin to see the importance of disrupting race and organizing ourselves
on different terms.
Note, however, that debunking accounts are employed strategically;
whether they are apt is highly sensitive to contextual factors. The goal, re-
call, is to challenge our investment in certain unjust practices whose injustice
is occluded or masked. The debunking attempts to highlight features of the
practices that make it hard for those of goodwill to continue enacting them.
Yet there are different kinds of racial practices, and people engage in them
with different degrees of awareness. Some practices we enact routinely, mind-
lessly. Others we enact in spite of knowing they harm us or others, for they
define the broad shape of life in our social milieu. And others are recuperative
practices that offer counter-hegemonic understandings and opportunities.
Because debunking has an epistemic and political aim, it may not be neces-
sary if the harm or wrong of the practices are transparent, or if the practices
have already been turned toward justice (Botchkina 2016). In defending the
SPR account, I offer it as an option to be taken up, or not, as a tool in moving
forward toward racial justice.
1.8. Conclusion
It will become clear to the reader that my methodology for answering the
question “What is race?” is different from that of my coauthors. According to
all of them, we should be seeking an understanding of what we are ordinarily
talking about when we talk of race, and with caveats mentioned earlier (i.e.,
that it isn’t all about our talk), I agree with that. But how do we determine
what that is?
In answering the question “What is race?” there are semantic constraints
on us. It would not be reasonable to answer, “Race is a type of furniture.”
But the semantic constraints don’t determine how we must go on. There are
different epistemic and pragmatic standards that may guide our interpreta-
tion of the representational tradition. And there are normative considerations
about what practices we should continue and the best route for maintaining
or discouraging them. I have argued that the SPR account is semantically
permissible, and that in some contexts it is morally and politically valuable,
depending on the practices that are being targeted and the epistemic position
of those engaged in them. One of the important functions of language is to
highlight features of the world that matter for coordination; the function of
34
34 • W h a t I s  R a c e ?
SPR is to highlight—in the relevant cases—how our racializing practices and

identities contribute to injustice.
References
Alcoff, Linda. 2000. “Is Latina/o Identity a Racial Identity?” In Jorge Gracia and
Pablo DeGreiff (eds.), Hispanics/Latinos in the U.S.: Ethnicity, Race and Rights.
New York: Routledge, pp. 23–4 4.
Alcoff, Linda. 2005. Visible Identities: Race, Gender and the Self. Oxford: Oxford
University Press.
Anderson, Elizabeth S. 1995. “Knowledge, Human Interests, and Objectivity in
Feminist Epistemology.” Philosophical Topics 23(2): 27–58.
Anderson, Elizabeth S. 2010. The Imperative of Integration. Cambridge, MA: Harvard
University Press.
Andreasen, Robin. 2000. “Race: Biological Reality or Social Construct?” Philosophy of
Science 67(3): 666.
Andreasen, Robin. 2004. “The Cladistic Race Concept: A Defense.” Biology and
Philosophy 19(3): 425–4 42.
Appiah, K. Anthony. 1996. “Race, Culture, Identity: Misunderstood Connections.” In
A. Appiah and A. Gutmann (eds.), Color Conscious: The Political Morality of Race.
Princeton, NJ: Princeton University Press, pp. 30–105.
Appiah, K. Anthony. 2002. “The State and the Shaping of Identity.” In The Tanner
Lectures on Human Values, vol. 23, ed. Grethe B. Peterson. Salt Lake City: University
of Utah Press, pp. 234–299.
Bernasconi, Robert, and Tommy L. Lott. 2000. The Idea of Race. Indianapolis and
Cambridge, MA: Hackett.
Blauner, Robert. 1972. Racial Oppression in America. New York: Harper & Row.
Blum, Lawrence. 2002. “I’m Not a Racist, But . . .”: The Moral Quandary of Race. Ithaca,
NY: Cornell University Press.
Botchkina, Ekaterina. 2016. “Issues in Objectivity and Mind- Dependence.”
Ph.D. thesis, MIT.
Botchkina, Ekaterina, and Jerome Hodges. 2016. “Objectivity and Conceptual
Change.” Unpublished manuscript.
Davis, F. James. 1995. “The Hawaiian Alternative to the One-Drop Rule.” In Naomi
Zack (ed.), American Mixed Race. Lanham, MD: Rowman and Littlefield, pp.
115–131.
DuBois, W. E. B. 1991[1987]. “The Conservation of Races.” In Philip S. Foner (ed.),
W. E. B. DuBois Speaks: Speeches and Addresses 1890–1919. New York: Pathfinder,
pp. 73–85.
Epstein, Brian. 2015. The Ant Trap. Oxford: Oxford University Press.
Espiritu, Yen Le. 1992. “Ethnicity and Panethnicity.” In Espiritu, Asian American
Panethnicity. Philadelphia: Temple University Press, pp. 1–18.
35
Fadiman, Anne. 2012. The Spirit Catches You and You Fall Down. New York: Farrar,
Straus, and Giroux.
Fraser, Nancy. 1989. “What’s Critical about Critical Theory? The Case of Habermas
and Gender.” In Fraser, Unruly Practices. Minneapolis: University of Minnesota
Press, pp. 113–143.
Garfinkel, Alan. 1981. Forms of Explanation. New Haven, CT: Yale University Press.
Glasgow, Joshua. 2009. A Theory of Race. New York: Routledge.
Gooding- Williams, Robert. 1998. “Race, Multiculturalism, and Democracy.”
Constellations 5: 18–41.
Gracia, Jorge J. E. 2007. Race or Ethnicity? On Black and Latino Identity. Ithaca,
Haslanger, Sally. 2012. Resisting Reality: Social Construction and Social Critique.
Oxford: Oxford University Press.
Haslanger, Sally. 2014. “Studying While Black: Trust, Opportunity, and Disrespect.”
DuBois Review 11(1): 109–136.
Haslanger, Sally. 2016. “What Is a (Social) Structural Explanation?” Philosophical
Studies 173: 113–130.
Haslanger, Sally. 2017a. “Racism, Ideology, and Social Movements.” Res Philosophica
94(1): 1–22.
Haslanger, Sally. 2017b. “Failures of Individualism: Materiality.” Presented at the First
Annual Critical Social Ontology Workshop, St. Louis, MO.
Haslanger, Sally. Forthcoming. “Going On, Not in the Same Way.” In Alexis Burgess,
Herman Cappelen, and David Plunkett, (eds.), Conceptual Engineering and
Conceptual Ethics. Oxford: Oxford University Press.
Hardimon, Michael. 2003. “The Ordinary Concept of Race.” Journal of Philosophy
100(9): 437–455.
Hardimon, Michael. 2017. Rethinking Race: The Case for Deflationary Realism.
Cambridge, MA: Harvard University Press.
Herzog, Donald. 1998. Poisoning the Minds of the Lower Orders. Princeton,
NJ: Princeton University Press.
Hochman, Adam. 2017. “Replacing Race: Interactive Constructionism about Racialized
Groups.” Ergo 4(3): 61–92.
Irwin, Neil, Claire Cane Miller, and Margo Sanger-Katz. 2014. “America’s Racial
Divide, Charted.” New York Times, August 19. https://www.nytimes.com/2014/
08/20/upshot/americas-racial-divide-charted.html
Jamal, Amaney, and Nadine Naber. 2008. Race and Arab Americans before and
after 9/11: From Invisible Citizens to Visible Subjects. Syracuse, NY: Syracuse
University Press.
Jeffers, Chike. 2013. “The Cultural Theory of Race: Yet Another Look at Du Bois’s ‘The
Conservation of Races,’” Ethics 123(April 2013): 403–426.
Kapusta, Stephanie. 2016. “Misgendering and Its Moral Contestability.” Hypatia
31(3): 502–519.
36
36 • W h a t I s  R a c e ?
Kendig, Catherine. 2011. “Race as a Physiosocial Phenomenon.” History and Philosophy

of the Life Sciences 33(2):191–222.
Kitcher, Philip. 2007. “Does ‘Race’ Have a Future?” Philosophy and Public Affairs
35(4):293–317.
Lear, Jonathan. 1986. “Transcendental Anthropology.” In Philip Pettit and John
McDowell (ed.), Subject, Thought and Context. Oxford: Oxford University Press,
pp. 267–298.
Mallon, Ron 2003: “Social Construction, Roles, and Stability.” In Frederick F. Schmitt
(ed.), Socializing Metaphysics: The Nature of Social Reality. Lanham, MD: Rowman
and Littlefield, pp. 327–353.
Mallon, Ron. 2006. “‘Race’: Normative, Not Metaphysical or Semantic.” Ethics
116(3): 525–551.
Marx, Karl. 1843. Letter to Ruge, Kreuzenach, September. http://marx.eserver.org/
1843-letters.to.arnold.ruge/1843.09-ruthless.critique.txt
McPherson, Lionel, and Tommie Shelby. 2004. “Blackness and Blood: Interpreting
African American Identity.” Philosophy and Public Affairs 32(2): 171–192.
Mills, Charles W. 1997. The Racial Contract. Ithaca, NY: Cornell University Press.
Mills, Charles W. 1998. Blackness Visible. Ithaca: Cornell University Press.
Mills, Charles W. 2017. Black Rights/White Wrongs: The Critique of Racial Liberalism.
Omi, M., and H. Winant. 1994. Racial Formation in the United States.
New York: Routledge.
Quine, W. V. O. 1953. “On What There Is.” In Quine, From a Logical Point of View.
Cambridge, MA: Harvard University Press, pp. 1–19.
Risjord, M. W. 2000. Woodcutters and Witchcraft: Rationality and Interpretive Change
in the Social Sciences. Albany: State University of New York Press.
Root, Michael. 2001. “The Problem of Race in Medicine.” Philosophy of the Social
Sciences 31(1): 20–39.
Root, Michael. 2003. “The Use of Race in Medicine as a Proxy for Genetic Differences.”
Philosophy of Science 70(5): 1173–1183.
Root, Michael. 2009. “Measurement Error in Racial and Ethnic Statistics.” Biology and
Philosophy 24: 375–385.
Rosen, Gideon. 2015. “Real Definition.” Analytic Philosophy 56(3): 189–209.
Saul, Jennifer M. 2012. “Politically Significant Terms and Philosophy of
Language: Methodological Issues.” In Sharon L. Crasnow and Anita M. Superson
(eds.), Out from the Shadows: Analytical Feminist Contributions to Traditional
Philosophy. Oxford: Oxford University Press, pp. 195–214.
Schroeter, Laura, and Francois Schroeter. 2015. “Rationalizing Self-Interpretation.”
In Chris Daly (ed.), The Palgrave Handbook of Philosophical Methods.
New York: Palgrave-Macmillan, pp. 419–4 47.
Sewell, William. 2005. “The Concept(s) of Culture.” In Gabrielle M. Spiegel (ed.),
Practicing History: New Directions in Historical Writing after the Linguistic Turn.
New York and London: Routledge, pp. 76–95.
37
Shelby, Tommie. 2005. We Who Are Dark. Cambridge, MA: Harvard University Press.
Shelby, Tommie. 2014. “Racism, Moralism, and Social Criticism.” DuBois Review
11(1): 57–74.
Spencer, Quayshawn. 2014. “A Radical Solution to the Race Problem.” Philosophy of
Science 81(5): 1025–1038.
Sundstrom, Ronald. 2002. “Race as a Human Kind.” Philosophy and Social Criticism
28(1): 91–115.
Stocking, George Jr. 1994. “The Turn-of-the-Century Concept of Race.” Modernism/
Modernity 1(1): 4–16.
Taylor, Paul. 2004. Race: A Philosophical Introduction. Cambridge: Polity Press.
Taylor, Paul. 2016. Black Is Beautiful: A Philosophy of Black Aesthetics. Hoboken,
NJ: Wiley-Blackwell.
Zack, Naomi. 2002. Philosophy of Science and Race. New York: Routledge.
38
2 C U LT U R A L C O N S T R U C T I O N I S M
Chike Jeffers
Human races are social constructions. What I mean by this is that,

while there are aspects of racial diversity among humans that may
be studied by natural science, the fundamental factors making it the
case that races exist are sociohistorical in nature. Racial distinctions
have come to be and continue to exist in the present as a result of
the ways that we as humans have interacted and organized our af-
fairs over time. To believe this, as I do, is to be a social construc-
tionist about race. Like Sally Haslanger in the first chapter of this
book, one of my aims in what follows is to explain why this is the
most attractive position on the metaphysics of race. More distinc-
tively, however, I want to argue that it is important to distinguish
between two kinds of social constructionism, which I will call po-
litical constructionism and cultural constructionism. In contrast with
Haslanger’s political constructionist account of race, I will defend a
cultural constructionist account.
The first section of the chapter provides a framework for com-
paring theories of race and then offers critiques of biological realist
and anti-realist positions on race, delivered from a perspective that
is neutral between the two kinds of social constructionism that
I will differentiate in the chapter’s second section. In the second
section, once I have distinguished political from cultural construc-
tionism, I discuss why political constructionism might be seen as
the default position for a social constructionist. The third and final
section then provides an explanation and defense of my version
of cultural constructionism. Among my tasks in this section will
be spelling out the normative underpinnings and implications of
my view.
39
Cultural Constructionism • 39
2.1. Why Social Constructionism?

People have used and still use the English word ‘race’ and its etymologically
related cognates in other languages in a variety of ways. This makes it impor-
tant to achieve some clarity about the subject matter to be dealt with before
fruitful metaphysical discussion can take place. For example, people some-
times treat the terms ‘race’ and ‘ethnicity’ as completely interchangeable. It
seems necessary to me, though, that we avoid paying too much attention to
this kind of usage if we are to see any reason for debating whether races are
real and what races fundamentally are. Few people debate whether ethnic
groups exist and, although saying what they are might be slightly more con-
tentious, most experts would agree that we are talking about a kind of social
category. Thus, if races were nothing other than ethnicities, there would be
little reason to debate whether they exist and whether they are fundamentally
biological or social. We get closer to understanding debates about the nature
and reality of race, though, when we acknowledge the widespread assump-
tion that people of different ethnic backgrounds can be members of the same
race (e.g., white people may be Irish, Italian, Russian, something else). This
belief is commonly held among those who agree that race is real, even if they
disagree about whether race is biological or social in kind, and those who
do not believe that race is real can be expected to agree as well, though they
might be more careful to say that this is a matter of what would be the case if
races were real. Shared assumptions of this sort evidently suggest that there is
some common ground, some useful starting point for discussion, that we can
isolate before going on to investigate why and how people disagree about the
nature and reality of race.
The most helpful attempt at making precise what this starting point is,
in my view, would be Michael Hardimon’s account of the “logical core” of
the concept of race.1 Hardimon claims this logical core can be summarized
as follows:
(1) The concept of race is the concept of a group of human beings distin-
guished from other human beings by visible physical features of the
relevant kind.
(2) The concept of race is the concept of a group of human beings whose
members are linked by a common ancestry.
1. Michael O. Hardimon, “The Ordinary Concept of Race,” Journal of Philosophy 100

(September 2003): 441.
40
40 • W h a t I s  R a c e ?
(3) The concept of race is the concept of a group of human beings who orig-
inate from a distinctive geographic location.2
Each subsequent thesis here can be seen as building upon and explaining what
comes before it. People possess visible physical features of various kinds for
various reasons (their sex, their lifestyle choices, etc.), but the kind of phys-
ical features that distinguish them as members of races are inherited from
their parents, as races are groups whose members are linked by common an-
cestry. These features at least somewhat reliably indicate where in the world
the ancestors of group members lived, as races are groups who originate
from a distinctive geographic location. Regarding this last point, Hardimon
notes: “[t]‌he aboriginal habitat of common-sense and conventional races is
fixed by the location the ancestors of the members of those groups occupied
immediately prior to the advent of European oceangoing transport, which is
to say around 1492.”3 Thus two Canadians, both of whose heritage in Canada
goes back many generations, may nevertheless be obviously racially different
because the physical features of one are indicative of ancestors located in
Europe at the time Hardimon suggests is relevant, while the features of the
other indicate ancestors located in sub-Saharan Africa.
According to Hardimon, “[o]‌ne of the most striking results of our account
of the logical core of the ordinary concept of race is that race turns out to be
relatively unimportant.”4 One way to explain what he means by this is to say
that, if one sees a man who happens to be Chinese and rightly guesses on the
basis of his appearance that most of his ancestors in the fifteenth century lived
somewhere in East Asia, it is not clear that one has recognized anything of
great significance. It is certainly the case that, historically, in the West, recog-
nizing someone as “Oriental” was often thought to license various inferences
about the character and capacities of the person in question, but Hardimon
argues that the three theses at the core of the concept of race do not, by them-
selves, necessarily imply that we can learn anything of interest about people
merely by noticing the connection between their appearance and their place
of ancestral origin.
One useful way to compare metaphysical positions in the philosophy
of race is to see them as diverging with regard to what significance, if any,
2. Ibid., 442, 445, 447.

3. Ibid., 447–4 48.
4. Ibid., 451.
41
they accord to the differences in human appearance and ancestral place of

origin underlying racial categorizations. The dominant view about race from
some time in the early modern era until at least the latter part of the twen-
tieth century has been a biologically essentialist realism, according to which
we inherit, along with racial membership, a set of distinctive traits broader
than the particular physical features that indicate this membership, including
mental and behavioral tendencies, moral and intellectual talents or deficien-
cies, and physiological characteristics beyond a distinctive appearance as well.
Thus we have a long tradition, especially in the West, of taking appearance
and ancestry to be very significant indeed. More recently, some philosophers
have jettisoned the essentialism of the past while maintaining that there are
nevertheless biologically significant divisions of the human species that we
can describe as races (a view that will be defended by Quayshawn Spencer in
Chapter 3).
An anti-realist about race, on the other hand, would take Hardimon to be
begging the question when he says he has revealed race to be unimportant, for,
insofar as he is describing something real, it is not clear that it should count as
race. It is impossible to deny that there are some regularities in how we look
that relate to where many of our ancestors lived, but anti-realists deny, first
of all, that these regularities are significant in the way they have traditionally
been taken to be significant and, second, they hold that to deny this implies
the nonexistence of races. They hold, in other words, that it is only if there is
some systematic and more broadly relevant significance to appearance and
ancestry of the type imagined by biological essentialists that we have reason
to affirm that races exist. From this perspective, which will be defended by
Joshua Glasgow in Chapter 4 of this volume, to discuss biologically defined
groupings of humans in the non-essentialist manner of some more recent bi-
ological realists is to leave the topic of race behind.
In contrast with both of these positions, social constructionists about race
take appearance and ancestry to be very significant, not as a matter of biolog-
ical reality that we discover by looking at the world, but as a matter of social
reality that we produce and maintain through widespread patterns of thought
and behavior. The remainder of this section will consist in arguments for
rejecting biological realism—whether essentialist or not—and anti-realism,
along with explanation and defense of the social constructionist alternative.
Reasons for dismissing traditional biological essentialism have by now be-
come familiar to many. Simply put, the view is baseless: the ways in which
it connects differences in outer appearance and ancestral place of origin to
imagined inner differences—differences in our blood assumed to cause
42
42 • W h a t I s  R a c e ?
systematic variation in how we think and act—are not currently and arguably
never could be supported by scientific research. Paul Taylor aptly notes that,
from the essentialist perspective (which he refers to as “classical racialism”),
the result of two races mixing can be compared to “diluting a potion.”5 This
idea that each race’s blood is like a potion, with the special characteristics of
the race being the powers of the potion, is at odds with (1) our overwhelming
genetic similarity as a species, (2) our genetic distinctness as individuals, and
(3) how much we still do not know, even regardless of race, about how genes
interact with each other and the environment to cause character traits. With
regard to this last point, whatever we learn as our understanding of the rela-
tionship between genes and personality increases, it is extremely unlikely that
it will validate the traditional essentialist ascription of stable sets of hereditary
traits to the wide and somewhat arbitrarily demarcated swaths of humanity
that common sense racial groupings are.
But what about non-essentialist biological realism? Whatever it is, it is not
baseless. Take Robin Andreasen’s cladistic account of race, which would have
us understand races as reproductively isolated breeding populations whose
biological relationships with each other can be depicted as an evolutionary
tree, the branches of which constitute different races.6 As evidence that we
can construct such a tree, Andreasen points us to the work that geneticist
Luigi Luca Cavalli-Sforza has done, with others, to map the history of human
evolution by measuring the genetic distance between populations. Cavalli-
Sforza’s work, while not without controversy, is credible scientific research.
Andreasen’s proposal for how to understand what races are thus leaves be-
hind biological essentialism’s implausible claims about character traits while
making use of fascinating evidence from the field of genetics.
But should we see groupings of humans like those represented by
the branches of Cavalli- Sforza’s tree as races? Anti- realists and social
constructionists agree that we should not.7 Note, first, Andreasen’s finding
that the tree appears to explode our normal idea of East Asians as a group, as
Southeast and Northeast Asians appear on “two distinct major branches.”8
5. Paul C. Taylor, Race: A Philosophical Introduction, 2nd ed. (Cambridge, UK: Polity,
2013), 50.
6. Robin O. Andreasen, “A New Perspective on the Race Debate,” The British Journal for the
Philosophy of Science 49 ( June 1998): 199–225.
7. Cavalli-Sforza himself also disagrees with calling them races, but I accept Andreasen’s claim
that we should not take this disavowal as decisive. See Andreasen, “A New Perspective,” 213.
8. Andreasen, “A New Perspective,” 212.
43
Southeast Asians branch off of a division that also includes Pacific Islanders,
New Guineans, and indigenous Australians, while Northeast Asians branch
off of a division that includes the indigenous peoples of the Americas,
Europeans, and “Non-European Caucasoids.”9 Thus, according to Andreasen,
it would be more accurate to represent Koreans as sharing a race with Germans
and people from Thailand as sharing with a race with Fijians than to represent
these two somewhat similar-looking peoples of eastern Asia as sharing a race
with each other! Note, second, something that Andreasen does not explic-
itly address: the categorization of South Asians as “Caucasoids.” This does
receive mention, however, in Quayshawn Spencer’s recent work arguing for
biological realism. Spencer’s argument relies on noting the overlap between
the racial categories of the US Census and what he takes to be a biologi-
cally significant division of the human species into five genetically clustering
partitions of populations, but he notes that one obstacle to complete overlap
is the lumping in the Census of “South Asians with Asians and not with
whites.”10
Anti-realists and social constructionists hold that these kinds of discrep-
ancies between common-sense racial classifications and biologically respect-
able accounts of how we may subdivide our species demonstrate that biology
ultimately undermines, rather than supports, our talk of races. There is, of
course, the option of seeing these discrepancies as a matter of natural science
correcting misconceptions in our common-sense classifications. Consider,
however, the way that such discrepancies lead us to abandon the core notion
that race involves how appearance is linked to ancestry, and think also of the
confusion this may engender. Imagine a young woman, born in England to
parents from Bangladesh, whose dark brown skin has marked her for her
whole life as a minority of foreign origin. What should she make of the idea
that it would be accurate to classify her as being of the same race as the ma-
jority? Faced with a choice between describing herself in relation to white
people as racially different in recognition of how her appearance has gener-
ated a particular experience and describing herself as racially the same on the
basis of the broadness of “Caucasian” or “Caucasoid” as a category, should
she see both options as equally reasonable? The anti-realist will reject both
options as misleading while, as a social constructionist, I would deem the first
9. Ibid.
10. Quayshawn Spencer, “A Radical Solution to the Race Problem,” Philosophy of Science 81
(December 2014): 1031. See Chapter 3 in this volume for Spencer’s most recent expression of
his view.
4
44 • W h a t I s  R a c e ?
option more illuminating. Either way, the point about the second option will
be that it conflicts with common sense in a way that is best addressed by giving
up the idea that it counts as a description of race and choosing to phrase its
insight into our development and diversity as a species in other terms.
If we therefore put aside biological realism as an approach to race, how do
we decide between anti-realism and social constructionism? Anti-realism cer-
tainly has much to be said for it. There is good reason to think that it is hard to
separate talk of race from traditional biological essentialism. Even Hardimon,
while asserting that the concept of race at its core is logically independent of
essentialist ideas, admits that, as a historical matter, the concept came into gen
eral usage already laden with essentialism. He adds six theses, which he calls
the “racialist development” and which amount to biological essentialism, to
the original three theses in order to give us what he calls “the ordinary concep-
tion of race,” and he acknowledges that “[w]‌hen the logical core first entered
the historical scene, it was already articulated by the racialist development.”11
If essentialism is thus a heavy historical legacy to be overcome, it should fur-
thermore be acknowledged that we are by no means yet near to overcoming
it. Many think that the fact that people today often dress up essentialist ideas
in talk of ‘culture’ rather than ‘race’ only extends the power of such thinking.
Lawrence Blum gives us an example of this kind of talk: “These people ( Jews,
whites, Asians) just are that way (stingy, racist, studious); it’s part of their cul-
ture.”12 It seems clear, from an anti-realist perspective, that we will not defeat
this insidious tendency to essentialize by encouraging continued belief in and
talk about races as real. We should instead expose all such belief and talk as
mistaken or confused (not to mention, in many cases, hateful and oppressive).
11. Hardimon, “The Ordinary Concept of Race,” 451, 453. According to the six additional
theses, a race is “(4) a natural division of the human species into a hierarchy of groups that sat-
isfy the conditions specified in (1)–(3); (5) a group of human beings satisfying the conditions
specified in (1)–(3) which is characterized by a fixed set of fundamental, ‘heritable,’ physical,
moral, intellectual, and cultural characteristics common and peculiar to it; (6) a group of
human beings satisfying the conditions specified in (1)–(3) whose distinctive visible phys-
ical features are correlated with the moral, intellectual, and cultural characteristics that are
common and peculiar to it; (7) a group satisfying the conditions specified in (1)–(3) that
possesses an ‘essence’ which explains why it is that the group has the distinctive visible features
that it does, why it is that the group has the particular moral, intellectual, and cultural charac-
teristics it does, and the correlation between the two; (8) a group of human beings satisfying
the conditions specified in (1)–(3) whose members necessarily share its “essential” characteris-
tics; (9) a group of human beings satisfying the conditions specified in (1)–(3) whose essential
characteristics constitute the essence of its members.” Ibid., 452.
12. Lawrence Blum, “I’m Not a Racist, But . . .”: The Moral Quandary of Race (Ithaca,
NY: Cornell University Press, 2002), 134.
45
Despite the attractions of anti-realism, however, I reject it. When evalu-

ated as a position on the significance or insignificance of appearance as related
to ancestral place of origin, I think it fails to accurately describe and explain
reality. Think once more of the Englishwoman of South Asian descent. How
she looks as a result of where her ancestors are from has indeed been very
significant, not because of differences of character rooted in her blood, but
because of the social situation of South Asians in England. While anti-realism
helps us appreciate ways in which differences of appearance and ancestry are
not significant, only social constructionism redirects our focus toward ways
in which they really are. Racial difference, it should be noted, is not wholly
unrelated to that which we may study by means of natural science because
it is partly a matter of physical, biological, and geographical differences: it
involves how distinctive physical appearances indicate biological connections
of descent that tie us to particular geographical regions of the world. It is, how-
ever, only through social and historical processes that the particular physical,
biological, and geographical differences that we recognize as racial have come
to gain some relatively stable significance. It is only because racial distinctions
are, fundamentally, significant social distinctions that we can say, in spite of
the falsity of biological essentialism, that racial difference is not an illusion.
One might object, from an anti-realist perspective, that we have here
the same problem we encountered with non- essentialist biological re-
alism: namely, the problem of changing the subject. I claimed earlier that
anti-realists and social constructionists alike see the mismatches between
common-sense racial categorizations and those invoked by non-essentialist
biological realists as evidence that what those biological realists are discussing
is not, in fact, race. The question then becomes why anti-realists cannot make
a similar charge against social constructionists, given that they too are at odds
with common sense. According to common sense, how we look and where we
are from determine our race, regardless of social and historical conditions. It
also remains to a great extent common sense that people of different races nat-
urally inherit different characteristics or tendencies in thought and behavior.
If social constructionists deny all this, why should they be seen as still talking
about race?
The first thing to be said in response is that it matters that the divergence
from common sense in the case of social constructionists does not involve ra-
cial categorizations, as in the case of non-essentialist biological realists. Social
constructionists accept common-sense racial categorizations, precisely be-
cause it is only by looking at what people as a matter of sociohistorical contin-
gency widely accept that we can determine what races there are from a social
46
46 • W h a t I s  R a c e ?
constructionist perspective. The pattern of overlaps and divergences in cate-

gorization that we see with non-essentialist biological realists suggests that
they are talking about a kind of distinction that is similar and perhaps closely
related to racial difference, but not the same thing. The lack of divergence
in categorization by social constructionists, on the other hand, suggests that
what is at stake from this perspective is the same distinction that is at stake in
ordinary talk of race, despite the fact that social constructionists offer an al-
ternative and disruptive understanding of the nature of this distinction.
As an analogy, compare, on the one hand, the divergences between the
known facts of a real-life murder case and the plot of a film loosely based on
the case, and, on the other hand, the divergence or shift in understanding
brought about by a new piece of evidence that appears to exculpate the
person convicted of the murder. The murderer in the film may be similar to
the person we take to be the murderer in real life, but they are not the same.
By contrast, the person convicted of the murder is the same being even after
we come to believe that he is innocent. Social constructionism about race
involves this kind of shift in understanding what something is, not a change
in subject matter.
In relation to this same objection, though, we should also note that anti-
realists are generally not so obtuse as to deny that there are significant so-
cial distinctions that people normally discuss using the language of race.
They claim, however, that to speak clearly and accurately about these social
distinctions, we must admit that there are no such things as races and use a
different term to refer to the socially differentiated groups we wish to discuss.
Thus Kwame Anthony Appiah at one point held that there are no races but
there are racial identities, and it has remained Blum’s position that races do
not exist but racialized groups do.13 Conceptual moves such as these allow
anti-realists to claim that the social constructionist confuses rather than
clarifies things by continuing to use the term traditionally associated with bi-
ological essences when referring to a social and historical phenomenon.
13. See K. Anthony Appiah, “Race, Culture, Identity: Misunderstood Connections,” in Appiah
and Amy Gutmann, Color Conscious: The Political Morality of Race (Princeton, NJ: Princeton
University Press, 1996), 30–105, and Blum, Chapters 7–8 (131–163). Appiah has not made
the same distinction in more recent work and has thus drifted toward social constructionism.
See, for example, Kwame Anthony Appiah, “Does Truth Matter to Identity?” in Jorge J. E.
Gracia (ed.), Race or Ethnicity? On Black and Latino Identity (Ithaca, NY: Cornell University
Press, 2007), 19–4 4, and Lines of Descent: W. E. B. Du Bois and the Emergence of Identity
(Cambridge, MA: Harvard University Press, 2013). For Blum’s continued commitment to his
position, see Blum, “Racialized Groups: The Sociohistorical Consensus,” The Monist 93 (April
2010): 298–320.
47
What would make switching terms most appropriate, however, is a belief

that everyday talk about races is talk about some nonexistent stuff that must
be distinguished from the real social groups worth discussing. If, as I have
already suggested, the best account of the social construction of races would
have us acknowledge the existence of the groups referred to in everyday talk
but then provide a different account of their nature, then it is, in fact, a matter
of clarifying things to continue to use the same term, while it would be at the
very least unnecessary and possibly even misleading to switch terms.
Everyday talk about black people, for example, is best understood as refer-
ring to a real group to which one can belong, even if such talk often involves
false assumptions about what comes naturally to people of recent sub-Saharan
African ancestry. What needs to be recognized in order to properly trans-
form everyday understandings of this group is that it is distinct as a group not
merely because some of us have ancestors who, in the fifteenth century, lived
in sub-Saharan Africa, but rather because being visibly of this ancestry or in
some other way being known to be of this ancestry has, particularly since the
sixteenth century, been socially significant. Through, among other things, the
horrors of slavery, the injustices of colonialism, and, contrastingly, the activism
of movements promoting black pride, to be black in our world has been to be-
long to a meaningfully distinct category of human beings. When we recognize
this as a social and historical phenomenon that forms one part among others
of the larger global story of race, I believe we describe and explain reality in an
informative manner without indulging in essentialist myths, without taking
detours through respectable biology toward a separate subject matter, and
without mistakenly suggesting that talk of race is talk of nothing real.
2.2. Political versus Cultural Constructionism

I would not be so bold as to assume that the account of what races are in the
previous section would receive assent in its every detail by all who would call
themselves social constructionists. Details may be disputed, and perhaps major
features of the account as well. Nevertheless, the account in the previous sec-
tion is neutral between two different kinds of social constructionism, and it is
my view that the divergence between these two is highly significant. Fruitful
philosophical discussion of race going forward, I would argue, will require se-
rious attention to and critical comparison of these two positions so that their
potential merits and disadvantages may become clearer and those participating
in the debate about race may develop views on why to prefer one or the other.
48
48 • W h a t I s  R a c e ?
One way to begin introducing the difference between the two views is to
note that there is a certain vagueness in saying that race is socially constructed.
What kinds of social relations, processes, or states of affairs are involved in the
construction of racial reality? One answer is that race is made real wholly or most
importantly by hierarchical relations of power. I call this political constructionism.
Stated abstractly, the position need not commit one to any particular under-
standing of history, but, as a matter of fact, social constructionists who think
this way generally believe that the specific hierarchical relations of power that
make race real are those constituted or brought about by European imperialism
and the various social structures it created—in other words, the global sociopo-
litical system of white supremacy. Taylor, for example, conceives of races in the
modern world as “the probabilistically defined populations that result from the
white supremacist determination to link appearance and ancestry with social lo-
cation and life chances.”14
Let us look now at Haslanger’s view of race, as she is especially explicit about
offering a form of what I have called political constructionism. She gives us this
account of what a race is:
A group G is racialized relative to context C iffdf members of G are (all and
only) those:
(i) who are observed or imagined to have certain bodily features presumed
in C to be evidence of ancestral links to a certain geographical region
(or regions);
(ii) whose having (or being imagined to have) these features marks them
within the context of the background ideology in C as appropriately
occupying certain kinds of social position that are in fact either subordi-
nate or privileged (and so motivates and justifies their occupying such a
position); and
(iii) whose satisfying (i) and (ii) plays (or would play) a role in their system-
atic subordination or privilege in C, that is, who are along some dimen-
sion systematically subordinated or privileged when in C, and satisfying
(i) and (ii) plays (or would play) a role in that dimension of privilege or
subordination.15
14. Taylor, Race, 89–90.

15. Sally Haslanger, Resisting Reality: Social Construction and Social Critique (New York: Oxford
University Press, 2012), 236–237. Note that Haslanger, unlike Blum, treats “groups that are
racialized” as synonymous with “races.”
49
Note, first, the relationship between (i) and Hardimon’s three theses.
Haslanger, like Hardimon, takes race to involve visible features that relate us
through ancestry to a certain part of the world, although there is the inter-
esting difference here that Haslanger speaks also of imagined features and pre-
sumed ancestral links, rather than simply observed features and actual links.
We see in (ii) and (iii) Haslanger’s commitment to social construc-
tionism. She does not take the connection between features of the body,
ancestral relations, and ties to particular geographic regions to be sufficient
for racial membership. In order to amount to something distinguishable as
such, these attributes must furthermore figure in widely shared patterns of
thought about how different kinds of people can be compared with one an-
other, and they must moreover serve as the ground for actual differences in
how people in society position themselves and find themselves positioned in
relation to one another. As she puts it, adapting the classic feminist explana-
tion of how gender relates to sex, “race is the social meaning of the geograph-
ically marked body.”16
But Haslanger will not count just any kind of commonly drawn distinc-
tion and associated relationship between different groups as racial. She holds
that these distinctions in thought and patterns in group relations must be
matters of subordination and privilege. This is, in fact, what explains why
imagined and presumed attributes count just as much as accurately perceived
ones for her. If you are not actually linked by descent to the place where many
guess that you have roots, but the mistaken perception that many have regu-
larly results in your experience of advantage or disadvantage comparable to
that which is experienced by people who are actually linked by descent to that
place, then Haslanger aims to capture your social reality in her account of
what it means to belong to a race.
Haslanger’s approach is distinctive and, in several respects, controversial,
but I take her to be representative of the norm among social constructionists
in being a political constructionist. If I am right about this, one irony is that
what many take to be the pioneering philosophical account of race as a social
construction rather than a natural kind is, in fact, not a political construc-
tionist account. In 1897, at the first meeting of the American Negro Academy,
W. E. B. Du Bois presented a paper entitled “The Conservation of Races,”
in which he attempts to answer the question of “the real meaning of race.”17
16. Ibid., 236.

17. W. E. B. Du Bois, “The Conservation of Races,” in The Oxford W. E. B. Du Bois Reader, ed.
Eric J. Sundquist (New York: Oxford University Press, 1996), 39.
50
50 • W h a t I s  R a c e ?
Du Bois argues that natural science has failed to clarify and distinguish the
criteria for race membership, but that it remains the case that “subtle forces”
have divided us into “races, which, while they perhaps transcend scientific
definition, nevertheless, are clearly defined to the eye of the historian and so-
ciologist.”18 Having thus committed himself to the view that races are funda-
mentally sociohistorical, he goes on to define a race as “a vast family of human
beings, generally of common blood and language, always of common history,
traditions and impulses, who are both voluntarily and involuntarily striving
together for the accomplishment of certain more or less vividly conceived
ideals of life.”19
With its emphasis on differing “traditions” and “ideals of life,” I would
classify this definition of race as a form of cultural constructionism. For the
cultural constructionist, participation in distinctive ways of life, rather than
positioning in hierarchical relations of power, is what is most important in
making race real. As I have argued elsewhere, Du Bois can be seen not merely
as offering a cultural theory of race, but also as explicitly distancing himself
from the political approach.20 Writing in the wake of Plessy v. Ferguson, the
US Supreme Court case that cemented the system of Jim Crow segregation,
Du Bois identifies his purpose in thinking about the nature of race as political
but claims that, when seeking the substance of race, we will have to look be-
yond political conditions:
It is necessary in planning our movements, in guiding our future devel-

opment, that at times we rise above the pressing, but smaller questions
of separate schools and cars, wage-discrimination and lynch law, to
survey the whole question of race in human philosophy and to lay, on a
basis of broad knowledge and careful insight, those large lines of policy
and higher ideals which may form our guiding lines and boundaries in
the practical difficulties of everyday.21
As this passage suggests, it is not only of theoretical but also practical sig-
nificance for Du Bois that he arrives at his cultural definition of race. He ends
18. Ibid., 40.

19. Ibid.
20. Chike Jeffers, “The Cultural Theory of Race: Yet Another Look at Du Bois’s ‘The
Conservation of Races,’” Ethics 123 (April 2013): 403–426.
21. Du Bois, “The Conservation of Races,” 39.
51
up arguing that the advancement of civilization has been a matter of the pur-
suit of different ideals by different races, and then claims on that basis that
it is incumbent upon African Americans to proliferate institutions that will
help them to develop black culture and thus contribute something special to
world, rather than buckle under the pressure of American racism by devaluing
and seeking to be rid of their racial distinctness.
Cultural constructionism is thus an alternative to political construc-
tionism with a long history. It has sometimes been recognized and discussed
as an alternative, even if not by the name I have given it, as when Taylor
describes the “racial communitarian” as believing that “we construct races
by creating cultural groups.”22 Taylor offers arguments against this position,
which I will consider in the following section. More often than not, though,
cultural constructionism is simply ignored as a distinct option. The reason
for this, I think, is that political constructionism is such a common position
among social constructionists that many seem not to consider the possibility
that one might be a social constructionist without being a political construc-
tionist. While unfortunate, this is not, in my view, mysterious. Indeed, I think
there is much that can be said in explanation of why, to many, it seems just ob-
vious that political constructionism is the right way to think about the social
construction of race.
Note, first, that political constructionism appears to offer the best way
from a social constructionist perspective to understand the historical devel-
opment of racial difference. While biological realists may be willing to envi-
sion races becoming distinct from each other as much as tens of thousands of
years ago, social constructionists tend to treat races as products of the modern
era (that is to say, the last five centuries or so), and the most obvious way to
explain how they came about given this assumption is to point to the hier-
archical social structures created by European imperialism. Taylor is notably
flexible in being willing to see “race-thinking” in a broad sense—defined as
“assigning generic meaning to human bodies and bloodlines”—as already ex-
isting in the ancient world, but he sharply distinguishes what existed before
from “modern racialism,” or the system of thought and practice that “relies
mainly on skin color, facial features, and hair texture to divide humankind
into four or five color-coded groups—black, brown, red, white, yellow.”23 His
22. Taylor, Race, 100.

23. Ibid., 16, 18. One form of pre-modern “race-thinking” that Taylor does not say much about
is that which can arguably be found in medieval thought, especially that of the Islamic world.
The most convincing philosophical account of how modern racialism differs from that which
52
52 • W h a t I s  R a c e ?
willingness to recognize antecedents does not prevent him from saying that
“modern Europe invented the concept of race,” thus emphasizing that how-
ever old the process of assigning meaning to bodies and bloodlines may be,
modern Europeans “developed a vocabulary that highlights certain aspects of
this process” and then “refined it, exported it, tried to make it scientific, and
built it into the foundation of world-shaping . . . developments in political
economy.”24
That last part is, of course, hugely important. The standard social construc-
tionist story is that Europe’s colonization of most of the rest of the world, with all
the voluntary and involuntary movements, new assortments, and reorganized
institutional relations of peoples this entailed, brought it about that differences
of appearance and ancestry gained significance in the modern era in a system-
atic and global manner unlike anything that came before. This is why, as we
have seen, Taylor argues for the current reality of races by defining them as
populations distinct from each other not merely in appearance and ancestry,
but also in probabilities of social location and life chances, with the distinctness
in these latter regards being directly or indirectly the result of modern European
efforts to establish the supremacy of white people over all others.
What alternative story might a cultural constructionist tell about how
races as we know them came to be? Unfortunately, it will not help to look to
“The Conservation of Races.” Du Bois tells a tale of nomadic groups settling
in cities and beginning to specialize in different ways of life, followed eventu-
ally by the coalescing of cities into nations that constitute racial groups and
which are characterized by “spiritual and mental differences.”25 As fascinating
as this story may be, it is frustratingly vague on matters of chronology and
geography, making it hard to evaluate, much less accept. Du Bois is specific
only when he gets to the modern age, congratulating “the English nation”
for its role in developing the ideals of “constitutional liberty and commer-
cial freedom,” the “German nation” for “science and philosophy,” and the
“Romance nations” for “literature and art.”26 Note here also that, while Du
Bois recognizes “whites” as constituting one of two or three “great families
precedes it must, in my view, pay serious attention to this part of the world’s intellectual history.
For a starting point, see Paul-A. Hardy, “Medieval Muslim Philosophers on Race,” in Julie K.
Ward and Tommy L. Lott (eds.), Philosophers on Race: Critical Essays (Malden, MA: Blackwell,
2002), 38–62.
24. Ibid., 19, 20.
26. Ibid., 42.
53
of human beings” from a scientific perspective (along with “Negroes” and

“possibly the yellow race”), he disaggregates them into smaller families, in-
cluding those just listed, when differentiating between sociohistorical races
as he understands them.27 Given that it was in his time and remains in our
time socially significant to be white, we might even say that the charge of
changing the subject as evidenced by leaving behind common-sense racial
categorizations—leveled by anti-realists and social constructionists against
non-essentialist biological realists—applies also to his view. A political con-
structionist explanation of white distinctness as resulting from the white
supremacist construction of social hierarchies based on appearance and an-
cestry thus seems clearly preferable.
Secondly, political constructionism appears most persuasive in
explaining how race is made real socially, partly through the impact of racial
categorizations on personal experience. Consider a young black man in the
United States who has grown up as the adopted son of white parents in a
nearly all-white suburban community. Depending on how he was raised and
the interests of his friends, such an individual may feel culturally quite dis-
connected from black people as a group and rather continuous with those
around him. This fact, however, may have little bearing on his feeling that
his appearance and ancestry are very significant because of how often he is
looked at suspiciously by strangers, how store owners sometimes harass him,
how encounters with the police go differently for him than for his friends, etc.
This would seem to suggest that cultural difference may be unimportant to
one’s experience of race, while social hierarchy is essential or at least especially
deep in its impact given its connection to our ability to feel respected and
valued in our social context.
Another way we can make the foregoing point is by considering, in sharp
contrast, how race might be experienced by some as having little to no impact
on their life. The concept of white privilege, for example, is the concept of a
condition of which it is characteristic that having it makes it more likely that
one will be unaware of its existence and unaware that one has it.28 Many white
people living in social contexts characterized by racial multiplicity neverthe-
less go through much of life reflecting comparatively seldom upon matters of
race, and it can seem particularly unimportant to their sense of who they are
as individuals. This is best explained from a social constructionist perspective
27. Ibid., 39.

28. In this respect, it is similar to some but not all mental illnesses. I thank Tina Roberts-Jeffers
for this point.
54
54 • W h a t I s  R a c e ?
not by accepting that race has done little to shape their lives and identities, but
rather by noticing that it is precisely one aspect of how race may shape us that
one’s whiteness may be systematically hidden from view even as the relative
privileges flowing from said whiteness are enjoyed. The enjoyment of white
privilege is not merely compatible with but, more importantly, facilitated by
a lack of consciousness about race.29 Recognizing how an imperative to justi-
fication may come along with the conscious possession of privilege thus helps
clarify much about how many white people experience their whiteness, and
this is best explained as a matter of social hierarchy rather than simply as cul-
tural difference.
Finally, moving from personal experience to consideration of the social
landscape at large, political constructionism can seem best attuned to how
race matters socially in terms of major events and trends. As I first wrote this
during the summer of 2016, some of the ways in which race had been a prom-
inent feature of current affairs in the recent past included: growing attention
to police violence against black people and the corresponding growth of the
Black Lives Matter movement; the racism associated with Donald Trump’s
campaign for presidency of the United States, especially in the form of xeno-
phobia directed at Latinos and Muslims; concern about the role of racism in
the Brexit vote in the United Kingdom; concern over the role of racism in the
water pollution crisis in Flint, Michigan; racist reactions to Syrian refugees in
various Western countries; and activism and turmoil concerning racial justice
on university campuses in the United States and South Africa. At issue in all
cases, arguably, was the problem or response to the problem of the classifica-
tion of non-white people as less valuable and, in many cases, as particularly
threatening to a social order. If this is how race matters, how could it not be
clear that race is fundamentally a matter of social hierarchy?
2.3. Why Cultural Constructionism?

I do not find it strange, then, that political constructionism has been taken to
be the default position among social constructionists. So why would I choose
to defend cultural constructionism? The first thing I should say in order to
clarify my position is that I take race, at present, to be both politically and
29. Charles Mills has famously addressed this by developing the epistemological concept of
“white ignorance.” For the most focused expression of his view, see Mills, “White Ignorance,” in
Shannon Sullivan and Nancy Tuana (eds.), Race and Epistemologies of Ignorance (Albany: State
University of New York Press, 2007), 13–38.
5
culturally constructed. If the only way we could make sense of the distinction
between political and cultural constructionism were viewing the former as
denying that race is in any way culturally constructed and the latter as denying
that race is in any way politically constructed, I would deny that we have to
choose one or the other and reject both positions as false. Happily, I think
many other social constructionists who I would normally identify as political
constructionists would agree with me on this point.
Charles Mills, for example, when explaining his popular model of the
system of white supremacy as a “Racial Contract,” clarifies that this contract
should be understood as “creating not merely racial exploitation, but race it-
self as a group identity.”30 This point about hierarchy not merely being based
upon but rather generating racial difference is, of course, characteristic of
political constructionism. Still, many of Mills’s efforts at exploring the var-
ious dimensions of white supremacy include insightful attention to cultural
difference. For instance, while imploring political philosophers to take seri-
ously how race brings up questions of personhood, he encourages reflection
on ways in which personhood is linked to cultural membership. First, he
notes: “Colonization has standardly involved the denigration as barbaric of
native cultures and languages, and the demand to assimilate to the practices
of the superior race, so that one can achieve whatever fractional personhood
is permitted.”31 Resistance to racism thus often involves affirming the worth of
indigenous languages or, as in the Caribbean, creoles that deviate from the im-
perial standard. In the United States, Mills argues, “the construction of an ex-
clusionary cultural whiteness has required the denial of the actual multiracial
heritage of the country,” which means that white people “appropriated Native
American and African technical advances, language use, cultural customs, and
artistic innovations without acknowledgment, thereby both reinforcing the
image of nonwhites as subpersons incapable of making any worthwhile contri-
bution to global civilization and burnishing the myth of their own monopoly
on creativity.”32 As Mills sums up this discussion, he suggests that thinking
about race necessarily requires thinking about cultural difference:
Culture has not been central to European political theory because cul-
tural commonality has been presupposed. But once cultures are in
30. Mills, The Racial Contract (Ithaca, NY: Cornell University Press, 1997), 63.
31. Mills, Blackness Visible: Essays on Philosophy and Race (Ithaca, NY: Cornell University
Press, 1998), 115.
32. Ibid.
56
56 • W h a t I s  R a c e ?
contestation, hegemonic and oppositional, and linked with personhood,

they necessarily acquire a political dimension.33
If thinking about race requires thinking about cultures in contestation, it seems

clear that Mills believes race involves both differential power relations and par-
ticipation in different ways of life.
Assuming I am right that it is common among social constructionists to
recognize that both of these kinds of difference are somehow involved in race,
one might begin to suspect that distinguishing between ‘political construc-
tionism’ and ‘cultural constructionism’ is misleading and unnecessary. I think
this suspicion is wrong. What is required for the distinction to be useful is
that there be some social constructionists who think that differential power
relations are somehow most fundamental in the social construction of race
and some who, by contrast, accord that status to participation in different
ways of life. Mills, like Taylor, Haslanger, and others, fits the first descrip-
tion.34 I have already suggested that those who fit the second description are
less common, but, among those who have been prominent in debates about
the metaphysics of race, one example would be Lucius Outlaw.35
Let us consider more carefully, then, what might be involved in treating
politics or culture as more fundamental to the social construction of race. We
can define a maximally robust political constructionism as a view that takes
politics, in the sense of power relations, to be what matters most for the re-
ality of race in all of the following ways: (1) differential power relations are
what first brought racial difference into existence and are thus fundamental
in being the origin of races; (2) differential power relations count as most im-
portant in the present to the reality of race, which is to say that properly under-
standing any event, process, or state of affairs that involves race always requires
understanding how power relations are at stake, whereas there is nothing else
33. Ibid.
34. That politics is more fundamental than culture to the social construction of race for Mills
is especially clear in his “Multiculturalism as/and/or Anti-Racism?” in Anthony Simon Laden
and David Owen (eds.), Multiculturalism and Political Theory (Cambridge, UK: Cambridge
University Press, 2007), 89–114.
35. See Lucius T. Outlaw ( Jr.), On Race and Philosophy (New York: Routledge, 1996), espe-
cially the Introduction (1–21) and Chapters 6–7 (135–182). Outlaw’s position is unique be-
cause he not only treats culture as fundamental in the social construction of race, but also takes
race to be both social and biological in nature in such a way that races are appropriately called
“social-natural kinds” (7). The opposition between social constructionism and biological re-
alism that I depicted in the first section thus leaves out his distinctive view.
57
that must similarly always be understood; and (3) differential power relations
are essential to race, making it the case that if an egalitarian state of affairs in
which appearance and ancestry do not correlate with positions in a hierarchy
is achieved, race will be no more. I think many social constructionists are ro-
bust political constructionists according to these criteria.
A maximally robust cultural constructionism would, by contrast, hold
that (1) the origin of racial difference is to be found in divergences in ways
of life; (2) only cultural difference must always be understood in order to
understand the reality of race in the present; and (3) cultural difference is
essential to race, such that the end of distinctive ways of life would mean the
end of race. Confrontation between this bold position and political construc-
tionism as I have described it would perhaps be the most exciting way for
things to go in the rest of this chapter, but, unfortunately for those awaiting
such excitement, this is not a version of cultural constructionism I would de-
fend. Indeed, I disagree with each of these three points, which might lead one
to think it doubtful that I could deserve the title of cultural constructionist!
The reason I take on the title, in spite of not holding the maximally ro-
bust version of the position, is because I reject political constructionism in
a way that is, I think, best expressed by calling my view a form of cultural
constructionism. Before explaining how and why I reject it, however, let me
first acknowledge the crucial respect in which I do not challenge the political
constructionist account. I completely concede the first point about the origin
of race. In other words, I take European imperialism and the hierarchical so-
cial structures it created to be what gave rise to racial difference as we know
it. If admitting this were all it took to be a political constructionist, I would
have to identify as one.
I disagree, however, with the second and third points of the robust polit-
ical constructionist account for reasons involving my belief in the significance
of cultural difference to race’s existence and functioning as a social distinc-
tion. I reject the claim that politics is more important than culture at present.
I hold that they are of equal importance and, though I do not go so far as to
claim that culture is more important, this disagreement nevertheless already
commits me to putting additional emphasis on the significance of culture in
opposition to the political constructionist’s relative disregard. I also reject
the claim that the end of social hierarchy based on appearance and ancestry
would mean the end of race. Race as a social construction could live on past
the death of racism, in my view, given that racial groups could continue to
exist as cultural groups. Here too we see how my disagreement with the po-
litical constructionist allows me to uphold culture as particularly significant.
58
58 • W h a t I s  R a c e ?
While the political constructionist sees the end of racism as a potential future
transition from social reality to nothingness, I see the potential for a transi-
tion from cultural difference being one component of a social reality to being
the entirety of that reality.
To see most clearly why it is useful to call my view cultural construc-
tionism, though, it is necessary to pay attention to the role of values and
ideals in thinking about the nature and reality of race. The more we try to
draw a very sharp distinction between ethics and metaphysics, the less reason
there will be to say that my metaphysical stance should be described as cen-
tering culture more than politics. Race is fundamentally social, in my view,
but I do not take either politics or culture to be more fundamental in the
sense of being what is essential for the social reality of race. Culture cannot be
essential in this way if, as I hold, race is political at its origin. Politics cannot
be essential if, as I believe, a future in which race is merely cultural is possible.
This comparison seems, however, to leave the two equal in status.
Think now, though, about the fact that getting rid of unjust social
distinctions clearly ought to be our shared goal as human beings in society.
We therefore have a duty to work toward the end of race as a social reality
insofar as it is constituted by a hierarchy based on appearance and ancestry.
It is not so clear, by contrast, that we have any duty to work toward ending
the existence of different cultures. People differ in how permissible and valu-
able they take the perpetuation of cultural distinctions to be, but I am among
those who value cultural diversity and think that, at least in many cases, the
preservation of distinctive cultural traditions is desirable and admirable. The
continued existence of racial diversity as cultural diversity after the end of
racism is therefore, in my view, something good. As a result, one normative
implication of my position on race is that we should be orienting ourselves in
the present toward the eventual achievement of a world in which races exist
only as cultural groups. This vision for the future and the concern for valuing
the cultural aspects of race in the present that it entails makes it sensible to
say that culture is indeed fundamental, on my view. Thinking of cultural con-
structionism as including perspectives according to which we ought to actively
continue constructing races as cultural groups makes applying that label to my
view perfectly apt.
I will provide in the remaining space of this chapter a preliminary defense
of my version of cultural constructionism. I mentioned earlier that Taylor
criticizes cultural constructionism, and I will first consider the trilemma he
devises in order to do so. Responding to Taylor will be useful for establishing
the basic coherence and plausibility of my position. Second, I will say more
59
about how and why my position contrasts with a robust political construc-
tionism, allowing me to defend the superiority of my view. Further develop-
ment and defense of my position will then come in Chapter 6.
Taylor’s trilemma is aimed at what he refers to as the “strong communi-
tarian” view that races are groups “composed of people who share or have
some claim on a common culture,” with the normative implication that these
groups “deserve a certain kind of reverence or commitment from their con-
stituent members, such that the people who don’t participate in the culture
ought to do so.”36 This is, as he also calls it, a “cultural-nationalist account of
race.”37 The question that arises in relation to this view is this: How might the
claim that races are made up of people who share not only a similar appear-
ance and ancestry but also a common culture be justified? Taylor imagines
three options for how the cultural constructionist (as I will refer to his target)
could reply, and deems each option unsatisfactory.
First, the view might be that it is simply how nature works that, just as
members of races inherit common physical features, they inherit common
cultural tendencies as well. But this is classical racialism (i.e., biological essen-
tialism), and so it must be ruled out as an option. Second, he imagines that
a non-essentialist version of the position might take as its starting point the
plausible idea that “people who are treated in similar ways might do well to
join forces to resist their common oppression.”38 From there, one might rea-
sonably conclude that people oppressed on the basis of their common racial
categorization will be aided in their struggles against oppression by cultivating
togetherness through a sense of cultural community. The problem with this
“practical cultural nationalism,” as Taylor calls it, is that it is a prescriptive view
about what members of races ought to do, not a descriptive view concerned
with what races are, and thus it is irrelevant to the debate over the nature and
reality of race.39 Finally, if the cultural constructionist is not being prescriptive
but rather is, in fact, making the descriptive claim that races are, like ethnic
groups, made up of people associated with a common culture, then the view is
“simply incorrect.”40 Races and ethnic groups are not the same things.
36. Taylor, Race, 100–101. Taylor takes Molefi Asante’s Afrocentrism to be an example of a
view of this sort.
37. Ibid., 101.
38. Ibid.
39. Ibid.
40. Ibid., 102.
60
60 • W h a t I s  R a c e ?
Taylor thus treats cultural constructionism as untenable because it is either

empirically baseless (because essentialist), irrelevant (because prescriptive,
not descriptive), or a confusion of two distinct categories (races and ethnic
groups). It is already obvious that I would reject the essentialist version of
the identification of races with cultures, so the question becomes whether my
position is captured by either of the two non-essentialist options and, if so,
what I might say in response to Taylor’s criticisms. Let us first notice what is
strange about Taylor’s conception of a practical cultural nationalist. In order
for his criticism of irrelevance to work, he must imagine the practical cul-
tural nationalist as refraining from describing races as being presently cultural
groups, even while prescribing that the members of at least some races work
to change this situation. But cultural nationalists, as a rule, argue for the pres-
ervation and cultivation of a culture that they believe already exists, at least
in some shape or form, and this is true for those whose cultural nationalism is
racial in scope as well.
Take Du Bois. In “The Conservation of Races,” he exhorts African
Americans to lead black people as a whole in making a distinctive cultural
contribution to civilization. This is a prescriptive, future-oriented view, but
Du Bois clearly suggests that the cultural contribution to come will build on
contributions that have already been made. As “members of a vast historic
race,” he claims, his people have brought unique gifts to America, for their
“subtle sense of song has given America its only American music, its only
fairy tales, its only touch of pathos and humor amid its mad money-getting
plutocracy.”41
The position Taylor depicts is certainly not logically impossible, but given
the goal of evaluating what has actually been held by proponents of cultural
theories of race, we should note that, in practice, prescriptions of the type
he is discussing are almost always inseparable from an understanding of the
race being exhorted as already a cultural group in some sense, whether this
understanding is biologically essentialist or not.42 Thus Taylor’s second op-
tion is not really worth distinguishing from the first and third options. Given

42. Whether Du Bois should be viewed as relying on biological essentialism in “The
Conservation of Races” is a famously controversial question among philosophers of race and
Taylor, as a matter of fact, is one of the major contributors to this debate who have written
in defense of Du Bois. See Appiah, “The Uncompleted Argument: Du Bois and the Illusion
of Race,” Critical Inquiry 12 (Autumn 1985): 21–37, and Taylor, “Appiah’s Uncompleted
Argument: W. E. B. Du Bois and the Reality of Race,” Social Theory and Practice 26 (Spring
2000): 103–128.
61
that the first option has been ruled out, the crucial question is thus whether
my view is, as Taylor’s trilemma suggests, a simple confusion of race with eth-
nicity. This would, of course, be an ironic result, in light of the fact that the
need to avoid treating ‘race’ and ‘ethnicity’ as synonymous in order to think
productively about race was among the first things I asserted in this chapter.
Before responding, I should clarify how Taylor takes these two terms to
differ in meaning. According to him, both terms refer to groups based par-
tially on descent, but “ ‘race’ points to the body while ‘ethnicity’ points to
culture.”43 In other words, while races are groups distinguished at least in part
by shared ancestry and distinctive physical appearances, ethnicities are groups
distinguished by shared ancestry and by cultural factors like language and re-
ligion. I happen to think that this is a very reasonable way of differentiating
between race and ethnicity, so it will not be my strategy to say that Taylor
has failed in drawing the distinction. It is compatible with and, I would say,
important to my view that physical appearance plays a key role in racial dif-
ference that it does not by necessity play in ethnic difference. Taylor seems to
suggest that recognizing races as cultural groups involves contradicting that
point, but this is not the case. Culture is not displacing the centrality of the
body, in my view, but rather serving as a key factor in explaining how the body
is socially meaningful in cases of racial difference.
Consider examples of people feeling cultural pride in both their race and
their ethnicity, where these are not the same thing. One can feel a sense of cul-
tural allegiance to the black race as a whole, for instance, while also proudly
identifying as a member of an ethnic group that is but a small component
of the race or one that overlaps multiple races, as in the cases of those who
identify ethnically as Latin American or Arab. An Afro-Cuban individual
may love being a Latino and yet simultaneously take great pride in being of
African descent, with the result that she feels a strong sense of kinship and
shared cultural ownership when witnessing or participating in forms of cul-
ture originating in sub-Saharan Africa or in places in the African diaspora
outside Latin America. This example fits well with Taylor’s claim that ethnic
belonging need not be associated with a shared physical appearance, given the
diversity of ways Latin American people can look, while racial membership is
linked with visible continuities, as in the case of those whose features indicate
sub-Saharan African ancestry.
43. Taylor, Race, 53.

62
62 • W h a t I s  R a c e ?
We can, however, go a step further in demonstrating the lack of confusion

between race and ethnicity on my part by recalling my agreement with Taylor
and other political constructionists that the origin of races as we know them
is to be found in the construction of white supremacist social hierarchies.
What this means is that, on my view, when the Afro-Cuban I described feels
culturally connected to other black people, the pride she experiences involves
valuing black bodies and their activities in the face of their historical deval-
uation, a devaluation that was part of the colonial projects that made being
of this particular descent a salient shared identity in the modern world in the
first place. Adding this historical specificity to the account helps to make it
obvious beyond a shadow of a doubt that race is not being confused here with
ethnicity or any other social category. This pride is racial pride, and it can be
felt without having a biologically essentialist understanding of racial identity,
for our Afro-Cuban friend may well recognize the historical contingencies
upon which her feelings of belonging within a larger whole rest, whether with
respect to her race or her ethnicity.
Cultural constructionism thus need not be seen as necessarily empirically
baseless, irrelevant, or confused. There is a coherent and plausible idea here.
In order to make it clear why it should be seen not only as plausible but also
as preferable to political constructionism, I will now systematically contrast
my position with the three tenets of the robust version of that view. As men-
tioned multiple times now, the first tenet of the view is one I accept: races
emerged out of political conditions that divided people into groups une-
qual in power. What I would add, though, is that as soon as you have races
emerging in this way, you have social categories shaping the identities of those
who are included in them in such a way that these members may plausibly
view these categories as culturally significant.
The cultural significance of races can be seen as coming about in at least
three ways (and I mean all of these ways, not just any or some). First, the
emergence of racial categories is itself a cultural shift, and thus a social context
in which people are viewed as being of different races is in that way cultur-
ally distinctive. This is a point that may seem subtle but which is ultimately
somewhat obvious. I am not yet describing how individual races might differ
in their respective ways of life. I am describing instead how being socialized
into a world where people conceive of each other as racially different means
being socialized into a particular way of life. If we think, for example, of the
difference between, on the one hand, someone growing up in a West African
village in the fifteenth century in an area where it was common to be aware
only of what we would now think of as ethnic differences and, on the other
63
hand, an enslaved descendant of this person on a Caribbean island a couple

of centuries later who is acutely aware of her place in a racial hierarchy, we are
noticing a difference in the cultures that shaped these two related individuals.
The social construction of race is therefore, from the beginning, a cultural
process, and race can accurately be described as being from the start both po-
litically and culturally constructed, even if we acknowledge that, at the point
of origin, it is the political circumstance (e.g., the social hierarchy of slavery)
that gives rise to the cultural condition of racial identification being common.
Second, there are the novel forms of cultural difference between groups
that arise in the wake of the development of racial difference. Once people
are being socialized into worlds in which they inhabit different racial catego-
ries, it is a necessarily common occurrence that these social distinctions lead
to inhabiting relatively different worlds and thus participating in different
ways of life. As people of different African ethnicities came in the Americas
to inhabit the category of ‘Negro’ or ‘black,’ for example, their new cultural
creations were products of black culture. They could not be products of the dis-
tinctive ethnic groups of old as these distinctions faded. One might associate
them with particular territories, calling jazz a product of the United States,
samba a product of Brazil, calypso a product of Trinidad, and so on, but to
say this alone is misleading, for no explanation of these musical cultures that
fails to acknowledge their initial development primarily by those of African
descent is adequate.
Third, we should not let the novelty that comes along with social orders
and relations brought about by colonialism cause us to miss the ways in which
racial groups are also shaped culturally by historical patterns and events pre-
ceding racial formation. It is true that there is some level of anachronism in
an African American boy alive today learning about Great Zimbabwe, the
large city whose ruins we can still visit, that served as the capital of a flour-
ishing kingdom from the eleventh to the fifteenth century, and experiencing
this as a moment of pride in ‘black heritage.’ On a social constructionist
view, we have reason to deny that the Shona people who lived in this place at
that time were part of something we can call the black race (in contrast with
Shona people today, as the development starting in the late nineteenth cen-
tury of Rhodesia as a settler colony and the struggle against white rule leading
up to Zimbabwe’s independence in 1980 make it especially obvious on this
view that they are black). But there is also something clearly right about him
describing his experience in that way. He is engaging in the common cultural
practice of taking pride in the past accomplishments of one’s people, but what
enables him to count those who constructed Great Zimbabwe as his people
64
64 • W h a t I s  R a c e ?
is an ancestral connection to Africa made significant by the social reality that

being black is significant. Note how whiteness as a cultural reality has been
similarly shaped by pride among Europeans and those of European descent in
the accomplishments of ancient Greeks and Romans.
It might be thought that, while it makes sense to call this boy’s feeling a
case of taking pride in black heritage, the accuracy of the description does
not mean that the practice itself makes sense. This boy’s ancestors in Africa
were likely to be found in West Africa, not the southern part of the conti-
nent where Great Zimbabwe is. With no known connection to Zimbabwe,
how could Shona people be his people? There is much that could be said in
response, including the political constructionist point that a world shaped by
white supremacy is a world in which it makes sense for this boy to see him-
self as in the same group as others with a similar appearance and sub-Saharan
African ancestry. An important point about culture that should be made in
response, though, is that to scoff at the idea that West African ancestry could
allow a sense of cultural connection with people from what is now Zimbabwe
is to wrongfully assume that geographic distance automatically means a lack
of cultural commonality. We can relate this to Kwame Gyekye’s complaint
about some of Appiah’s work. Gyekye claims that Appiah expends great en-
ergy emphasizing Africa’s cultural diversity at the expense of recognizing
common “threads visible in the cultural tapestry of the African peoples.”44
Without denying Africa’s “pluralism,” Gyekye points out a number of “hor-
izontal relationships” (i.e., similarities) between traditional African cultures
in metaphysical, epistemological, moral, and sociopolitical matters.45 We can
discern premodern horizontal relationships in other racial groups as well, and
this is best understood, of course, not as the result of inborn impulses, but
rather sociohistorical processes. Think, for instance, of how we can tell the
story of Buddhism arriving from India in East Asia and gradually spreading
throughout the region, centuries before East Asians were classed together by
Westerners as “Orientals” or the “yellow” race.
These three forms of cultural significance—racial consciousness itself
as cultural, racial consciousness as facilitating new cultural developments,
and racial consciousness as shaped by prior cultural developments—are key
44. Kwame Gyekye, An Essay on African Philosophical Thought: The Akan Conceptual Scheme,
revised ed. (Philadelphia: Temple University Press, 1995), xxx.
45. Ibid., 195–210. His examples include: an ontology including a Supreme Being and ancestral
spirits; divination, witchcraft, and spirit mediumship as sources of knowledge; and commu-
nalism in social thought and practice.
65
aspects of a proper account of the social construction of race, on my view.

They are not central to standard political constructionism. That being said,
it is not clear that the committed political constructionist has reason as yet
to deny anything I have said. As long as I admit the political origin of race,
my talk of cultural significance arising in the wake and as part of this process
of origination may be seen as helpful detail, rather than harmful challenge.
A maximally robust cultural constructionism might claim that premodern
cultural commonalities represent in themselves the origin of racial difference.
I would not claim that. Cultural commonalities across large expanses are not
hard to find, and if one tried to divide up the world on that basis, many dif-
ferent sets based on alternative choices of how to divide would be possible.
Premodern commonalities matter in the case of races only because they ac-
cord with divisions based on appearance, but given the various continuities in
how we look, there are different sets based on alternative divisions possible on
this basis as well. Races as we know them, on my view, are appearance-based
groups that initially result from the history of Europe’s imperial encounters.
Where cultural constructionism as I conceive of it moves from offering a
change in emphasis to seriously challenging the dominant view among social
constructionists, then, is in its opposition to the second and third tenets of a
robust political constructionism. I reject the idea that cultural difference is
less important than differences in power relations for understanding racial
phenomena in the present. This forces me to address what we should make
of hypothetical examples like the adopted young man who experiences little
cultural attachment to blackness but whose experiences of social inequities
nevertheless render his blackness very significant. Does this not show that the
political constructionist is right and I am wrong about what is important?
The mistake here is the expectation that the cultural construction of race
would involve a uniformity of experience across individuals, whereas this is
the case neither for race’s cultural nor for its political aspects. For any partic-
ular way in which disadvantage can be experienced within a social hierarchy,
one cannot assume that because a person is non-white that he or she has per-
sonally experienced that form of racism or racism’s effects. This is why Taylor’s
political account of race is structured around probabilities—around the fact
that, in the United States, for example, “non-whites are much more likely to
be unemployed, to commit and become victims of violent crime, to receive
substandard medical care, and to live in inadequate housing.”46 To point out
46. Taylor, Race, 82–83. Emphasis mine.

6
66 • W h a t I s  R a c e ?
a non-white individual in the United States today who is employed, who has
not committed or been a victim of violent crime, who has received quality
medical care, and who lives in a very nice house does nothing to disprove his
theory.
Similarly, it is not my claim that all members of races have uniform cul-
tural experiences. Indeed, there is no such thing as uniformity of experience
within cultural groups, especially when the group in question is large and ge-
ographically dispersed. In any group that can be described as a cultural group,
it will be normal for some individuals to be more familiar with certain aspects
of the culture, less familiar with others. Once we are talking about groups that
are associated with cultures but whose members are also connected by other
ties—such as descent in the case of ethnicities, or citizenship in the case of
countries, geographic location in the case of regions, and so on—then there
can be not only differential familiarity with various aspects of the culture, but
also the common occurrence of some group members having little to no in-
vestment in the group’s culture. What makes it the case that there is a culture
of the group to speak of is not all group members being equally invested and
engaged in reproducing a specific set of customs, but rather there being many
group members whose identification with the group is connected with invest-
ment and engagement in practices that they take to be distinctively related to
the group’s existence, which is a state of affairs compatible with a significant
amount of diversity in what is taken to be distinctive and in how invested and
engaged group members are.
I have provided an argument, then, against taking cases of individual expe-
rience as evidence that politics is more fundamental than culture to race, but
this defensive move is, of course, not enough to show that politics and culture
are, at present, equally important to understanding racial phenomena. This is
not the kind of claim for which it is easy to provide definitive proof, but I will
use three examples of issues involving race to motivate the idea that paying
attention to matters of social hierarchy without also paying attention to how
people often identify on the basis of appearance and ancestry with distinctive
ways of life generally leads to confusion about what is going on.
Consider, first, the example of education. This is a topic that can be ra-
cially fraught, especially in majority-white countries with sizable non-white
minorities (although I have mentioned already that racial dynamics in educa-
tion have been recently controversial in South Africa). The political account
of race is certainly helpful in illuminating many problems with education,
such as inequality in basic access, inequality in funding, and unequal treat-
ment of students by teachers and other staff with regard to discipline and the
67
provision of opportunities. One will badly misunderstand mobilization over

racism in education, however, if one ignores matters of curriculum design,
and such matters can only be understood through attention to cultural differ-
ence. For many concerned with racism in education, it would be manifestly
unsatisfactory for us to achieve equality in access, funding, discipline, and
opportunities while doing nothing about the traditional privileging of white
people, their accomplishments, and their perspectives in school materials
and teaching methods. This traditional Eurocentric bias is viewed by many
anti-racists not only as problematic in itself, but also as a significant factor in
non-white students—especially black students—doing less well than others,
dropping out at higher rates, and other such manifestations of inequality.
Concern about these effects strongly motivates demand for the availability of
black-focused schools, where emphasis on the value of black cultural heritage
is an essential component of the pedagogical approach.47
While there can be no mistaking the centrality of cultural difference in
demands for and efforts to provide black-focused education, it might be
objected that this mode of dealing with the problem of racism in schools is
controversial and that, if put aside, it remains possible to understand the issue
at hand without any focus on cultural differences between students. After all,
it seems clear that students of every racial group benefit from a curriculum
that does not minimize the contributions or ignore the perspectives of non-
white people. This is indeed true but, if used by a political constructionist as
a reason for not paying attention to cultural difference, this is self-defeating.
The only way this point could be seen as eliminating the role of cultural differ-
ence would be if the suggestion were that an inclusive curriculum benefits all
students in the same way, but that would mean denying that the Eurocentric
approach being targeted for change hurts non-white students differently than
white students, thus denying that there is a problem of racial hierarchy here
in the first place! To admit that students are placed in unequal positions by
a Eurocentric curriculum is to admit that some students—white ones—are
culturally affirmed by such a curriculum while others—non-white students—
are not. An inclusive curriculum in a racially integrated school benefits eve-
rybody but benefits different students in different ways. Most importantly,
47. For an example of such a school, see the Africentric Alternative School in Toronto,
Ontario: http://schoolweb.tdsb.on.ca/africentricschool/Home.aspx. Political philosopher
Will Kymlicka sympathetically considers the case for black-focused schools in the United
States and in Canada in his essay, “A Crossroads in Race Relations,” in Kymlicka, Politics in
the Vernacular: Nationalism, Multiculturalism, and Citizenship (New York: Oxford University
Press, 2001), 177–199.
68
68 • W h a t I s  R a c e ?
for our purposes, if it is effective with regard to non-white students, such a

curriculum encourages them to positively value their racial group and what is
unique about it, instead of leading them to covet whiteness as the authorita-
tive source of goodness and progress.
Consider, second, the issue of interracial marriage. Opposition to it by
white people dedicated to keeping their race pure is, by now, an uncontro-
versial example of irredeemable racism. But what about opposition to it by
non-white people? Is this similarly worthy of automatic dismissal and dis-
dain? A number of black philosophers have demanded that we take seriously
the concerns motivating such opposition among black people, especially
as expressed by black women. Anita Allen points out that many African
Americans think of Loving v. Virginia, the Supreme Court case that struck
down bans on interracial marriage, as an important blow to racial injustice
while still remaining “morally troubled by marriage between blacks and
whites.”48 What could explain this mix of attitudes? Notice that the attitude
toward the law here is sensitive to how marking racial difference is often a
matter of affirming social hierarchy. It is therefore telling and useful from my
perspective that Allen foregrounds the ideal of commitment to one’s cultural
community as often underlying the moral concern.
To be clear, Allen does not conclude that those who hold this ideal are
ultimately right to oppose interracial marriage, but she defends the moral in-
nocence of interracial marriage not by dismissing the values invoked against
it, but by outlining a position that “reconciles interracial marriages between
blacks and whites with black community-centered concerns about respect
and care.”49 Taylor also takes concerns that non-white people have about in-
terracial marriage seriously, but he appears to go even further than Allen in his
conclusion by not only rejecting a duty to marry within the race, but finding
little justification for anyone having a right to “color-conscious endogamy.”50
48. Anita Allen, “Interracial Marriage: Folk Ethics in Contemporary Philosophy,” in Naomi
Zack (ed.), Women of Color and Philosophy: A Critical Reader (Malden, MA: Blackwell,
2000), 183.
49. Ibid., 193.
50. Taylor, Race, 167. Besides Allen and Taylor, the other classic philosophical engagement
with this topic is Mills’s “Do Black Men Have a Moral Duty to Marry Black Women?” Of
the three, Mills ends on the note most friendly to black opponents of interracial marriage,
answering the question in his title by saying that there are enough at least partially strong
arguments to yield a “presumptive duty,” while leaving it open how easy it may be to defeat the
presumption. See Mills, “Do Black Men Have a Moral Duty to Marry Black Women?” Journal
of Social Philosophy 25 ( June 1994): 150.
69
When rejecting arguments that rely on preserving culture, he repeats his re-
frain that “races aren’t cultural groups.”51 Later on, however, when clarifying
that he is not saying we should immediately eradicate all racial endogamy,
Taylor concedes that “racial populations may serve as incubators for ethnic
communities, whose members may choose to relate to each other more closely
than to other groups” and, in this way, racial endogamy may be “the conse-
quence, sign, and mechanism of some benign segregation.”52 This sounds to
me like an argument from the cultural dimension of race to the permissibility
of having a preference for marrying within one’s race, which is support from
a surprising source for my claim about the importance of culture to race.53
Consider, finally, stereotyping. It is undoubtedly one of the things one
must understand in order to know how race works in the world at present.
Simple interactions are affected, such as when acquaintances wrongfully as-
sume that you must know how to dance because you are black or you must be
good at math because you are East Asian. It is a major issue in art and media,
requiring critical analysis of when, how, and to what purpose stereotypes
function in individual cases and in patterns across various works, genres, and
forms of representation. It is also a source of danger to people’s bodies and
lives, given the way that racial profiling, especially by the police, promotes
the influence of stereotypes over when people are detained and how they are
treated, including whether they might be subjected to deadly force. A polit-
ical constructionist account of race can help us understand all of this, from the
small slights to the serious harms, by pointing out how stereotypes rob us of
our individuality and obscure our humanity by flattening us into caricatures
on the basis of appearance and ancestry—caricatures that correspond to par-
ticular placements in a social order.
The logic of the political constructionist account, however, can lead us to-
ward an untenable position. We are apt to respond to stereotypes by affirming
our individuality—“I am not just some black guy, I am Chike, and Chike is
not good at basketball”—or our humanity—“Stop portraying us in servile
roles only, we are fully human and we can be heroes”—or sometimes both.
These responses certainly have merit, but the more strongly we cling to our
51. Ibid., 161.

52. Ibid., 169.
53. Surprising but not extremely so. As I have noted elsewhere, this is not the only case of
Taylor making comments about the relationship between racial and cultural identity that seem
strikingly compatible with my perspective. See Jeffers, “The Cultural Theory of Race,” 420 n52,
and Taylor, Race, 114–115.
70
70 • W h a t I s  R a c e ?
uniqueness as individuals or to our shared humanity, the more we move

in the direction of wrongly suggesting that nothing of interest can be said
about groups. Consider the example of someone being stereotyped as par-
ticularly in touch with nature because she belongs to one of the indigenous
groups of North America. This is racial stereotyping and, like all stereotyping
of individuals, involves a false, essentialist assumption. One has learned the
wrong lesson, however, if one concludes that all perceptions of the first peo-
ples of this land as distinct in how they conceive of and interact with the nat-
ural environment must be equally mistaken. Indeed, it is necessary to learn
something about cultural difference in this regard in order to properly under-
stand debates about land ownership and use, special rights to hunt and fish,
and various other matters involving indigenous peoples.
Stereotypes, then, are problematic distortions, but not by virtue of
representing differences in thought and behavior between races. The
problems are in how they exaggerate differences, in how intrinsic to group
membership they represent differences, in how they reduce groups to specific
differences, thus obscuring inner complexity and diversity, and, at times, how
the differences they represent are completely made up. Opposing stereotypes
should not be equated, then, with opposing the representation of racial groups
as having different ways of life. Sometimes a stereotype will be related to real
differences between races that are political in nature, such as stereotypes about
criminal behavior that can be connected to disproportionate convictions for
crimes because of the disadvantaged socioeconomic status of the group and
bias within the justice system. Other times, a stereotype will relate to a cul-
tural difference with roots older than racial difference as we know it, as in
the indigenous case just mentioned. Often enough, there will be relations of
both kinds: stereotypes about black people being good at dancing may be re-
lated both to the significance of dance in traditional African cultures as well
as a white supremacist willingness to recognize physical but not intellectual
talents among black people. And, finally, some stereotypes may have nothing
but the most tenuous relation to anything real, whether political or cultural.
What relation or lack of relation there may be, though, cannot be predicted
in advance of attention to the racial group’s past and present.
Throughout the preceding discussions of education, interracial mar-
riage, and stereotyping, a crucially important theme emerges. In each case,
a political account of race can be credited with insight into the ways that
the issue involves unfair distinctions and divisions between people that
ought to be overcome. In each case, however, paying attention to culture
is necessary for appreciating how the issue also involves legitimate forms
71
of difference, that is, differences that need not be overcome but rather
affirmed and appreciated. My claim that culture and politics are equally
fundamental to race in the present can be understood as the claim that
there can be no justification in assuming, before examining an issue
involving racial difference, that the ways in which people are different
from each other in this case will be the unfair kind or the legitimate kind,
although the best bet will be that both kinds are involved. The norma-
tive implication of this view is that dedication to fighting racism requires
sensitivity to racism’s ability to operate in two seemingly contradictory
ways: it creates and sustains difference where there ought to be none, and
it disparages and suppresses difference where it ought to be respected and
valued. Achieving victory over racism involves arriving at the point where
the first kind of difference is no more, while the second kind of difference
is uninhibited in flourishing.
This brings up my final point: my opposition to the political construc-
tionist tenet that racial difference would no longer exist if equality were to
be achieved. If races are at present partly cultural constructions, then the end
of racial hierarchy has the potential to usher in a condition of racial equality,
where races as cultural groups coexist in an egalitarian manner, rather than a
post-racial era in which there are no more races. I do not deny that the latter
outcome is possible—as a social constructionist, I accept that just as races
came into existence, they may cease to exist. I deny, however, that their ceasing
to exist is a necessary condition for or consequence of the end of racism.
Furthermore, as someone of sub-Saharan African descent, I personally desire
the indefinite persistence of black people as a cultural group.
Despite political constructionism’s dominance, I take my position here to
be the more intuitive one. What could make the political constructionist view
true? Even if you suspect that it is most likely that we would need to mix to-
gether until appearance gave no clue as to ancestry before racism would end,
that is not yet to say that it is simply impossible that we could eliminate relative
advantage and disadvantage while still identifying as black, white, Polynesian,
etc. The political constructionist can make the move, however, of refraining
from claiming that that would be impossible and claiming instead that, under
conditions of equality, identifications of these sorts would no longer count
as identifications with races. Haslanger, for example, has considered the pos-
sibility that groups that are otherwise like races but are not “hierarchically
organized” might be called “ethnicities.”54
54. Haslanger, Resisting Reality, 245.

72
72 • W h a t I s  R a c e ?
I reject this move, both Haslanger’s specific version (as ‘race’ ought not to
be confused with ‘ethnicity’) and the general idea that we should call races
something else after the achievement of equality. I take it to be both intui-
tive from an everyday perspective and expressive of a social constructionist
highlighting of historical development to hold that people’s continued at-
tachment to being black, for example, in a post-racist world would remain
an attachment to a race. It would remain, that is, identification with a group
distinguished by appearance and ancestry but made distinguishable in these
ways through social significance.
73
3 HOW TO BE A BIOLOGICAL RACIAL REALIST
Quayshawn Spencer
3.1. Introduction
“There’s an echogenic intracardiac focus (EIF) on the ultrasound
image.” That is what I heard during the second trimester ultrasound
exam for the mother of my first child. Those are not words that any
parent wants to hear. An EIF is a small bright spot on an ultrasound
image that represents a calcification in the heart of a fetus. The scary
thing about spotting an EIF is that EIF is correlated with having a
fetus with an abnormal number of chromosomes in all or some of
its cells, a state called ‘aneuploidy’ in medical jargon. Furthermore,
aneuploidy usually (but not always) causes a genetic disorder in the
child. For instance, a fetus with an extra chromosome 21 in all of its
cells (a state called ‘trisomy 21’) usually develops Down Syndrome.1
Other genetic disorders that arise from aneuploidy are Patau
Syndrome (caused by trisomy 13), Edwards Syndrome (caused by
trisomy 18), and Turner Syndrome (caused by monosomy X).2
The next step after spotting an EIF is to assess whether the
chance of having an aneuploidal fetus is high enough to warrant
doing an amniocentesis, which is a procedure where amniotic fluid
is extracted from the mother and the fetal cells are tested for an-
euploidy. But an amniocentesis is not risk-free. Doing an amnio-
centesis during the second trimester will result in a miscarriage in
2.5% of instances for women 20–34 years old (Papantoniou et al.
1. However, there are benign aneuploidies. For instance, a fetus with trisomy 21
in some of its cells instead of all of them will develop “mosaic Down Syndrome,”
which is benign if few enough cells are affected.
2. A person has monosomy X just in case she has one X chromosome and no other
sex chromosome in all of her non-reproductive cells.
74
74 • W h a t I s  R a c e ?
2001, 1055). So, an amniocentesis is inadvisable if the chance of having an

aneuploidal fetus is lower than the chance of having a miscarriage from an
amniocentesis. Here is where things got interesting. Our risk assessment was
very short. The obstetrician said, “I wouldn’t recommend an amniocentesis
because she’s Asian. EIF is a common occurrence for Asian mothers.”
While I was delighted to hear the recommendation, I was also skeptical.
How good was our obstetrician’s reasoning? In particular, was she justified
in using race as a relevant factor in her risk assessment? After all, what we are
trying to do is assess the risk of aneuploidy in a fetus, and aneuploidy is a purely
biological condition. What does that have to do with the mother’s race? So
I did some research. It turns out that T. D. Shipp et al. (2000) conducted
a landmark study on whether there are racial differences in EIF frequencies
among expectant mothers, and whether any such differences (if they exist) are
caused by racial differences in having an aneuploidal fetus among expectant
mothers.
Shipp et al. (2000, 461) divided mothers into Asian, Black, White, and
Unknown. Next, the authors found that the EIF rates for Asian, Black,
White, and Unknown mothers were 30.4%, 5.9%, 10.5%, and 11.1%, respec-
tively, but that only one fetus had aneuploidy and it was from a White mother
(Shipp et al. 2000, 461).3 Given the sample sizes for each race, it follows that
the average EIF rate for the sample was 12.1%, which is much lower than the
30.4% seen in Asian mothers.4 Furthermore, using the definition of a condi-
tional probability, a frequentist interpretation of the probability of an event,
and the results from this study, it follows that the probability of having an
aneuploidal fetus given that an EIF is observed on the mother’s second tri-
mester ultrasound image (call it ‘Pr{Aneuploidy | EIF}’) is 1 out of 59, or
≈ 1.7% , and that the probability of having an aneuploidal fetus given that
an EIF is observed on an Asian mother’s second trimester ultrasound image
(call it ‘ Pr{ Aneuploidy | EIF ∩ Asian} ’) is less than or equal to 1 out of 14
( ≤ 7.1% ).5
3. The fetus had monosomy X in some, but not all, of its non-reproductive cells and was diag-
nosed with mosaic Turner Syndrome.
4. Shipp et al. (2000, 461) sampled 46, 34, 400, and 9 mothers from the Asian, Black, White,
and Unknown races, respectively.
5. I say “less than or equal to” instead of “less than” here because Shipp et al. were unable to
follow up with one of the Asian mothers to determine whether her child had aneuploidy. See
Shipp et al. (2000, 461).
75
How to Be a Biological Racial Realist • 75
Furthermore, these probabilities are not unique to Shipp et al.

(2000). S. H. Tran et al. (2005) did a follow- up study on 7,480
mothers and found that Pr{ Aneuploidy|EIF} = 9 309 ( ≈ 2.9% ) and
Pr{ Aneuploidy|EIF ∩ Asian} = 3 83 ( ≈ 3.6% ). While Tran et al.’s
Pr{ Aneuploidy|EIF} value is slightly higher than Shipp et al.’s,6 their
Pr{ Aneuploidy | EIF ∩ Asian} values and the pattern that the race of the
mother matters are consistent with Shipp et al.’s study.
Given that research, two things became clear to me. First, the probabi-
listic reasoning of our obstetrician was flawed. While our obstetrician was
correct that EIF is a more common occurrence for Asian mothers com-
pared to mothers overall, the latter is because aneuploidal fetuses are more
common in Asian mothers! Moreover, it takes a large sample of expectant
mothers to see that. Second, our obstetrician was correct that race matters
in calculating the risk of having an aneuploidal fetus. So, I did a calculation
of my own using the Pr{ Aneuploidy|EIF ∩ Asian} value from Tran et al.
(2005) and determined that an amniocentesis was unwarranted, and not be-
cause of a miscarriage risk, but because of the test’s false-positive rate for
detecting aneuploidy!
Stories like the preceding raise the interesting philosophical question
of whether race is biologically real. While I—as a concerned parent—
interpreted the research as showing that race matters in medical genetics,
many medical scholars would discourage such an interpretation. For instance,
Michael Yudell et al. (2016, 564–565) have argued that “racial classifications
do not make sense in terms of genetics,” and, thus, to use race as an indicator
of human genetic diversity in any way is “problematic at best and harmful at
worst.” In truth, there are three routes that one can take to explain the higher
occurrence of aneuploidal fetuses in Asian mothers in the medical studies
I discussed.
One route is to look for a purely biological explanation, such as differences
in medically relevant allele frequencies between Asian mothers and mothers
of other races.7 Another route is to look for a purely social explanation. For
6. This might be explained by the fact that 57% of the mothers in Tran et al. sample were 35 or
older, which is itself a risk factor for having an aneuploidal fetus. See Tran et al. (2005, 159).
7. For instance, one could look at the alleles that affect spindle checkpoint. Spindle checkpoint
is a series of checks during gametogenesis that reduce the probability of chromosomal nondis-
junction (the most frequent cause of aneuploidy) (May and Hardwick 2006).
76
76 • W h a t I s  R a c e ?
instance, neither Shipp et al. (2000) nor Tran et al. (2005) report the average
age of Asian mothers in their samples. Since we know that a woman’s risk of
having an aneuploidal fetus increases with age, the reason why Asian mothers
in these studies displayed a higher risk for having an aneuploidal fetus might
have been because they were, on average, getting pregnant at a much later
age than mothers of all other races. Yet a third route is to look for a biosocial
explanation. For instance, Shannon Sullivan (2013) has highlighted how epi-
genetic processes—such as inheritable DNA methylation acquired from diet,
pollution, or stress—can explain some racial disparities in health.8 So, that
could be what is happening in this case.
Hence, we have an interesting and unsettled philosophical question
about whether (and, if so, how) race matters in calculating someone’s risk
for being born with a genetic disorder. Furthermore, answering that question
encourages a position on the biological reality of race.9 If you think that race is
not biologically real, then it probably would not make sense to you to include
race in a calculation of someone’s risk for developing a genetic disorder. For
instance, people who think that race does not exist or that race is wholly so-
cially real and not at all biologically real would be baffled by such a risk assess-
ment. However, if you think that race is biologically real, then whether race
is relevant in such calculations is a sensible question to ask. Of course, there
are other good reasons for asking whether race is biologically real, but its rel-
evance to medical genetics is sufficient to warrant philosophical attention.10
What does the question “Is race biologically real?” mean? Well, first,
I want to engage with my coauthors, and second, I want to engage with
people in the medical profession struggling with whether race should be used
in genetic disorder risk assessments and in other ways relevant to medical ge-
netics. Since both groups are interested, to some extent, in ‘race’ as it is used
8. An epigenetic process is any inheritable process in an organism that alters its gene activity
without altering its genetic sequence (Weinhold 2006, A163). There are three paradigm
examples of epigenetic processes: histone acetylation (which causes DNA to unwrap itself
from histones, making genes available for expression), DNA methylation (which involves
methylation at the cytosine bases in front of a gene, thus preventing that gene’s expression),
and mRNA silencing from microRNA (which is when non-protein-coding RNA halts gene
expression by deactivating protein-coding RNA).
9. I say “encourages” instead of “presupposes” because it is possible for something to not be bio-
logically real but to be a reliable indicator for something that is biologically real.
10. In fact, my personal interest in whether race is biologically real came from reading The Bell
Curve and wondering whether the authors were confused when they posited a “genetic compo-
nent” to the average IQ score differences among Blacks, Whites, and East Asians (Herrnstein
and Murray 1996, 299).
7
to classify people in current, ordinary American English, that is the way I will
understand ‘race’ in the question.11
For instance, in Joshua Glasgow’s A Theory of Race, he explicitly states
that he is interested in what ‘race’ means according to “competent English
speakers in the United States” (Glasgow 2009, 3). He also focuses on “con-
temporary mainstream discourse” in that linguistic group (Glasgow 2009,
8). Also, in Sally Haslanger’s Resisting Reality, she states that she is interested
in the “single or dominant public meaning (or folk concept) of ‘race’ ” as it
is used among “competent users of English” (Haslanger 2012, 304). While
Haslanger does not limit her focus to American English speakers, she is cer-
tainly interested in how people are “currently racialized in the United States”
(Haslanger 2012, 308).
Finally, in Chike Jeffers’s “The Cultural Theory of Race,” he assumes
a combination of Paul Taylor’s and Michael Hardimon’s definitions for
‘race’ ( Jeffers 2013, footnote 62). Furthermore, Taylor (2013, 20) is upfront
about his primary interest in “contemporary US conceptions of race” and
its “English” roots. Also, Hardimon (2017, 27) has recently clarified that his
focus is “ordinary uses of the English word ‘race’ and its cognates.”
As for engaging with people in the medical profession, there are certainly
many medical scientists and healthcare providers who do not care about how
‘race’ is used in American English. However, many of them do. For instance,
both Neil Risch et al. (2002, 5) and Esteban Burchard et al. (2003, 1171) have
argued that the racial scheme used on the “2000 US Census” is relevant to
studying and treating human genetic diseases.
But there is a second ambiguity lurking here, namely, what I mean by a
“biologically real” entity. All I will say right now is that I intend to use the
term ‘biologically real entity’ in a way that adequately captures all of the
entities that are used in empirically successful biology (e.g., the monophy-
letic group, the TYRP1 gene, the hypothalamus, etc.) and that adequately
rules out all of the entities that are not (e.g., the monobaramin, the feeble-
mindedness gene, the destructiveness organ, etc.).12 However, I will offer a
11. For the rest of this chapter and Chapter 7, I will drop the phrase “to classify people” when
talking about ‘race’ usage in current and ordinary American English. Instead, I’ll just presup-
pose that the usage of ‘race’ in this context is about classifying people. I’ll also stop modifying
the noun ‘American English’ with “current” and “ordinary” as well for the rest of this chapter
and Chapter 7, and, instead, I will just presuppose these modifiers when I talk about American
English.
12. The monobaramin is the fundamental unit of classification in baraminology, which
is a creation-science version of taxonomy. See Wood (2006, 151) for its definition. The
78
78 • W h a t I s  R a c e ?
particular conception of a biologically real entity when I defend my answer to

the question of interest. So, for clarity, the question I will answer is whether
race is biologically real, and, more specifically, whether race—in any way that
‘race’ is used in American English—is real in the same way as entities like the
monophyletic group, the TYRP1 gene, and the hypothalamus. My answer to
this question is a highly qualified ‘yes.’
3.2. OMB Race Talk as a US Race Talk

Suppose a race talk is a discourse that uses ‘race’ (or a synonym) to classify
people into subgroups. Suppose the subgroups picked out in a race talk are
races and the names of races are race terms. Also, for ease of discussion, I will
call any race talk that occurs in American English a US race talk.13 While
this jargon is new, I consider it to be a thinner version of Taylor’s (2013,
28) “race-talk.” According to Taylor (2013, 16–18), a race-talk is any discourse
that utilizes “race thinking,” and race thinking is “a way of assigning generic
meaning to human bodies and bloodlines,” by which he means the activity of
drawing “distant” inferences about a group of people from “bodily appear-
ance and ancestry.” While I like Taylor’s definition of ‘race-talk,’ we will soon
see why it is too thick to capture the diverse ways in which groups of people
are called ‘races’ in American English.
One US race talk that is widely used by current Americans is OMB race
talk. OMB race talk is any race talk that uses the meaning of ‘race’ that’s cur-
rently adopted by the Office of Management and Budget (OMB), which is
the largest office in the executive branch of the US government. Also, by ‘cur-
rently adopted’ I mean the race talk that the OMB endorses on the date that
I’m writing this chapter, not the race talk that the OMB happens to endorse
when this chapter is being read. In OMB race talk, the races are American
Indians, Asians, Blacks, Pacific Islanders, and Whites. Hispanics are not a
race in OMB race talk, but rather, are an ethnicity composed of people from
feeble-mindedness gene is a fictional gene that was often referred to by eugenicists. For ex-
ample, see Davenport (1917, 365). The destructiveness organ is a fictional organ in animal
brains that was believed to exist by phrenologists. See Combe (1853, 256–276) for a discussion
of this organ.
13. Note that I am using ‘US race talk’ differently here than how I used it in Spencer (2014).
In Spencer (2014, 1026), I used ‘US race talk’ to name the race talk that has the widest-used
meaning of ‘race’ in the US that is also used by a majority of US citizens.
79
Table 3.1 The OMB’s “Definitions” for Each of Its Races According
to Federal Register Document 97-28653
American Indian or Alaska Native—A person having origins in any of the

original peoples of North and South America (including Central America) and
who maintains tribal affiliation or community attachment.
Asian—A person having origins in any of the original peoples of the Far East,
Southeast Asia, or the Indian subcontinent including, for example, Cambodia,
China, India, Japan, Korea, Malaysia, Pakistan, the Philippine Islands, Thailand,
and Vietnam.
Black or African American—A person having origins in any of the black racial
groups of Africa.
Native Hawaiian or Other Pacific Islander—A person having origins in any of the
original peoples of Hawaii, Guam, Samoa, or other Pacific Islands.
White—A person having origins in any of the original peoples of Europe, the
Middle East, or North Africa.
multiple races.14 Furthermore, according to the OMB, people can belong to

more than one race at a time. In Table 3.1, I have listed what the OMB calls
its “definitions” for each of its race terms according to the federal register
document where the OMB introduces its racial scheme; a document called
“97-28653.”
OMB race talk usually occurs in formal communication among Americans
and usually involves one or more persons self-reporting their race(s) to an-
other party. For example, it is not uncommon for Americans to engage in
OMB race talk when applying to college, applying for a job, applying for a
mortgage loan, applying for a birth certificate, filling out a health provider
survey, filling out a child-care registration request form, or so forth. See the
following figures for some evidence.
Figure 3.1 is a screenshot of the race and ethnicity questions on the 2016
college application for Penn State. Figure 3.2 is a screenshot of the race and
ethnicity question on the 2016 registration request form for a child-care
center in Pennsylvania. Figure 3.3 is a screenshot of the race and ethnicity
questions on a 2016 Starbucks’ job application for a barista position.
14. While ‘Latino’ is a synonym for ‘Hispanic’ in OMB race talk, I will primarily use ‘Hispanic’
to talk about Hispanics in this chapter.
80
80 • W h a t I s  R a c e ?
Figure 3.1. Question 8 on Pennsylvania State University’s 2016 undergraduate

application.
Figure 3.2. The race and ethnicity questions on Today’s Child Learning Centers’ 2016
child-care registration request form.
The OMB began regulating race talk among federal agencies in 1977 with
the introduction of Directive No. 15, which is a statistical policy directive that
requires any federal agency in the United States that uses race talk in official
business to classify people into races in a way that is translatable into OMB’s
racial scheme. From the Department of Education to the Centers for Disease
Control and Prevention, all federal agencies in the United States must follow
Directive No. 15.15
15. Incidentally, Directive No. 15 is one reason why the OMB’s racial scheme is used outside
of the US government. For instance, because the US Department of Education (USDE) has
to comply with Directive No. 15, it requires all educational institutions that receive USDE
funding to use OMB’s racial scheme when reporting racial and ethnic data to the USDE. This
is why many American colleges and universities use OMB’s racial scheme on their college
applications. See document E7-20613 in the federal register.
81
Figure 3.3. The race and ethnicity questions on Starbucks’ 2016 job application for a
barista position.
In 1997, the OMB revised its race talk to include only the five races that it
uses today. In that revision, the OMB clarified that the purpose of Directive
No. 15 is, first, “to provide consistent data on race and ethnicity throughout
the Federal Government,” and second, “to enforce civil rights laws” (OMB
1997, 58782). Also, the OMB (1997, 58782) said that it revised its race talk in
1997 in order to deal with concerns about its 1977 race talk as being outdated
due to a significant rise in “immigration” and “interracial marriages” in the
United States since 1977. To deal with these concerns, the OMB included the
people indigenous to Central and South America in its American Indian race
(e.g., Maya, Pima, Quechua, etc.), recognized Asians and Pacific Islanders
as two distinct races, dropped its Asian or Pacific Islander race, and allowed
people to be a member of more than one race.
Despite the empirical support that I have provided for the claim that
OMB race talk is a US race talk, this claim is not uncontroversial. For
instance, someone might object to OMB race talk as being an ordinary
race talk. Since a US race talk must be an ordinary race talk (given how
I have defined it), the objection would imply that OMB race talk is not a
US race talk. Perhaps the motivation for such an objection is that in order
to be an ordinary race talk, it is not sufficient to be a race talk that occurs
82
82 • W h a t I s  R a c e ?
in ordinary discourse (which is how I have defined it in the preceding).

Rather, the race talk must be “how ordinary people conceive of race,”
which may differ from how a group of “experts” conceive of race (Glasgow
2009, 48).
This is a good concern. It is important to distinguish between ordinary
race talk and “specialist” race talk among experts because, first, the two can
harbor different meanings of ‘race,’ and second, I am interested in exploring
ordinary race talk in this chapter (Glasgow 2009, 48). Furthermore, OMB
race talk is a specialist race talk. It is the default race talk that agencies in the
US government use. With that said, I am not convinced that we should limit
what an ordinary race talk is to only those race talks that embody “how ordi-
nary people conceive of race,” and that is because it assumes that ordinary race
talk does not partake in a linguistic division of labor.
Notice that limiting what an ordinary race talk is to what ordinary people
conceive about race implicitly assumes that ordinary people are the cor-
rect people to consult to find out the meaning of the terms they are using.
Sometimes the latter is not a bad assumption. For instance, most English-
speaking Americans should be able to define ‘foot’ (the unit of measurement),
at least in terms of inches. However, the latter assumption is false for a large
portion of terms used by ordinary people. In cases where
(3.1) ordinary speakers intend a term t to refer,

(3.2) ordinary speakers intend t to refer to the same object that a group of
experts on t intends t to refer to, but
(3.3) ordinary speakers do not know or do not agree on what t means
it turns out that the meaning of t is whatever that group of experts means
by t . The fact that some terms used by ordinary speakers have a meaning
determined by a group of experts was first recognized by Hilary Putnam
(1973, 704), and he called this sociolinguistic phenomenon a “division
of linguistic labor.” For example, consider the term ‘DNA’ in American
English.
The term satisfies (3.1) because Americans do not use ‘DNA’ as if it
is a term with no referent, like, say, ‘unicorn.’ Americans also talk about
DNA with an intention to talk about the same stuff as biochemists,
geneticists, and other scientists who are experts on DNA, thus satisfying
(3.2). However, most Americans do not know what DNA is. For instance,
if you think that DNA is just the genetic material of living things, you’re
wrong. For one, the genetic material of all living things on earth used to
83
be RNA.16 But also, it is possible for something to be a strand of DNA

without ever having played the role of being genetic material (e.g., DNA
synthesized in a lab). Rather, biochemists, geneticists, and other DNA
experts define ‘DNA’ as ‘a polymer of deoxyribonucleotides’ (Stryer 1995,
75–76).
While it is true that some terms used in ordinary discourse seem to be
involved in a division of linguistic labor but are not actually involved in a
division of linguistic labor,17 there is no reliable way to know which ones are
and which ones are not without empirical investigation. So, as long as it is
possible that terms used in ordinary discourse (including ‘race’) are involved
in a linguistic division of labor, we should not require ordinary race talk to
be “how ordinary people conceive of race” (Glasgow 2009, 49). Rather, we
should define ordinary race talk as race talk used in ordinary discourse, and
pay attention to “how ordinary people conceive of race” in an ordinary race
talk only after ruling out the possibility that ‘race’ is involved in a division of
linguistic labor. However, one interesting fact about OMB race talk is that it
is involved in a division of linguistic labor.
3.3. The Meanings of ‘Race’ and Race Terms in OMB

Race Talk
Remember that Putnam’s conditions for a term t having a meaning that is de-
termined by a group of experts on t , call it ‘ e ,’ are as follows: (3.1) ordinary
speakers intend t to refer, (3.2) ordinary speakers intend t to refer to the
same object that e intends t to refer to, and (3.3) ordinary speakers do not
know or do not agree on what t means. It turns out that ‘race’ and race terms
in OMB race talk satisfy (3.1)–(3.3), and here is why.
First, some solid evidence that American English speakers intend ‘race’
and race terms to refer in OMB race talk is that the overwhelming majority
of Americans self-report one or more race when queried for their race in that
race talk. For example, on the 2010 US Census questionnaire, there were
299.7 million respondents, and a whopping 93.8% self-reported one or more
OMB race, while just 6.2% reported “Some Other Race” (Humes et al. 2011,
16. This is known as the RNA world hypothesis, and it was independently invented by Francis
Crick, Leslie Orgel, and Carl Woese in the late 1960s. See Robertson and Joyce (2012) for a
discussion of the hypothesis.
17. For some examples, see Dupré (1981, 74–75).
84
84 • W h a t I s  R a c e ?
4). That statistic would be hard to explain if American English speakers did
not intend ‘race’ and race terms to refer in OMB race talk.
Next, there are strange patterns in how American English speakers self-
report their OMB race that would be hard to explain if (3.2) were not true
for ‘race’ and race terms in OMB race talk. First, on the 2000 US Census
questionnaire—which is the most recent one that collected data on Arab
ancestry—80–97% of Arab Americans self-reported ‘White’ (de la Cruz and
Brittingham 2003, 8). This result might seem strange, but it is not strange if
Arab Americans intend to use ‘White’ (in OMB race talk) in the same way
that the OMB uses it. After all, in OMB race talk, White is not a narrow
group limited to Europeans, European Americans, and the like. Rather,
White is a broad group that includes Arabs, Persians, Jews, and other ethnic
groups originating from the Middle East and North Africa.
Second, on the 2010 US Census questionnaire, the majority of Hispanic
Americans self-reported in a way that corresponded to their primary ancestry
in three continental groups.18 The most populous Hispanic American na-
tional origin groups are Mexicans (58.7%), Puerto Ricans (15.1%), Cubans
(3.3%), Salvadorians (3.0%), and Dominicans (2.7%).19 Furthermore, we
know from genetic studies that Cuban Americans, Puerto Rican Americans,
Dominican Americans, and Mexican Americans have, on average, 73%, 62%,
50%, and 47% “Caucasian” ancestry, respectively (Manichaikul et al. 2012,
4).20 Moreover, what is interesting here is that the average Caucasian ancestry
of a Hispanic American national origin group nicely correlates with the pro-
portion of that group that self-reports ‘White’ alone in OMB race talk.
18. Actually, the correct term to use here is ‘genomic ancestry.’ I will explain why later. Also,
the continental groups I’m referencing are “Caucasian, African, and Native American”
(Manichaikul et al. 2012, 1). Finally, I’m looking at how Hispanic Americans’ racial self-
reporting correlates with their primary ancestry in these three continental groups because just
looking at racial self-reporting for Hispanic Americans as a group is likely to be misleading
(due to confounding), and it’s plausible to think that Hispanic Americans’ racial self-reporting
is correlated to this particular kind of ancestry.
19. These are all of the Hispanic national origin groups that composed ≥ 2.5% of total
Hispanic Americans according to 2010 US Census data, including Puerto Rican residents. See
Ennis et al. (2011, 14) and USCB (2010).
20. The estimate for Salvadorian Americans is missing because they have not yet been singled
out in genetic studies of Hispanic Americans. Also, while I am just reporting estimates from
Manichaikul et al. (2012), their estimates fall within the 95% confidence interval of estimates
from other studies, such as the “European ancestry” estimates for Mexican and Puerto Rican
Americans in Risch et al. (2009, 3).
85
For instance, on the 2010 US Census questionnaire, the proportion of Cuban

Americans, Puerto Rican Americans, Mexican Americans, and Dominican
Americans who self-reported ‘White’ alone was 85.4%, 63.2%, 52.8%, and
29.6%, respectively (Ennis et al. 2011, 14). Conducting a linear regression anal-
ysis shows that the average Caucasian ancestry of a Hispanic American national
origin group positively and highly correlates ( r = +0.864 ) with the proportion
of that group that self-reported ‘White’ alone on the 2010 US Census question-
naire.21 This pattern would be hard to explain if (3.2) were not true for ‘race’ and
race terms in OMB race talk.
Now, one could worry that the statistic that I reported about the racial self-
reporting of Arab Americans in OMB race talk is outdated.22 After all, a lot has
changed for Arab Americans since September 11, 2001 (or ‘9/11’). Most impor-
tantly, Arab Americans have experienced many more hate crimes since then due
to being stereotyped as Muslim terrorists23—so much so that in February 2015,
the Federal Bureau of Investigation (FBI) added a new uniform crime reporting
bias code—code 31—to track hate crimes against Arab Americans (FBI 2015,
table 1).
While a lot has changed for Arab Americans since 9/11, whether and how
much those changes have affected their racial self-reporting in OMB race talk
is testable. For instance, if we look at the “Some Other Race” respondents
to the 2010 US Census questionnaire in the 50 states and the District of
Columbia, and compare that number to the USCB’s 2010 estimate for the
number of Arab Americans, we can estimate that the maximum percentage
of Arab Americans who wrote in some other race (e.g., Arab, Middle Eastern,
etc.) on the 2010 US Census questionnaire was 36.7%.24 For context, the
21. The linear regression equation I used to make this calculation is: Y = 1.6812 X − 39.761 .
22. For instance, I would expect Joshua Glasgow, Linda Alcoff, and Paul Taylor to have this
concern. See Glasgow (2003, 472; 2009, 96), Alcoff (2006, 258), and Taylor (2013, 146–147).
23. What is so absurd about this stereotype is that the overwhelming majority of Arab
Americans are not even Muslim! For instance, in Alia Malek’s myth-busting book A Country
Called Amreeka, she reports that just 24% of Arab Americans are Muslim (Malek 2009, ix–x).
Rather, the overwhelming majority of Arab Americans are Christian (Malek 2009, x).
24. I arrived at this estimate in the following way. I started by using the USCB’s 2010 American
Community Survey one-year estimate for the number of Arab Americans in 2010 (1,698,570).
Next, I assumed that the percentage of Arab Americans who self-reported as Hispanic on the
2010 US Census questionnaire was the same as the percentage who did so on the 2000 US
Census questionnaire, which was 3.2% (de la Cruz and Brittingham 2003, 8). Thus, there
should have been 54,354 Hispanic Arab Americans in 2010. Next, I assumed that the percentage
of Arab Americans who self-reported as Hispanic on the 2010 US Census questionnaire had
86
86 • W h a t I s  R a c e ?
“Some Other Race” write-in rate for Mexican Americans on the 2010 US
Census questionnaire was 39.5% (Ennis et al. 2011, 14).
However, remember, the latter is a maximum estimate. In fact, it assumes
that all of the non-Hispanic “Some Other Race” write-ins on the 2010 US
Census questionnaire came from Arab Americans, which is almost cer-
tainly false. So, as it turns out, the aftermath of 9/11 has not affected the
racial self-reporting of most Arab Americans ( ≥ 63.3% ) in OMB race talk.
Furthermore, this result should not be too surprising. According to the Pew
Research Center, 94% of Jewish Americans self-report as “non-Hispanic
white,” and this is despite the fact that the rate of anti-Semitic hate crimes is
very high in the United States (Lugo et al. 2013, 46; FBI 2014).
So far, I have provided empirical support for (3.1) and (3.2) holding for
‘race’ and race terms in OMB race talk. All that remains to be done to show
that OMB race talk is involved in a division of linguistic labor is to show that
American English speakers do not know or do not agree on what ‘race’ and
race terms mean in OMB race talk. But this will be easy.
While there are lots of empirical studies that are relevant for supporting
the claim that Americans do not share a common meaning for ‘race’ and
race terms when engaging in OMB race talk, my favorite study is the focus
group portion of the Alternative Questionnaire Experiment (AQE), which
was conducted by Elizabeth Compton et al. (2013) for the USCB. The AQE
focus group study is unusually informative for three reasons. First, it uses
focus groups instead of surveys, and lots of useful, qualitative information
can arise in focus groups that are hard to obtain from surveys. Second, it was
explicitly designed to study how Americans use ‘race’ and race terms in OMB
race talk (Compton et al. 2013, 68–69). Last, and most importantly, it is one
of the few studies on how Americans use ‘race’ and race terms that uses a na-
tionally representative sample of US adults.25 So, what did they find?
the same “Some Other Race” reporting rate as Hispanic Americans overall, which was 36.7%
(Humes et al. 2011, 6). Thus, there should have been 19,948 Arab Americans who reported
both ‘Hispanic’ and ‘some other race’ on the 2010 US Census questionnaire. Next, I assumed
that all of the non-Hispanic “Some Other Race” respondents on the 2010 US Census question-
naire (a total of 604,265 people) were Arab Americans (Humes et al. 2011, 6). Next, I added
19,948 and 604,265 to obtain a maximum value for the number of Arab Americans who self-
reported “Some Other Race” on the 2010 US Census questionnaire, which, of course, turns
out to be 36.7% of the number of Arab Americans in 2010.
25. For instance, while Hirschfeld (1996), Glasgow et al. (2009), Morning (2011), and Guo
et al. (2014) have conducted relevant empirical studies for this topic, their samples of US adults
are not nationally representative. However, see OMB (2000) for another nationally represen-
tative empirical study on how Americans use ‘race’ and race terms in OMB race talk.
87
One major finding was that there was “no consensus” on the definition of
‘race’ in OMB race talk (Compton et al. 2013, 70). Rather, “race was defined
as skin color, ancestry, culture, etc.” among focus group participants (Compton
et al. 2013, 70). Another major finding was that many participants expressed con-
fusion about why the White race included Arabs and why Hispanics were not
a race (Compton et al. 2013, 70). But what was most fascinating was that the
participants “recommended that these terms should be defined so respondents
could better understand how to report” (Compton et al. 2013, 71). The first two
major findings suggest that (3.3) is true for ‘race’ and race terms in OMB race
talk, and the last major finding removes all doubt about whether ‘race’ and race
terms are operating by a division of linguistic labor in OMB race talk. Here, the
respondents are basically saying, “We are trying to racially self-report in the way
the OMB wants us to, but we need more guidance!”
Now that we have solid evidence that ‘race’ and race terms are involved in
a linguistic division of labor when used in OMB race talk, we can move on to
figuring out what ‘race’ and race terms mean in OMB race talk by scrutinizing
what the OMB intends these terms to mean. But let me back up a bit and
talk about meaning. ‘Meaning’ is understood in different ways by academics.
However, since I am interested in linguistic meaning, linguistic meaning is
a prime area of research for philosophers of language, and since 76.6% of
“specialists” in philosophy of language adopt a truth-conditional approach to
the linguistic meaning of a name, I will adopt the truth-conditional approach
to meaning to figure out what ‘race’ and race terms mean in OMB race talk
(Bourget and Chalmers 2014, 483).26 The truth-conditional approach to the
meaning of a name is to see a name’s meaning as the “contribution” it makes
to the truth-conditions of propositions in which the name occurs (Perry
2001, 18).27
26. The operational definition used for a specialist in philosophy of language in this study was
that of a “regular” faculty member in a “leading” department of philosophy in the English-
speaking or analytic philosophy world who lists ‘philosophy of language’ as an area of special-
ization (Bourget and Chalmers 2014, 468). Also, “leading” was determined by having a score
of 1.9 or above in the Philosophical Gourmet Report (PGR) or by being judged to be “com-
parable” to such schools by the editor of PGR, which, at that time, was Brian Leiter (Bourget
and Chalmers 2014, 468).
27. To be clear, Perry (2001, 17) considers meanings to be the rules that assign content to types
of expressions or subsentential expressions (e.g., names). However, Perry (2001, 18) does say
that “ordinary” meaning is the same thing as content. So, what I’m calling linguistic meaning
is what Perry calls ordinary meaning or content, not meaning. However, what philosophers of
race are interested in when they talk about the meaning of ‘race’ is the content of ‘race,’ not the
rules for assigning content to names.
8
88 • W h a t I s  R a c e ?
For instance, suppose I want to know what ‘Fab Five’ means in the specialist
English discourse of NCAA basketball talk, and suppose I want to know its
truth-conditional meaning. Then, what I should do is figure out what I can
substitute for ‘Fab Five’ in all of the propositions that include ‘Fab Five’ in the
relevant context while maintaining the same truth-values. Historically, there
are two ways of going about doing this. One way is to use a set of “identifying
conditions” (conditions that competent users of a term use to pick out the ref-
erent of the term) (Perry 2001, 4). For example, we could define ‘Fab Five’ as
“the 1991 recruited class for the Michigan Wolverines men’s basketball team.”
But another way is to use the object that the term designates. For example, we
could define ‘Fab Five’ as the set consisting of Juwan Howard, Ray Jackson,
Jimmy King, Jalen Rose, and Chris Webber. The first approach is known as
descriptivism among philosophers of language, while the second approach is
known as referentialism.
There is an ongoing debate in the philosophy of race about whether de-
scriptivism or referentialism is the best way to model an ordinary meaning of
‘race.’28 However, I do not want to take sides in this debate. Rather, I will as-
sume that both approaches are respectable options, but that the best approach
to use for ‘race’ and race terms in OMB race talk is the one that works best
for these names. For instance, it is widely acknowledged among philosophers
of language that non-referring names (e.g. ‘feeble-mindedness gene,’ ‘Santa
Claus,’ etc.) are poorly modeled by referentialism (Perry 2001, 6–7).29 So, it
will be prudent to model the meanings for ‘race’ and race terms in OMB race
talk as their referents only if these names refer.
Also, it is widely acknowledged among philosophers of language that
descriptivism is a poor model for a name’s meaning if assuming that the
name’s identifying conditions are its meaning results in getting the wrong
truth-values for a large number of counterfactual or modal propositions in
which the name occurs (Perry 2001, 5).30 For instance, Saul Kripke (1980,
117) argued that ‘yellow metal’ is not the meaning of ‘gold’ because taking it to
28. For example, see Glasgow (2009, 20–26), Haslanger (2012, 429–4 45), and Glasgow
(forthcoming).
29. ‘Feeble-mindedness gene’ was a name used in eugenics for what is now known to be a
nonexistent gene.
30. A counterfactual proposition is a conditional where the antecedent intentionally states
something that is false, such as “If the Golden State Warriors had won the 2016 NBA finals,
then they would have had a better season.” A modal proposition is a proposition that says
something is or is not necessary or possible, such as, “LeBron James could have been the NBA’s
MVP in the 2015–16 season.”
89
be so leads to several counterfactual conditionals with the wrong truth-value.

One example that Kripke (1980, 118) gave was, “If the substance in South
Africa that we call ‘gold’ were not actually yellow due to an optical illusion
brought about from the South African atmosphere, then there would be no
gold in South Africa.” The correct truth-value for this counterfactual con-
ditional is false according to Kripke (1980, 118), but if the meaning of ‘gold’
is ‘yellow metal,’ then this counterfactual conditional is true. So, it will be
prudent to model the meaning of ‘race’ and race terms in OMB race talk as
their identifying conditions only if doing so captures the correct truth-values
for a large number of counterfactual and modal propositions in which these
names occur.
Furthermore, I will judge whether ‘race’ and race terms refer in OMB race
talk and how well the identifying conditions and referents (if there are any)
for ‘race’ and race terms serve as truth-conditional meanings by appealing to
what the OMB presently intends to pick out with ‘race’ and its race terms,
both in the actual world and in non-actual, accessible possible worlds.31
First, let’s look at what the OMB calls its “definitions” for its race terms.
These are the identifying conditions that many American English speakers
use to figure out how to self-report in OMB race talk. However, given what
the OMB intends to pick out with its race terms, these identifying conditions
are anything but meanings. Before the OMB introduced its revised racial
scheme in 1997, it adopted 13 “principles” to guide that revision (OMB
1997, 58782).32 According to principle 4, OMB race terms should pick out
“population groups” in humans that are “comprehensive in coverage” and
“nonduplicative” (OMB 1997, 58783). In other words, in the OMB’s ra-
cial scheme, there are not supposed to be any unnecessary races, and every
single member of the human species should belong to one or more races.33
31. Thus, I am adopting Kripke’s (1980, 163) view that the referent of a name is fixed by the
“present intentions” of the speaker (or speakers) that control its meaning. Also, I will be using
quantified modal first-order free logic with a T interpretation of necessity and necessary iden-
tity to assign truth-values to propositions in all possible worlds accessible to the actual world.
See Girle (2009, 14, 107) for its syntax rules and Girle (2009, 14–15, 39, 108, 133) for the
meanings of important types of expressions in the language (e.g., its propositions, its logical
constants, its necessary truths, etc.).
32. These principles were developed by a committee of more than 30 federal agencies put to-
gether by the OMB in 1993 whose job it was to explore various options for changing OMB’s
racial scheme (OMB 1997, 58782).
33. It is easy to see why the OMB wants this. Obtaining a racial classification like this would
solve the problem of how to classify any US immigrant and any child born from an interracial
mating in the US.
90
90 • W h a t I s  R a c e ?
However, given what we know about human evolutionary history, the “defi-
nition” that the OMB provides for ‘Black’ makes all of the other OMB races
unnecessary!
Remember that the OMB claims that the “definition” for ‘Black’ is “A
person having origins in any of the black racial groups of Africa” (OMB
1997, 58789).34 Also, the OMB has explicitly or tacitly recognized all of
the following ethnic groups as examples of Blacks: African Americans,
Afro-Brazilians, Cape Verdeans, Ethiopians, Haitians, Jamaicans, Louisiana
Creoles, and Nigerians (OMB 1995, 44682; OMB 1997, 58789; OMB 2000,
28).35 However, the problem here is that it is not just African Americans,
Ethiopians, Jamaicans, and the like that are Black according to this definition.
Rather, all humans are Black according to this definition given what we know
about human evolutionary history.
First, according to the widest accepted theory on the evolution of human
populations, all current human populations descend from a single popula-
tion (of about 1,000 people) that resided in East Africa about 100,000 years
ago (Cavalli-Sforza and Feldman 2003, 270). Second, according to the most
widely accepted theory on the evolution of human skin pigmentation, all
humans had dark skin until about 40,000–60,000 years ago, when we first
left Africa and found ourselves in environments with low ultraviolet B light
( Jablonski and Chaplin 2010, 8962). Together, these two facts imply that all
living humans—every single one of us—descend from black-skinned people
in Africa, and, thus, all of us are Black according to the OMB’s “definition”
for ‘Black.’ While that result makes the OMB’s racial scheme “comprehensive
in coverage,” it also makes all OMB races except Blacks unnecessary, which is
something that the OMB does not want.
Now, we could try to fix this problem by offering a more nuanced iden-
tifying condition for ‘Black.’ For instance, we could add ‘recent’ in front of
‘origins’ in the OMB’s “definition” in order to try to fix the problem. However,
adding such tweaks creates counterfactual problems. For example, suppose
34. The OMB is notoriously vague about what it means by ‘racial groups.’ However, it does not
mean ‘races,’ since the OMB only acknowledges five races in its racial scheme and its “defini-
tion” for ‘Black’ is an attempt to define one of those races. However, given how the OMB uses
‘racial groups,’ I will interpret it as interchangeable with ‘ethnic groups.’
35. For instance, the OMB rejected requests from Cape Verdean Americans and Louisiana
Creole Americans to be recognized as distinct races because they can self-report as mixed
Blacks (OMB 1995, 44682; OMB 1997, 58786).
91
that by ‘recent’ we mean ‘before twenty-one generations ago.’36 Furthermore,

suppose that, contrary to how events actually unfolded, the English settlers
who created the thirteen colonies that would eventually become the United
States brought just one installment of black-skinned people from Africa
(hereafter, black Africans) to the colonies for slave labor and forced that pop-
ulation to exclusively inbreed for twenty generations up until a generation t
. Suppose we call these highly inbred people American Africans to contrast
with African Americans, who are a mixed people. Also, suppose that, just as
in the actual world, all human populations in this non-actual, accessible pos-
sible world descended from a single black African population. Then, since no
American African at t has “recent origins” in any black African people, no
American African at t is Black according to our revised definition for ‘Black’!
But there is more. Since no American African at t has any recent origin in
any OMB race at all, this tweak prevents OMB’s racial scheme from being
“comprehensive in coverage” as well, which is a clear violation of principle 4.
Suppose we call the possible world in the preceding the American African
world. The American African world is not “wholly metaphysical” (Hardimon
2013, 27; 2017, 45).37 There are lots of human populations that are similar to
American Africans in the actual world. For example, there are many unmixed
Aboriginal Australians who do not possess “recent origins” from any of the
original people to any of the geographic regions that the OMB mentions in
its race term “definitions.” Instead, unmixed Aboriginal Australians exclu-
sively descend from the original people to Sahul, who arrived in Sahul 46–60
kya (where 1 kya is equal to 1,000 years) (McEvoy et al. 2010, 297).38 Some
examples of such populations are the Karryarra people of Western Australia,
the Kuranda people of Queensland, and the Gunganji people of Queensland
(Bergström et al. 2016, 810). Thus, there are many Aboriginal Australians
who can go back at least 1,840 generations without finding a single ancestor
36. This is not an arbitrary number. Assuming that an average human generation is 25 years
(which is standard in population genetics), twenty-one generations back from 2017 is 1492,
the year that Europeans first colonized the Americas.
37. This is a phrase that Michael Hardimon uses to respond to a thought experiment of
Glasgow’s that attempts to show that sharing a common ancestry is not necessary to being
a race in the ordinary English sense. See footnote 13 in Hardimon (2013) and Hardimon
(2017, 45).
38. Sahul was a continuous landmass including present-day Australia, New Guinea, and
Tasmania from at least 100 kya to about 10 kya.
92
92 • W h a t I s  R a c e ?
from one of the original human populations to any geographic region that the
OMB mentions in its race term “definitions.”39
The problems that I have raised for considering the OMB’s identifying
condition for ‘Black’ as a meaning can be generated in an analogous way for
each OMB race term. By that I mean, each of the “definitions” that the OMB
provides for its race terms are inadequate to pick out the intended referents
of those terms. Furthermore, while we could continue to try to tweak these
identifying conditions to avoid each concern, a simpler explanation for what
is going on here is that the OMB intends to pick out ancestry groups with
its race terms, and since everyday American English is ill-equipped to artic-
ulate the essences of ancestry groups, we are better off taking the meanings
of OMB race terms to be the objects they designate and leaving the task
of articulating the nature of each race in OMB race talk to the experts on
ancestry: geneticists.40
As for ‘race’ in OMB race talk, the OMB does not even attempt to give
a definition for that term.41 Furthermore, when names are used without
any identifying conditions, but rather, as just tags for objects, that itself is
some evidence that the name’s meaning is just its referent.42 For instance, in
the city of Philadelphia, the name ‘Penn’ is just a tag for the University of
Pennsylvania.43
Now, one could object here and try to offer a descriptive definition for
‘race’ in OMB race talk. For instance, perhaps the OMB is assuming what
Hardimon (2003, 437; 2017, 27) calls “the ordinary concept of race,” which
is supposed to be a very thin concept of race that captures “ordinary uses of
the English word ‘race’ and its cognates” (at least in the dominant use of ‘race’
in ordinary English). However, according to the ordinary concept of race,
39. Here, I’m making a conservative assumption of an average Aboriginal Australian genera-
tion of 25 years.
40. The term ‘ancestry group’ is not mine. It was coined by Marcus Feldman (2010, 151).
41. For evidence that the OMB does not attempt to provide a definition for ‘race’ in any of its
publications on its racial scheme, see OMB (1995), OMB (1997), Wallman (1998), and OMB
(2000).
42. I’m borrowing the locution “tags for objects” from Perry (2001, 4). However, the conven-
tion of talking about names with referential meanings as merely “tags” originates with Ruth
Barcan Marcus (1961, 310).
43. I learned this fact the hard way when I first moved to Philadelphia and misinterpreted the
name ‘Penn’ as a nickname for Pennsylvania State University in a casual conversation. I was
quickly corrected!
93
“visually indistinguishable” races are impossible, but that situation is not im-
possible in the OMB’s racial scheme (Hardimon 2003, 442).44
For instance, in the OMB’s racial scheme, Melanesians are a Pacific
Islander subgroup (OMB 1997, 58789). However, in biological anthro-
pology, it is well known that Melanesians, on average, share the same visible
racial traits as black Africans (e.g., dark skin, black hair, very curly hair, full
lips, etc.).45 Suppose that American Indians, Asians, Blacks, Whites, and
Melanesians exist in a non-actual possible world accessible to ours, but that
no non-Melanesian Pacific Islanders exist in that world. Suppose we call this
world the black Pacific Islander world.
It’s worth pointing out that the black Pacific Islander world is not a world
that clashes with biological facts. The world could easily be generated from
non-Melanesian Pacific Islanders engaging in enough interbreeding with un-
mixed White people to make all non-Melanesian Pacific Islander subgroups
go extinct. Now, an important observation about the black Pacific Islander
world is that it contains two OMB races that are visibly indistinguish-
able: Blacks and Pacific Islanders. Thus, Hardimon’s ordinary concept of race
is not as ordinary as he thought!
While we could tweak Hardimon’s ordinary concept of race to attempt
to achieve an adequate descriptive definition for ‘race’ in OMB race talk, that
strategy is no more likely to work than our previous attempt to tweak the
identifying condition for ‘Black.’ Rather, the simplest explanation for the way
the OMB uses ‘race’ is that the term’s meaning is just its referent. But now the
question arises, what is that referent? In my previous work on OMB race talk,
I discovered a surprising fact about how the OMB uses ‘race’ (Spencer 2014,
1028). The OMB never calls race a kind or a category, but rather, always calls
race a set of categories or population groups. For instance, the OMB calls race
a “set of categories” six times in 97-28653. This observation leads me to believe
that the meaning of ‘race’ in OMB race talk is just the set of five races used in
that race talk.
44. It’s worth noting that many philosophers of race besides Hardimon think that the way ‘race’
is used in American English requires that races are not visibly indistinguishable. Some of these
other proponents are Naomi Zack (2002, 37), Lawrence Blum (2002, 132), Glasgow (2009,
33), and Taylor (2013, 16). So, what I will say next applies equally well to these philosophers’
theories of race as well.
45. See Spencer (2015, 50) for a discussion of this interesting fact. Also, by a “racial” trait
I mean what Glasgow (2009, 86) means, which is one’s skin color, facial features, hair type,
and, sometimes, hair color.
94
94 • W h a t I s  R a c e ?
While the result that the meaning of ‘race’ in OMB race talk is a set might
be surprising at first, there are lots of names in American English that are used
as tags for sets. For example, consider ‘Fab Five.’ That name is used as a tag in
sports lingo for Juwan Howard, Ray Jackson, Jimmy King, Jalen Rose, and
Chris Webber. Likewise, ‘Twin Towers’ is a name used as a tag in sports lingo
for Tim Duncan and David Robinson. But more importantly, assuming that
the meaning of ‘race’ in OMB race talk is just the set of five races in that race
talk provides us with a large number of correct truth-values for related modal
propositions. For example, given the referential approach, the following
modal propositions possess the correct truth-value of true: “It is possible for
there to be two visibly indistinguishable races” and “Pacific Islanders could be
visibly indistinguishable from Blacks.”46
Even though the referential approach has been fruitful so far, its utility
will disappear if the things that I have been calling “referents” for ‘race’ and
race terms in OMB race talk do not actually exist. While it is possible to de-
fend the view that non-referring names have referential meanings, that de-
fense is going to be a tough sell to many philosophers of language.47 Thus, to
convincingly defend my use of the referential approach, I need to show that
the relevant terms refer, and, moreover, refer to what I have claimed they refer
to. So, I need to show that OMB race terms refer to real ancestry groups in
the human species, and I need to show that ‘race’ in OMB race talk refers to a
real division of humans into ancestry groups.
3.4. The Nature and Reality of Race and the Races

in OMB Race Talk
Before I begin, I should say more about how I will establish the reality of race
and the races in OMB race talk. I will show that all of these entities are real
in virtue of being biologically real entities. Unlike many philosophers of race,
I will not require a biologically real entity to “exist objectively” or “independ-
ently of human interest” (Andreasen 1998, 209; Sundstrom 2002, 93).48 Also,
I will not require a biologically real entity to be a “primary or fundamental
46. I am assuming that the correct truth-value for these propositions is determined by how the
OMB intends to use ‘race’ and its race terms, not my intuitions about what these truth-values
should be.
47. Nevertheless, for one attempt to do so, see Braun (1993).
48. For other proponents of this way of defining ‘biological racial realism,’ see Mills (1998,
45–4 6), Zack (2002, 4–5), and Maglo et al. (2016, 2).
95
category in human population genetics,” or otherwise be very important to

biology (Maglo 2011, 363; Hochman 2013, 347). Rather, what I will mean by
a biologically real entity is an epistemically useful and justified entity in a well-
ordered research program in biology, which I will call a genuine biological en-
tity. Furthermore, I am adopting this conception of a biologically real entity
not because I want to defend a version of biological racial realism. Rather,
I am adopting it because I think it adequately captures the collection of enti-
ties that are actually used in empirically successful biology (e.g., monophyletic
group, TYRP1 gene, hypothalamus, etc.).
For instance, if we restrict the realm of biologically real entities to only
those entities that exist independently of human interest, then we would
have to tell population geneticists that they are wrong that “ethnic groups”
in the human species (which exist only because of human interest) are real
biological populations, such as the Han people of China, the Yoruba people
of Nigeria, and the Maya people of Central America (Cavalli-Sforza 2005,
338–339). Also, if we restrict the realm of biologically real entities to only
those that are very important to biology, then we would have to tell molecular
geneticists that trivial alleles, such as the 93C allele from the TYRP1 gene, are
not real because they are not important enough to biology.49
While the theory of a genuine biological entity is complex, the part of
the theory that I will use is the part that designates an entity e as biologically
real if
(3.4) e is useful for generating a theory t in a biological research program p,

(3.5) using e to generate t is warranted according to the epistemic values of p
to explain or predict an observational law of p, and
(3.6) p has coherent and well-motivated aims, competitive predictive power,
and frequent cross-checks (Spencer 2012, 193).50
I will assume that population genetics satisfies (3.6). Population genetics

has been such an empirically successful research program in biology, it is
not worth our time to detail exactly how it satisfies (3.6). Thus, in order
to show that race and the races in OMB race talk are biologically real, all
49. The only function of the 93C allele is coding for blond hair in some Melanesian people.
See Kenny et al. (2012).
50. For all of the details of this theory, see Spencer (2012) or Spencer (2016).
96
96 • W h a t I s  R a c e ?
I need to do is show that they satisfy (3.4) and (3.5) for population ge-
netics. To do this, I will use recent results from human population structure
analysis.
A common research project in population genetics is to figure out all of
the ways that a species subdivides into biological populations. This is called
an analysis of “population structure” for that species, and each subdivision is
called a “population subdivision” of that species (Hartl and Clark 2007, 275).
There are many ways that population geneticists go about conducting popula-
tion structure analysis, but a common method today is to use patterns in allele
frequencies across a species’ organisms to detect that species’ demes (which
are its randomly mating groups of organisms), and then to use patterns in
allele frequencies across a species’ demes to detect all other levels of popula-
tion structure in that species. In essence, the method is to use different types
of “genetic structure” to infer all of the population subdivisions in a species
(Cavalli-Sforza 2005, 338).
In a landmark study by Noah Rosenberg et al. (2002), which was cross-
checked by Rosenberg et al. (2005), five levels of genetic structure were
detected among putative human demes. Furthermore, one of those levels is
relevant for us because it is where we find both a human population subdivi-
sion and the referents for ‘race’ and race terms in OMB race talk. In Figure 3.4
are the genetic structure results from Rosenberg et al. (2005).
In Figure 3.4, Rosenberg et al. are reporting five levels of genetic structure
among putative human demes. They discovered these levels from analyzing
993 loci in the human autosome that lack protein-coding alleles from 1,048
people in 52 ethnic groups that represent our entire geographic range, using
a fuzzy genetic clustering algorithm in a computer program known as struc-
ture.51 Each level is named according to the number of “genetic clusters” in
the subdivision (Rosenberg et al. 2005, 660). So, for example, K = 3 is the
level with three genetic clusters. Also, genetic clusters (represented as colors in
Figure 3.4) are nothing more than fuzzy groups of organisms (organisms are
represented as colored horizontal lines in the figure) such that an organism’s
degree of membership in any genetic cluster is equal to the proportion of its
51. Any human’s genome is divided into three parts. A human’s allosome is her set of sex
chromosomes. A human’s autosome is her set of non-sex chromosomes. Finally, every human
has a set of mitochrondrial DNA that composes part of her genome. Note that the autosome
is what Rosenberg et al. (2002, 2005) are studying. Keep this in mind when interpreting their
results.
97
K=6 K=5 K=4 K=3 K=2
Figure 3.4. The genetic structure of putative human demes according to Rosenberg
et al. (2005, 663).
98
98 • W h a t I s  R a c e ?
genome that originated from that cluster.52 Multicolored horizontal lines rep-
resent “mixed” organisms, and monochromatic horizontal lines represent un-
mixed organisms (Rosenberg et al. 2005, 660). Now, let us turn our attention
to K = 5 genetic clusters: Africans, Eurasians, East Asians, Oceanians, and
Native Americans.
While some of the K levels in the figure do not reflect a human population
subdivision (e.g., K = 2), population geneticists have provided compelling ev-
idence that K = 5 does. First, even though Rosenberg et al.’s K = 5 result does
not always appear in similar studies, it is robust.53 In particular, Rosenberg
et al.’s K = 5 result has appeared in ~ 70% of all human genetic clustering
studies that use a worldwide sample of human ethnic groups (Spencer 2015,
48).54 Furthermore, these studies have used different samples of people,
ethnic groups, and loci, and genetic clustering computer programs with dif-
ferent clustering algorithms (e.g., structure, frappe, admixture, etc.).
Second, even though Rosenberg et al.’s sample of ethnic groups is far from
perfect due to its abundance of isolated populations and unmixed people (see
Figure 3.4), there is ample evidence that Rosenberg et al.’s K = 5 result is not
merely an artifact of that sample.55 For one, Rosenberg et al. (2005, 663) have
shown that even after controlling for the geographic distance among sam-
pling locations, their K = 5 result still holds. But more importantly, Trevor
Pemberton et al. (2013) have also obtained Rosenberg et al.’s K = 5 result
using the largest and most diverse sample of human ethnic groups to date.
They used 5,795 people from 267 ethnic groups from all over the world,
52. For clarity, qki is a model parameter of the admixture mode of structure that represents
the proportion of an individual’s genome that originated from a cluster, where that individual
is i and that cluster is k (Pritchard et al. 2000, 948). However, qki can be interpreted in ad-
ditional ways depending on the data set. For instance, sometimes qki values are interpreted as
membership grades when the clusters are plausibly viewed as biological populations or ancestry
groups. For instance, in Rosenberg et al.’s (2002, 2382) study, qki values are called “member-
ship fractions.”
53. For critics who worry about the variation in which genetic clusters appear at K = 5 given dif-
ferent background assumptions used in the analysis, see Hochman (2013, 348) and Barbujani
et al. (2013, 157). However, I have addressed this concern elsewhere and in depth. In particular,
see Spencer (2014, 1034–1035) and Spencer (2015, 48).
54. By a worldwide sample, I mean a sample that includes every “major area” in the United
Nations’ 2011 classification of countries. These areas are Africa, Asia, Europe, Latin America
and the Caribbean, Northern America, and Oceania.
55. For critics who worry that Rosenberg et al.’s K = 5 result is an artifact of their sample of
ethnic groups, see Kittles and Weiss (2003), Serre and Pääbo (2004), Bolnick (2008), Maglo
(2011), and Hochman (2013, 2014).
9
including dozens of non-isolated populations and hundreds of mixed people,

such as African Americans, Coloured South Africans, Latin Americans, and
Polynesians (Pemberton et al. 2013, 891, 897).56
Finally, we have adequate reason to believe that the genetic structure at
K = 5 in humans is caused by underlying human population structure be-
cause each K = 5 human genetic cluster is anchored in a region circumscribed
by major geographic barriers to human interbreeding, such as “oceans, the
Himalayas, and the Sahara” (Rosenberg et al. 2005, 663).
Now, even if Africans, East Asians, Eurasians, Native Americans, and
Oceanians form a human population subdivision, the latter does not imply
that this subdivision is biologically real. As Koffi Maglo (2010, 362) has as-
tutely pointed out, the utility of an entity in biology does not entail its bio-
logical reality. Thus, we need to argue for the biological reality of the human
population subdivision at K = 5 directly.
In the medical genetics and population genetics literature, Africans, East
Asians, Eurasians, Native Americans, Oceanians, and other continent-level
human populations are known as “continental populations” (Cooper et al.
2003, 1167; Zhao et al. 2006, 399). However, for ease of reference, I will call
Africans, East Asians, Eurasians, Native Americans, and Oceanians, and only
these five populations, the human continental populations. The evidence that
the set of human continental populations is biologically real is the following.
First, the set of human continental populations satisfies (3.4) because it is
useful in population genetics for generating a theory about human population
structure—namely, the theory that the set of human continental populations
is the population subdivision at level K = 5 in humans.
Second, the set of human continental populations satisfies (3.5) be-
cause the theory in which the entity is posited is warranted according to the
population-genetic epistemic values of empirical accuracy, completeness, and
quantitative precision to predict a population-genetic observational law.57
That observational law is that humans have K = 5 genetic structure that is
largely geographically clustered in the following regions: the Americas, Sub-
Saharan Africa, Oceania, Eurasia east of the Himalayas, and Eurasia west of
the Himalayas and North Africa. Given our assumption that population ge-
netics satisfies (3.6), it follows that the set of human continental populations
56. Also, Rosenberg et al.’s K = 5 result was replicated again by Mallick et al. (2016, 9) using the
second largest and second most diverse sample of ethnic groups to date.
57. These epistemic values are discussed in Pierre Duhem’s The Aim and Structure of Physical
Theory. See Duhem (1906/1981, 19–30).
01
100 • W h a t I s  R a c e ?
is biologically real. An analogous line of reasoning can be used to show that

each human continental population is biologically real as well.
Now that we have established the biological reality of the set of human
continental populations and all of its members, I think you can predict where
I am headed. From looking at the set of human continental populations and
the set of races in OMB race talk, it is not absurd to think that the two sets
are identical, which is what I will call the identity thesis. In fact, several med-
ical geneticists realized early on that the set of human continental populations
and the five major races used on the “2000 US Census” were at least “aligned
nearly perfectly” (Risch et al. 2002, 5–6).58 Of course, there’s a big difference
between being nearly identical and identical. Nevertheless, one way to de-
fend the identity thesis is to show that adopting it provides us with solutions
to the puzzles that led us to reject the OMB’s “definitions” as definitions as
well as the best predictive power—which are usually marks of a true empirical
theory.59
For one, the identity thesis solves the puzzle of how to define a person’s
ancestry in a way that makes the OMB’s racial scheme “nonduplicative” as the
OMB intends. Remember that the OMB’s “definition” for ‘Black’ is insuffi-
cient to yield a nonduplicative racial classification of people because human
evolutionary history unfolded in such a way that every living human is a Black
person according to the OMB’s “definition” for ‘Black,’ thus making all non-
Black races in the OMB’s racial scheme unnecessary. Also, remember that
temporal qualifiers for ancestry (e.g., ‘recent’) don’t fix this problem due to
counterfactual scenarios like the American African world and actual outliers
like unmixed Aboriginal Australians. But also, comparative qualifiers for an-
cestry (e.g., ‘primary’) are dead ends as well for a different reason.60
What solves the preceding puzzle is that racial ancestry in the OMB’s ra-
cial scheme is all and only ancestry that contributes to an individual’s genome,
58. For other medical geneticists who made this observation, see Burchard et al. (2003,
1171). For some population geneticists who made it, see Sarah Tishkoff and Kenneth Kidd
(2004, S21).
59. Here and elsewhere in this book, I will be assuming a rather weak view of truth for empir-
ical theories that comes from Arthur Fine’s natural ontological attitude. The view is “referen-
tial” and simply states that “a sentence (or statement) is true just in case the entities referred to
stand in the referred-to relations” (Fine 1984, 98).
60. Specifically, given what we know about the tree of life, the primary ancestry (understood as
the majority of one’s ancestors) of any human lies outside of the human species, thus making
every human race-less. Furthermore, tweaking the qualifier to ‘primary human ancestry’ will
yield unintended results, such as no Polynesian being a Pacific Islander due to the majority of
any Polynesian’s human ancestors not being Pacific Island natives.
10
called “genomic ancestry” in the population-genetic literature (Weiss and

Long 2009, 707). For instance, according to the identity thesis, the meaning
of ‘Black’ is the African population. Thus, a Black person is a person with
genomic ancestry from the African population. That’s it. In other words, if
any allele in a person’s genome originated from the African population, that
person is Black. Furthermore, the degree to which a person is Black is equal
to the proportion of her alleles that originated from the African popula-
tion. Hence, according to the identity thesis, there are plenty of people who
aren’t Black.
For instance, a Taiwanese American who has 100% genomic ancestry from
the East Asian population is exclusively Asian, and a European American
who has 100% genomic ancestry from the Eurasian population is exclu-
sively White, which makes the Asian and White races useful in OMB’s ra-
cial scheme, just as they were intended to be. Also, an Aleutian Islander with
100% Native American genomic ancestry is exclusively American Indian,
and a Native Hawaiian with 100% Oceanian genomic ancestry is exclusively
Pacific Islander, which makes the American Indian and Pacific Islander races
useful in OMB’s racial scheme, just as they were intended to be.
Second, the identity thesis solves the puzzle of how to make the OMB’s
racial scheme “comprehensive in coverage.” For instance, anyone from the
American African population in the American African world is exclusively
Black, since all of her alleles originated from the African population. Also,
unmixed Aboriginal Australians—a group with no recent ancestors from
the original people to any geographic region mentioned in the OMB’s
“definitions”—are exclusively Pacific Islander (McEvoy et al. 2010, 300).61 So,
solving the puzzle of which race unmixed Aboriginal Australians belong to is
a concrete accomplishment of the identity thesis.
In addition, notice that there are several geographic regions that are not
mentioned in the OMB’s “definitions” despite there being indigenous people
to these regions. For instance, in addition to Australia, the OMB neglects
to mention the Andaman Islands, Central Asia, and Madagascar, to name a
few. So, it’s unclear how to racially classify the indigenous people to these
forgotten lands. However, the OMB needs to racially classify each of these
61. I say “unmixed” because many Aboriginal Australians have recent European ancestors
due to the colonization of Australia by European settlers. In fact, a recent genetic study by
Duncan Taylor et al. (2012, 534) showed that 59% of Aboriginal Australian males possess a
Y chromosome inherited from a European male. Thus, a substantial proportion of Aboriginal
Australians are, at least, White.
012
102 • W h a t I s  R a c e ?
indigenous people in order to have a “comprehensive” racial classification.

The identity thesis solves this puzzle effortlessly.
According to current genetic clustering results, the Malagasy are a mixed
people who, on average, belong to the African and East Asian populations
(Kusuma et al. 2016, 5). As for Andaman Islanders, they are an assortment of
different ethnic groups with different genomic ancestry mixture averages. For
example, the Jarawa are mostly unmixed and Oceanic (Aghakhanian et al. 2015,
1210). However, some studies of the Onge show that they are mostly mixed and
belong primarily to the East Asian and Oceanic populations (Mallick et al. 2016,
9). Finally, indigenous Central Asians are a mixed people that primarily belong
to the Eurasian and East Asian populations. However, their primary racial mem-
bership varies by ethnic group. For instance, the Tajiks are primarily Eurasian,
while the Uzbeks are primarily East Asian (Martínez-Cruz et al. 2011, 221).
Next, if the identity thesis is true, then we not only can solve lots of
puzzles about OMB race talk, but we can make many predictions with “very
high” accuracy (Burchard et al. 2003, 1172). For instance, if the vast majority
of US adults have a primary human continental population membership, are
competent in OMB race talk, self-report a single race, and racially self-report
the human continental population in which they have primary membership,
then using knowledge about primary human continental population mem-
bership alone, geneticists should be able to predict the self-reported OMB
race of most US adults with very high accuracy. Interestingly, this is exactly
what geneticists are able to do.
For example, using a nationally representative sample of US college
students (N = 2, 065), Guang Guo et al. (2014) tested the extent to which
they could predict the self-reported OMB race of subjects who reported a
single race using only each subject’s primary genomic ancestry in a human
continental population. After finding no self-reported Pacific Islanders and
just four self-reported American Indians in the sample, the authors decided
to focus on self-reported Asians, Blacks, and Whites (Guo et al. 2014, 153).
Next, looking at just the subjects who reported a single race (which was 1,773
subjects) and using only structure and a sample of each subject’s genome, the
authors were able to predict each subject’s race with 98.8% accuracy (Guo
et al. 2014, 153).62 While this is an amazing feat, Guo et al.’s result is not
unique. Hua Tang et al. (2005, 271) were able to predict the self-reported
OMB race of 2,657 US adults with 99.8% accuracy using primary human
62. This statistic includes the self-reported Hispanic subjects.

013
continental population membership alone.63 While there are other instances

of predictive power that I could talk about to lend further support to the
identity thesis, perhaps I should wrap up.
3.5. Conclusion
In this chapter, I have defended a nuanced biological racial realism as an ac-
count of how ‘race’ is used in one US race talk. I will call the theory OMB race
theory, and the theory makes the following three claims:
(3.7) The set of races in OMB race talk is one meaning of ‘race’ in US race talk.
(3.8) The set of races in OMB race talk is the set of human continental
populations.
(3.9) The set of human continental populations is biologically real.
I argued for (3.7) in sections 3.2 and 3.3. Here, I argued that OMB race
talk is not only an ordinary race talk in the current United States, but a race
talk where the meaning of ‘race’ in the race talk is just the set of races used in
the race talk. I argued for (3.8) (a.k.a. ‘the identity thesis’) in sections 3.3 and
3.4. Here, I argued that the thing being referred to in OMB race talk (a.k.a.
the meaning of ‘race’ in OMB race talk) is a set of biological populations in
humans (Africans, East Asians, Eurasians, Native Americans, and Oceanians),
which I’ve dubbed the human continental populations. Finally, I argued
for (3.9) in section 3.4. Here, I argued that the set of human continental
populations is biologically real because it currently occupies the K = 5 level of
human population structure according to contemporary population genetics.
Before I end, it will be interesting to see how much OMB race theory
sheds light onto the problem that motivated this chapter. While I will not
pretend that OMB race theory has the power to settle the debate about
whether (and, if so, how) race matters in medical genetics, the theory does
provide some helpful insight that may inch us closer to a resolution. For one,
OMB race theory implies that medical scientists who investigate whether
there are genetic explanations for racial disparities in heath are not making a
63. This statistic only includes the self-reported non-Hispanic Blacks and non-Hispanic
Whites, and leaves out the Chinese, Japanese, and Hispanics in Tang et al.’s sample. I’m leaving
out the Chinese and Japanese because they didn’t self-report ‘Asian’ in the study, but rather, as
‘Chinese’ or ‘Japanese.’ See Tang et al. (2005, 269). I’m leaving out the Hispanics because it’s
well known that Tang et al.’s sample of Hispanics was unrepresentative. See Glasgow (2009, 95).
014
104 • W h a t I s  R a c e ?
metaphysical mistake provided that the races they are using are OMB races.
The latter is because OMB race theory gives us the result that OMB races
are biological populations that are essentially genomic ancestry groups, and
it is metaphysically possible for such populations to non-accidentally differ
in medically relevant allele frequencies. So, for instance, Eric Jorgenson et al.’s
(2004, 276) study that searched for medically relevant differences in genetic
maps among “African Americans,” “East Asians,” and “whites” was not a met-
aphysically confused research project.64
However, my result conflicts with Yudell et al.’s (2016, 564–565) claim
that “racial classifications do not make sense in terms of genetics.”65 While I am
sympathetic to Yudell et al.’s claim, it turns out that some racial classifications
do make sense in terms of genetics, namely, the OMB’s racial classification.
However, Yudell et al. are absolutely right that some racial classifications do
not make sense in terms of genetics, such as any racial classification based on
what Anthony Appiah (1996, 54) has called “racialism.”66
A third result that’s relevant for whether (or how) race matters in med-
ical genetics is that OMB race theory does not imply that OMB races differ
in medically relevant allele frequencies, and it does not imply that OMB
races don’t differ in medically relevant allele frequencies. Likewise, OMB race
theory does not imply that OMB races differ in any socially important traits
(e.g., intelligence, beauty, moral character, etc.), and it does not imply that
OMB races don’t differ in any socially important traits. Determining whether
OMB races differ in any phenotypic ways requires a separate empirical in-
vestigation. Furthermore, I am not saying this out of political correctness.
It turns out that the DNA evidence that supports the existence of human
continental populations comes from non-protein-coding and non-functional
DNA in the human genome. Nevertheless, we now know that it’s metaphysi-
cally possible for some races to matter in medical genetics because some races
are biologically real.
64. Note that the authors are using ‘African American’ in this study as a term that is synony-
mous to the OMB’s ‘Black.’ Also, a genetic map is a map of the relative position of each gene in
a genome. The first genetic map was constructed by Alfred Sturtevant in 1913.
65. For other scholars who hold the same view, see Root (2003), Graves and Rose (2006),
Kaplan (2010), and Roberts (2011, 129).
66. According to Appiah (1996, 54), “racialism” is the view that humans naturally divide into a
small number of groups called ‘races’ in such a way that the members of each race share certain
fundamental, inheritable, physical, moral, intellectual, and cultural characteristics with one
another that they do not share with members of any other race.
015
References
Aghakhanian, F., Y. Yunus, R. Naidu, T. Jinam, A. Manica, B. Hoh, and M. Phipps.
2015. “Unravelling the Genetic History of Negritos and Indigenous Populations of
Southeast Asia.” Genome Biology and Evolution 7(5): 1206–1215.
Alcoff, L. M. 2006. Visible Identities: Race, Gender, and the Self. Oxford: Oxford
University Press.
Andreasen, R. O. 1998. “A New Perspective on the Race Debate.” The British Journal for
the Philosophy of Science 49(2): 199–225.
Appiah, K. A. 1996. “Race, Culture, Identity, Misunderstood Connections.” In K. A.
Gutmann (ed.), Color Conscious. Princeton, NJ: Princeton University Press, pp.
30–105.
Barbujani, G., S. Ghirotto, and F. Tassi. 2013. “Nine Things to Remember about Human
Genome Diversity.” Tissue Antigens 82: 155–164.
Barcan Marcus, R. 1961. “Modalities and Intensional Languages.” Synthese
13(4): 303–322.
Bergström, A., N. Nagle, Y. Chen, . . . C. Tyler-Smith. 2016. “Deep Roots for Aboriginal
Australian Y Chromosomes.” Current Biology 26: 809–813.
Blum, L. 2002. I’m Not A Racist But . . . : The Moral Quandary of Race. Ithaca,
Bolnick, D. 2008. “Individual Ancestry Inference and the Reification of Race as a
Biological Phenomenon.” In B. Koenig, S. Lee, and S. Richardson (eds.), Revisiting
Race in a Genomic Age. New Brunswick, NJ: Rutgers University Press, pp. 70–85.
Bourget, D., and D. Chalmers. 2014. “What Do Philosophers Believe?” Philosophical
Studies 170(3): 465–500.
Braun, D. 1993. “Empty Names.” Noûs 27(4): 449–4 69.
Burchard, E., E. Ziv, N. Coyle, S. Gomez, H. Tang, A. Karter, . . . N. Risch. 2003. “The
Importance of Race and Ethnic Background in Biomedical Research and Clinical
Practice.” The New England Journal of Medicine 348(12): 1170–1175.
Cavalli-Sforza, L. 2005. “The Human Genome Diversity Project: Past, Present, and
Future.” Nature Genetics 6: 333–340.
Cavalli-Sforza, L., and M. Feldman. 2003. “The Application of Molecular Genetic
Approaches to the Study of Human Evolution.” Nature Genetics 33: 266–275.
Combe, G. 1853. A System of Phrenology, 5th edition, revised, vol. 1.
Edinburgh: MacLachlan and Stewart.
Compton, E., M. Bentley, S. Ennis, and S. Rastogi. 2013. 2010 Census Race and Hispanic
Origin Alternative Questionnaire Experiment. Washington, DC: US Census Bureau.
Cooper, R., J. Kaufman, and R. Ward. 2003. “Race and Genomics.” The New England
Journal of Medicine 348(12): 1166–1170.
Davenport, C. 1917. “The Effects of Race Intermingling.” Proceedings of the American
Philosophical Society 56(4): 364–368.
de la Cruz, G., and A. Brittingham. 2003. The Arab Population, 2000. US Census
Bureau. Retrieved from http://www.census.gov/prod/2003pubs/c2kbr-23.pdf
016
106 • W h a t I s  R a c e ?
Dupré, J. 1981. Natural Kinds and Biological Taxa. The Philosophical Review
90(1): 66–90.
Ennis, S., M. Ríos-Vargas, and N. Albert. 2011. The Hispanic Population, 2010. US
Census Bureau. Retrieved from http://www.census.gov/prod/cen2010/briefs/
c2010br-0 4.pdf
FBI. 2014. 2014 Hate Crime Statistics: Table 1. Washington, DC: Federal Bureau of
Investigation.
FBI. 2015. Hate Crime Data Collection Guidelines and Training Manual, version 2.0.
Washington, DC: Federal Bureau of Investigation.
Feldman, M. 2010. “The Biology of Ancestry: DNA, Genomic Variation, and Race.”
In H. Markus and P. Moya (eds.), Doing Race: 21 Essays for the 21st Century.
New York: W. W. Norton, pp. 136–159.
Fine, A. 1984. “The Natural Ontological Attitude.” In J. Leplin (ed.), Scientific Realism.
Berkeley: University of California Press, pp. 83–107.
Girle, R. 2009. Modal Logics and Philosophy, 2nd edition. Montreal: McGill-Queen’s
University Press.
Glasgow, J. 2003. “On the New Biology of Race.” The Journal of Philosophy
100(9): 456–474.
Glasgow, J. 2009. A Theory of Race. New York: Routledge.
Glasgow, J. Forthcoming. “‘Race’ and Descriptivism.” In Q. Spencer, The Race Debates
from Metaphysics to Medicine. New York: 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.
Graves, J., and M. Rose. 2006. “Against Racial Medicine.” Patterns of Prejudice
40: 481–493.
Guo, G., Y. Fu, H. Lee, T. Cai, K. Harris, and Y. Li. 2014. “Genetic Bio-Ancestry and
Social Construction of Racial Classification in Social Surveys in the Contemporary
United States.” Demography 51(1): 141–172.
Hardimon, M. 2003. “The Ordinary Concept of Race.” The Journal of Philosophy
C(9): 437–455.
Hardimon, M. 2013. “Race Concepts in Medicine.” Journal of Medicine and Philosophy
38: 6–31.
Hardimon, M. 2017. Rethinking Race: The Case for Deflationary Realism. Cambridge,
MA: Harvard University Press.
Hartl, D., and A. Clark. 2007. Principles of Population Genetics, 4th edition.
Sunderland: Sinauer Associates.
Haslanger, S. 2012. Resisting Reality. Oxford: Oxford University Press.
Herrnstein, R., and C. Murray. 1996. The Bell Curve: Intelligence and Class Structure in
American Life. New York: Free Press.
Hirschfeld, L. A. 1996. Race in the Making. Cambridge, MA: MIT Press.
017
Hochman, A. 2013. “Against the New Racial Naturalism.” The Journal of Philosophy
CX(6): 331–351.
Hochman, A. 2014. “Unnaturalised Racial Naturalism.” Studies in History and
Philosophy of Biological and Biomedical Sciences 46: 79–87.
Humes, K., N. Jones, and R. Ramirez. 2011. Overview of Race and Hispanic Origin
2010: 2010 Census Briefs. Washington, DC: US Census Bureau.
Jablonski, N., and G. Chaplin. 2010. “Human Skin Pigmentation as an
Adaptation to UV Radiation.” Proceedings of the National Academy of Sciences
107(2): 8962–8968.
Jeffers, C. 2013. “The Cultural Theory of Race: Yet Another Look at Du Bois’s ‘The
Conservation of Races.’” Ethics 123(3): 403–426.
Kaplan, J. 2010. “When Socially Determined Categories Make Biological
Realities: Understanding Black/White Health Disparities in the U.S.” The Monist
93(2): 281–297.
Kenny, E., N. Timpson, M. Sikora, M. Yee, A. Moreno-Estrada, C. Eng, . . . S. Myles.
2012. “Melanesian Blond Hair Is Caused by an Amino Acid Change in TYRP1.”
Science 336: 554.
Kittles, R., and K. Weiss. (2003). “Race, Ancestry, and Genes: Implications for Defining
Disease Risk.” Annual Reviews: Genomics and Human Genetics 4: 33–69.
Kripke, S. A. 1980. Naming and Necessity. Cambridge, MA: Harvard University Press.
Kusuma, P., N. Brucato, M. Cox, D. Pierron, H. Razafindrazaka, A. Adelaar, . . . F.
Ricaut. 2016. “Contrasting Linguistic and Genetic Origins of the Asian Source
Populations of Malagasy.” Scientific Reports 6: 260–266.
Lugo, L., A. Cooperman, G. Smith, E. O’Connell, and S. Stencel. 2013. A Portrait
of Jewish Americans: Findings from a Pew Research Center Survey of U.S. Jews.
Washington, DC: Pew Research Center.
Maglo, K. 2010. “Genomics and the Conundrum of Race: Some Epistemic and Ethical
Considerations.” Perspectives in Biology and Medicine 53(3): 357–372.
Maglo, K. 2011. “The Case against Biological Realism about Race: From Darwin to the
Post-Genomic Era.” Perspectives on Science 19(4): 361–390.
Maglo, K., T. Mersha, and L. Martin. 2016. “Population Genomics and the Statistical
Values of Race: An Interdisciplinary Perspective on the Biological Classifcation
of Human Populations and Implications for Clinical Genetic Epidemiological
Research.” Frontiers in Genetics 7: Article 22.
Malek, A. 2009. A Country Called Amreeka: U.S. History Retold Through Arab-
American Lives. New York: Free Press.
Mallick, S., H. Li, M. Lipson, . . . D. Reich. 2016. “The Simons Genome Diversity
Project: 300 Genomes from 142 Diverse Populations.” Nature 538(7624): 201–206.
Manichaikul, A., W. Palmas, C. Rodriguez, C. Peralta, J. Divers, . . . J. Mychaleckyj.
2012. “Population Structure of Hispanics in the United States: The Multi-Ethnic
Study of Atherosclerosis.” PLoS Genetics 8(4): e1002640.
018
108 • W h a t I s  R a c e ?
Martínez-Cruz, B., R. Vitalis, L. Ségurel, F. Austerlitz, M. Georges, . . . E. Heyer. 2011.

“In the Heartland of Eurasia: The Multilocus Genetic Landscape of Central Asian
Populations.” European Journal of Human Genetics 19: 216–223.
May, K., and K. Hardwick. 2006. “The Spindle Checkpoint.” Journal of Cell Science
119: 4139–4142.
McEvoy, B., J. Lind, . . . A. Wilton. 2010. “Whole-Genome Genetic Diversity in a
Sample of Australians with Deep Aboriginal Ancestry.” The American Journal of
Human Genetics 87: 297–305.
Mills, C. 1998. Blackness Visible: Essays on Philosophy and Race. Ithaca, NY: Cornell
University Press.
Morning, A. 2011. The Nature of Race: How Scientists Think and Teach about Human
Difference. Berkeley: University of California Press.
OMB. 1995. “Standards for the Classification of Federal Data on Race and Ethnicity.”
Federal Registrar 60(166): 44674–4 4693.
OMB. 1997. “Document 97-28653: Revisions to the Standards for the Classification
of Federal Data on Race and Ethnicity.” Federal Register 62(210): 58782–58790.
OMB. 2000. Provisional Guidance on the Implementation of the 1997 Standards for Federal
Data on Race and Ethnicity. Washington, DC: Office of Management and Budget.
Papantoniou, N., G. Daskalakis, J. Tziotis, S. Kitmirides, S. Mesogitis, and A. Antsaklis.
2001. “Risk Factors Predisposing to Fetal Loss Following a Second Trimester
Amniocentesis.” British Journal of Obstetrics and Gynaecology 108: 1053–1056.
Pemberton, T., M. DeGiorgio, and N. Rosenberg. 2013. “Population Structure in a
Comprehensive Genomic Data Set on Human Microsatellite Variation.” G3: Genes,
Genomes, Genetics 3(5): 891–907.
Perry, J. 2001. Reference and Reflexivity. Stanford, CA: CSLI Publications.
Pritchard, J., M. Stephens, and P. Donnelly. 2000. “Inference of Population Structure
Using Multilocus Genotype Data.” Genetics 155: 945–959.
Putnam, H. 1973. “Meaning and Reference.” The Journal of Philosophy 70(19): 699–711.
Risch, N., E. Burchard, E. Ziv, and H. Tang. 2002. “Categorization of Humans in
Biomedical Research: Genes, Race and Disease.” Genome Biology 3(7): 1–12.
Risch, N., S. Choudhry, M. Via, M. Basu. R. Sebro, C. Eng, . . . E. Burchard. 2009.
“Ancestry-Related Assortative Mating in Latino Populations.” Genome Biology
10(11), R132: 1–16.
Roberts, D. 2011. Fatal Invention: How Science, Politics, and Big Business Re-Create Race
in the Twenty-First Century. New York: The New Press.
Robertson, M., and G. Joyce. 2012. “The Origins of the RNA World.” Cold Spring
Harbor Perspectives in Biology 4(a003608): 1–22.
Root, M. 2003. “The Use of Race in Medicine as a Proxy for Genetic Differences.”
Philosophy of Science 70(5): 1173–1183.
Rosenberg, N., S. Mahajan, . . . M. Feldman. 2005. “Clines, Clusters, and the Effect
of Study Design on the Inference of Human Population Structure.” PLoS Genetics
1(6): 660–671.
019
Rosenberg, N., J. Pritchard, . . . M. Feldman. 2002. “Genetic Structure of Human

Populations.” Science 298(5602): 2381–2385.
Serre, D., and S. Pääbo. 2004. “Evidence for Gradients of Human Genetic Diversity
within and among Continents.” Genome Research 14: 1679–1685.
Shipp, T., B. Bromley, E. Lieberman, and B. Benacerraf. 2000. “The Frequency of the
Detection of Fetal Echogenic Intracardiac Foci with Respect to Maternal Race.”
Ultrasound in Obstetrics & Gynecology 15: 460–4 62.
Spencer, Q. 2012. “What ‘Biological Racial Realism’ Should Mean.” Philosophical
Studies 159(2): 181–204.
Spencer, Q. 2014. “A Radical Solution to the Race Problem.” Philosophy of Science
81(5): 1025–1038.
Spencer, Q. 2015. “Philosophy of Race Meets Population Genetics.” Studies in History
and Philosophy of Biological and Biomedical Sciences 52: 46–55.
Spencer, Q. 2016. “Genuine Kinds and Scientific Reality.” In C. Kendig (ed.), Natural
Kinds and Classification in Scientific Practice. Abingdon: Routledge, pp. 157–172.
Stryer, L. 1995. Biochemistry, 4th edition. New York: W. H. Freeman.
Sullivan, S. 2013. “Inheriting Racist Disparities in Health: Epigenetics and the
Transgenerational Effects of White Racism.” Critical Philosophy of Race
1(2): 190–218.
Sundstrom, R. 2002. “Race as a Human Kind.” Philosophy & Social Criticism, 28(1):
91–115.
Tang, H., T. Quertermous, . . . N. Risch. 2005. “Genetic Structure, Self-Identified
Race/Ethnicity, and Confounding in Case-Control Association Studies.” American
Journal of Human Genetics 76(2): 268–275.
Taylor, D., N. Nagle, K. Ballantyne, . . . R. Mitchell. 2012. “An Investigation of
Admixture in Aboriginal Australian Y-Chromosome STR Database.” Forensic
Science International: Genetics 6: 532–538.
Taylor, P. 2013. Race: A Philosophical Introduction, 2nd edition. Cambridge:
Polity Press.
Tishkoff, S., and K. Kidd. 2004. “Implications of Biogeography of Human Populations
for ‘Race’ and Medicine.” Nature Genetics 36(11): S21–S27.
Tran, S., A. Caughey, and M. Norton. 2005. “Ethnic Variation in the Prevalence
of Echogenic Intracardiac Foci and the Association with Down Syndrome.”
Ultrasound in Obstretrics & Gynecology 26: 158–161.
USCB. 2010. Race and Hispanic or Latino Origin for Puerto Rico: 2010 Census Summary
File 1. Washington, DC: US Census Bureau.
Wallman, K. 1998. “Data on Race and Ethnicity: Revising the Federal Standard.” The
American Statistician 52(1): 31–33.
Weinhold, B. 2006. “Epigenetics.” Environmental Health Perspectives 114(3):
A160–A167.
Weiss, K., and J. Long. 2009. “Non-Darwinian Estimation: My Ancestors, My Genes’
Ancestors.” Genome Research 19: 703–710.
10
110 • W h a t I s  R a c e ?
Wood, T. 2006. “The Current Status of Baraminology.” Creation Research Society

Quarterly 43(3): 149–158.
Yudell, M., D. Roberts, R. DeSalle, and S. Tishkoff. 2016. “Taking Race Out of Human
Genetics.” Science 351(6273): 564–565.
Zack, N. 2002. Philosophy of Science and Race. New York: Routledge.
Zhao, Z., N. Yu, Y. Fu, and W. Li. 2006. “Nucleotide Variation and Haplotype Diversity
in a 10-kb Non-Coding Region in Three Continental Human Populations.”
Genetics 174: 399–4 09.
1
4 IS RACE AN ILLUSION OR A (VERY) BASIC

REALITY?
Joshua Glasgow
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 re-
peatedly engaged in race riots that targeted Chinese and Filipino
lives and property. Following an 1882 ban on Chinese immigrants,
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.
12
112 • W h a t I s  R a c e ?
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 cit-
izens, and while this policy would take different forms over time, naturaliza-
tion was racially restricted until the 1940s, when the Nazi specter shamed any
nation that still stained itself with racialized population policies. In the mean-
time, 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 tes-
timony. 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 indig-
enous 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
13
Is Race an Illusion or a (Very) Basic Reality? • 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 en-
tire 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, peti-
tioned to be legally recognized as white for the purpose of becoming a cit-
izen. 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 ex-
pand 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 citi-
zenship, 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 codi-
fied whiteness from the bench: for legal purposes, Thind’s evidence was beside
14
114 • W h a t I s  R a c e ?
the point. In this context, the words ‘white’ and ‘Caucasian’ were to be un-
derstood 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 cit-
izens, 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 suffo-
cating 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 fol-
lowing 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.
15
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 ex-
clusion 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 regu-
larly 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 actu-
ally 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 ordi-
nary, 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 conver-
sation 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
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).
16
116 • W h a t I s  R a c e ?
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 dis-
cussion (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 fo-
cused on. Even narrowing the scope of our investigation to the context of ordi-
nary 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 under-
stood from many different perspectives. Nonetheless, to the extent that people
from those different standpoints are at least sometimes having a coherent conver-
sation about race, and at least sometimes participate in a shared system of racial
identification and practice with each other, underlying that common conversa-
tion 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 com-
plicit 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
17
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
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 con-
versation 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.
18
118 • W h a t I s  R a c e ?
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 irrele-
vant 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 prin-
cipled biological reason to put one racial boundary here and another racial
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.
19
boundary there based on visible traits. From the perspective of (good) biolog-
ical 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 cate-
gories (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 road-
block that the populations with these conditions do not correspond to or-
dinary 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 bi-
ological about being black gives rise to this disease susceptibility. But black
people who are recent immigrants to the United States do not suffer dispro-
portionate 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
210
120 • W h a t I s  R a c e ?
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 intel-
ligence, 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: sci-
entifically provable genealogical populations are not races.10 The source of
the mismatch is that genealogical populations are ultimately determined by
8. Variations on the Mismatch Objection can be found, for example, in Appiah (1996, 71–
74); Atkin (2017); Blum (2002, 143–144); Condit (2005); Feldman et al. (2003); Glasgow
(2003; 2009, Ch. 5); Hirschfeld (1996, 4); Hull (1998); Jorde and Wooding (2004); Keita
et al. (2004); Keita and Kittles (1997, 538); McPherson and Shelby (2004); Millstein (2015);
Montagu (1964, 7); Pigliucci (2013); Pigliucci and Kaplan (2003, 1166); Witherspoon et al.
(2007); Yudell et al. (2016); Zack (2002).
9. For variations on genealogy theory, see, for example, Andreasen (1998, 2000, 2004); Arthur
(2007, Ch. 2); Burchard et al. (2003); Dobzhansky (1941, 162); Kitcher (1999); Mayr (2002);
Risch et al. (2002); Sarich and Miele (2004); Spencer (2014); Templeton (1998).
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 represen-
tative samples (Witherspoon et al. 2007).
12
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 vis-
ibly 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 posses-
sion 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 ances-
tral 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
21
122 • W h a t I s  R a c e ?
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 genealog-
ical 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 genealog-
ical 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 ra-
cialize 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 an-
swer to population science.
Now one reply to the Mismatch Objection is that it doesn’t matter if sci-
ence 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
213
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 con-
ceptually 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
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.
214
124 • W h a t I s  R a c e ?
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 ref-
erent for ‘races’ is the set of populations recognized by the most plausible ver-
sion 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 con-
cede 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 co-
vert 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 incom-
patible with other commitments, as we have seen. We are committed to ra-
cial 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
215
with someone of a different race. Many are committed to certain racial iden-
tities, 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 con-
ceptual 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, surpris-
ingly. 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 con-
ceptual 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: ‘were-
wolf ’ surprisingly refers to person with rabies. But we don’t say that; ‘were-
wolf ’ 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
216
126 • W h a t I s  R a c e ?
lie. But we have to be careful here: this is just evidence; it is not a conclu-
sion. Some say that each folk theory of race reveals a separate operative con-
cept 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 ul-
timately 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 incom-
patible 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 po-
tential 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 ex-
perimenter 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 biolog-
ical traits and skin color in particular. Even if everyone prioritized biology,
it still would not be decisive. After all, if a representative poll found an-
cient Europeans univocally saying that whales are fish, this would not de-
cisively 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,
217
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 pre-
dict 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 tre-
mendous 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 fantas-
tical things, we are more willing to stop believing in the existence of bio-
logical 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 insurmount-
able 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 or-
ganized 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 de-
liver 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 be-
havior. (Or perhaps, vice versa.) Even then, biology seems unlikely to deliver
race in that sense.
218
128 • W h a t I s  R a c e ?
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 di-
viding 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 construc-
tionist 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 am-
nesia 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 our-
selves and others racially.
Rudimentary constructionism says that, because we no longer recog-
nize 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.
13. For elaboration, see Glasgow (2009, esp. Ch. 6).

219
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 con-
structionism that we need to capture. Let’s call groups that society recognizes
as races racialized groups (Blum 2002; Shelby 2005). Racialization, as dis-
tinguished from race, is socially real. Because differential treatment follows
racialization, we bend ourselves into groups that come with their own possibil-
ities, 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 lock-
step 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 exper-
iment 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 gov-
ernment 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 rudimen-
tary 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
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.
310
130 • W h a t I s  R a c e ?
in racial classification. And because these more durable facts might not dis-
appear during a bout of collective racial amnesia, a more sophisticated con-
structionism 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 ac-
counts, 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 par-
ticular 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 institution-

alized 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
13
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 construc-
tionist 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 dis-
appear. 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
312
132 • W h a t I s  R a c e ?
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 con-
cept 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 pessi-
mistic 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 construc-
tionism, 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 anal-
yses of racialized practice are crucial to addressing injustice, but that is dif-
ferent from the task of identifying what exists.
All forms of constructionism say that race disappears if and when the rel-
evant social facts, the ones that the constructionist thinks are race-making
31
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 cen-
turies of unjust racist behavior. Wanting their children to avoid the same ra-
cial 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 cul-
ture 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 con-
structionism, that baby was firmly Asian, but her Asian-ness somehow instan-
taneously 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,
314
134 • W h a t I s  R a c e ?
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 cor-
rect. 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 opera-
tive 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 or-
ganized 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 biolog-
ical 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 an-
other 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
315
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 reve-
lation 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, pre-
serving 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 “ref-
erence 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 gene-
alogy 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’ success-
fully 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 can-
didate 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 refer-
ence. ‘Werewolf ’ does not refer to a person with rabies. When we discov-
ered 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
316
136 • W h a t I s  R a c e ?
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 sur-
prising 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 construc-
tionism 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 vi-
olence, about privilege, about pain, about culture. We build the fences, and
lives are consistently and unjustly taken and made worse off. Redress is re-
quired. 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 se-
mantic 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 vin-
dicate 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.
317
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 under-
mine 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
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).
318
138 • W h a t I s  R a c e ?
genogroup (Montagu 1964, 23). I have argued that it would be better to re-
pair the mistake and preserve something very close to our present racial iden-
tities 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 revolu-
tionary 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 revolu-
tionary 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 be-
come 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 con-
ceptual 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 compel-
ling 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.
319
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 busi-
ness of finding everything that exists. For example, everyone agrees that sci-
ence 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 soci-
ology 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 sci-
entifically 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 irrele-
vant. 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 sci-
entific 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
410
140 • W h a t I s  R a c e ?
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 biolog-
ically 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 re-
veal such dangers (Shulman and Glasgow 2010), and just because a fact is dan-
gerous, 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.
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).
41
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 rele-
vant. 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 igno-
rance 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 per-
ceptual 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 ob-
viously 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
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).
412
142 • W h a t I s  R a c e ?
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 an-
other, 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 iden-
tity 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 mul-
tiple 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 rele-
vance. 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
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.
413
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 pres-
ident 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 an-
swer 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 unscien-
tific 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 bio-
logical racial groups, but there are traits that could be the basis for real, basic
races. Is that enough to have race?
41
144 • W h a t I s  R a c e ?
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 de-
ploy 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 inde-
terminate 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.
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: 425–4 42.
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 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.
415
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,
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/2014/0 4/14/is-barack-obama-black/
Cillizza, Chris. 2014b. “The Next America Poses Challenges, Opportunities for
Politicians.” The Washington Post, April 13. https://www.washingtonpost.com/pol-
itics/the-next-america-presents-challenges-opportunities-for-politicians/2014/0 4/
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/2015/0 7/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.
416
146 • W h a t I s  R a c e ?
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, Joshua, and Jonathan Woodward. 2015. “Basic Racial Realism.” Journal of the
American Philosophical Association 1: 449–4 66.
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.
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.
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.
417
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:// w ww.pewresearch.org/ f act- tank/ 2 014/ 03/ 14/ u- s - c ensus- l ooking- 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.
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,
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.
418
148 • W h a t I s  R a c e ?
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
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.
419
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.
510
5 H A S L A N G E R ’ S R E P LY T O G L A S G O W,
JEFFERS, AND SPENCER
5.1. Returning to the Question

Each of the four accounts of race offered in earlier chapters captures
something important about the phenomenon of race and offers a
credible account of what race is. How do we decide between them?
Do we have to decide between them?
Philosophical disagreements are often hard to evaluate. It is
easy to wonder: What are the data? What are the agreed-upon
methods? What is at stake? Moreover, answers to questions such as
“What is race?” can easily inherit disagreements from other related
topics. For example, we saw in Chapter 1 that if we try to answer
the question “What is race?” by considering what the term ‘race’
means (using what Mallon [2006] calls the “semantic strategy”), we
often end up with disagreement about race because of background
disagreements about meaning. And disagreements about meaning
can themselves rest on disagreements about such matters as how
the mind works, what we can know, and whether abstract entities
exist. Do we need to have a complete philosophical account of the
world in order to determine which account of race to accept? That
would be discouraging!
An initial, and somewhat straightforward, way of evaluating
an answer to a “What is X?” question considers if it gets the cases
right: we have a pre-theoretic sense of at least core examples of X
and not-X , and an acceptable answer must apply to those cases.1 For
example, suppose we answer “What is a bachelor?” by saying that a
bachelor is an unmarried male. That is not good enough. The Pope
is an unmarried male, but is the Pope a bachelor? What about my
1. This is the “Matching Criterion” that Mallon’s (2006) and Glasgow’s (Chapter 4
in this volume), “Mismatch Objection” relies on.
15
Haslanger’s Reply • 151
dog Sparky? I don’t think so. But just getting the core cases right doesn’t seem
to be enough either, for the answer to such questions are supposed to give us
insight not just into which things are X, but also into what it is to be X. Plato’s
classic example in the Euthyphro is piety. He points out that all and only those
things that are pious are loved by the gods. But, he argued, being loved by the
gods isn’t what it is to be pious. Why not? Because piety is in some important
sense prior to the gods’ love: the gods love prayer, for example, because it is
pious. So their love cannot be what makes prayer pious.
How do we answer questions about what it is to be X? What is this sort of
question even about? Let’s consider why we might ask such questions. Two
kinds of contexts come to mind. One is when there is a conflict over whether
something is, or is not, X. For example, I say that mold is a plant, and you
disagree. We then look up what it is to be a plant and discover that plants are
living organisms that produce their own nutrients through photosynthesis.
This is what it is to be a plant. So I’m wrong. Molds are not plants, they are a
kind of fungi. Cases of this kind don’t actually need to involve conflict. They
may be more a matter of uncertainty: Is a mold a plant or an animal? (Neither,
we discover, because molds don’t satisfy what it is to be a plant or what it is
to be an animal; figuring this out may also require knowledge of what it is to
be a mold.) Another context for asking such questions is when we want to ex-
plain why something characteristically behaves in a certain way. For example,
I want to understand why the trees in my yard are dropping their leaves, and
you explain what it is to be deciduous, which may, in turn, involve an expla-
nation of what it is to photosynthesize, etc. The twofold role of what it is to be
X—settling uncertainty over cases and seeking explanation—are both impor-
tant in answering such questions.
This suggests that to answer questions about what it is to be X, we must
situate Xs in a broader frame of understanding. The questions carry implicit
contrasts, for example, is a mold a plant (or an animal or a fungi)? Why is
my tree dropping its leaves (rather than keeping them all winter)? Explaining
what it is to be X characterizes the phenomenon in ways that provide explan-
atory links to related phenomena.2 But it is not always clear what sort of ex-
planation is called for, or what phenomena are relevantly related.
With this in mind, it is worth considering the idea that the four chapters
are asking and answering somewhat different questions. As I read them,
2. I want to allow here that there are different kinds of explanations, e.g., explanations in terms
of composition, function, structure, etc. See Garfinkel (1981) on explanation as a response to
contrastive questions.
512
152 • W h a t I s  R a c e ?
Glasgow and Spencer focus primarily on race as a classification of humans: Is

this person Black? Is that person Asian? If so, then what features underlie
those attributions? Jeffers and I, however, focus on a broader range of so-
cial and cultural practices: What are racial identities? What makes a group
a racial group? Race is not assumed to be an intrinsic fact about us or our
bodies, by virtue of which we can be classified. Race is a conceptual tool for
understanding the social world; race is attributed to individuals derivatively
by virtue of facts about their social milieu.
Our inquiries have different starting points, so our questions about the
practices in question reasonably differ. Glasgow and Spencer ask whether the
practices of racial classification in question are warranted, or whether the clas-
sificatory terms fit the world. (Spencer says yes, Glasgow, no.) Jeffers and I ask
whether and why certain social and cultural practices are properly thought
of (and should continue to be thought of ) as racial. Jeffers and I both think
that some social practices are racial, but differ in which ones we take to be
relevant to this determination and whether the practices we focus on should
continue. How does one go about comparing and evaluating answers to dif-
ferent questions?
In spite of the differences, one might nevertheless think that there is a
background question that we all must answer: What makes something (any-
thing!) racial, for example, what makes a practice (of classification, or other
social practice) a racial practice? And here the four of us do seem to differ.
Spencer takes it at face value that the OMB is employing a racial classifica-
tion: this is what it claims to be doing and what people grant it the authority
to do. Glasgow takes the reliance on biological visible features to be a nec-
essary condition for racial classification based on a careful examination of
ordinary judgments, hypothetical cases, psychological studies, and reflective
equilibrium. Jeffers takes a shared form of life to be necessary for a group to
be a racial group, based on historical research, sociological studies, and nor-
mative considerations concerning the value of racial practices. And I take so-
cial hierarchy to be a necessary condition on racial groups based also on an
interpretation of our past and present practices and normative considerations
about how we might usefully deploy ‘race’ in political debate. It would seem
that we disagree not only in our conclusions, but also in our methods. We are
engaged in different projects that overlap, but shouldn’t be expected to yield
the same results; there is room for many projects to flourish. That being said,
there are reasons why we are engaged in different projects, and it is worth con-
sidering some of those reasons. I will consider Spencer and Glasgow first, and
then turn to Jeffers, whose views and methods are closest to my own.
513
5.2. Reply to Spencer
On Spencer’s view, races are genomic ancestry groups corresponding to the
OMB classification of people into American Indians, Asians, Blacks, Pacific
Islanders, and Whites. He provides evidence that the OMB classification is
authoritative from the point of view of ordinary Americans by reference to
the census and other legal documents. He bases the claim that there are ge-
nomic ancestry groups corresponding to the OMB classification on recent
population genetics.3
What question or questions is this theory of race answering? As he sets
it out, the heart of Spencer’s project lies in asking whether there are biolog-
ical races that might be relevant for medical research. Is such research legit-
imate, or is the belief in such races misguided? More specifically, he wants
to determine whether there are genetic differences between the five groups
formally classified as races in the United States. The OMB classification is
significant because it is the standard for medical and other governmental re-
search; and ordinary residents of the United States, when interacting with the
state, knowingly rely on it.
Spencer makes clear why his project is important: if there is even a chance
that there are significant biological differences corresponding to the Rosenberg
K = 5 genetic ancestry groups, we should be doing research that explores such
possibilities. We can facilitate such research if there is a standard classification
system that can be assumed across research projects, a system that individuals
know and the state enforces, and that matches the Rosenberg groups. It turns
out that the OMB racial classification system meets these conditions, so it has
a legitimate claim to being, in some sense, ordinary and—given the potential
relevance for medical research—a valuable classification system. Moreover, it
is one that most Americans consider a “racial” classification system.
It seems that a social constructionist about race, myself in particular, could
grant all this. It is compatible with my view that there are genetic ancestry
groups corresponding to the OMB categories and that we should keep an
open mind about whether there are significant medical results to be had by
doing genetic research on these groups. Given the challenges of gathering data
and the significance of racial identities in the United States, it may also be
fruitful to continue to use the term ‘race’ for these categories. Such questions
3. I’m not going to discuss the “mismatch” between the OMB classifications and the Rosenberg
K = 5 genomic ancestry groups. This is something that Glasgow discusses at length in his
chapters.
514
154 • W h a t I s  R a c e ?
and concerns are very different from my own, however, so if concession on

these points is what he is after, I see no conflict with my view (with the sub-
stantial caution that it is an empirical question whether using the term ‘race’
for these categories is, on balance, politically wise in the long term, or not).
Nevertheless, there are a number of places where I find Spencer’s
arguments unconvincing. First, if the issue is really about promoting bio-
medical and other research, it is not clear to me that the Rosenberg K = 5
level of classification is the best one for research purposes (see Koenig et al.
2008; Hochman 2013); there is little evidence that the K = 5 classification
is the most interesting or medically significant. On the assumption that the
goal is to organize ourselves into meaningful (or potentially meaningful)
biological categories for the purposes of medical research, wouldn’t a more
fine-grained (or differently grained) classification system be better? Why not
divide up human populations along the lines of historical exposure to malaria
(so linked to sickle cell disease), or those descended from pastoral tribes (so
less prone to lactose intolerance), or annual exposure to sunlight (relevant to
vitamin D absorption and other diseases such as multiple sclerosis), or many
other possible classifications?4 Given that the state can undertake to enforce
just about any system, why shouldn’t the OMB use a more obviously bio-
logically fruitful one? Or why include racial categories on the census (and
other legal documents) at all and instead leave the classification to the med-
ical professionals?
I also find implausible Spencer’s arguments that the OMB is the “ex-
pert” on race that ordinary speakers defer to. Would the general population
defer to any classification tagged as “racial,” by the OMB; for example, if the
OMB chose the Rosenberg K = 12 classification, would that become the or-
dinary concept of race? (I doubt it.) It might be that such a change in the
4. I’m also confused about Spencer’s claim that racial terms are “tags” or “names” (as in the
sense of proper names) for the K = 5 continental populations. On one interpretation, the tag
or name is attached to a particular—just as the name ‘Sally’ is attached to me. What is the
particular for which ‘Black’ is a name? Perhaps it is the set of individuals who have a certain
genetic profile. But which set is that? It isn’t the set of currently existing (living) people with
that profile, because someone born in 10 months of two Black parents should also be in the set.
Is it the set of all future and past individuals with that genetic profile? Note that we are not in
a position to determine what set that is—it may not even be determinate which set it is. Is it
the set of all possible individuals who might have had that genetic profile? At this point we are
nearing the second option, which is to take racial terms to be names, not for particulars, but for
properties, e.g., having genetic profile G. But then it is confusing to emphasize that racial terms
are names for populations, rather than just claiming that the term functions as an ordinary
predicate. To say that x is Black is to say that x has a certain genetic profile. Racial populations
are those who have one of the K = 5 genetic profiles.
51
state-mandated classifications is not in the offing anyway because the OMB

is not primarily responsive to biologists or doctors, and is both historically
and currently responsive to opinions of the American population (Nobles
2000).5 Because the OMB seeks input from and defers to the general popula-
tion about what race is in setting up its categories, it is hard to claim also that
it is functioning as the expert.
Spencer relies on the broad use of the census and other legal documents to
obtain information about race to support the claim that most Americans “defer”
to the state, and in particular the OMB, to determine what race is. However,
Americans are required by law to fill out forms in which they designate their
race (for the census, school, employment, etc.) and are given a mandated set
of options to choose from. There is considerable evidence that millions of
Americans change their race and/or ethnic categories from census to census;6
so many, at least, do not find such forms a meaningful exercise to report the
facts about their ancestry (and why, if it were, would self-report be taken as suf-
ficient?). There is a sense, of course, in which I defer to the state in determining
what the state’s racialOMB categories are, just as I defer to the state in determining
what a class-A felony is (or, as in my earlier chapter, what a sheriff is).7 But this
is no evidence that I defer to the state in what racial or ethnic practices I engage
in, or not. I also defer to astrologists in determining what astrological sign I am,
and if I am asked to identify my sign, I give the answer based on my birthdate.
But my ability and willingness to do this says nothing about how I live my life,
what matters to me, or whether I take the classifications to be warranted. I take
astrology to be ridiculous. If I were required by law to complete a census (or
other legal document) that asks me to indicate my astrological sign, not based
on my birthdate, but on self-identification, I would give an answer, but my an-
swer would say nothing about what I mean, how I live, or to whom I grant au-
thority. Those of us interested in racial and racializing practices see the OMB as
part of the system that creates and enforces race as we know it, not a source of
expertise with respect to what race is.
5. The Census Bureau National Advisory Committee on Racial, Ethnic, and Other Populations
(previously the National Advisory Committee) includes members from diverse populations,
conducts regular public meetings, and makes recommendations based on input from academic
and non-academic sources.
6. http :// w ww.pewresearch.org/ f act- t ank/ 2 014/ 0 5/ 0 5/ m illions- o f- a mericans-
changed-their-racial-or-ethnic-identity-from-one-census-to-the-next/
7. For example, the fact that state governments use a binary sex distinction on state documents
and that most people assume, as a result, that sex is binary, does not make it true. The fact that
most people have either an XX or XY chromosomes does not make it true either.
516
156 • W h a t I s  R a c e ?
There are clearly different legitimate priorities in considering the huge

range of racial phenomena—biological, medical, cultural, political, legal,
aesthetic, normative—and the theorist comes to them with a particular set
of practices in mind, a particular set of questions, and a particular theoret-
ical and/or political purpose (see Haslanger 2016). The social and political
practices that I want to describe and explain concern patterns in human in-
teraction, racial ideology, and durable forms of social stratification, and for
this, neither the OMB classification, nor the Rosenberg research, is espe-
cially useful. Biology and the state are only some of the many factors to draw
on in answering legitimate questions about race. I am happy to grant that
Spencer has pointed to one way of understanding race in response to one set
of questions and concerns. We should be cautious, however, in concluding
that he has captured “our” concept of race, or that that we should use the term
as he proposes in all (or even some) contexts.
5.3. Reply to Glasgow
Glasgow provides us with a hard choice. Either race is not real—our uses of
the term ‘race’ are vacuous and are based on illusion—or race is real in only
a “basic” sense. In the latter case, the term ‘race’ picks out somewhat random
groups of individuals and has no meaningful role in natural or social science.
My comments here will focus on the first, anti-realist, option. In fact, I’m
sympathetic to Glasgow’s “basic realism,” that is, a permissive ontology that
allows him to say that the term ‘race’ has an extension, even though that ex-
tension is neither a natural or social kind (Haslanger 2012, Ch. 6). He and
I disagree, however, about what constitutes a social kind. On my view, social
kinds are not necessarily the subject matter of social science. For example, a
city animal commission could stipulate a distinction between big dogs and
little dogs (e.g., by weight) for the purposes of designating different dog parks
as open to different sized dogs. Such a distinction, I believe, is a social distinc-
tion, even though it would not play a role in social science (Haslanger 2016).
Nevertheless, races, as I understand them, are important kinds in social sci-
ence. (I will say more about how I understand race as a social kind in my reply
to Jeffers.) The core difference between Glasgow’s view and my own lies in our
methods for determining what it is to be a race. And I believe this difference
can be traced back to differences in our broader projects.
How should we understand Glasgow’s project and the question he is asking?
As I read him, Glasgow is primarily interested in how we—understood as
the general population—think about race. Our thinking about race, however,
517
need not be conscious and self-aware. To elaborate this, he draws on the dis-
tinction between manifest and operative concepts. A concept is manifest to
us, if we can introspect its content (e.g., by a priori reflection on application
of the term to potential cases and intuited connections to other concepts). So
the concept of chair is plausibly manifest. I don’t need to conduct empirical
research to figure out what a chair is, and it is unlikely that I would be sur-
prised to find out what chairs really are. Operative concepts, on his view, are
not fully accessible through a priori reflection. We may need empirical input,
and we may be surprised about what we learn. So, for example, at one point
the concept of fish seemed to involve nothing much more than an animal that
swims in oceans, rivers, lakes, and such. But then we found out about whales.
Whales swim in oceans, but aren’t fish. That was a surprise! Similarly, it can
be surprising to learn that tomatoes and avocados are fruits. The operative
concept, then, takes our ordinary judgments as central, but allows for some
degree of correction both to the content and the extension.
But, Glasgow argues, there is a limit to what we can learn and still be
working with the same concept. If our empirical investigation yields that
werewolves are just people with rabies (no magic involved), do we keep the
concept of werewolf, or drop it? In fact, when this hypothesis emerged, we
stopped applying the concept of ‘werewolf ’ to actual cases, because the em-
pirical results were incompatible with the core commitments underlying
how we viewed and treated alleged werewolves. As he describes them, oper-
ative concepts are what we rely on in ordinary discourse, and consist “of the
commitments people would stick to when they are forced by consistency to
abandon part of a set of incompatible commitments” (Glasgow, Ch. 4).8
According to Glasgow, one core commitment about race is that “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” (Glasgow, Ch. 4) Since humans display a continuum
of biological, visible traits, the condition is not satisfied. Not all Asians have
eyelids with epicanthic folds. Not all Blacks are dark skinned or have tightly
8. Note that although Glasgow relies on my characterization of an operative concept (Haslanger

2012, Ch. 2), we disagree about this characterization of operative concepts. I don’t think there
need be a set of shared commitments that we (the general population) rely on when faced with
inconsistencies or that, even if there are, that the ones we choose to hold on to in one context
or at one time would have been the same in others, e.g., before we learned the relevant empirical
facts. This disagreement stems, in part, from our background philosophies of language. I am
a semantic externalist and Glasgow, as I understand him, is a neo-Fregean. See: https://plato.
stanford.edu/entries/two-dimensional-semantics/
518
158 • W h a t I s  R a c e ?
curled hair. Many people in the world are such that they cannot be placed into
a race by virtue of appearance; but according to our pre-theoretic judgements,
we all have a race. Given the inconsistency in our judgments, we have to give
up something. Glasgow argues that the ordinary response is to hold on to
the appearance criterion for race (the core commitment), and reject exist-
ence: races don’t exist.
As I interpret Glasgow’s project, the crucial issue is how to understand our
classificatory practices. We are inveterate classifiers and constantly rely on clas-
sification in thought and action. On Glasgow’s view, a crucial and ineliminable
part of what we do when we classify by race is to classify by (visible) appearance.
But it is misguided to think that there are a small number of human groups that
can be differentiated by appearance, and that these map onto our pre-theoretic
selection of racial groupings. It is an illusion that has had terrible consequences,
and we should, on his view, disrupt the illusion and all it has produced.
I am sympathetic to Glasgow’s effort to debunk our ordinary thinking
about race and to find ways to convince people that their classificatory
practices purporting to group humans into races are misguided. Yes, people
do rely on visual cues to differentiate races and to attribute race to people
(Alcoff 2005). And this is a source of much wrongdoing and injustice. If we
could stop people from engaging in racial attributions based on appearance,
we would be better off.9
I have some concerns, however, about Glasgow’s arguments for his view.
In particular, I think that the phenomenon of racial classification is more
complex and the common practices are less consistent than he suggests.
We surely attribute race to individuals based on factors other than appear-
ance: this is why there is resistance to racial “passing.” For example, there are
individuals who appear and represent themselves as White, but are still con-
sidered Black (see Hobbs 2016; Ginsberg 1996); there are transracial efforts
(e.g., Rachel Dolezal appeared Black but many think she remained White10);
John Howard Griffith in Black White Me (1961) remained White in spite of
his successful disguise; Gregory Howard Williams (who became president
9. My agreement with Glasgow here concerns an issue on which I disagree with Jeffers. If we
understand race as culture—with no associated appearance expectations or norms—then I am
not opposed to race—though I would call it ethnicity rather than race. But I don’t trust culture
that brings with it and enforces appearance and ancestry rules.
10. In 2015, Rachel Dolezal, who identifies as African American, or “trans-Black,” was revealed
to have White parents. For more information, see: http://www.nbcnews.com/news/nbcblk/
rachel-dolezal-why-she-can-t-just-be-white-ally-n738911, also Dolezal and Reback (2017).
519
of the University of Cincinnati) “learned” he was Black at age 10 and to

most appears White (Williams 1995).11 More generally, in my experience,
appearance is given much more importance in the White community than
the African American community for determining racial membership. Given
this history of the sexual exploitation of Black slaves and the one-drop rule,
African Americans are well aware that there is a broad spectrum of appear-
ance in the Black community, that some Blacks have “white” features and,
more generally, that appearance is a poor guide to race. Such examples strike
me as important evidence of “our” commitments and should be factored in as
inputs to deliberation regarding whether and how we go on.
One strategy for managing these examples is to suggest that races (the
groups) are associated with a particular set of visible traits, but race need not
be attributed to individuals based on appearance. For example, one might
count as a member of the Black race, even if one doesn’t “appear” Black, by
virtue of having a recent ancestor (within n generations?) who appears (stere-
otypically?) Black. This would also help address some of the mismatch caused
by the spectrum of human features. But, importantly, one doesn’t get to be
White in the United States by virtue of having a recent ancestor who appears
White. And as far as I can tell, there is no canonical appearance for Native
Americans or Asians. This is partly why, in the United States, individuals
who are hard to interpret visually in racial terms are often asked “Where are
you from?” with the hope of eliciting a racial cue. (This is also reasonably
interpreted as an insult—the implication is that if you aren’t, or don’t appear,
White then you aren’t American.) Such questions are often addressed to those
who are born in the United States (and sometimes several generations before
them are as well), so it isn’t about birthplace or current residence.
Glasgow suggests, however, that there must be a shared concept of race—a
shared set of core conditions on the application of the term—in order for
us to communicate, even to disagree. One might extend his argument by
suggesting that if there isn’t a shared set of (consistent) conditions, then that
is another reason to think there are no races. Nothing can satisfy a set of in-
consistent conditions.
However, it isn’t true that one must share a core set of commitments re-
garding the meaning of one’s terms in order to communicate with others. In
11. For a summary of his story, and a slide show, see http://magazine.uc.edu/issues/0310/pres-
ident.html. Williams’s autobiography (1995) goes into more detail. Critics have questioned his
narrative of identity: https://adpowellblog.wordpress.com/2014/06/14/the-problem-with-
gregory-howard-williams-poster-child-for-the-one-drop-myth-of-white-racial-purity/
610
160 • W h a t I s  R a c e ?
ordinary cases, one only needs enough common ground for the purposes at
hand (Stalnaker 2002), and not for all conversations in which I might use the
term in question. For example, I can talk with my colleague about the person
in the next room, even though he believes that persons are essentially psycho-
logically unified and I think they are unified by the continuity of their body.
His core commitments include psychological continuity and not somatic
continuity; mine are the opposite; our intuitions about hypothetical cases
may diverge. Yet the people we are talking about aren’t going to be transferred
to new bodies, so we communicate perfectly well.12 Most of us have no idea
what our core semantic commitments are, and it is likely that we vary those
commitments depending on context and with whom we are communicating.
Moreover, if you and I are using a term in different ways, it is fairly easy to ad-
just one’s use for the conversation at hand. For example, I have an articulated
set of commitments with respect to the term ‘race’ (see Chapter 1), but I know
that these are not the conventional commitments and that most others I talk
to don’t share those commitments. This does not prevent me from communi-
cating with them, agreeing and disagreeing with them about empirical claims,
and generally managing in (and often resisting) racial practices.13
Nevertheless, I agree with Glasgow that there are many errors in ordi-
nary thinking about race, and that we sometimes think we are engaged in a
meaningful classificatory process, when we aren’t. We should aim to correct
these errors, and doing so will be a step in promoting justice. For example,
in August 2017, as I am writing this reply, the US Justice Department has
announced that it will reopen an affirmative action lawsuit against Harvard
University, alleging discrimination in admissions against Asian Americans.14
In the wake of this decision, National Public Radio hosted a call-in discus-
sion of affirmative action during which a caller, “Sean from Dayton,” pas-
sionately expressed that in college admission, consideration of “the color
of our skin should be non-negotiable. That shouldn’t even be an admissible
12. For many more examples, see Williamson (2003).

13. There is much controversy over word meaning, and multiple distinctions are relevant here.
Gasparri and Diego (2016) offer an excellent overview. For example, even if terms have a lexical
meaning, it is not clear what determines that meaning, and lexical meaning (plus syntax) is not
enough to determine a truth-evaluable proposition; moreover, semantics and syntax certainly
don’t determine what is said on a particular occasion (Saul 2012).
14. The recent Justice Department action is described here: http://www.thecrimson.com/ar-
ticle/2017/8/11/justice-department-intervene-analysis/ and useful background on the case
can be found here: http://www.thecrimson.com/article/2016/11/7/harvard-admissions-
lawsuit-explainer/. The New York Times also offers commentary: https://www.nytimes.com/
2017/08/02/us/affirmative-action-battle-has-a-new-focus-asian-americans.html?_r=0
16
part [of the criteria for admission].”15 This is an excellent example of how
race is often equated, at least verbally, with visible features such as skin color.
(Though what “skin color” is Sean from Dayton thinking distinguishes
Asian Americans, who are supposedly at issue in this discussion?16) Like,
Glasgow, I think the assumption that race is a matter of skin color (or other
visible markers) should be rejected, and it would help in discussions such
as the one just mentioned to point this out. But Glasgow and I differ in the
next step. He argues that we should conclude that races don’t exist (though
racialized groups do). I maintain that there is a way of reading the history
of our understanding of race such that it is apt to claim that races exist as
social groups defined in part by a projection of “color” markings onto cer-
tain (assumed) lineages, and systematic subordination along those lines (see
Chapter 1). What does this disagreement between us amount to? Does it
really matter which approach one takes?
In fact, I think that there are costs and benefits to both an anti-realist
approach and a social constructionist approach. In some contexts, one ap-
proach will be more fruitful, and in other cases, another approach will be
effective. As I’ve emphasized, negotiation of meaning in conversation is a
fluid and socially complex matter, and is not limited to insisting that one’s
interlocutor conform their usage of a term to the lexical or conventional def-
inition. Context-free semantic facts, if there are any, cannot be wheeled in to
settle our disagreement.
Nevertheless, let’s return to Sean from Dayton’s comments. As I read his
comments, reference to “skin color” is not reflecting a core commitment
about the visibility of race (or the “skin color” of Whites or Asians), but in-
stead functions as a code for race that purports to reveal its moral irrelevance.
Skin color, like eye color or earlobe shape, is a paradigm of something that
should (in an ideal context) be morally irrelevant. If race is just skin color,
then it too is morally irrelevant. But this equation with race and skin color
deflects attention from what’s at stake. Once skin color and other such visible
markers have been used as a basis for unjust treatment, ignoring that history is
either staggeringly naïve, or disingenuous.
15. See On Point, August 7, 2017. “Rethinking Affirmative Action,” with Jane Clayson, Anemona
Hartocollis, Stuart Taylor, and Nancy Leong. http://www.wbur.org/onpoint/2017/08/07/
rethinking-affirmative-action In the podcast, Sean’s comments occur from 23:01–24:03; the
comment about skin color quoted in the text begins at 23:48.
16. One way of interpreting Sean from Dayton’s points is to see him as complaining specifically
about “black” and “brown” people “taking” admissions slots; this is why “skin color” is at issue.
The fact that Asian Americans are also being displaced by Whites is obscured.
612
162 • W h a t I s  R a c e ?
Glasgow’s approach invites us to change Sean’s thinking: your belief in

races as essentially distinguished by skin color (or as visibly marked groups)
is false. Instead, think of Asian Americans (etc.) as racialized groups. You can
then see that differential racialized group treatment is warranted. Racialized
groups have been treated unjustly. But racialized groups are not races.
My strategy, instead, is to focus on Sean’s actions: look at what you are
doing. Your awkward reaching to skin color as a proxy for race—a move that
you may not be aware of or even be prepared to reflectively endorse—is part
of a racializing process that creates and sustains the very racial groups we are
talking about. Who are these groups? They are groups of individuals who
may or may not get into Harvard (which was his concern), not by virtue of
differences in skin color, but rather, due to actions like yours. Your apparently
true and innocent comments about the irrelevance of skin color mask the
(White) power you have to deny that race is morally significant, and this,
in turn, buttresses your power. (He explicitly states in the full version of his
comments that he is White.)
As I see it, my approach is focused on the actual people who will be af-
fected by the discussion and the policies—the ones who are the subject matter
of the controversy—rather than intuitions about how to apply terms in a
broad range of remote possibilities. Let’s talk about these people right here—
you, Sean, call them a ‘race’—what do they really have in common? Not skin
color! You are wrong about that. And you are wrong about the moral aptness
of differential treatment of races, for race isn’t all about skin color (or visible
biological features, more generally).
There is a sense in which I don’t really care what Sean from Dayton is
thinking, but who he is talking about and what his comments do to the groups
being referenced (and those who make policy affecting them). Racialization
does not happen elsewhere and elsewhen, but in our everyday conversations.
My doubts about the importance of core semantic commitments (etc.) of the
sort Glasgow explores leads me to focus more on racial practices other than
racial thinking and racial classification. Racial language and concepts are not
irrelevant to racial practices. However, invocation of “color”—whether it be
skin color or other markers—is not an innocent classification scheme gone
wrong, but is an instrument in constituting, tracking, policing, abusing, glo-
rifying, and celebrating forms of social hierarchy.17
17. There is so much literature on this issue. A powerful example is in Lawrence Blum (2002,
Ch. 6), though perhaps ironically, Blum agrees with Glasgow that there are no races, but only
racialized groups. I believe that the heart of the differences between Glasgow and Blum, and
613
All that said, I am completely open to the idea that in some contexts, it is
useful to challenge how people think about race, and assumptions they make
about the necessary conditions (visible or otherwise) for being a race, or a
member of a race. Intervening in political debate is like intervening in a con-
versation. There isn’t a general strategy that will work in all contexts. My goal
in this book has been to argue that the project of trying to understand what
“the ordinary person” means by race is not really a well-defined project, given
the controversies in the background about meaning, and that even if we are
trying to identify what “the ordinary person” thinks about race, the phenom-
enon is more complex than Glasgow’s account captures. Finally, I grant that
it is not at all obvious what is at stake in distinguishing races from racialized
groups, but have offered some reasons for thinking that not only do the full
range of considerations suggest that races are social groups (see Chapter 1),
but also that focusing on the groups that people are trying to talk about when
they use the term ‘race,’ rather than supposed core commitments of their
concepts, is not only semantically permissible, but also politically valuable.
5.4. Reply to Jeffers
Recall that in my discussion of “What is X?” questions at the beginning of
this reply, I suggested that such questions, more specifically, “What is it to be
X?” typically arise in context in which we are uncertain about (or encounter
conflict in) adjudicating instances of X, or when we are trying to situate Xs in
an explanatory framework. As suggested earlier, I read Glasgow and Spencer
as primarily concerned with race as a system of classification. When con-
sidering systems of classification, two sets of questions are pressing: (1) On
what basis does the system differentiate between kinds of things? This helps
us sort out controversies over cases. And, (2) are we warranted in using the
classifications; do they give us a fruitful way to understand the range of phe-
nomena? Are there things that fall into each of the various categories in the
system? Answers to these questions help us evaluate the explanatory potential
of the classifications within a broader explanatory framework.
me, lies in the philosophy of language. I’m attempting here, however, to articulate what is po-
litically at stake in the linguistic controversy. In effect, what people think is not obviously ac-
cessible to us (or to them) and doesn’t really capture what we are talking about with racial
language: what we are talking about is a group of people, and what matters is how we treat these
people, not some obscure intuitions about how we would use a term under bizarre conditions.
614
164 • W h a t I s  R a c e ?
Jeffers and I don’t deny that racial categories provide ways of classifying
people. However, our discussions in this book are guided by an interest in
race as a set of social practices. What does that mean?
Consider a different question. What is a family? As I understand this ques-
tion, it is not about our concept of family, or what the term ‘family’ means, or
how families, as a category, should fit within a broader classification system.
Rather, asked in the context of the contemporary United States, it is a ques-
tion about a particular social formation: the (nuclear) family. People organize
themselves into families; the state recognizes families; families have effects on
their members and their neighbors; families can be evaluated as functional
and dysfunctional. In other words, families are part of a social ontology.
To say that families should be included in a social ontology is not to say
that they are things. It is more illuminating to say that families are small sys-
tems that fit within bigger systems.18 Families are made up of individuals and
particular sorts of relationships between them (e.g., spouse of, parent of, child
of, sibling of, pet of ). They play a functional role in the society; for example,
they provide a space of intimacy and lasting bonds, a site for dependent care,
and a financial unit for cooperation and taxation.
When there are disagreements about what families are, the disagreements
are plausibly about the sorts of relationships that constitute families (Do
same-sex spousal relationships constitute a family? Are birthparents part
of an adoptive family?), and the social function of families (How much de-
pendent care should the family, as opposed to the state, be responsible for?).
These questions are not best answered by consulting intuitions about possible
cases or by considering the legal (OMB?) definition of a family. The question
invites us to investigate a variety of families and similar groups, to come up
with a model that has explanatory potential, and to evaluate whether there
might be meaningful adjustments to how the formation works that would
constitute an improvement (along some lines or other). For some time, the
model (nuclear) family has consisted of a married heterosexual couple and
their offspring. However, this model has never fully fit the facts, though the
facts differ from place to place, culture to culture. Groups of individuals who
live together in intimate, long-standing relationships that involve financial
and material interdependence include unmarried (and now married) straight
and gay couples, adopted children, extended biological family, single parents,
18. Note that systems are not just groups of individuals, e.g., their parts are specified in terms of
functional roles and they are multiply realizable. See Epstein (2015); Haslanger (2017).
615
groups of unmarried (coupled or un-coupled) friends, etc. The question,

“What is a family?” is too open-ended to answer on its own because the data
are too complex and could be organized in different ways. We need more in-
formation about the point of the question (is there a background controversy
about cases?), the relevant contrasts, the explanatory (or legislative) project
to which an answer contributes, and background normative considerations
(Garfinkel 1981). For example, am I a sociologist attempting to explain the
racial achievement gap by reference to family structure? Or a psychologist
planning a treatment plan for a depressed child? Or am I a legislator making
a recommendation for revision of the tax code? Or am I deciding whom to
invite to a family reunion?
When I ask “What is race?”—and I believe this is true of Jeffers as well—
the goal of the project is to understand a distinctive set of social formations
generally understood as racial, for example, patterns of association, stable
positions within a social hierarchy, values and cultural differences (e.g., in re-
ligion, music, literature, and art). One issue is whether these social formations
have a biological basis; for example, is the racial wealth gap, or the academic
achievement gap, or the marriage rate, best explained by biological differences
between the races? All of the authors in this book would reject this hypothesis.
In short, no. A related question that also raises issues of biology is whether
humans are naturally disposed to mistrust or feel hostile to others who “look
different” from us (Hirschfield 1998; Mallon 2019). My guess is that the four
of us all reject this hypothesis as well. (However, note that even if this is a
human predisposition, that does not mean that it is inevitably expressed—
socialization can override it.) On my view, the very notion of people “looking
different” is socially shaped. As a mother once mentioned to me (jokingly) in
commenting that I look like my (adopted, Black) son: “People say I look like
my sons. But how could that be? They are boys!” Her point was to raise the
question of why the judgment of “looking like” in such cases is supposed to be
evaluated mainly in terms of facial features, skin color, hair texture, and other
racial markers, and not by age, size, sex.
In general, social constructionists about race look to history to explain
the origin of racial formations and to social and psychological dynamics to
explain their persistence. In this respect, Jeffers and I are engaged in very sim-
ilar projects, and as far as I can tell, I think we would agree on the broad out-
line of the explanations of race as we know it. But there are also some telling
differences.
On Jeffers’s view, race is the product of European voyages of exploration
and conquest (cf. Mallon 2013). The minimal conception of race sketched by
61
166 • W h a t I s  R a c e ?
Hardimon (2003) in terms of appearance, ancestry, and geography has been

used as a basis for dividing humans into groups, and according them different
social, political, and legal positions. The groups, being differently positioned
in society, each developed a culture that both protected them against harms
and offered affirming interpretations of their histories, their lives, and their
bodies. As a result, although the origin of racial groups lies in subordination
along lines of minimal race, racial formations have evolved past this and are
sources of value. In effect, one might argue that a social-political conception
of race, such as my own, stops the analysis too early: although it captures a
stage in the history of race, it does not do justice to how races offer valuable
ways of life that, although rooted in a history of subordination and injustice,
have surpassed that. As a result, we should not seek to undermine the racial
order, but to support and build upon the existing racial formations.
I am deeply sympathetic with Jeffers’s view. I agree that racial groups have
developed valuable ways of life and that racial identity is a meaningful part
of many people’s lives. I also do not intend, nor is my view committed to,
eliminating or resisting cultural differences; I want to eliminate White su-
premacy and other forms of unjust social hierarchy. There are two points,
however, where I think Jeffers and I diverge. The first concerns the scope of
our projects. The second concerns the normative considerations that guide
our different inquiries.
In undertaking to provide an account race, the task is complicated by the
fact that races are formed differently in different contexts. For example, it
appears that there are both different races, and different criteria for being a
race, in Brazil, South Africa, the United Kingdom, the United States, and
other parts of the world. As I understand the four authors in this book, we
are primarily concerned with racial formations (and racial classification) in
the United States. However, this does not mean that we are only concerned
with African Americans, Asian Americans, Native Americans, European
Americans, and Latin Americans (the last of course being a controversial
case of race). Considering how race functions in the United States, I would
think that an African American’s distant cousin who lives in Cameroon is
Black, though not an African American; a Hmong American’s grandmother
living in Laos is also Asian, though not Asian American. That races extend
beyond the United States is, for certain purposes of explanation, important.
Attitudes and decisions within the United States (e.g., concerning immigra-
tion, legal standing, stigma, and cultural appropriation) have an effect on
racial formations, not only within the boundaries of the United States, but
around the world. And race has always been an international matter, involving
617
the capture, enslavement, transport, and exploitation of geographically dis-

tant peoples, justified in many cases by reference to race (and admittedly also
supposed “degrees of ” culture). The problem is, however, that it is not clear
how to define a “way of life” that is shared by all Asians, or all Blacks, or all
Whites, or all Native Americans (I mention this also in Chapter 1). Even
Latin Americans, who are sometimes counted as an ethnicity rather than a
race, do not share a way of life (Alcoff 2000).
So here is one issue of scope: In analyzing racial formations in the
United States, are we just considering the formations that occur within
US borders? In other words, are we just considering African Americans,
Native Americans, etc.? In my own account of race, I do not want to re-
strict myself to formations that occur within the United States, but include
how racial formations in the United States extend beyond our boundaries
to impact the rest of the world and affect the racial formations globally.
For example, Blacks, around the world, are racialized in the United States,
and this affects international relations, development efforts, and global
capitalism.
But suppose we restrict ourselves to racial formations within the United
States. Is there a basis for claiming that those who participate in these
formations “share a culture” in a very thin sense? Perhaps they are cultures
of the sort discussed by Espiritu as pan-ethnicities. Pan-ethnic groups have
a shared identity as being racialized by the dominant culture, and develop
shared practices of resistance (among others). (We would have to adjust this
to include Whites as a pan-ethnic group—perhaps based on dominance?) I’m
not wholly convinced that US racial groups even constitute pan-ethnicities.
I would guess that there are complex disasporic populations who would con-
stitute the relevant pan-ethnicities for non-White groups (Gilroy 1993), and
these aren’t coextenisve with African Americans, or Asian Americans, (etc.),
culturally or geographically. Nor are they coextensive with racial or racialized
groups that span original and diasporic communities.
One approach to address this problem—again focusing on US racial
formations—would be to start from shared cultures within American racial
groups, and take whomever the members of those groups count as members as
part of the race. For example, if there is an African American way of life, and
those who participate in that way of life count anyone with relatively recent
ancestry in Africa as a participant in that way of life, then because African
Americans regard Black Africans as members of their racial group, they are.
I am not suggesting that Jeffers would endorse such an interpretation. I offer
it because I think it illuminates some of the issues. In particular, it prompts the
618
168 • W h a t I s  R a c e ?
question: What counts as a shared “way of life”? If the extension of a race is deter-
mined simply by being thought of in a certain way by core members, even if there
is no shared language, shared cultural traditions, or substantive contact, then it is
not clear how a “way of life” within the group is a source of unity and meaning.
I am a whole-hearted believer in the value of shared ways of life. I’m raising here
a version of the Mismatch Objection: I don’t think that meaningful shared ways
of life correlate with races understood as racial formations.
But perhaps I’m still interpreting the scope of the project too widely. Perhaps
we should not be thinking of all African Americans, or all Asian Americans, or
all Whites, as the relevant groups, but more culturally coherent and meaningful
subsets. Immigrant Haitians in the Boston area share a “way of life” that is dif-
ferent from the “way of life” of African Americans in rural Mississippi. Should
these count as different racial groups? Among American Whites also, there are
substantial differences in culture. Should we conclude that American Whites are
not a single racial group? These questions raise, I think, an important question
for both Jeffers and me: Is it really useful to think of racial formations at all? Is
race a plausible way of understanding how people organize themselves across the
entire United States? Perhaps this is a false presupposition at the heart of the so-
cial constructionist project.
Let’s return to the analogy with the family. Families are very diverse. Even in
the United States there are different ways of forming families (e.g., love marriages,
arranged marriages, no marriage); there are different norms for the behavior of
partners to the marriage; there are different norms for child-rearing, household
and financial matters, and inclusion of non-biologically related members. Not all
family members love each other. Not all families live together. Not all families in-
volve children or even sex. What makes a family, we might say, is structural: there
are certain relationships of interdependence internal to the family, and certain
functional roles that the family plays in the larger society.
Races are different from families, but they can also be understood struc-
turally. Here is one way of capturing the structure: There is set of “anchoring”-
ancestors who were (or were thought to be) from a particular part of the
world at a certain point in history,19 and a set of features that these anchoring-
ancestors and their “same race” descendants have (or are thought to have).
Having (or being thought to have) an anchoring-ancestor of this kind, and the
features that are supposed to pass through the lineage, positions one in a so-
cial hierarchy. The hierarchical structure created and sustained by referencing
19. Jeffers draws on Hardimon (2003, 442, 445, 447) to suggest that for American races that
it is ~1492.
619
the anchoring-ancestors and their assumed descendants positions individuals

in races. Sameness of culture within a race is not required, so we avoid the
mismatch between culture and race. These structures, however, are helpful in
explaining social and historical phenomena.
However, I grant that there are mismatches, unintuitive consequences,
and plenty of vagueness in this sort of social-political account. For example,
on my view, a group can count as a race even though it is a total fiction that
there are anchoring-ancestors from a particular region. For example, race and
caste hierarchies are similar, but different. In caste hierarchies, differences be-
tween groups are not linked to different geographical origins, but typically
are social role or function based. So, for example, one caste may be identified
with priests, another with warriors, and so on. But there is no assumption
that the priests, or warriors, originated in a particular geographical region.
(Sometimes caste is explained by analogy with the body: the body depends
on there being hands and feet, heart and mind; likewise, society’s body also
depends on such distinctions.)20 Suppose, however, that an ideology develops
according to which one of the castes (or sub-castes) is imagined to have orig-
inated elsewhere and to be physically distinguishable; suppose further that
social formations emerge that exploit this, for example, by questioning the
group’s belonging, questioning whether the members count as “one of us,”
or their rights to remain. Then, on my view, this group is racialized—it is in
the process of becoming a race—even if the ideological claims are false. The
same might occur in the case of an ethnicity, a class, or other social divisions.21
I grant that this is does not capture how many of us think of race, but it is val-
uable as a tool for understanding how racial formations work: what attitudes
to expect, what policy changes to be alert to, and the evolution of practices
both within a group and in response to it.
Another complaint about my view is that it doesn’t capture the phenom-
enon of “passing.” If an individual has a race by virtue of being a member of a
racialized group, and if racialized groups consist of those who occupy a par-
ticular social position—possibly based on misreadings of visible features, or
20. For example, “The body also provides metaphors that fund religious conceptions of ideal
social relationships. Hinduism’s Vedic literature, for example, explains that the four succes-
sively ‘higher’ social castes derived from a cosmic sacrifice separating a primordial being’s
mouth, arms, thighs, and feet” (Fuller 2015). See Rig-Veda Bk 10 (http://www.sacred-texts.
com/hin/rigveda/r v10090.htm). This line of thought supports the idea that castes have been
considered endogenous (in contrast to race as exogenous) and by analogy with the body.
21. Arguably, this happened in Rwanda between the Hutus and Tutsis, and in the former
Yugoslavia between the Serbians and Croatians.
710
170 • W h a t I s  R a c e ?
fictions of origin—then one who “passes” as (is viewed as, treated as, and lives
as) Asian is a member of the Asian race (mutatis mutandis for other races).
This consequence is one that, at least to me, seems apt for some purposes,
but not all. As mentioned in my reply to Glasgow, I think that how we use
language is a complicated and context-sensitive matter, and we should not at-
tempt to legislate how a term should be used in all contexts.22 How we want
to think about “passing” and claims of being transracial are not to be decided,
I think, by consulting our current intuitions, but by a reading of our past and
our normative commitments.
However, one might accept my account for races, as groups, but com-
plicate what it is to be a member of a race. For example, let races be social
formations within a broader sociopolitical structure (formations that dis-
tribute rights, access to social goods, material resources and such, based on
“color”), and add that to be a member requires both that one is socially posi-
tioned in such a formation, and that one is accepted as a member of the race
by others in the group, were the facts of one’s ancestry known. In cases of
“passing,” there would at least be some cases in which one would not satisfy
this further condition (consider again Rachel Dolezal). Adding the condi-
tion would potentially contribute some explanatory power to the account,
but it is also an important normative question whether we should add it, and
who are the appropriate “we” to add it.
This brings us back to the second difference between Jeffers’s approach and
mine. Both of our projects are importantly normative. Jeffers emphasizes the
value of racial solidarity and racial practices, both for those who participate in
them, and for the cause of justice. He eloquently articulates the importance of
maintaining races as cultural groups. I take it that this would involve not just
sustaining racial cultural practices, but using the minimal conception of race
as necessary conditions on membership. So the normative recommendation
is that we continue to group ourselves according to the minimal conception
of race (appearance, ancestry, geography), and support the development and
continuity of those cultures within these groups that celebrate the distinctive
histories, traditions, and bodies of their members.
I agree with Jeffers that we need, for the foreseeable future, to promote the
development of racial practices that protect members of racial groups from
the effects of injustice and that discourage others from perpetrating injustice.
Racial solidarity is valuable under conditions of injustice, both among the
22. I also discuss this in my original paper (Haslanger 2012, published first in 2000).
17
subordinated, and also among those who reject racial privilege (Shelby 2005).
Both Jeffers and I also recognize that an account of race raises questions about
how we envision the world we are aiming for, and what prefigurative practices
it is important to embrace (Leach 2013). A prefigurative politics embodies
in the movement for social justice the values and practices it aims to achieve.
What are those values and practices?
Suppose that we want to promote cultural forms and practices that pro-
vide people ways of understanding their histories, their life experiences, and
their embodiment in ways that are affirming and empowering. As I see it, the
question on which Jeffers and I mainly disagree is whether drawing divisions
between humans on the basis of the minimal conception of race, that is, along
lines of “color” in my sense of the term (marked bodies that carry the meaning
of geographically framed ancestry), is valuable in the long run. I think it is not.
And moreover, I think there are risks in treating “color” divisions between
humans as an ideal, that is, as something that we should embrace not just in
response to current injustices, but as something to be maintained “after the
revolution,” so to speak.
As some evidence that actual shared “color” is not necessary for par-
ticipation in empowering cultural forms, there is no doubt that there are
individuals who do not have the relevant racial ancestry for a racial group,
say Asian, and do not appear to have the relevant ancestry, but nonethe-
less participate in racial practices and find meaning in them. For example, a
child might be told incorrectly that he had an Asian grandparent (perhaps
he, or one of his parents, was adopted, and was not given correct infor-
mation), and that even though he doesn’t appear Asian, this is part of his
background. The child might make this a central part of his identity and
participate fully in an Asian American community. Actual ancestry and
appearance are a poor guide to who finds what meaningful, not only by
“mixed”-race individuals, but even those who fall squarely within the cur-
rent racial formations.
But, more generally, how do we want to “carve” histories, and distinguish
bodies, and prioritize ancestry? What does it mean for me to understand
“my history”? Is my history the history of White people? Why that history?
Is my past reduced to my ancestral background? My maternal ancestors
sided with the British in the American Revolution and fled to Canada.
Does that tell me anything about myself (other than the mere fact that this
is true of my great-great-great-great-grandfather)? Should I consider myself
Canadian? English? Royalist? Am I less American than a descendant of a
family who fought with the revolutionaries? And what of my Black children
712
172 • W h a t I s  R a c e ?
(who, as mentioned before, were adopted). Is this history not theirs, too? Is
their Jewish grandfather’s emigration from Poland not part of their history?
Is one’s history really to be traced by “blood”? Why? Even if we do find value
in tracing biological ancestry, how should that factor into our identities and
group formations?
Moreover, what are the cultural practices that inform my embodiment?
Skin cancer runs in my family, so I apply sunblock. But I also am amazingly
adept at braiding cornrows, maintaining dreadlocks, putting in extensions
(either by braiding or crochet), doing Senegalese twists, and other hair
techniques. (I’m not so good at doing a weave, but have managed to do so a
few times.) I’ve studied hair designs, learned through practice the geometry
of the head, can distinguish different hair products and make an informed
choice. I am known for light hands. My embodiment is not just a first-person
experience. It is a way of being in relation to others, in both intimate and
public ways. The phenomenology of embodiment is deeply affected and
enhanced by appreciation of, and being appreciated by, others (Haslanger
2012, Ch. 9). Sometimes this is achieved through encounters with sameness,
sometimes encounters with difference.
More generally, why should we assume that empowering cultural practices
are ones that group those of the same “color”? Shouldn’t we be interested in
disrupting the assumption that “color” should divide us and seek ways to be
embodied and form identities across “color”? It is extremely important to
note that I am not claiming that we should deny any groups of people (large
or small) opportunities to construct narratives or to engage in embodied
practices that celebrate their understandings of their past or their bodies. This
includes groups who have the same “color.” However, I reject the idea that
“color”-based narratives and practices are necessarily constitutive of what it
is to be of a race.
I am well aware that “blood” relations are valued more highly than “non-
blood” relations in the contemporary United States. I’m also aware that racial
markings are taken to be of great significance to identity. These are back-
ground assumptions that I believe need to be questioned and that should
not be enforced through norms of racial authenticity. In the world I envision
“after the revolution,” we will not be bound primarily by narratives about the
significance of “blood” or appearance, for I believe these narratives misguide
us to focus on relationships that narrow our capacity to live together well. The
narratives may be important for now and for long into the future. But I work
on a daily basis to prefigure a world in which they no longer limit how we
might live in justice.
713
5.5. Conclusion
At the beginning of this reply, I suggested that there was much right and im-
portant about all four of the accounts of race sketched in this book. I posed
the question: Do we have to decide between them? I have suggested that the
four authors might reasonably be read as asking different questions and as un-
dertaking projects of different scope. To the extent that this is correct, there
isn’t pressure to pick one as “the right view.” We are all making important
contributions to the understanding of race—whether we are talking about
racial classification or racial formations.
My own approach differs from Spencer’s and Glasgow’s because I am not,
first and foremost, attempting to understand race as a system of classification,
or what the term ‘race’ means. Instead, I am trying to understand and explain
a distinctive social formation that, I believe, shapes not only how the United
States is socially organized, but how the United States functions in global ec-
onomic and cultural relations.
I agree with a tremendous amount of Jeffers’s discussion: our views are
close both methodologically and substantively. We differ, perhaps, on the
scope of our inquiry, and on the normative considerations that shape our vi-
sion of justice. However, I have not argued for the vision I’ve just sketched
here. I am prepared to believe that mine is not well-grounded in the expe-
rience of being racialized in North America, and that it is ultimately naïve
about what form of justice is possible. But I also think that we are not in a
position to tell, yet, what justice will involve in the long run. The best we can
do is engage in a prefigurative politics that constitutes experiments in living
(Anderson 2014). In effect, I doubt that philosophy can tell me I’m wrong (or
right!) about what justice will be, in the long run. That is something we must
learn by doing, and doing together.
References
Alcoff, Linda Martín. 2000. “Is Latina(o) Identity a Racial Identity?” In Jorge Gracia
and Pablo De Greiff (eds.), Hispanics/Latinos in the United States: Ethnicity, Race,
and Rights. New York: Routledge, pp. 23–4 4.
Alcoff, Linda Martín. 2005. Visible Identities: Race, Gender and the Self. Oxford: Oxford
University Press.
Anderson, Elizabeth. 2014. “Social Movements, Experiments in Living and Moral
Progress: Case Studies from Britain’s Abolition of Slavery.” The Lindley Lecture,
University of Kansas.
174
174 • W h a t I s  R a c e ?
Blum, Lawrence. 2002. I’m Not a Racist, But. . . . Ithaca, NY: Cornell University Press.
Dolezal, Rachel, and Storms Reback. 2017. In Full Color: Finding My Place in a Black
and White World. Dallas TX: Benbella Books.
Epstein, Brian. 2015. The Ant Trap. Oxford: Oxford University Press.
Fuller, Robert. 2015. “Religion and the Body.” Oxford Research Encyclopedia on Religion.
http://religion.oxfordre.com/view/10.1093/acrefore/9780199340378.001.0001/
acrefore-9780199340378-e-18#
Garfinkel, Alan. 1981. Forms of Explanation: Rethinking the Questions in Social Theory.
New Haven, CT: Yale University Press.
Gasparri, Luca, and Diego Marconi. 2016. “Word Meaning.” The Stanford Encyclopedia
of Philosophy (Spring 2016 edition), Edward N. Zalta (ed.), https://plato.stanford.
edu/archives/spr2016/entries/word-meaning/.
Gilroy, Paul. 1993. The Black Atlantic: Modernity and Double Consciousness.
New York: Verso.
Ginsberg, Elaine K. 1996. Passing and the Fictions of Identity. Durham, NC: Duke
University Press.
Gooding- Williams, Robert. 1998. “Race, Multiculturalism, and Democracy.”
Constellations 5(1): 18–41.
Griffith, John Howard. 1961. Black Like Me. New York: Haughton, Mifflin, Harcourt.
Hardimon, Michael. 2003. “The Ordinary Concept of Race.” Journal of Philosophy
100(9): 437–455.
Hardimon, Michael. 2017. Rethinking Race: The Case for Deflationary Realism.
Haslanger, Sally. 2016. “Theorizing with a Purpose: The Many Kinds of Sex.” In
Catherine Kendig (ed.), Natural Kinds and Classification in Scientific Practice.
New York: Routledge, pp. 129–144.
Haslanger, Sally. 2017. “Failures of Individualism: Materiality.” Presented at the First
Annual Critical Social Ontology Workshop, St. Louis, MO. Draft on Academia.
edu.
Hirschfield, Lawrence A. 1998. Race in the Making. Cambridge, MA: MIT Press.
Hobbs, Allyson. 2016. A Chosen Exile: A History of Racial Passing in the United States.
Hochman, Adam. 2013. “Against the New Racial Naturalism.” Journal of Philosophy
6: 331–351.
Koenig, Barbara A., Sandra Soo-Jin Lee, and Sarah S. Richardson. 2008. Revisiting
Race in a Genomic Age. New Brunswick, NJ: Rutgers University Press.
Leach, Darcy. 2013. “Prefigurative Politics.” In The Wiley-Blackwell Encyclopedia of
Social and Political Movements. http://onlinelibrary.wiley.com/doi/10.1002/
9780470674871.wbespm167/abstract
715

116(3): 525–551.
Mallon, Ron. 2013. “Was Race Thinking Invented in the Modern West?” Studies in
History and Philosophy of Science Part A 44(1):77–88.
Mallon, Ron. 2019. “Naturalistic Approaches to Social Construction”, The Stanford
Encyclopedia of Philosophy, Edward N. Zalta (ed.), forthcoming URL = <https://
plato.stanford.edu/archives/spr2019/entries/social-construction-naturalistic/>.
Nobles, Melissa. 2000. Shades of Citizenship: Race and the Census in Modern Politics.
Stanford: Stanford University Press.
Saul, Jennifer. 2012. Lying, Misleading, and What Is Said: An Exploration in Philosophy
of Language and Ethics. Oxford: Oxford University Press.
Shelby, Tommie. 2005. We Who Are Dark. Cambridge, MA: Harvard University Press.
Stalnaker, R. 2002. “Common Ground.” Linguistics and Philosophy 25(5): 701–721.
Williams, Gregory Howard. 1995. Life on the Color Line: The True Story of a White Boy
Who Discovered He Was Black. New York: Penguin Random House.
Williamson, Timothy. 2003. “Blind Reasoning.” Proceedings of the Aristotelian Society,
Supplementary Volumes 77: 249–293.
716
6 J E F F E R S ’ S R E P LY T O G L A S G O W,
HASLANGER, AND SPENCER
Useful debate concerning the metaphysics of race seems to me to

require a healthy balancing of conciliatory and critical approaches,
which might be true of all philosophical debate, but I also have
reasons related to this particular topic in mind. With regard to the
virtue of being conciliatory, there are the general values of being
charitable and avoiding nitpicking forms of disagreement, but there
is also the important question of how much distance we ought to
perceive between the major philosophical positions on what races
are. Ron Mallon’s seminal challenge in his article, “Race: Normative
Not Metaphysical or Semantic,” forces us to consider the possibility
that metaphysical disagreement about race among philosophers is
largely an illusion, as there is widespread agreement among them
on what exists and on the nature of what exists, with the appear-
ance of disagreement still sustained only by differences in defining
terms.1 Mallon concludes that this appearance obscures the real de-
bate worth having over normative questions concerning the value of
various kinds of talk of race. One lesson that can be drawn from his
challenge, even if we hold that there is a substantive metaphysical
debate about race to be had, is that we should not exaggerate how
different anti-realism, social constructionism, and non-essentialist
biological realism are when they are all united in denying the exist-
ence of the biological essences traditionally ascribed to races and it
remains unclear how much they diverge with regard to those enti-
ties or relations whose existences they positively affirm.
Nevertheless, willingness to be critical, sometimes sharply so, is
warranted not only by the general wisdom of proving the worth
1. Ron Mallon, “‘Race’: Normative, Not Metaphysical or Semantic,” Ethics 116

(April 2006): 525–551.
71
Jeffers’s Reply • 177
of ideas by putting them to the test of confrontation with opposing ideas,

but also by the high stakes underlying the discussion of race. Distinctions
between groupings referred to as races are of massive social importance in
today’s world. What therefore ought to undergird a critical orientation when
doing metaphysics of race is an acute sense of the social significance of getting
things right—indeed, I would go so far as to say that one ought to have a
conscious commitment to one’s metaphysical musings on race having useful
implications for figuring out problems of social life (and thus also a commit-
ment to forthrightly opposing positions one believes will lead us astray in
that enterprise). This is, to be clear, not a matter of sacrificing truth for social
utility, but rather prioritizing the derivation of practical implications when
thinking about a matter of social importance.
Reflecting on the need to say something useful with regard to social
problems can also help us address Mallon’s challenge and maintain that, de-
spite what he says, metaphysical debate about race will necessarily continue.
Ideas of race among the general public are no doubt currently in flux, but
biological essentialism remains very much an important part of the mix.
This means that, even assuming it is true that philosophers of race disagree
on absolutely nothing concerning what exists in the world, there is still the
question of how best to encourage the wider public to revise their metaphys-
ical understandings of humankind. Consider now, in connection with this,
Mallon’s examples of what should be factored in when engaging in normative
debate about racial discourse:
the epistemic value of ‘race’ talk in various domains, the benefits and
costs of racial identification and of the social enforcement of such
identification, the value of racialized identities and communities fos-
tered by ‘race’ talk, the role of ‘race’ talk in promoting or undermining
racism, the benefits or costs of ‘race’ talk in a process of rectification
for past injustice, the cognitive or aesthetic value of ‘race’ talk, and the
degree of entrenchment of ‘race’ talk in everyday discourse.2
Philosophers who come to the conclusion that, say, the costs of racial
identification outweigh the benefits must reflect on how we convince people
to stop taking for granted and let go of an aspect of their lives that seems
as normal and real as their occupation or height. Philosophers who decide,
2. Ibid.., 550.
718
178 • W h a t I s  R a c e ?
by contrast, that the benefits outweigh the costs must reflect on how we en-
courage the embrace of racial identities while simultaneously encouraging the
stripping away of essentialist beliefs about race. Debate between philosophers
holding these two positions will thus be, in significant part, debate about
how it makes sense to try to change or not change common sense metaphys-
ical pictures of the world. Adding the other factors Mallon mentions to the
conversation only multiplies the possible metaphysical positions that can
be held, as someone could perhaps promote an anti-realist stance in relation
to personal racial identification, a social constructionist stance in relation to
rectificatory justice, a non-essentialist biological realist stance in relation to
certain medical matters, and so on.
In what follows, I will engage with my coauthors’ chapters, attempting to
balance conciliatory and critical moves. In the first section, I will respond to
Spencer’s non-essentialist biological realism, suggesting that we can see bi-
ology as potentially illuminating matters of race without agreeing that races
are fundamentally biological. The second section will be about Glasgow’s two
competing positions, arguing that both his basic realism and his anti-realism
collapse into social constructionism when plausibly reconstructed. Finally,
in the third section, I will address Haslanger’s political constructionism,
highlighting our similarity and then explaining and defending how my cul-
tural constructionism differs.
6.1. On Spencer’s Non-Essentialist Biological Realism

One motivation for non-essentialist biological realism is the fact that race
has been used by biologists as a subdividing principle for species other than
humans. Philip Kitcher points out, for example, that Theodosius Dobzhansky,
a central figure in the development of evolutionary genetics, discussed in his
classic work on that subject differences between races among beetles, fruit
flies, and snails.3 To differentiate between races of fruit flies does not imply
that some fruit flies are superior to others or even that we can expect stable
behavioral differences between these races beyond higher rates of endoge-
nous mating. The thought therefore arises that biological race should be rec-
ognized as a phenomenon very distinct from both the essentialist myths and
the complicated social matters many of us associate with talk of race. Kitcher,
3. Philip Kitcher, “Race, Ethnicity, Biology, Culture,” in In Mendel’s Mirror: Philosophical

Reflections on Biology (New York: Oxford University Press, 2003), 232.
719
in his “Race, Ethnicity, Biology, Culture,” explores the possibility that social
differentiation in the United States may lead in some cases to reproductive
isolation sufficient to be counted as biological racial difference. Even with
his linking of biology with social matters, it is the measurable fact of mating
patterns, and not the social pattern of perception of outward difference as-
sociated with differing geographical origins, that makes race a concept he is
willing to employ here.
Andreasen’s use of Cavalli-Sforza’s tree to distinguish between cladistic
races among humans, which I described in Chapter 2, severs the tie be-
tween racial differentiation and social differentiation as completely as pos-
sible. Interaction between groups thought of as races in the modern period
is not only not the basis for racial differentiation on this account but such
interaction is, in fact, the cause of what Andreasen takes to be the gradual
disappearance of races: “Ever since the voyages of discovery, colonization
and immigration have been blurring racial distinctness.”4 The reduction of
reproductive isolation between different branches of the tree collapses the
branches, so the massive relocations and increasingly frequent interactions
of people in the modern era threaten the distinctness of races, as Andreasen
understands them, in comparison to the ways in which geographical barriers
fostered their distinctness over the long period of human evolution prior to
the modern era. Andreasen thus argues that social constructionism and her
brand of biological realism are “not in competition” but rather are “comple-
mentary” because they address two different subjects.5 When thinking about
what “race” means, we may look to social constructionism to “aid our under-
standing of social and political implications of current uses of the term” and
to a biological realist view of the kind that Andreasen offers to help us “under-
stand the patterns and processes of human evolution.”6
Especially given the goal of being conciliatory, there is a tough challenge
here to the idea that social constructionists must reject race as a biological
reality in order to affirm that it is fundamentally social in nature. Why not
simply allow that “race” is a term that can be used in different ways in dif-
ferent contexts of research and discussion, and thus that it sometimes refers
4. Robin O. Andreasen, “A New Perspective on the Race Debate,” The British Journal for the
Philosophy of Science 49 ( June 1998): 215.
5. Ibid., 218.
6. Ibid., 220, 219.
810
180 • W h a t I s  R a c e ?
to something social and sometimes something biological?7 Despite the temp-

tation to relent here, it strikes me as important that we resist and emphasize
the confusion bound to result from encouraging among the general public
the simultaneous acceptance of two strongly divergent concepts of race. Even
if we imagine that the resulting confusion will be relatively minimal, the ques-
tion to ask is why we should court any such confusion at all when we can
choose other specialized terms to refer to the kinds of populations that non-
essentialist biological realists like Andreasen have in mind.
Spencer, in Chapter 3, provides social constructionists with a different
sort of challenge. By tying his defense of the biological reality of race to the
social currency of the racial categories of the US Census, Spencer suggests
that the task of the non-essentialist biological realist is to reveal the biolog-
ical underpinning of some commonly accepted notion of what races there
are, rather than offer up a novel scientific division of our species that may
have no connection whatsoever to what is commonly accepted. Spencer thus
meets social constructionists on their own turf, so to speak. Furthermore, by
taking as his starting point the question of whether race is medically rele-
vant, he aptly highlights the practical significance of the possibility that our
common sense racial categories represent real biological diversity. This shared
focus with social constructionists on what has social currency and on what
makes a practical difference to people’s lives makes it a powerful challenge
that Spencer demonstrates significant overlap between what are counted as
races by the US Census and genetic clusters among human populations as
found in the research of Noah Rosenberg and colleagues.
By the same measure, though, this shared focus increases the bite in the
standard objection to all forms of non-essentialist biological realism: the
Mismatch Objection, according to which discrepancies between biologi-
cally respectable distinctions and common sense racial distinctions show that
we should not describe the biologically respectable distinctions as racial. In
Chapter 2, I used the example of a woman born in England to parents from
Bangladesh finding out that South Asians ought to be in the white rather
than Asian category on Spencer’s account. There is, I should admit, an in-
teresting way in which Spencer’s choice of Census categories helps him deal
with certain Mismatch Objections. I would argue, for example, that to think
7. There are, to be sure, criticisms of Andreasen’s specific proposal for a biological concept of
race that could lead one to reject it regardless of one’s position on the relationship between
biological and social concepts. For a recent example, see Zinhle Mncube, “Are Human Races
Cladistic Subspecies?” South African Journal of Philosophy 34(2) (2015): 163–174.
18
of Arab Americans as white will result in a failure to understand both anti-

Arab prejudice as well as anti-racist forms of Arab pride, both of which take
the non-European origin of Arabs to be significant.8 Thus I would take the
fact that Arabs are the same race as those of English or German heritage on
Spencer’s account to be a mismatch. Spencer can respond, however, that we
do not have a mismatch here between biological and social categories because,
as he shows, the overwhelming majority of Arab Americans report themselves
as white on the Census. My reply to this is that it is not at all odd that Arab
Americans are responding in accordance with the definitions provided by
the Census, and it is worth remembering that simplicity (e.g., less rather than
more categories) and stability (e.g., not rushing to add or subtract categories)
are valued traits of bureaucratic schemes. I therefore take the current classifi-
cation of Arab Americans as evidence of weakness in the ability of the Census
to track racial difference, rather than as a demonstration of no mismatch.
The South Asian mismatch, though, remains probably the clearest ex-
ample, and it is important, given the point just made, that its clarity is not
dependent on seeing the Census as correct in its lumping of South Asians
together with other Asians. It is interesting to compare the approach of the
Office of Management and Budget with that of Statistics Canada: with the
exception of those counted as “Aboriginal,” the demographics of non-white
Canadians are captured through the category of “visible minority” and the
current list of options for reporting such status are Chinese, South Asian,
Black, Filipino, Latin American, Southeast Asian, Arab, West Asian, Korean,
and Japanese.9 If the OMB can be seen as misleadingly lumping too many
kinds of Asians together, StatsCan arguably does too much unnecessary split-
ting of Asians, but both recognize the social reality that South Asians are un-
derstood to be non-white.
The lesson of this mismatch and of any other is that biological continuities
and discontinuities between human populations is a subject independent of
the social recognition of racial differences. This is, of course, the same point
that Andreasen directly and centrally affirms. My view, then, is that some
non-essentialist biological realists, like her, recognize this independence
and court unnecessary confusion by advocating that we use “race” to refer
to both subjects, while others, like Spencer, recognize the need to avoid such
8. For useful discussion of these matters, see Amaney Jamal and Nadine Naber (eds.), Race
and Arab Americans before and after 9/11: From Invisible Citizens to Visible Subjects (Syracuse,
NY: Syracuse University Press, 2008).
9. http://www.statcan.gc.ca/eng/concepts/definitions/minority01a
812
182 • W h a t I s  R a c e ?
confusion and offer a unified account of race but then wrongfully downplay
the independence of the two subjects.
Now, in the spirit of conciliation, I would say that to call these subjects in-
dependent is not to deny that they are, in some ways, closely related. Indeed,
I think social constructionists should never be afraid to admit that one cannot
tell the story of racial distinctions without biological diversity entering
the picture. The forms of physical difference involved in racial distinctions
are necessarily at least partially related to forms of reproductive isolation,
whether as a result of people being geographically separated going back to the
distant past or through more recent social distinctions. Once we admit this,
it becomes easy to admit that study of the genetics of human populations of
the kind Spencer invokes has at least the potential to be illuminating with
regard to the study of race. We can also be open-minded, as he is, about what
such research might uncover regarding medical matters and how we might
connect what is uncovered to our usage of common sense racial categories in
medical settings.
Nevertheless, race is fundamentally social and not fundamentally biolog-
ical. The strength of this position—or at least the strength of my adherence to
it—can be illustrated with a thought experiment. Imagine if any mismatches
between the groups picked out by the study of genetic clustering and the
groups known as races in everyday thought and social practice were to disap-
pear in the following remarkable way: people with influence over education
policy, media outlets, and other means of knowledge dissemination came to
be convinced by Spencer’s non-essentialist biological realism and, over the
course of a few generations, the identification of the genetic clusters at K = 5
as races became common sense, at least in the United States. In this scenario,
we would have not just extensive overlap with some bureaucratic categories
but a tight match between a set of biologically real groupings and the ideas
of what races there are in all major forms of public discourse. This is, impor-
tantly, not impossible. It is part of how social norms work that scientific and
philosophical ideas can shift and reshape them. Were this to happen, would it
not then be the case that biology had become the foundation of race?
My answer is no. It is not merely possible but almost certain that biological
research in this scenario could provide us with interesting facts about races,
but this would be completely contingent on the prevailing social situation.
A subsequent shift in popular ideas could result in much less connection to
anything of biological significance (for example, grouping together European
and East Asian under a broad Eurasian category determined by lightness of
skin while splitting off dark-skinned South Asians, etc.). This second shift
813
would not be a move away from race toward something else, but simply a
reorganization of racial designations and identifications. The previous con-
nection to something more biologically significant would thus be revealed as
inessential, for what is essential to race is that people’s looks and lineages as
tied to places of origin gain social significance.
6.2. On Glasgow’s Basic Realism and Anti-Realism

My stance with respect to Glasgow’s chapter can be described as simulta-
neously conciliatory and critical because I reject both of the two positions
on race that he describes and defends as live possibilities while holding that
both of them can be easily reconstructed and accepted as versions of social
constructionism. Indeed, I think Glasgow provides us with the grounds for
not merely a social constructionist but a cultural constructionist perspective,
in particular! Perhaps surprisingly, it is his anti-realism that collapses much
more quickly into social constructionism than his basic realism, so I will start
with my criticisms of the latter.
There is no doubt that Glasgow and his former student Jonathan
Woodward have provided an intriguing challenge to philosophy of race with
their introduction of basic racial realism.10 A form of realism about race ac-
cording to which race is neither biological nor social? This is, if nothing else,
admirably creative and bold. Important to whether one can take it seriously is
one’s willingness to accept Glasgow’s focus on visible traits as the core of the
concept of race. Spencer, as readers may recall, specifically rejects the idea that
races must be distinguishable from each other on the basis of visible traits,
using the case of Melanesians resembling sub-Saharan Africans as an example.
I differ with Spencer on this and see two possible ways of resisting his
point. Lacking Spencer’s commitment to Census categories, one could resist
the idea that we are talking about two different races here, since many who
use the term ‘black’ and apply it to sub-Saharan Africans would be completely
comfortable applying it to Melanesians as well. Alternatively, as I would
prefer, one could caution against diminishing to nothing the physical distinc-
tiveness of Melanesians. While it is certainly the case that one could choose
an individual Melanesian and an individual sub-Saharan African and stump
people on which is which, it seems easily possible to me to gain familiarity
with Melanesian traits to the point of being able to do well at picking them
10. See Joshua Glasgow and Jonathan M. Woodward, “Basic Racial Realism,” Journal of the
American Philosophical Association 1(October 2015): 449–4 66.
814
184 • W h a t I s  R a c e ?
out when shown a number of randomly chosen comparisons (and I am not

only referring here to the noteworthy high incidence of natural blondes in
the Solomon Islands, although this provides a nice example). I thus remain
attached to difference in appearance of some sort as a necessary component
of racial difference.
I part with Glasgow, on the other hand, about whether ancestry is a neces-
sary component. He thinks it is not and uses his thought experiment in which
a chemical agent is put into the water supply that makes everyone look like
the Dalai Lama to show this. According to Glasgow, a world in which we all
look the same is a world without race, even if, for a few generations, geograph-
ical and cultural factors work to keep genealogical lines about as distinct as
they were before. It is true that I can imagine someone saying, in that circum-
stance, “we are all the same race.” I can also imagine someone saying, in that
circumstance, “we appear to be the same but this is only true at a superficial
level—it remains the case that we are divided into races and our habits of asso-
ciation and our continued reproduction along these particular ancestral lines
demonstrate that.” I think the second hypothetical reaction is the more per-
ceptive one and it is, of course, congenial to me that Glasgow evokes culture
as a means of the persistence of reproduction along the old racial lines. Note
also that I am not giving up on the importance of appearance by allowing that
racial difference could persist in a world in which we all look like the Dalai
Lama because difference in appearance was essential in distinguishing the old
racial lines. Genealogical distinctions unrelated at any point in time to differ-
ence in appearance would not, I think, count as racial.
There is an insight in basic realism, however, that can be appreciated even
among those of us who maintain the importance of ancestry alongside ap-
pearance. Glasgow writes: “There are multiple, cross-cutting ways of dividing
us up by visible traits” (142). Similarly, I write in Chapter 2: “given the var-
ious continuities in how we look, there are different sets based on alternative
divisions possible on this basis” (65). The ways of dividing up the world ra-
cially that we are familiar with are contingent and sit alongside other pos-
sible ways that are equally real in at least one important sense: there really
are differences and similarities in how humans look—based, I would add, on
ancestral connections—that can be discerned and used to differentiate us into
broad groupings. These differences and similarities are not imaginary, even if
boundaries we draw or could draw in accord with them will always necessarily
be fuzzy.
Thus, if basic realism is the view that races are real groups of people uni-
fied by similar visible traits, we have reason to acknowledge that, first of all,
815
it identifies something real about human beings and, second, this reality is
worthy of our attention when trying to understand the contingent nature of
racial difference as we know it or are familiar with it. These two points can be
easily incorporated into a social constructionist account of race. Here is how
such an account might proceed: groups that are the same in physical char-
acteristics to the groups we call races in this world could exist without the
social practices and common modes of thought that make them races in our
world. Indeed, we might even want to refrain from saying that social practices
and common modes of thought brought the groups we know as races into
existence, because it is reasonable to think that, when considered from the
vantage point of their delimitation by physical similarities inherited biolog-
ically, most of them pre-existed these practices and modes of thought. Races
are thus social groups whose racial nature is a product of social construction
but whose existence as sets of similar-looking people is not dependent upon
anything social.
As it turns out, such an account is not only possible but has recently been
defended by Jeremy Pierce in his book, A Realist Metaphysics of Race. According
to Pierce,
whatever groups get organized by our conceptual scheme into races

would exist apart from such social organizing, and the social practices
make those groups significant in a way that they would not be if we stuck
with pure biology, since the biological features that tend to be common
within each race are not any more important biologically than any traits
we do not see as significant markers of racial distinctions.11
Glasgow notes the similarity between basic realism and Pierce’s view (142n18).
He also points out what he takes to be a terminological difference between
Pierce and himself, namely, that Pierce appears to think that, for a view to be
identified as social constructionist, “it is enough if it says that races are made
socially significant,” while Glasgow holds that social constructionists “must
say that society makes races exist” (142n18). As I see it, Glasgow is responding
to a problem of equivocation in Pierce’s account. Pierce writes, for example,
“it is better to conceive of the racial social construction as not generating the
existence of racial groups to begin with but instead as drawing attention to
11. Jeremy Pierce, A Realist Metaphysics of Race: A Context-Sensitive, Short-Term Retentionist,

Long-Term Revisionist Approach (Lanham, MD: Lexington Books, 2015), 97.
816
186 • W h a t I s  R a c e ?
already-existing groups, giving them social salience and moral importance.”12

The wording here suggests that the groups that preexist the advent of their
social salience can be meaningfully referred to as races or, at least, as “racial
groups.” But, as Glasgow notes, it seems clear at other times that Pierce’s view
is that “without the social relevance, those groups are not races” (142n18).
Pierce cannot count as a social constructionist, I think, unless he believes
there are no races prior to the social phenomena that generate races. I believe
we ought to be charitable and take as the clearest statement of his view the
claim that “social constructions single out existing entities and make them
into races.”13 It is telling, however, that he takes himself to be reaffirming
this very claim when, a few pages later, he writes: “I do think races exist, and
I think they exist independent from the social processes that give them sig-
nificance.”14 This latter claim is, as far as I can tell, indistinguishable from
basic realism, which is a rejection of social constructionism. While one lesson
here is the obvious point that we should avoid equivocation, I think it is a
happy thing for basic realism that it is so very close to what we are taking to be
Pierce’s considered view that Pierce himself appears to mix them up. I say this
because I take Pierce’s considered view to be a credible position on the nature
of race. I am hesitant to embrace it wholeheartedly, as there are challenges
in thinking about the way racial concepts have played a role in influencing
the social and reproductive interactions of people that complicate the idea
of a set of pre-existing groups made into races only by social salience (think,
for example, of a racial category like mestizo). It may be, however, that the
position can be easily adjusted to deal with such challenges (note that, when
elaborating it prior to attributing it to Pierce, I spoke of most, rather than all,
of the groups we know as races pre-existing their transformation into races).
Basic realism is thus clearly very close to being a credible form of social
constructionism. If not collapsed into something like Pierce’s considered
view, however, it is completely implausible. Glasgow attempts to make basic
realism sound not only plausible but more commonsensical than social con-
structionism by pointing out that, according to basic realism, we are what race
we are regardless of social practices. If this were really all that basic realism
says about the conditions of racial membership, this would indeed be an ad-
vantage over social constructionism with regard to the need to not deviate
12. Ibid.
13. Ibid., 92.
14. Ibid., 95.
817
too far from common sense, as we social constructionists must accept that our
position does deviate from this standard assumption. Basic realism does not,
however, actually hold this advantage. According to basic realism, as Glasgow
tells us, whatever social practices are or are not in place, Barack Obama is not
just black (as he and many others would identify himself ) or mixed (as some
would say, perhaps even interchangeably), but rather he belongs simultane-
ously to at least three different races, and this is not a matter of those races
being mixed within him (indeed, “mixed-race” is one of the three!) but rather
a matter of the different ways the spectrum of humanity can be carved up.
This evidently does not accord with common sense.
Indeed, if basic realism is true, I take it to be the case that Barack Obama
and each of us are members of an infinite number of races. After all, given
the point that humanity is a spectrum, are there not an infinite number of
ways of carving up the spectrum? I am willing to grant basic realism the point
that, however many carvings are possible, the resulting groups can be consid-
ered real groups of similar things. I am even willing to grant that we could
treat each group as a possible race. To say that they are all actual races and
we are right now members of an uncountable number of them seems to me
simply nonsensical. It undoubtedly leaves common sense far behind, and it
is not clear what advantage it seeks to achieve in compensation for this de-
parture. Non-essentialist biological realism departs from common sense with
its mismatches in order to carve up humanity in a biologically significant
way. Social constructionism departs from common sense in holding that so-
cial phenomena create and sustain the existence of races in order to reveal
how, here as elsewhere, we have reason to question whether what seems nat-
ural and independent of our thoughts and practices is, in fact, what it seems.
What does basic realism teach us with its departure from common sense?
As mentioned earlier, only when reconstructed as the social constructionist
claim that there are real groups based on similarity of appearance that can
become races through social salience does it teach us something useful about
the contingency of our racial categories. Otherwise, as far as I can see, it drops
common sense and picks up nothing of value to replace it.
Glasgow’s version of anti-realism departs from common sense as well, but
only in the understandable and productive way that all forms of anti-realism
do, that is, by denying that there are races in light of the problems of biolog-
ical essentialism that make common discourse about race so often untrust-
worthy. If we must choose between Glasgow’s basic realism, unreconstructed,
and his anti-realism, I think the choice is extremely easy. We should obvi-
ously go with the latter. The reason why, to put the point bluntly, is that when
81
188 • W h a t I s  R a c e ?
Glasgow’s anti-realism is evaluated in combination with what he calls his rev-

olutionary reconstructionism about race, there is, in fact, no meaningful dif-
ference between what he offers us and social constructionism. The collapse is
pretty much automatic.
Let us review the steps in Glasgow’s argument. Races are ordinarily un-
derstood to be groups distinguished by certain visible traits. Humanity is a
spectrum with regard to visible traits, and biology will not lead us to carve
humanity up in this or that particular way on the basis of visible traits.
Non-essentialist biological realism is defeated by the Mismatch Objection.
Thought experiments show that we can imagine people still being this or that
race even in the absence of any social facts that could be viewed as dividing
us along racial lines, so social constructionism fails. Being neither biologically
nor socially real, race is not real at all (or, at least, so we will conclude if we do
not consider or are not convinced by basic realism). We have, however, built
identities and communities on the mistake of believing race is real, and we
should address this error not by abandoning the term ‘race’ but by redefining
it. We should change the meaning of the term ‘race’ from “groups distin-
guished by certain visible traits” to “social groups distinguished by being dif-
ferently racialized,” where racialization refers to being socially recognized as
belonging to a particular race. Once we make this change, race will be real.
At what point along the way here should the social constructionist pro-
test? I think one does not capture the ordinary understanding of race when
including appearance but leaving out ancestry and geographical origin, but
this is not a disagreement that draws a line between anti-realists and social
constructionists. The points against biological realism about race will, of
course, be affirmed rather than denied. So what about the alleged refutation
of social constructionism? As a social constructionist, I believe people are
raceless in the absence of any social facts dividing them into groups distin-
guished by appearance as related to ancestral place of origin. Am I thereby
disagreeing with Glasgow? No, because he is not saying this is a sentiment that
would be incoherent under any circumstances. His claim is that people who
conceive of race in this way do so in a manner that does not accord with the
ordinary understanding of the term. This is not something a social construc-
tionist should feel the need to protest. Why should we deny that it remains
common for people to understand race as something naturally inherited and
not dependent upon any social facts? To suggest that denying this is part of
social constructionism is to set up a straw man.
What social constructionists believe is that ordinary discourse about
race, insofar as it implies that race is biological and not social, is mistaken,
819
but it remains the case that ordinary discourse about race refers to something
real, namely, certain social distinctions. Thus social constructionists have no
reason to protest against what Glasgow wrongfully presents as a refutation of
their position, although we will obviously still want to disagree with his con-
clusion that race is not real. We should not rush to do so, however, because, as
it turns out, Glasgow’s position is that social constructionists are wrong about
whether race is real but everyone ought to become a social constructionist
and, if that happens, it will become the case that social constructionists are
right. As a result of this paradoxical stance, what appears to be a meaningful
difference turns out to be meaningless. This is perhaps best shown by drawing
a contrast with the disagreement we have with an anti-realist like Lawrence
Blum, who holds that races are not real but racialized groups are. Is there
a meaningful metaphysical disagreement here? As we have seen, Mallon
suggests there is not, but I think he is wrong. Why?
Blum and I have much agreement on what there is—socially distinct
groups generally referred to as ‘races’—and on what there is not—biologically
distinct groups taken to be real by many, perhaps most, who speak of ‘races’—
but we disagree on how to promote change among the general public with
regard to common ideas of what there is and what there is not. It is impor-
tant to Blum that we speak only of “racialized groups” when trying to speak
of what is real because we need to carefully avoid “the implication that the
groups being referred to are actual races (in the classic sense)—that they pos-
sess group-specific, biologically-based inherent behavioral and psychological
tendencies and characteristics.”15 It is helpful, from this point of view, that
“racialization refers to a process, largely imposed by others (but sometimes self-
generated also), that a group undergoes.”16 This is a powerful position worthy
of serious consideration. As stated in Chapter 2, however, I take the position
that it is unnecessary and possibly misleading to switch from talking about,
say, “the black race” to talking about “the group of people racialized as black”
as if those talking about the black race were or are describing something non-
existent as opposed to, as I would describe it, referring to a real group while
misunderstanding aspects of its nature. I view the difference between Blum
and myself here as meaningful (and, contrary to Mallon, I think it counts as
metaphysical).
15. Lawrence Blum, “Racialized Groups: The Sociohistorical Consensus,” The Monist 93 (April
2010): 300.
16. Ibid.
910
190 • W h a t I s  R a c e ?
What difference is there between what I recommend to the general public

and what Glasgow recommends? None, because he proposes a “semantic rev-
olution,” namely, the redefinition of “a race” as “a socially racialized group,”
which will make it possible in the future for race to be “as the constructionists
think it is now” (138). If the goal is for social constructionists to be right about
race in the future through promoting their position to the general public, the
only possible meaningful difference could be disagreement about whether
they are right in the present. But there is no meaningful disagreement here.
It is not as if there are groups whose present existence social constructionists
affirm but whose existence Glasgow denies. He accounts for the existence of
these groups using the same language as Blum: “there are racialized groups”
(129). It is not as if social constructionists are right about what we should
want the concept of race to be in future, as Glasgow says they are, but wrong
about what people ordinarily take it to be right now, because no social con-
structionist believes that social constructionism has already achieved the
status of common sense. In short, there is nothing for a social constructionist
to criticize but the pretense that this version of anti-realism is meaningfully
different from social constructionism. Rightly understood, Glasgow’s anti-
realism is not only not the enemy of social constructionism—it is a plea for a
concerted effort at singing its praise!
A final point I will make is that, for all I have said so far, Glasgow’s anti-
realism might be seen as equally compatible with both of the kinds of social
constructionism I described in Chapter 2. As a matter of fact, though, I think
his view supports cultural constructionism, in particular. Notice that his pri-
mary reason for encouraging the reconstruction of the concept of race involves
the importance of identities and communities. In his book, A Theory of Race,
he talks about how the embrace of racial identity can, first of all, be “a psy-
chologically and materially healthy response to living in a racist society,” and
this is certainly a point that political constructionists can accept.17 He goes on
to note, though, that there are some for whom “race would be valuable even
if racism were no more” and he invokes Lucius Outlaw, whom I mentioned
in Chapter 2 as perhaps the most prominent example of a cultural construc-
tionist.18 The harm that people like Outlaw (and myself, we can add) would
perceive in the elimination of racial identities is among the concerns that
Glasgow takes into account in arriving at his reconstructionist position.
17. Joshua Glasgow, A Theory of Race (New York: Routledge, 2009), 134.
18. Ibid.
19
Now consider this telling comment in Chapter 4: “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” (137). Glasgow believes, it seems,
that anti-realist talk of racialized groups will not fail us insofar as we wish to
attend to injustice, but the implication of such talk that racial categorizations
are imposed mistakes will prove unhelpful when it comes to preserving
our racial identities and communities as sources of value (and, once again,
he directs us to read Outlaw). It is not unreasonable, I think, to interpret
Glasgow as saying that political constructionism ultimately holds little, if any,
advantage over anti-realism of the kind that Blum offers, but that cultural
constructionism makes it clear that anti-realism of that kind is inadequate. If
I am right in claiming that this leads him to a give us a version of anti-realism
that is ultimately indistinguishable from social constructionism, then it seems
fair to me to say that Glasgow’s anti-realism collapses into not just social but
specifically cultural constructionism.
6.3. On Haslanger’s Political Constructionism

Naturally, there is much in common between Haslanger’s position and mine,
as we represent two forms of social constructionism. Beyond our obvious
agreement that racial difference is fundamentally social in nature, it is notable
that we both find it important to understand this claim as a matter of trying
to speak with more clarity and insight about the same phenomena that has
been spoken about in biologically essentialist terms, rather than as a matter of
changing the subject. Haslanger emphasizes this, first, by using the example
of Aristotle on slavery. Aristotle, we can say, was speaking about something
we would agree is real when he offered his account of what slaves are, but this
account was “badly mistaken” and a social explanation (“to be a slave is to be
owned by someone according to the laws or customs of one’s social milieu”)
represents “an improvement on the naturalistic account that defines slaves in
terms of their cognitive capacities” (22-23). This example is very instructive,
given the centrality that social constructionists tend to accord slavery in the
development of modern ideas of race. Haslanger also uses the examples of
Hippocrates’ investigations of cancer, other shifts in understanding medical
conditions, and changes in how people think about various divisions between
castes, classes, and ethnic groups, such as the move from theological accounts
of the power of monarchs to social accounts. Do all these conceptual shifts
count, in Glasgow’s eyes, as revolutions? Whether they do or not, to be a
912
192 • W h a t I s  R a c e ?
social constructionist about race, as Haslanger and I understand the position,

is to be actively engaged in promoting this type of shift with respect to race. If
this makes us revolutionaries and he is on our side, he is welcome.
Haslanger and I part ways, of course, on the role of culture in the expla-
nation of the social nature of race. Before dealing with this disagreement,
however, I should note that we are sometimes more or less on the same page
even where Haslanger takes us to be notably different. She suggests, in her
conclusion, that she differs from me no less than Spencer and Glasgow in
methodology, because she approaches the question “What is race?” taking
into consideration not merely semantic constraints from ordinary usage
but also epistemic and pragmatic standards of interpretation and, per-
haps most importantly, “normative considerations about what practices we
should continue and the best route for maintaining or discouraging them”
(34). She defends her political constructionism as appropriate, in part, be-
cause it is “morally and politically valuable,” for it helps us highlight “how
our racializing practices and identities contribute to injustice” (34). The
lesson she draws from Laura and François Schroeter with respect to our
interpretations of terms ideally helping us to see the point or rationale of our
practices reminds me of Ronald Dworkin’s interpretive model of adjudica-
tion, according to which judges must combine positive law with “principles
of political morality that taken together provide the best interpretation of
the positive law” because they give “the best justification available for the po-
litical decisions the positive law announces.”19 I find much to appreciate in
such ways of thinking about interpretation. When comparing my approach
to answering the “What is race?” question to Haslanger’s, I see little reason
to think that consciously adopting her method would make a noticeable dif-
ference to my deliberation and no reason at all to think it would make any
difference to my conclusions.
What is my method? I take Hardimon’s three theses representing the log-
ical core of race to be a relatively neutral starting point, a move which might
be thought of as recognition of a semantic constraint and which Haslanger
seems to agree with, since the contents of the three theses are evidently in-
corporated into her definition of a race (26). I then suggest that we can use-
fully compare different metaphysical stances on the nature and reality of race
by asking what significance they accord to our differences in appearance on
the basis of ancestral place of origin. My conclusion in the first section of
19. Ronald Dworkin, “Law’s Ambitions for Itself,” Virginia Law Review 71 (March 1985): 176.
913
Chapter 2 is that social constructionism best captures what significance these

differences can be reasonably said to have and, by the end of the chapter,
my conclusion is that cultural constructionism, in particular, is the most in-
sightful kind of social constructionism. To investigate whether and how much
I diverge methodologically from Haslanger, it must be noted that central to
my notion of capturing significance is the challenge of capturing moral and
political significance. When I use my South Asian British example to reveal
weakness first in non-essentialist biological realism and then in anti-realism,
I rely, admittedly implicitly, on moral and political imperatives, as I would say
the importance of recognizing this woman as non-white and of recognizing
her non-whiteness as socially significant stems most of all from the need to
do justice to her vulnerability to racism and to reasons she can and should
take pride in her South Asian ancestry in spite of this racism. Is this not a
case of normative considerations playing a role in conceptual clarification in a
manner similar to the way they do for Haslanger?
Normative considerations figure more explicitly in my criticisms of polit-
ical constructionism for downplaying the importance of culture in the pre-
sent and for leading us to conclude that race cannot survive the end of racism.
I even claim that the reason to name my position “cultural constructionism” in
the first place involves its normative implications. If I were to somehow make
the goal of a moral and political value in conceiving race more explicit and
central, it would only lead me to the same position. I share with Haslanger the
aim of revealing ways in which “our prior thinking is false or misguided,” but
I foreground, among those ways, the problem of the Eurocentric devaluations
of the histories and cultures of non-white peoples (33). I too hope “to dis-
rupt our ways of thinking, to motivate a new relationship to our practices,”
but, unlike Haslanger, I aim not to move us away from race toward “orga-
nizing ourselves on different terms” but rather to disrupt all ways of thinking
that allow those of one race to set themselves over and above or in hostile
competition with those of other races and to motivate relating to practices
of racial differentiation as ways of expressing and embodying diversity in all
its potential innocence, fecundity, and reciprocal gain (33). I fail to see how
there is methodological difference of great significance between Haslanger’s
approach and my own, despite our divergent conclusions.
Let us now clarify this divergence. Haslanger writes in Chapter 1 that “the
cultural account requires that races, as a group, share a culture” (25). This is
misleading when taken as a description of my view, for readers will recall that
I do not offer a maximally robust cultural constructionism, which would in-
deed take cultural difference to be strictly required for racial difference. On
914
194 • W h a t I s  R a c e ?
my view, diversity in ways of life is not essential to race, but neither are imbal-
anced relations of power. I can even envision a circumstance in which both
racial hierarchy and racial cultures have faded away but race lives on as a legal
distinction that is mainly of bureaucratic significance, engendering no ine-
quality between members of racial groups but also representing nothing of
great significance to the identities of members. My point is that race, on my
view, is fundamentally social and will live on as long as racial distinctions are
socially recognized in some form, just as it will die if they cease to be socially
recognized in any form, whether the form is political, cultural, or something
else. The reason my view can be identified as a kind of cultural constructionism
is because it takes culture to be fundamental from a normative standpoint, for
I hold that the value of cultural difference is the reason we may value race and
hope to see it live on indefinitely, rather than take its destruction to be our
goal, at least in the long run.
As Haslanger further evaluates the cultural account, as she calls it, she
attempts to be both conciliatory and critical. She sharply differentiates
ethnicities from races, defining the former as cultural groups and denying
that the latter can count as such because they are generally made up of people
from multiple cultures. On the other hand, she accepts that racial identities
do sometimes give rise to “pan-ethnicities,” which “emerge when multiple
groups are racialized and treated as one group, and form an identity and way
of life as a result” (28-29). She recognizes that pan-ethnicities can involve
people bonding together in “celebration and resistance” and that the “racially
identified artistic movements, cultural norms, and forms of association” that
can come about as a result may be valued even beyond the ways in which they
constitute responses to oppression (29). I take Haslanger to be trying here to
acknowledge the kinds of social phenomena I highlight while nevertheless
remaining critical of my position, because it remains the case, in her view, that
“pan-ethnicities are not races” (29).
I would agree with this, as Haslanger makes it clear that pan-ethnicities are
often the result of local processes of racialization and may thus be restricted in
scope to people in that locale. She gives Asian Americans as an example (and
perhaps it is my Canadian bias showing, but when she lists peoples of Asia
that would presumably fit into this pan-ethnicity, it once again strikes me as
so curious—and so clearly out of accord with the idea of distinctive appear-
ance as central to race—to think of Gujarati people from western India and
Japanese people as literally the same race). She allows that this scope might
widen in light of some people in Asia taking themselves to be Asian, but this
self-identification in response to the racialization of Asians in the United
915
States and elsewhere would not widen it so far as to include all who would
count as racially Asian. Pan-ethnicities therefore represent, like ethnicities,
subsets of races, at least as Haslanger conceives them. This reassures me that
I am not intending to talk about pan-ethnicities when I speak of races and,
without meaning to doubt the usefulness of Yen Le Espiritu’s work with the
concept, I do not currently see any role for talk of pan-ethnicity in my own
attempts to clarify race and culture.
The critical question is how I can make the claims I do for the funda-
mental importance of culture to race if I am not relying on a concept such as
pan-ethnicity. Haslanger’s idea that people in Asia may be included within
an Asian pan-ethnicity if they identify as Asian in response to racialization
in the United States and elsewhere seems to make sense as a model of the
social construction of a cultural group. If I am not talking about this sort of
connection to influential ideas and practices, then what is social, much less
cultural, about how people are included within racial groups on my account?
As useful a critical question as this is, the first thing to note is that it applies
just as much to Haslanger’s account of racial membership as it applies to my
account. She writes: “An individual ethnic Hmong living in China or Laos is,
I would maintain, Asianus, even if there is nothing distinctively ‘Asian’ about
Hmong culture, and she does not identify as Asian (and maybe has not even
heard of the designation)” (29-30). What is social about this? Why should
someone who has never heard of the category of Asians be classified as one if
race is a social rather than natural category?
Haslanger explains: “That she counts as Asianus is clear, however, by how
Hmong are viewed and treated within the United States . . . and how she
would be viewed and treated if she came here” (30). The existence of the
United States as a social context in which she would be recognized as Asian
is enough to secure the inclusion of this woman in the category. There is per-
haps something strange about practices in social settings you have never vis-
ited being sufficient to slot you into a social group, but I ultimately think
counterfactual considerations of this sort are appropriate. Knowing how
you stand in relation to others in the world involves not only what you have
consciously experienced, but also what you would be likely to experience if
you happened to interact with others in this or that way. I can make use of
counterfactual considerations as Haslanger does, although it is important for
me not to tie the existence of races to single locations like the United States.
I take race to be a global phenomenon, which does not mean I ignore the
way popular conceptions of which races there are can differ between different
places. If we are indeed talking about race and not the variety of the world’s
916
196 • W h a t I s  R a c e ?
ethnic differentiations, though, conceptions of race in different places must

necessarily overlap more than they diverge because they all represent broad
divisions of the world according to appearance as related to place of origin
and they are all shaped by the historical force of the modern West’s imperial
antagonisms. What divergences there are often have to do with the concep-
tual flexibility that comes with the mixing of races, although there are un-
doubtedly also divergences that stem from the variety of ways we can carve up
the spectrum of humanity.
I will admit that counterfactual considerations do less for me than for
Haslanger. She might grant me that we can share a story about how people
can be part of races without knowing it, but she will continue to ask how
culture can be upheld as centrally important to what races are when, unlike
ethnic groups, races do not “share a form of life” (28). But what does it mean
to share a form of life? Must we share a language? Must we share a religion?
Must we share a very particular bounded set of customs governing all the
major facets of public and private life? Haslanger does not tell us what would
have to be missing for us to not share a form of life. She treats ethnic groups as
sharing forms of life, but would she say we cannot speak of countries sharing
forms of life if they are ethnically heterogenous? What about regions made up
of multiple ethnically heterogenous countries? I think we can meaningfully
speak, for example, not only of Fanti (and, more broadly, Akan) culture, but
also of Ghanaian culture and of West African culture. I would not stop there,
though. I think we can speak meaningfully of Western culture, even though
this category covers an amazing variety of ethnic, national, and regional iden-
tities. Races, understood as cultural groups, have this kind of broad scope
(and, of course, Western culture is inextricably linked although not reduc-
ible to the racial category of whiteness). This broadness does not render them
formless. Their forms of life derive from historical developments involving
encounters between previously isolated groups, the circulation of people and
ideas, and modes of interaction made possible by various technologies, all de-
veloping within the context of various attempts at establishing, protecting, or
resisting white supremacy.
Jemima Pierre’s historical and ethnographic work in her book, The
Predicament of Blackness: Postcolonial Ghana and the Politics of Race, nicely
exemplifies important aspects of the approach to thinking about race I am
defending. Opposing the treatment of articulations of race as “exported
products of the contemporary West (and the United States in particular),” she
explores how Ghana’s past and present involve various forms of racialization
and racial projects as a result of and in response to the history of colonialism
917
and present-day realities.20 Referencing the work of Charles Mills, she

announces that she takes “as a point of departure the interdependence of cul-
tural, political, and economic terrains in a modern world constructed in and
through the key distinctions around race and the apparatuses of global white
supremacy.”21 When analyzing the racialization of Africans during the colo-
nial period as “native,” she exposes the way that policies of indirect rule in
a place like the Gold Coast (as it was then called) emphasized and institu-
tionalized ethnic differences in consequential ways, even as urban spaces were
starkly segregated in ways that parallel the more widely acknowledged racial
divisions of a settler colony like South Africa. Moving to the contemporary
period, as she locates the persistence of a privileged position for whiteness in
contemporary Ghana, examines the phenomenon of skin bleaching, considers
the ambiguities of slavery-related tourism, and discusses the perception of
some Ghanaians that those whose ancestors were enslaved in the Americas
are lucky, we have ample opportunity to consider how the globality of race is
in many ways a matter of the globality of that which political constructionists
take to be fundamental: racial hierarchy.
Something like skin bleaching, however, constitutes not merely evidence
of race as hierarchy but also the cultural connection between people in Ghana
and black populations elsewhere, such as Jamaica, where skin bleaching is also
a major issue. Not all cultural connections are healthy, and the widespread
practice of glorifying lighter skin is an aspect of black culture that ought not
be preserved and maintained. Pierre’s chapter on skin bleaching ends, how-
ever, with discussion of a poster found in a girls’ bathroom in a high school.
Issued by Ghana’s Food and Drug Board, it features a smiling dark-skinned
woman and the message: “BLACK IS BEAUTIFUL. DO NOT DESTROY
YOUR BEAUTY THROUGH BLEACHING.”22 This is black cultural re-
sistance and pride, expressed in a way that has resonated throughout the black
world, especially its English-speaking part, since the 1960s.
Black cultural enjoyment is, however, not just a matter of consciously
resisting that which denigrates blackness. Among Pierre’s most interesting
ethnographic accounts is her discussion of First Fridays Accra. First Fridays is
20. Jemima Pierre, The Predicament of Blackness: Postcolonial Ghana and the Politics of Race
(Chicago: University of Chicago Press, 2013), xi.
21. Ibid., xv.
22. Ibid., 121. Although Pierre does not discuss it, it is worth noting that the woman on
the poster sports straight hair, either because she is wearing a wig or has used some hair-
straightening process.
918
198 • W h a t I s  R a c e ?
a kind of networking event aimed at black professionals that was first devel-
oped in the United States in the late 1980s. When Pierre describes the event
as she experienced it in Ghana’s capital, Accra, what emerges is more than
just the convergence of an African American institution with a Ghanaian
location:
By the end of the evening, I have met diverse groups of Black

people: Ghanaians native to Accra; those who immigrated from other
parts of the country; Nigerians or Liberians who now live in Accra;
Black South Africans visiting Ghana for the first time; Haitians and
Haitian Americans who have moved permanently to Accra; second-
generation Ghanaian Americans who are either visiting or have repatri-
ated to their parents’ home; Jamaican American and African American
graduate students conducting research; and African American
professionals who have either decided to set up businesses in Accra,
are working for companies stationed in Accra, or are on short-term
volunteer trips. Clearly, this First Fridays Accra event is where people
from all over Africa and the African diaspora converge in the making
of a modern Black cosmopolitanism in Ghana.23
This black cosmopolitanism, as Pierre calls it, demonstrates how cultural

blackness is not exhausted by its relationship with white supremacy. It is
true and important that First Fridays originated in large part as a response to
the isolation felt by African Americans in white-dominated professions, but
something like First Fridays Accra tells us not only about the role of black cul-
tural institutions in providing a refuge from white domination but, perhaps
more so, about the way black cultural institutions can serve as testaments to
and celebrations of the diversity of black people without this diversity thereby
seeming incompatible with unity.
In light of the themes of celebrating racial identity and cultivating racial
unity just explored, it is worthwhile to consider at this point an important
possible objection to my view that Haslanger hints at but does not press. She
talks about how thinking of white people as a pan-ethnicity alerts us to ways
in which being white offers “creative opportunities” but also “resources for
domination—or even more often, escape from subordination” (29). What
does it mean to think of races as cultures and of cultures as worth celebrating
when we turn our focus to white people? Is it not a clear consequence of my
23. Ibid., 178–179. Pierre herself was born in Haiti and raised in the United States.
91
view that white people should cherish white culture? Does this not mean that
I encourage the promotion of “white pride,” thereby providing surprising but
convenient philosophical support to those who are dedicated to the principle
of white supremacy and who seek to deny that whiteness currently holds an
unfairly privileged position? This concern is, of course, an especially pressing
one in the wake of the rise of the so-called alt-right.
I take it to be obvious that my view ought to be rejected if it gives non-
accidental support to white supremacist calls for white pride, that is, if taking
it to support such calls involves logical implications of the view rather than
misreadings based on mere appearances. I also admit that there is at least the
appearance of support, because my view can indeed make sense of the notion
of white pride as something worth having. While there can be no doubt that
white racial identity occupies a unique position in the racial landscape rela-
tive to others, I do not hold that the end of racism requires the end of white-
ness, just as I do not hold that it requires the end of any other racial category.
While whiteness could fade away, my view is compatible with and can explain
people holding on to whiteness as a cultural identity after the end of racism.
This is why white cultural pride could be permissible on my view, although
the very idea of a post-racist future makes it clear that the possible persistence
of white cultural identity that I countenance is necessarily divorced from the
widespread treatment of whiteness as supreme.
There is a practical difficulty here. I can envision people in the future
taking pride in white European heritage in an uncomplicatedly justifiable
manner because I imagine as a prior condition for a post-racist future signifi-
cant amounts of white people actively working and collaborating with others
over a long period time in order to destroy white supremacy in all its forms.
Thus this future white pride would be partly inspired by positive feelings
about the process of eradicating the kinds of institutions and sentiments that
currently inspire calls for white pride. This is importantly different from the
way that orienting ourselves toward a post-racist future validates rather than
negates current calls for pride in non-white racial identities. How then should
anti-racist white people relate to white cultural identity in the present? For
example, might it not be prudent to avoid championing pride in whiteness in
the present given how easy it would be for one to be taken as championing
white supremacy? I do think that at least the specific term “white pride” is too
deeply associated with racism at present to be used as a rallying cry, although
I also nevertheless think that some forms of pride in European heritage must
be permissible and even healthy in the present if, as I believe, it will be pos-
sible to affirm the value of this identity in a racially egalitarian future.
02
200 • W h a t I s  R a c e ?
Fortunately, there has been something of a wave of work in philosophy

addressing the practical difficulties of how white identity may be positively
affirmed by anti-racist white people. Most prominently, Linda Martín
Alcoff ’s The Future of Whiteness argues for “the need for whites to have a
place in the rainbow, as whites.”24 Alcoff ’s central point in the book is that
whiteness as a social identity must not be seen as fixed and unable to change,
but rather as capable of evolving in valuable ways. Shannon Sullivan’s Good
White People: The Problem with Middle-Class White Anti-R acism advocates
white self-love, but not in the kind of uncritical and reactionary way that
white supremacists do: “A white person’s loving herself as a white person
means her critically caring enough about the effects whiteness has in the
world to work to make it something different and better than what it is
today.”25 We can also look to Outlaw, whose essay, “Rehabilitate Racial
Whiteness?”—a question he answers in the affirmative—precedes these
books by about a decade.26
Still, Haslanger worries about the ethical implications of holding on to
racial identity as a form of cultural identity even if racism is defeated. She
claims, with some justification, that I focus on “the benefits of racial cultural
unity, but not the costs of racial segregation” (31). The costs she has in mind
include unfair pressure on those within the race to conform to cultural ex-
pectations and unfair limitations on those outside the race with regard to
their ability to participate in cultural practices. Her concerns here are un-
derstandable, but it should first be made clear that they are not concerns
about things I actually endorse. Especially in a post-racist future in which
there is no need to find ways to support forms of cultural resistance in the
face of white supremacy’s cultural dimension, it is not my view that the good
of the survival of distinctive racial cultures generates a duty upon each in-
dividual member of each race to consciously aim to secure said survival.
Even in the present, it is important to me that those who are committed to
cultural resistance recognize that others may be equally committed to that
goal even when they diverge widely with regard to what they take to be cen-
tral to the culture, what they believe counts or does not count as part of
24. Linda Martín Alcoff, The Future of Whiteness (Cambridge, UK: Polity, 2015), 188.
25. Shannon Sullivan, Good White People: The Problem with Middle-Class White Anti-R acism
(Albany: State University of New York Press, 2014), 10.
26. Lucius Outlaw, Jr., “Rehabilitate Racial Whiteness?” in George Yancy (ed.), What White
Looks Like: African-American Philosophers on the Whiteness Question (New York: Routledge,
2004), 159–171.
210
it, and so on.27 Also, despite the deep concern that many who seek to fight
racism have about cultural appropriation, nothing I have said implies that it
is wrong to be influenced by or participate in the cultural practices of those
of other races.
One might worry, however, that whether or not I intend to endorse such
restrictions on individuality and cultural exchange, the preservation of racial
cultures that I promote makes these problems inevitable or at least very likely.
One might even worry that, while the problems just mentioned could argu-
ably arise in a world characterized by equality and mutual respect between
races, it is hopeless to imagine the preservation of racial cultures without the
persistence or re-emergence of negative stereotyping, chauvinism, maneu-
vering to achieve a higher position, and other destructive modes of thinking
and interacting. In response, I would say that I reject the idea that any of these
problems are truly inevitable—I cannot see why this would be the case—but
I cannot deny that they would be dangers to watch for, temptations to be
resisted, perhaps challenges to be overcome. They are not, however, problems
unique to racial difference, but challenges related to group difference in gen
eral. We cannot avoid facing them.
Thinking of how a critic might respond to this, I would reject the idea
that humans should seek to rid themselves of all forms of group difference
as too implausible and clearly undesirable for serious consideration (think,
for example, of the implication concerning linguistic difference), but a more
plausible claim might be that we should seek to minimize how many forms of
group difference we have and that something with the track record of badness
that racial difference has definitely ought to be abandoned. I disagree. I think
we recognize the value of our shared humanity best when we treat every form
of group difference that is meaningful to at least some of its members and
which could possibly be benign as an opportunity to prove our ability to ben-
efit from rather than be torn apart by group difference. Whenever we shrink
from taking on the challenges of group difference, opting not to promote
equality in diversity and unity without uniformity, but rather treating same-
ness as the only guarantee of peace, we demonstrate a demoralizing lack of
faith in ourselves and a fear of complexity that is bound to inhibit our prog-
ress as moral and social beings. Racial identities are meaningful to many of us
and are not reducible to positions in a hierarchy. Their preservation should
be—and, I believe, is—possible.
27. For more on this, see my “The Ethics and Politics of Cultural Preservation,” The Journal of
Value Inquiry 49 (March 2015): 205–220.
20
202 • W h a t I s  R a c e ?
As a final point, Haslanger is not mistaken in thinking that races as cul-

tural groups, on my understanding, are not “porous” in the sense that you can
be of any ancestry whatsoever and just make the choice to join (31). Once
it is clear that my view does not imply the wrongness of individuals being
influenced by and participating in the cultures of others, I do not see this as a
problem. Freedom to learn from and engage in many of the cultural practices
of others is not only compatible with my view, but essential to its claim that
cultural diversity ought to be valued. Freedom to be considered a full-fledged
member of any cultural group if one so chooses is a kind of luxurious freedom
the desire for which I do not really understand. Cultural constructionism as
I conceive it does not offer this latter kind of freedom, but I think what it
offers instead is more attractive.
Take, for example, the difference between political constructionism and
cultural constructionism regarding how we might think in a post-racist future
about mixed-race families. On a political constructionist view, what would be
considered a mixed-race family in the present will not only be unremarkable
but completely indistinguishable from other families in a post-racist future
because racial difference will no longer exist. Cultural constructionism, as
I conceive it, holds that, in a post-racist future, such a family will be distinct
but in a way that is worth appreciating, for it will exemplify within itself di-
versity, which is valuable, and, at that point, it will also symbolize the way that
racial difference, while still real, is no longer a source of antagonism. This sym-
bolic power is dependent on the continued social relevance of distinctions
of appearance as related to ancestral place of origin, and it is true that this
means that valuing as much mixing as possible would have the negative ef-
fect of leading directly to the end of racial difference. It is no part of my view,
however, that mixing as such should be vilified or forbidden and, again, even
beyond the importance of freedom of choice in family construction, it is pos-
sible to value mixed families on a cultural constructionist view in a way that
political constructionism makes impossible by pointing us toward making
mixing itself impossible.
230
7 S P E N C E R ’ S R E P LY T O G L A S G O W,
HASLANGER, AND JEFFERS
7.1. Introduction
So far you’ve read four philosophical views on race. You’ve also read
Haslanger’s reply to Glasgow, Jeffers, and myself. And you’ve read
Jeffers’s reply to Glasgow, Haslanger, and myself. In this chapter,
I will provide my reply to Glasgow, Haslanger, and Jeffers. I will also
clarify the race theory I presented in Chapter 3. The outline of this
chapter is as follows. First, I’ll clarify the race theory I presented in
Chapter 3. Second, I’ll address one major concern that two of my
coauthors have about my race theory. Third, I’ll advance an objec-
tion that applies to Glasgow’s, Haslanger’s, and Jeffers’s race theo-
ries. Fourth, I’ll advance an objection aimed only at Glasgow’s and
Jeffers’ race theories. Finally, I’ll provide conclusive remarks.
7.2. A Few Clarifications

In Chapter 3, I used the pressing situation of apparent racial
disparities in genetic disorders—such as the higher incidence of
fetal aneuploidy among Asian mothers—to motivate a metaphys-
ical investigation of what race is and whether it’s real given how
‘race’ is used to classify people in current and ordinary American
English.1 What I discovered was that there is a way that American
English speakers talk about race such that race is a biological and
biologically real thing. I discovered that there’s a widely and fre-
quently used race talk in American English that was started in 1997
1. Like in Chapter 3, I will refrain from using the modifiers “to classify people,”
“current,” and “ordinary” from now on for convenience. However, note that I’m
always presupposing these modifiers in this chapter when I talk about ‘race’ and
American race talk.
024
204 • W h a t I s  R a c e ?
by the US Office of Management and Budget (OMB) that uses a biological

thing that’s also biologically real as its meaning of ‘race’—where meaning was
understood referentially.
From further investigation, I discovered that race in OMB race talk (as
I call it) is nothing more than a specific set of five biological populations in the
human species that I dubbed the human continental populations. The human
continental populations are the following groups, and their OMB aliases are
in parentheses: Africans (Blacks), East Asians (Asians), Eurasians (Whites),
Native Americans (American Indians), and Oceanians (Pacific Islanders).2 In
other words, I argued that what ‘race’ means in OMB race talk is just the set of
human continental populations, and what each race term means in OMB race
talk is just the biological population it picks out. For instance, according to
OMB race theory, what ‘Black’ means is the African population. Also, since
it’s sufficient to have genomic ancestry from the African population in order
to be a member of that population, it follows from OMB race theory (and
substitution) that it’s sufficient to have genomic ancestry from the African
population in order to be Black in the OMB’s use of ‘Black.’3
In Chapter 3, I didn’t delve into any deep metaphysics about the essences of
the human continental populations. However, following the population genet-
icist Marcus Feldman (2010, 151), I did say that each one is an “ancestry group”
(Spencer, Chapter 3, 92). I also said that each population is capable of having
partial members, or, to put it more precisely, each human continental popula-
tion consists of a fuzzy set of people at any given time.4 At this point, I should
2. Just like in Chapter 3, I’ll use ‘Caucasian’ interchangeably with ‘Eurasian.’ This is common in
human genetics, but more common in medical genetics than population genetics.
3. However, I didn’t argue that having genomic ancestry from a human continental popula-
tion is necessary in order to be a member of that population. This is because that member-
ship condition doesn’t work for the first members of a human continental population. For
instance, how can the first people in the Native American population possess genomic ancestry
from the Native American population if that population didn’t exist before those people? For
how I think membership conditions work for all members of a human continental population
and other biological populations like them, see Spencer (2016). Nevertheless, I do think that
possessing genomic ancestry from a human continental population is a necessary and sufficient
membership condition right now and since the last time that each human continental popula-
tion was entirely composed of unmixed members, which some geneticists put at 1492, but the
date could have been in the more distant past (Ramachandran et al. 2010, 606).
4. Suppose crisp sets are the objects called ‘sets’ in Zermelo-Fraenkel set theory. Let an object
space X be a crisp set of objects. Suppose a membership function µ is a function such that
µ
X → [0,1]. Then, a fuzzy set A� is a pair ( X  , µ  ) . Unlike crisp sets, ‘∈’ has no meaning for
A A
fuzzy sets. The analogous relation is belonging. Suppose µ A ( x ) is x’s grade of membership (or
205
Spencer’s Reply • 205
clarify that I intended those claims to be claims about the essential properties
of human continental populations.5 However, while I said that each group of
people that currently belongs to a human continental population is “geograph-
ically clustered” (Spencer, Chapter 3, 99), I didn’t say that each one, necessarily,
has a distinctive geographic origin. Furthermore, I’m still thinking about what
the essential geographic properties of human continental populations are.6
So, my view is that each human continental population (and thus each
OMB race), by its very essence, originated as a fuzzy set of people some-
where in the world. For example, Blacks originated in sub-Saharan Africa,
and Asians originated in East Asia. Furthermore, at every point in time after
each human continental population originated, each population, by its very
essence, consisted entirely of people who formed a fuzzy set, and the whole
population itself (which exists through time) formed an ancestry group based
on common genomic ancestry. Given this view of what an OMB race is, it
should be obvious that I believe in racial essences, but not the kind of essences
that involves intrinsic properties and other tenets of racialism, which has been
widely refuted by philosophers of race as untenable for any folk race that
exists (Mallon 2006, 528–529).7 Rather, I believe that the essential properties
of OMB races are all relational and extrinsic.
strength) in A� . Then, x belongs to A� just in case x ∈ X A and µ A ( x ) > 0. These definitions

are from Zadeh (1965). Also, notice that while these definitions imply that objects can partially
belong to fuzzy sets, they don’t imply that any object that belongs to a fuzzy set must partially
belong to that set. Rather, these definitions allow for objects to wholly belong to fuzzy sets
as well.
5. Interestingly, a philosopher of social science, Brian Epstein, thinks about (human) social
groups in a similar way that I think about OMB races. In a fascinating book, Epstein (2015,
149) argues that social groups are, essentially, things that coincide with a “set of people” at any
given time. Since I’m arguing that OMB races are, essentially, biological groups that coincide
with a fuzzy set of people at any given time, there’s an interesting parallel in our independently
developed views which may, I think, reveal a deep connection between human social groups
and human biological groups.
6. Nevertheless, see Hardimon (2017, 27–31, 93) for a philosopher who thinks that the human
continental populations and the OMB races do, necessarily, have a “distinctive” geographic
location of origin. However, one reason why I’m hesitant to make this claim is that it’s not
clear that every human continental population has a distinctive geographic location of origin.
For example, at this point, it’s still unsettled whether the Oceanian population originated in
Southeast Asia and moved to Oceania or originated in Oceania (McEvoy et al. 2010, 303). If
the former occurred, then East Asia would not be the distinctive geographic location where
East Asians originated. Rather, it would be the shared location where both East Asians and
Oceanians originated.
7. For review, racialism is the view that humans naturally divide into a small number of groups
called ‘races’ in such a way that the members each race share certain fundamental, inheritable,
026
206 • W h a t I s  R a c e ?
For clarity, by a relational property I mean a property that must hold be-
tween two or more things when it’s instantiated, such as having a certain
weight. For example, when ordinary English speakers say that a person weighs
110 pounds on Earth, what we actually mean is that this person is gravitation-
ally accelerating toward Earth with a specific magnitude. On the other hand,
when a physicist says that a person has a rest mass of 49.9 kilograms, she’s
attributing a non-relational property to that person.8
By an extrinsic property of a thing I mean a property of that thing that
doesn’t arise entirely from that thing itself. It contrasts with an intrinsic
property, which arises entirely from the thing itself.9 For example, having
(naturally) blond hair is an extrinsic property that some people have. It’s ex-
trinsic because blond is a color, and all colors arise from interactions between
observers, surfaces, light, the light medium, and a few other factors (Hatfield
2003). However, having negative electric charge is an intrinsic property that
some fundamental particles possess, such as electrons, muons, and strange
quarks.
Another fact I discovered in Chapter 3 was that the set of human conti-
nental populations is biologically real, at least in the sense of being an “epistemi-
cally useful and justified entity in a well-ordered research program in biology”
(Spencer, Chapter 3, 95). In particular, I showed that the set of human conti-
nental populations is useful as a human population subdivision in population
genetics (a well-ordered research program in biology) to explain why humans
subdivide into five genetic clusters that correspond to five major geographic
regions whose boundaries (e.g., the Sahara, the Himalayas, oceans, etc.) are
significant obstacles to human interbreeding (Spencer, Chapter 3, 99). Also,
the set of human continental populations is justified for this use because the
aforementioned five genetic clusters have been reproduced in genetic clus-
tering studies that have used a worldwide sample of human ethnic groups, and
in the largest and most comprehensive human genetic clustering study to date
(Spencer, Chapter 3, 98-99).
Next, I should clarify that the evidence for OMB race theory (as I call
it) is entirely abductive. In other words, I take OMB race theory to be the
physical, moral, intellectual, and cultural characteristics with one another that they do not
share with members of any other race (Appiah 1996, 54).
8. Notice that I’m talking about rest mass, which is an invariant quantity of bodies in physics.
The mass that ordinary people measure is known as relativistic mass in physics, which is another
relational property.
9. These distinctions are from Weatherson and Marshall (2017).
027
theory that best explains both what the OMB means by ‘race’ and various
phenomena in population genetics. One rival I considered was Michael
Hardimon’s theory of what ‘race’ means in OMB race talk, which is what he
calls ‘the ordinary concept of race.’ This race concept, among other things,
bars races from being “visually indistinguishable” (Hardimon 2003, 442).10
One way in which I said OMB race theory was “best” was insofar as it
solves more puzzles than any of its rivals about what race is and what races
are in OMB race talk. Some examples are being able to define ‘Black’ using
ancestry and without making every living person Black (Spencer, Chapter 3,
100-101), and being able to racially classify unmixed Aboriginal Australians
(Spencer, Chapter 3, 100). Another way in which I said OMB race theory
was “best” was insofar as it’s more predictively powerful than any of its rivals.
For example, I discussed how geneticists can predict with 98.8%–99.8% accu-
racy the OMB race self-reports of US adults (when they report a single race)
using one’s primary human continental population membership (Spencer,
Chapter 3, 102). But also, I used OMB race theory to predict the truth-values
of relevant modal propositions with greater accuracy than Hardimon’s ordi-
nary concept of race and other similar rivals. Some examples are: “It is pos-
sible for there to be two visibly indistinguishable races” and “Pacific Islanders
could be visibly indistinguishable from Blacks” (Spencer, Chapter 3, 93).
Now, I’ll offer a further clarification about the membership conditions for
belonging to an OMB race according to OMB race theory since I know
OMB race theory is probably counterintuitive.
To be clear, a person with any genomic ancestry from a human conti-
nental population is a member of that race. So, for example, a person’s parents
need not be entirely Caucasian and entirely African in order to be partially
White and partially Black. Also, OMB race theory, together with some well-
confirmed genomic and demographic data, entails that millions of people in
the United States are racially mixed according to the OMB’s racial scheme.
For instance, there are 42 million African Americans in the United States,
and they’re, on average, 73.2% African, 24% Caucasian, and 0.8%* Native
American.11 Also, there are 32 million Mexican Americans in the United
10. Also, remember that in Chapter 3, I mentioned that since visible distinguishability is part
of the US race theories developed by Naomi Zack (2002), Lawrence Blum (2002), Paul Taylor
(2013), Joshua Glasgow (2009; this volume), and, I should include, Chike Jeffers (this volume),
these are all also rivals to OMB race theory for explaining what ‘race’ means in OMB race talk.
11. Here and for the rest of the chapter, I’ll use asterisk superscripts to flag statistics that are not
statistically significant (at 95% confidence) according to the authors’ assessment of survey or
sampling error. Also, I’ll use dagger superscripts to flag statistics with no error bounds provided
208
208 • W h a t I s  R a c e ?
States, and they’re, on average, 48% Native American, 47% Caucasian, and
4%† African.12 Another racially mixed group of Americans are South Asian
Americans. There are about 3.8 million South Asians in the United States, and
they’re, on average, 68.4% Caucasian, 26.5% East Asian, and 5.1%† African.13
Here are two more examples. There are 1.4 million Dominican Americans in
the United States, and, on average, they’re 50% Caucasian, 43% African, and
6%† Native American.14 Finally, our good friends the Native Hawaiians, who
total 180,000 people in the United States, are, on average, 56% Oceanian,
30% East Asian, and 14% Caucasian.15
With that said, some ethnic groups of Americans are, on average, very
nearly racially unmixed, and so there are almost certainly millions of racially
unmixed people in the United States as well. Most racially unmixed people
in the United States are likely to be non-Hispanic European Americans.
There are currently 195 million non-Hispanic European Americans in the
United States, and, on average, they’re 98.6% Caucasian, 0.19%* African, and
0.18%* Native American.16 There are also at least 12.8 million Eastern Asian
Americans in the United States, and they’re, on average, 95.5% East Asian,
from the authors and that are low enough that one should be skeptical as to whether they’re
statistically significant (at 95% confidence). These demographic data are from Rastogi et al.
(2011, 6), and the ancestry data are from Bryc et al. (2015, 42).
12. These demographic data are from Ennis et al. (2011, 14), and the ancestry data are from
Manichaikul et al. (2012, 4). It’s also worth clarifying that the 95% confidence intervals for
most of these statistics are so wide that sometimes geneticists don’t know which genomic an-
cestry is the primary one. For example, Risch et al. (2009, 4) obtained statistically indistin-
guishable genomic ancestry results from Manichaikul et al. (2012, 4) for Mexican Americans.
Furthermore, Risch et al.’s 95% confidence intervals for Mexican Americans’ Caucasian and
Native American ancestries, respectively, were (49.4%, 64.9%) and (50.5%, 65.8%). So, we re-
ally don’t know whether Mexican Americans are, on average, primarily Caucasian or primarily
Native American.
13. These demographic data are from Hoeffel et al. (2012, 15), and the ancestry data are from
Guo et al. (2014, 155).
14. These demographic data are from Ennis et al. (2011, 14), and the ancestry data are from
Manichaikul et al. (2012, 4).
15. These demographic data are from Hixson et al. (2012, 14), and the ancestry data are from
Kim et al. (2012, 4).
16. The genomic ancestry facts cited are from Bryc et al. (2015, 42). Also, I estimated the cur-
rent number of non-Hispanic European Americans by subtracting current counts of US Arabs,
US Iranians, US Jews, and US Hispanic Whites from a current count of US Whites. The
counts for US Whites and US Hispanic Whites are from Hixson et al. (2011, 3). The count
of US Jews is from Tighe et al. (2013, 1). Finally, the counts for US Arabs and US Iranians are
from the “Total Ancestry Reported” file from the 2010 American Community Survey (1-year
estimates), which is available at http://factfinder.census.gov.
029
4.0%† Caucasian, and 0.5%† African.17 Another group of Americans with low
racial mixture is American Jews. There are currently 6.8 million American
Jews in the United States, and, on average, they’re 96% Caucasian and 4%
African.18
What’s interesting here is that even though millions of Americans are ra-
cially mixed, the majority of Americans are probably racially unmixed due
to demographic and genealogical facts. For one, the overwhelming majority
of Americans (63.1%) are non-Hispanic European Americans.19 Second, as
I mentioned previously, non-Hispanic European Americans are, on average,
98.6% ( ±1.4% ) Caucasian. These two facts alone provide strong evidence
that most Americans are racially unmixed, which, if true, undermines the
myth that the US is a “melting pot” as opposed to a “mixed salad.”20
At this point, I should say that even if my theory of OMB racial mem-
bership seems strange, it’s the right result. Remember, the OMB (1997,
58789) embraces the existence of people with “more than one race” and
starts off attempting to define each of its race terms with the phrase “A person
having origins in . . .” Although OMB race theory tweaks the OMB’s focus on
“origins” to a focus on genomic ancestry, the OMB’s idea that people can have
lots of racial memberships and to differing degrees as a function of one’s an-
cestry is maintained in OMB race theory. So, this result may seem strange if,
for example, one thinks the OMB’s racial scheme is supposed to be capturing
Americans’ racial self-identifications instead of their racial memberships in an
ancestry-based racial scheme. However, while the OMB (1997, 58782) wants
their racial scheme to receive “broad public acceptance” for practical reasons
(since they want to use self-reporting as the default method of collecting ra-
cial membership data), they’re simply not in the business of trying to pin
down Americans’ racial self-identifications. This is for two reasons.
17. These demographic data are from Hoeffel et al. (2012, 15), and the ancestry data are from
Guo et al. (2014, 155). Also, by Eastern Asians I mean people with Northeast Asian descent
(e.g., Chinese, Japanese, Taiwanese, etc.) or Southeast Asian descent (e.g. Cambodians,
Filipinos, Thai, etc.).
18. These demographic data are from Tighe et al. (2013, 1), and the ancestry data are from
Moorjani et al. (2011, 6).
19. I calculated this percentage from dividing my count of non-Hispanic European Americans
by the US Census Bureau’s 2010 count of US residents, which is close to 309 million. See
Hixson et al. (2011, 3).
20. These sayings are from a former student of mine from Canada who told me that she was
shocked to hear these statistics because she had been taught in Canadian schools that the US is
a “melting pot” and Canada is a “mixed salad.”
120
210 • W h a t I s  R a c e ?
First, the OMB was not able to find a consensus (or even a near consensus)
among Americans on what race is and which groups are races (OMB 1995).
Second, and as I discussed in Chapter 3, OMB race talk was invented to serve
the interests of the US government, not the American people’s racial self-
identification desires. In particular, it was invented to prevent communica-
tion breakdown when federal agencies share racial data and to enforce civil
rights legislation (OMB 1997).
One final clarification that I’ll make before moving on is clarifying how
I see myself as disagreeing with my coauthors. At this point, even if you are
convinced that OMB race theory is true, you may be wondering how on earth
OMB race theory contradicts any other race theory defended in this book. To
be specific, Sally Haslanger (Chapter 1, 25, 27) defends the “racialized group”
(a sociopolitical construct) as the “dominant” meaning of ‘race’ among
English speakers in the United States, and she argues that the racialized group
is real as well. Joshua Glasgow (Chapter 4, 117) argues that the most “fre-
quent” meaning of ‘race’ used in ordinary communications among compe-
tent English speakers in the United States is a “concept” that requires races to
be “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 dispro-
portionate extent,” and, furthermore, that because of this requirement, race
(in this sense) is a biological thing but “neither biologically nor socially real.”
Finally, Chike Jeffers (Chapter 2, 39) argues that races are real sociocultural
groups according to the meaning of “the English word ‘race’ and its etymo-
logically related cognates.”21
So, Haslanger, Glasgow, and Jeffers all believe that there’s a single domi-
nant meaning of ‘race’ among (at least) American English speakers, and they
focus their efforts on articulating what race is and whether race is real as-
suming that dominant meaning of ‘race.’22 Given these facts, one could rightly
worry that I am not disagreeing with any of my coauthors because I have not
21. While Jeffers emphasizes that races (for users of the English word ‘race’ and its cognates)
originated through “the global sociopolitical system of White supremacy,” races are not essen-
tially political groups according to Jeffers (like they are for Haslanger) because races according
to Jeffers can survive the end of racism by being merely cultural groups ( Jeffers, Chapter 2).
22. Of course, Haslanger and Jeffers claim they are capturing more than just the dominant
meaning of ‘race’ among American English speakers. They’re also both interested in other rel-
evant English speakers (e.g., British, Canadian, etc.) and other linguistic communities with
cognate terms (e.g., some European linguistic communities). However, for now, I’ll focus on
the common core among us, which is American English speakers.
21
argued that the OMB’s meaning of ‘race’ is the single dominant meaning of
‘race’ among American English speakers.
To this concern, I will say up front that OMB race theory does not con-
tradict any of my coauthors’ race theories.23 However, OMB race theory, to-
gether with a few empirical facts, implies a race theory that does contradict
every other race theory in this book because the theory has a radically pluralist
form. But before I sketch what that theory looks like, I will say a bit about
what I think the common ground is among Glasgow, Haslanger, Jeffers, and
myself with respect to what a dominant meaning of ‘race’ is.
First of all, a dominant meaning of ‘race’ need not be the only meaning of
‘race’ that a linguistic community uses. Furthermore, it would be imprudent
to require it to be so, since, given the empirical data we already have on how
American English speakers use ‘race,’ a debate about what race is and whether
it’s real given the only meaning of ‘race’ that American English speakers
use would be over the minute we started the debate! That’s because there’s
ample empirical evidence that American English speakers do not use a single
meaning of ‘race’ for all of their communications about race. Since I discussed
this research in section 3.3 of Chapter 3, I won’t repeat it here. However, I will
say that, in my opinion, the sociologist Ann Morning summarizes the empir-
ical data best when she says:
Perhaps the clearest theoretical proposition to emerge from this re-

search is that we cannot assume that individuals hold a single defini-
tion of race. Instead, they may carry around a “tool kit” (Swidler and
Arditi 1994) of race concepts from which to draw depending on their
reading of the situation to be deciphered. (Morning 2009, 1186)
Now, even though ‘race’ is undoubtedly polysemous in American English,

there still may be a single dominant meaning of ‘race’ among American
English speakers that Americans default to when engaging in race talk in
order to prevent communication breakdown. Here’s an analogy using lan-
guage to illustrate how this situation is possible.
While US residents who are five years old or older speak 381 distinct
languages, 95.2% of them speak English “very well” or “well” (Shin and
Kominski 2010, 1–2). Furthermore, the second most widely spoken language
23. However, in the past, I have defended the set of human continental populations as the
single dominant meaning of ‘race’ among American English speakers. See Spencer (2014).
21
212 • W h a t I s  R a c e ?
among US residents who are five years old or older is Spanish, but only 8.7%
of US residents who are five years old or older speak Spanish “very well” or
“well” (Shin and Kominski 2010, 2). These statistics show that while English
is not the only language that Americans speak, English is, plausibly, the
single dominant language spoken among Americans—where, roughly, that
means it’s the language that’s the widest and most frequently spoken among
Americans that’s spoken by at least a majority of Americans and has a com-
manding margin of victory compared to its closest rival. So, if ‘race’ use among
American English speakers is analogous to language use among US residents,
then it’s appropriate to say that there’s a single dominant use of ‘race’ among
American English speakers.
However, there’s another live possibility for how American English
speakers use ‘race,’ and this is that there are multiple, distinct, and dominant
meanings of ‘race’ in American English. This possibility can also be cap-
tured with another language use analogy. For example, while Republic of
China (Taiwan) residents speak multiple languages (e.g., Mandarin Chinese,
Taiwanese, Hakka, English, etc.), there are two that most Taiwan residents
speak: Mandarin and Taiwanese. Among Taiwan residents six years old or
older, 83.5% of them speak Mandarin and 81.9% of them speak Taiwanese.24
Also, no other language is spoken anywhere near as widely as Mandarin and
Taiwanese in Taiwan. The next closest fluently spoken language in Taiwan is
Hakka at 6.6% frequency. So, it’s not obvious at all that Taiwan has a single
dominant language. Rather, it’s more appropriate to say that Taiwan has two
dominant languages. One (Mandarin) is dominant in formal communica-
tions (e.g., government, school, etc.), whereas the other (Taiwanese) is dom-
inant in informal communications (e.g., speaking with friends and family).
The question now becomes whether ‘race’ use among American English
speakers is more like language use among Americans or language use among
the Taiwanese?
For reasons that I will provide later in this chapter, I believe that ‘race’
use among American English speakers is more like language use among the
Taiwanese than language use among Americans. In any case, the appropriate
24. I should clarify that these data are actually for Taiwan resident nationals, not all Taiwan
residents. However, I’ll drop the ‘national’ qualifier for ease of communication. Also, the data
actually report the frequency of languages spoken at home. But, of course, if you can speak a
language at home, you can speak that language. Finally, these data and all language use data
for Taiwan that I’m citing in this chapter are from the “General Statistical Analysis Report”
that Taiwan published online after analyzing the data from its 2010 Population and Housing
Census. The report is available at https://eng.stat.gov.tw.
213
way to view OMB race theory is not as a theory about what race is and
whether it’s real given that the OMB’s meaning of ‘race’ is the single dominant
meaning of ‘race’ in American English. Rather, the appropriate way to view
OMB race theory is as a part of a larger radically pluralist theory about what
race is and whether it’s real in the American English context. In other words,
I think that the OMB’s meaning of ‘race’ is one dominant meaning of ‘race’
among American English speakers, but not the only one, because there is no
such thing as the only one. I’ll wrap up this section by clarifying what I mean
by a radically pluralist race theory.
I’ll say that a pluralist race theory is one that presupposes there to be no
single correct answer to the question of what race is and whether race is real
in the relevant context, but that presupposes there to be at least one correct
answer to this question.25 For instance, given what Glasgow says at the be-
ginning of Chapter 4, it’s accurate to say that he’s defending a pluralist race
theory with respect to how American English speakers use ‘race.’ Of course,
that doesn’t stop Glasgow from defending a single dominantly correct answer
to the question of what race is and whether it’s real for American English
speakers.26
This brings me to radical pluralism. I’ll say that a radically pluralist race
theory goes further than a pluralist race theory by accepting pluralist tenets
and adding that there isn’t even a single dominantly correct answer to the
question of what race is and whether it’s real in the relevant context, but
there’s still at least one dominantly correct answer to this question. So, to be
crystal clear, my race theory for what race is and whether it’s real relative to
how ‘race’ is used in American English is radically pluralist, but one crucial
part of that theory is OMB race theory. Admittedly, this is a complex US race
theory, but at least now you see how I’m disagreeing with my coauthors.27
7.3. Addressing the South Asian Mismatch Objection

So far, all I’ve done is clarified OMB race theory, its evidence, and how it fits
into a larger US race theory that contradicts the race theories of my coauthors.
25. My characterization of a pluralist race theory is intended to be similar in form to Gilbert

Harman’s definition of ‘moral relativism.’ See Harman (2000).
26. Another example of a US race theory like Glasgow’s is Michael Hardimon’s. For evidence
of Hardimon’s embrace of pluralism, see Hardimon (2017, 173).
27. By a US race theory I mean a theory about the nature and reality of race according to how
‘race’ is dominantly used in American English.
214
214 • W h a t I s  R a c e ?
Now, I’ll do my best to reply to one major objection that two of my coauthors
offered me. I fully acknowledge that there’s plenty more objections against
OMB race theory that are worth a reply. However, due to space limitations,
I can only address one with care, and this objection is a good one.
A historically popular objection against any attempt to biologically vin-
dicate a folk racial classification is to claim that there’s a “mismatch” between
the groups called ‘races’ in the biological theory of race and the groups called
‘races’ in the folk discourse (Mallon 2006, 533). If the mismatch is serious
enough, it shows that the meaning of ‘race’ in the biological theory and the
meaning of ‘race’ in the folk discourse are different meanings. Both Jeffers and
Glasgow offer unique Mismatch Objections to any attempt to define ordi-
nary race terms in American or British English in a biological way.
In Chapter 2, Jeffers uses a thought experiment involving a dark-skinned
Bangladeshi English woman to show that “it conflicts with common sense”
to racially classify her with Whites ( Jeffers, Chapter 2, 44). In Chapter 4,
Glasgow brings up the concern that Hispanics and Arabs “find no home
in the flow charts of population genetics, but they are routinely racialized
in the United States” (Glasgow, Chapter 4, 122). Also, elsewhere, Glasgow
has raised doubts about South Asians being correctly racially classified with
Whites in other attempts to defend biological racial realism relative to a folk
meaning of ‘race.’28
Now, as you may have guessed, I don’t find the Arab or Hispanic Mismatch
Objection challenging. This is because OMB race theory only claims that the
OMB’s meaning of ‘race’ is biological, and the OMB is very clear that Arabs
are White and Hispanics are not a race. Also, this isn’t a hollow victory be-
cause, as I showed in Chapter 3, OMB race talk is an ordinary race talk in
American English. It’s just that this particular ordinary race talk operates by a
division of linguistic labor whereby the OMB fixes the meanings of ‘race’ and
race terms. So, even if there are contexts in current American life where Arabs
and Hispanics are treated as races by ordinary people, there are also contexts
in current American life where Arabs and Hispanics are not treated as races
by ordinary people, namely, when ordinary people use OMB race talk.
South Asians, on the other hand, are another story. Given everything that
the OMB has written about its racial scheme, it’s not clear at all that it’s correct
to say that South Asians are, on average, part White and part Asian instead
28. For example, see Glasgow (2009, 96). However, Glasgow hints at this objection again in
this book. See Glasgow (Chapter 4, 114).
251
of Asian alone, at least in the OMB’s racial scheme. Rather, the OMB always
talks about South Asians as paradigm Asian people. So, one challenging mis-
match concern is whether it’s the case that the OMB truly intends ‘Asian’ to
be synonymous with ‘East Asian’?
The seriousness of this concern is immediately obvious once one scrutinizes
a few facts about how the OMB uses the term ‘Asian.’ First, when the OMB
introduced ‘Asian’ as an OMB race term in 1997, it explicitly named “India”
and “Pakistan” as examples of countries that many Asian people live in (OMB
1997, 58786). Second, in the OMB’s attempt to define ‘Asian,’ the “Asian”
geographic range includes “the Indian subcontinent” (OMB 1997, 58786).
Furthermore, South Asian Americans tend to follow the OMB’s lead and ra-
cially self-report ‘Asian’ instead of ‘White’ in high numbers. However, you
wouldn’t know this from looking at what geneticists report.
For instance, Tang et al. (2005) didn’t sample any South Asians, and Guo
et al. (2014) sampled South Asians but separated them from the remaining
self-reported Asians before attempting to predict racial self-reports! In truth,
if one were to have added South Asians to the rest of the self-reported Asians
in Guo et al.’s sample, their accuracy for predicting self-reported Asians would
have fallen from 97.7% to 66.1% (Guo et al. 2014, 153). All of these facts to-
gether suggest that there’s a mismatch between what ‘Asian’ means in OMB
race talk and what ‘East Asian’ means in OMB race theory.
Suppose we call this the South Asian Mismatch Objection for ease of ref-
erence. For clarity, the South Asian Mismatch Objection states that what
‘Asian’ means in OMB race talk and what ‘East Asian’ means in OMB race
theory are different because the group of people that the OMB intends to
pick out with ‘Asian’ (Chinese, Japanese, Filipinos, Indians, Pakistanis, etc.) is
simply different from the group of people that count as East Asian (Chinese,
Japanese, Filipinos, etc.). While the South Asian Mismatch Objection is for-
midable, it’s not a fatal objection to OMB race theory.
First, I should clarify that, according to OMB race theory, almost all living
South Asians are East Asian. This is because South Asians are, on average, a
very racially mixed people. For example, remember that in the United States,
South Asian Americans are, on average, 26.5% East Asian. That’s not nothing!
So, the OMB is not wrong to racially classify almost all living South Asians
with Asians. Also, note that the OMB cannot, on pains of inconsistency, deny
that almost all living South Asians are White as well. Remember, the OMB
(1997, 58789) wants it to be the case that any person with “origins in any of
the original peoples of Europe, the Middle East, or North Africa” is White.
However, we know from the Out of Africa theory of human migration
126
216 • W h a t I s  R a c e ?
history that all South Asians descend from the original peoples of the Middle
East (Cavalli-Sforza and Feldman 2003, 270).
So far so good. However, we can see the real problem when we look at how
the OMB racially classifies unmixed South Asians. From everything the OMB
has written, it’s reasonable to think that the OMB would racially classify un-
mixed South Asians as Asian instead of White. And that’s the real challenge
of the South Asian Mismatch Objection. Here’s a concrete example to illus-
trate the challenge. From recent human genetic clustering studies, geneticists
have discovered that the Kalash people of Pakistan are a very unmixed group
of people. For example, Rosenberg et al. (2002) found that, on average, Kalash
Pakistanis are 99% Caucasian.29 So, there’s bound to be hundreds of unmixed
Kalash people out there whom the OMB would racially classify as Asian
alone, while OMB race theory would racially classify unmixed Kalash people
as Caucasian alone. That’s a mismatch! Furthermore, it’s hard to imagine how
the OMB could be wrong here. How can the OMB be wrong about who its
own race terms pick out in the world? If the OMB truly intended to pick out
all South Asians with ‘Asian,’ including unmixed South Asians, how can the
OMB be wrong about classifying unmixed South Asians with Asians?
While it may be counterintuitive, the OMB can be wrong about who its
own race terms pick out in the world. This is because, sometimes, when we fix
the referents of English terms, we get some of what that term picks out wrong.
This was always part of the referential theory of (English) name meanings.30
Also, there are plenty of examples of this happening in the English language.
Here’s one.
In botany, it’s widely acknowledged that the initial sample of the species
that botanists now call ‘watermelon,’ and, officially, ‘Citrullus lanatus,’ consists
of the plants in the Mediterranean that Carl Linnaeus named ‘Cucurbita
citrullus’ in 1753, and the plant in South Africa that Carl Thunberg, a stu-
dent of Linnaeus’s, named ‘Citrullus lanatus’ in 1773. Since Linnaeus left
no remnants of C. citrullus, but Thunberg preserved some of his C. lanatus
plant, in the 1930s, botanists placed the remnants of Thunberg’s plant in a
museum as a paradigm member of C. lanatus to settle all disputes about the
watermelon’s characteristic genomic and phenotypic properties. These spe-
cial paradigm members of species are known as type specimens in systematic
biology.
29. This statistic is from supplementary Table 2 in Rosenberg et al. (2002).

30. For example, see Kripke (1980, 136).
217
However, interestingly, in 2014, two botanists decided to run a phyloge-

netic analysis on the type specimen of C. lanatus. What they found was puz-
zling. They found that the watermelon’s type specimen is not closely related
to Egusi melon (Citrullus mucosospermus), even though all other phyloge-
netic analyses of C. lanatus plants place the Egusi melon as a sister species to
the watermelon.31 They also found that the watermelon’s type specimen is dis-
tantly related (a first cousin) to all other samples of watermelon they studied
(which were from Benin). It was a real head scratcher. The researchers could
have interpreted their results in two different ways.
First, the researchers could have concluded that C. lanatus is not a ge-
nealogically tidy species (or perhaps not a species at all!) since the term
‘C. lanatus’ must, at least, pick out its type specimen as a species member.
Alternatively, the researchers could have concluded that C. lanatus is a gene-
alogically tidy species, but it just so happens that the watermelon’s type spec-
imen isn’t a watermelon plant! Interestingly, the researchers concluded the
latter, not the former, and the community of botanists have not rejected this
conclusion, but rather, have embraced it.32
Since English-speaking botanists can reject a type specimen for a species
as a member of that species, it’s simply not true that the OMB cannot be
wrong about racially classifying unmixed South Asians with Asians instead
of Whites. The question now becomes, what evidence is there (beyond what’s
presented in Chapter 3) that the OMB really did intend to pick out the set of
human continental populations with ‘race’ as opposed to a different classifica-
tion of people that’s similar to the set of human continental populations, but
that classifies all South Asians with Asians.
This is a good concern, so here’s some additional evidence that the OMB
intended to pick out the set of human continental populations with ‘race’
in 1997. In 2013, the director of the US Census Bureau from 1998 to 2000,
Kenneth Prewitt, came out and said that the OMB’s 1997 racial classification
was a deliberate attempt to mimic “Blumenbach’s racial taxonomy” (Prewitt
2013, 17). In Prewitt’s (2013, 17) words, “An extraordinary thing happened
two hundred years after Blumenbach announced that the world’s population
31. In systematic biology, two species A and B are sister species iff each one immediately
descended from the same species. Thus sister species in systematic biology are analogous to
siblings in ordinary English.
32. These researchers were Guillaume Chomicki and Susanne Renner, and this story comes
from Chomicki and Renner (2015).
218
218 • W h a t I s  R a c e ?
should be divided into five race groups distinguished by skin color. The
United States government agreed.” Prewitt (2013, 18) even calls the OMB’s
races “Blumenbachian races.”
Furthermore, Prewitt should know what the OMB’s true intentions were
in 1997 because he worked closely with the OMB demographers who revised
the OMB’s race talk in 1997. It was Prewitt’s responsibility to figure out how
to best incorporate the OMB’s new racial scheme into the 2000 decennial
census. But let me back up a bit and talk about Blumenbach’s racial classifica-
tion and its link to the OMB’s racial scheme.
J. F. Blumenbach was an eighteenth-century physical anthropolo-
gist who, in 1795, published a new, fivefold way of classifying people
into races using mostly visible physical features of the face and body,
but also linguistic attributes. Blumenbach’s five races were Americans,
Caucasians, Ethiopians, Malays, and Mongolians. Table 7.1 is a summary
of Blumenbach’s description of each race and the groups of people he
thought belonged to each race.
What was truly original about Blumenbach’s racial classification was
its comprehensiveness. Somehow, Blumenbach managed to classify every
living human into a race, which no race scholar before him had accom-
plished. According to Prewitt, OMB demographers liked this feature of
Blumenbach’s racial scheme and decided to adopt a racial scheme very similar
to Blumenbach’s in order to be able to racially classify any potential US immi-
grant. In Prewitt’s words,
Blumenbach’s taxonomy was universal and totally inclusive. Every

living person on earth could be assigned to one of its five categories.
A consequence of employing a universal classification system in the
American census is that any new immigrant arriving from any corner
of the world will be put into this preexisting taxonomy—whether or
not he or she seems to fit. Recently arrived Ethiopians, for example,
are counted in the census as African American, though the former
are Nilotic in ways unfamiliar to the African Americans (Prewitt
2013, 18).33
33. In the preceding quote, Prewitt is, unfortunately, using ‘African American’ in two different
ways. In the first occurrence, ‘African American’ is a race term that’s synonymous with ‘Black.’
In the second occurrence, ‘African American’ names the largest ethnic group of Blacks in
the US.
219
Table 7.1 A Summary of Blumenbach’s Five Races
Race Description Members Textual Evidence
American Copper-colored skin, All indigenous people Blumenbach

black and stiff and of the Americas except (1795/2000, 29)
straight hair, broad Eskimos
faces, chiseled cheeks,
etc.
Caucasian White-colored skin, Europeans (except Blumenbach
rosy cheeks, brown or the Sami and Finns), (1795/2000, 28)
chestnut-colored hair, North Africans, South
oval faces, straight and Asians, West Asians,
narrow noses, etc. and Western Siberians
west of the Obi River
Ethiopian Black-colored skin, Aboriginal Blumenbach
black and curly hair, Australians, sub- (1795/2000,
thick noses, puffy lips, Saharan Africans, 28–29, 37)
etc. Melanesians, etc.
Malay Tawny-colored skin, Malayo-Polynesian Blumenbach
black soft and thick language speakers (1795/2000, 29,
hair, full and wide except non-Papuan 36–37)
noses, . . . , and use Melanesians (e.g.,
“the Malay idiom” Filipinos, Malagasy,
Micronesians,
Sunda Islanders,
Polynesians, etc.)
Mongolian Yellow-colored skin, All remaining East Blumenbach
black and stiff and Asians (except (1795/2000, 28)
straight hair, broad Malays), Eskimos,
faces, small noses, Finns, and Sami
narrow eyelids, etc.
Of course, the OMB wanted its races to be ancestry groups instead of phe-
notypic and linguistic groups like Blumenbach’s races. As a consequence, the
OMB’s races do not perfectly align with Blumenbach’s races. For instance,
the OMB’s Black race lacks the Aboriginal Australians and Melanesians that
fall into Blumenbach’s Ethiopian race, because while Aboriginal Australians,
Melanesians, and sub-Saharan Africans look very similar, it’s been known
20
220 • W h a t I s  R a c e ?
since at least the 1980s that Aboriginal Australians and Melanesians are dis-
tantly related to sub-Saharan Africans.34
In any case, even though Blumenbach’s races and the OMB’s races do not
perfectly align, there’s definitely a one-to-one correspondence between the
two sets of races (e.g., American corresponds to American Indian, Caucasian
corresponds to White, etc.). Furthermore, the OMB (1997, 58782–58783)
did say that it wanted a “comprehensive in coverage” racial scheme and a ra-
cial scheme that can deal with “growth in immigration.” So, Prewitt is prob-
ably right that the OMB wanted to adopt a human classification scheme very
similar to Blumenbach’s racial scheme.
Now, given that Prewitt is right that the OMB wanted to adopt a human
classification scheme very similar to, but not identical to, Blumenbach’s ra-
cial scheme, and given that the OMB wanted to adopt a human classifica-
tion scheme based on ancestry instead of phenotype and language like
Blumenbach’s racial scheme, and, furthermore, given that the set of human
continental populations is an ancestry-based Blumenbach-like human classi-
fication scheme, it’s hard to deny that the thing in the world that the OMB
intended to pick out with ‘race’ in 1997 was, in fact, the set of human conti-
nental populations.
7.4. The Empirical Adequacy Objection for Glasgow,

Haslanger, and Jeffers
I will now turn my attention away from OMB race theory and toward the
race theories of my coauthors. Remember that Glasgow, Haslanger, and
Jeffers are all offering a theory about what race is and whether it’s real relative
to the single dominant meaning of ‘race’ in, at least, American English—but
in other linguistic contexts as well in Haslanger’s and Jeffers’s cases. However,
now it’s time to seriously question whether modeling dominant race talk
in American English as stemming from a single dominant meaning of ‘race’
provides us with a more empirically adequate US race theory than a radically
pluralist US race theory.
The term ‘empirical adequacy’ was coined by the philosopher of science
Bas van Fraassen in 1980 as a minimal condition for accepting a scientific
theory. Without using any of van Fraassen’s jargon, very roughly, we can say
that a theory is empirically adequate relative to all of the phenomena that the
34. For example, see Cavalli-Sforza et al. (1988, 6003).

21
theory is designed to explain or predict just in case the theory, in fact, explains
or predicts all of that phenomena (van Fraassen 1980, 12). Van Fraassen is
careful to note that a theory doesn’t achieve empirical adequacy just for
explaining all of the relevant observed phenomena, but rather, all of the phe-
nomena (observed and unobserved) that the theory is supposed to account
for. While it’s true that no scientific theory has ever been shown to be empir-
ically adequate, scientists have used closer proximity to empirical adequacy
(or what I’ll call the empirical adequacy standard) as a useful way to select one
theory out of a group of rivals. For example, biologists who studied heredity
eventually accepted Mendel’s theory of genes after it became obvious that
all of its serious rivals (esp. Spencer’s theory of physiological units, Darwin’s
theory of gemmules, and Weismann’s theory of ids) simply failed the empir-
ical adequacy standard relative to Mendel’s theory (Morgan 1926, 26–31).35
In addition to being a constraint on theory acceptance that’s widely
adopted in science, the empirical adequacy standard is also widely accepted
among philosophers of science as being a minimal condition for reliably
picking one scientific theory over another as being true or closer to the
truth.36 It’s important to note that van Fraassen didn’t construe his empirical
adequacy standard as being truth-tracking because he was a scientific anti-
realist. However, scientific realists and scientific non-realists other than scien-
tific anti-realists have embraced the empirical adequacy standard as being at
least what’s required to reliably say that one scientific theory is true or more
true than another.37 Also, while the term ‘empirical adequacy’ was coined by
van Fraassen, the idea of empirical adequacy and its important link to reliably
accepting true scientific theories has been known by philosophers of science
since at least as early as Pierre Duhem’s idea of “complete” scientific theories
in 1906.38
Now, since philosophical race theories (or at least the ones offered in this
book) are types of scientific theories (given that they’re descriptions about
35. Of course, ironically, after genetics got started in the early 1900s, it was quickly realized that
Mendel’s theory had to be revised considerably in order to account for genetic linkage, chro-
mosomal recombination, chromosomal non-disjunction, and other hereditary phenomena
that Mendel’s theory didn’t account for.
36. By a reliable pick I mean one that’s arrived at from an inferential method with low false-
positive and low false-negative error. This notion of epistemic reliability is widely accepted
among philosophers of science. See, for example, Godfrey-Smith (2003).
37. Some examples are Longino (1990, 93–94) and Fine (1999).
38. See Duhem (1906/1981, 19).
2
222 • W h a t I s  R a c e ?
what race is and whether race is real given a specific linguistic context), it’s
appropriate to hold them up to the same minimal epistemic standards as
scientific theories, including the empirical adequacy standard. So, let’s see
whether the race theories that Glasgow, Haslanger, and Jeffers have offered us
are closer to being empirically adequate compared to a radically pluralist US
race theory, such as one that includes OMB race theory.
Remember in Chapter 3 when I said that 93.8% of US residents self-
reported an OMB race on the 2010 US Census questionnaire (Chapter 3,
83), and that geneticists can predict US adults’ OMB race self-reports with
98.8%–99.8% accuracy given a few reasonable background assumptions
(Spencer, Chapter 3, 102). It turns out that these statistics can be used to esti-
mate the extent to which US residents are competent in using OMB race talk.
The estimate turns out to be that, approximately, 92.7% of US residents are
competent in using OMB race talk.39 Also, while this estimate assumes that
almost everyone who provides at least her primary human continental popu-
lation membership as her race on an official form (e.g., US Census question-
naire) is competent in using OMB race talk, this assumption is not that risky.
Remember that the overwhelming majority of Americans are people with
a single and clearly primary human continental population membership,
such as non-Hispanic European Americans (63.1% of US residents, 98.6%
Caucasian on average), African Americans (13.6% of US residents, 73%
African on average), Eastern Asian Americans (4.4% of US residents, 95.5%
East Asian on average), American Jews (2.2%, 96% Caucasian on average),
and so forth. Given these demographic and genealogical facts, it should not
be surprising that so many Americans are competent in using OMB race talk.
It’s not a hard language game for most Americans to play!
Now, couple the fact that about 92.7% of US residents are competent in
using OMB race talk with the additional fact that OMB race talk is the default
39. This calculation was made by simply multiplying 0.988 by 0.938 and converting the product
to a percentage. I say ‘suggest’ and ‘approximately’ because it’s debatable whether everyone who
has self-reported her primary OMB race (according to genomic ancestry estimates) is compe-
tent in using OMB race talk. For one, some of these respondents could have made a lucky guess.
Second, it’s unclear how many US Census racial self-reports were self-reports in the ordinary
sense. Many people don’t know this, but ‘self-reported’ (at least with respect to 2010 census
data) is jargon for the US Census Bureau. It includes reports made by the respondent (who may
not be the person whose race is being communicated) and assignments from the US Census
Bureau itself when data are missing (a.k.a. imputations) (Humes et al. 2011, footnote 8). Third,
and finally, there’s the matter of error. The US Census Bureau sometimes counts people more
than once, counts dead people (who will have imputed census data), omits people, etc., which
adds up to a small amount of error for any statistic that the US Census Bureau reports.
23
race talk on any form in the US that has US government oversight or any form
in the US that for some reason is tethered to the OMB’s racial scheme, and
we have a reasonable basis for saying that the OMB’s meaning of ‘race’ is, at
least, one dominant meaning of ‘race’ among American English speakers.
Some examples of these forms are mortgage loan applications, birth certificate
applications, food stamp applications, college admissions applications, day-
care enrollment forms, health provider enrollment forms, health insurance en-
rollment forms, job applications at colleges and universities, federally funded
or administered scholarship and fellowship applications, and so forth.40
Now, given that OMB race talk harbors one dominant meaning of ‘race’
among American English speakers and given our presupposition of what a
dominant meaning of ‘race’ is, it turns out that Glasgow’s, Haslanger’s, and
Jeffers’s race theories fail the empirical adequacy standard when compared to
a radically pluralist US race theory, and especially one that includes OMB
race theory. In particular, the race theories that Glasgow, Haslanger, and
Jeffers have offered us are empirically inadequate in two important respects.
First, each of these race theories misdiagnose what, essentially, race is or what,
essentially, the races are in situations where the OMB’s racial scheme is being
presupposed. Second, each of these race theories oversimplifies the complexity
of American communications about race. Here are two examples that nicely
illustrate my points.
Remember that I said the OMB’s racial scheme is the default racial scheme
used by American colleges and universities to collect racial data on their col-
lege applicants.41 Well, this fact has consequences because it makes it highly
40. Some of these examples have interesting histories. For instance, the reason why OMB race
talk is dominant on American day-care enrollment forms is because American day cares usu-
ally offer parents the option of applying for the US Department of Agriculture’s Child and
Adult Care Food Program and that application always uses the OMB’s racial scheme to ask
about the applicant’s race. Also, as of December 2007, American colleges and universities are
legally required to collect and report the racial and ethnic data of their students and employees
to the US Department of Education if they want any funding from the department, and you
can guess which racial scheme that racial data must be reported in. Also, while health care and
health insurance providers are not legally required to report any racial data to the US govern-
ment, they are, nevertheless, anchored into using the US government’s racial scheme. In short,
the enormous database of health information that the US government has (e.g., data from the
NIH, CDC, NCHS, etc.) is only useful if you’re using a racial scheme that’s at least compat-
ible with the US government’s. Obviously, the easiest thing to do here is simply use the US
government’s racial scheme. I must thank Ronald Copeland for the last insight, who’s the chief
diversity officer for Kaiser Permanente.
41. However, there are a few US colleges and universities that don’t use the OMB’s racial
scheme when collecting racial data on their applicants. Some states legally ban the consider-
ation of race in the admissions decisions of their public colleges, which has led to the public
24
224 • W h a t I s  R a c e ?
likely that any particular discussion or dispute among Americans about the
lawfulness of race-based preferential affirmative action in college admissions
today presupposes the OMB’s racial scheme, whether Americans know it
or not.
For example, as I’m writing this chapter, the US Department of Justice
(DOJ) is reconsidering a complaint against Harvard University that was
originally submitted to the US Department of Education (ED) and the
DOJ in May 2015 by the Coalition of Asian-American Associations (or “the
Coalition”). The complaint was dismissed under the Obama administration,
but the Trump administration is reconsidering the complaint on the grounds
that “[t]‌he Department of Justice is committed to protecting all Americans
from all forms of illegal race-based discriminations” (Chakraborty 2017).42
In the complaint, the Coalition accuses Harvard of engaging in racial dis-
crimination for years that has been repeatedly upheld by the US Supreme
Court to be unlawful.43 In particular, the Coalition accuses Harvard of using
a “de facto racial quota” on Asian admits to Harvard College (Coalition 2015,
41). Since Harvard has been using the OMB’s racial scheme to collect racial
data about its applicants since 2010, this particular national discussion about
race is anchored by the OMB’s racial scheme whether Americans know it
or not.44
colleges in these states not asking for their applicants’ OMB race(s) in the admissions pro-
cess. For example, since California is one of these states, no University of California campus
(e.g., UC Berkeley, UCLA, UCSD, UC Davis, etc.) asks about their applicants’ race(s) in the
admissions process. However, Caltech, Claremont McKenna, Harvey Mudd, Pepperdine,
Pomona, Scripps, Stanford, USC, USD, USF, and just about every other private college or
university in California ask about their applicants’ race(s) in the admissions process and use
the OMB’s racial scheme to do it.
42. Interestingly, the Coalition filed a second and similar complaint with the DOJ and ED
against Brown, Dartmouth, and Yale in May 2016. However, I’ll focus on the Harvard com-
plaint since it’s currently getting reinvestigated by the Trump administration.
43. Given how I’ve worded this sentence, you might be wondering what kind of racial dis-
crimination federal courts consider to be lawful. Well, lots! In federal courts, racial discrimi-
nation is understood to be any differential treatment based on race. So, for example, a person
who chooses to only intra-racially date would be racially discriminating according to federal
courts. In addition, unlawful racial discrimination, for federal courts, is not just any racial dis-
crimination that violates a federal law or the US constitution. If that were the case, race-based
preferential treatment in college admissions would be obviously unlawful since it violates Title
VI in the Civil Rights Act of 1964. Rather, federal courts consider racial discrimination that
violates a federal law or an article of the US Constitution to be merely presumptively unlawful.
However, such discrimination can be lawful if it passes a threefold test known as strict scrutiny.
44. I should clarify that Harvard used the OMB’s 1977 racial scheme (which had no Asian
race) to ask about race on its college applications before 2010. This actually raises an interesting
25
So, now the question is, what is the correct way to understand the nature
and reality of the Asian race relative to this particular national debate about
affirmative action? Is it correct to say that the Asian race, in this context, is,
essentially, a visible-trait grouping as Glasgow (Chapter 4, 119) would say?
I don’t think so. In this context, Asians roughly divide into South Asians ~
1.8 billion people) and Eastern Asians (~ 2.3 billion people).45 But these two
groups of people do not form a visible-trait grouping at all. Rather, about half
of them (South Asians) look like moderately or darkly pigmented Europeans
and the other half (Eastern Asians) possess a different distinctive look (e.g.,
light to moderate skin pigmentation, epicanthic folds, round noses, etc.).46
Is it correct to say that the Asian race, in this context, is, essentially, a cul-
tural group as Jeffers would say (Chapter 2, 58)? I don’t think so. Again, in
this context, Asians roughly divide into South Asians and Eastern Asians, but
these two groups of people do not share a cultural essence. There’s no combi-
nation of language, religion, food, music, literature, or other cultural property
that unites South and Eastern Asians into a distinctive cultural group.47
Finally, is it correct to say that the Asian race, in this context, is, essentially, a
“racialized group,” as Haslanger would say? Again the answer is ‘no.’ Even if it’s
true that the group of people consisting of, mostly, South Asians and Eastern
Asians is currently racialized in the United States, the latter is merely a con-
tingent fact about the group, not a necessary fact—and thus not an essential
semantic point. Harvard cannot be guilty of unlawful racial discrimination against Asian
applicants before 2010 because they didn’t recognize Asian as a race before 2010! Nevertheless,
Harvard could have been engaging in unlawful racial discrimination against Asian applicants
in 2010 and after.
45. These population estimates are from the United Nation’s 2015 world population estimates.
See United Nations (2017). Also, notice that I left out Central Asians. I did this for two
reasons. First, they make up just 1.7% of Central, Eastern, and South Asians according to the
United Nations (2017). But also, Central Asians are about evenly split with respect to how
they look. Some of them (e.g., Tajiks) look more Caucasian due to their primarily Caucasian
genomic ancestry, while the rest (e.g., Uzbeks) look more East Asian due to their primarily East
Asian genomic ancestry (Martínez-Cruz et al. 2011, 221).
46. It’s important to note that here and elsewhere in this chapter, I’m not assuming that the act
of naming or attributing properties to an object commits one to the actual existence of that ob-
ject. So, for example, when I said earlier that Glasgow’s theory of race commits him to the claim
that Asians form a visible-trait grouping if they’re a race, that claim does not commit Glasgow
to Asians existing if they’re a race (which is a good thing since Glasgow is sympathetic to racial
anti-realism). This might sound bizarre, but remember, I said in Chapter 3 that the background
logic that I’m adopting in my chapters is a free logic, and all free logics reject that naming an
object or attributing a property to an object commits one to that object’s actual existence.
47. A generic version of this point has been made before by Anthony Appiah. See Appiah
(1985, 36).
26
226 • W h a t I s  R a c e ?
fact—about the group. The reason why is because, as Jeffers (Chapter 2, 71) and
Glasgow (Chapter 4, 132) have pointed out earlier, “racial equality” is impossible
if races are racialized groups. However, the OMB talks about its races in such a
way that it’s at least possible for racial equality to come about.
For example, remember that one of the two main reasons why the OMB
(1997, 58782) created its race talk was to help federal agencies “enforce civil
rights laws.” Also, after the OMB revised its racial scheme in 1997, it published
a long document detailing exactly how its new racial scheme would be useful
in detecting racial gerrymandering, monitoring equal employment opportunity,
protecting equal opportunity in education, enforcing Title VI of the Civil Rights
Act of 1964, and improving FBI hate crime statistics (OMB 2000, 62–70, 83).
That doesn’t look like the behavior of an agency that rejects the possibility of
racial equality.
As you might have guessed by now, I think the most plausible way to de-
fine ‘Asian’ in this particular national affirmative action debate is as an ancestry
group, and, specifically, the genealogical population of East Asians (as OMB race
theory does). In this way, one can make sense of all of the phenomena discussed
in the preceding, plus many more, such as why the Coalition itself claims that
Harvard is discriminating against Asian applicants in virtue of their “Asian an-
cestry” (Coalition 2015, 31).
In addition to misdiagnosing what, essentially, race is or what, essentially, the
races are in particular American communications about race, the race theories
that Glasgow, Haslanger, and Jeffers have constructed also oversimplify the com-
plexity of some American communications about race. An excellent example of
this oversimplification is the 2015 American obsession with Rachel Dolezal’s
race.48
Dolezal is an American citizen who, in her own words, “was biologically
born white, to white parents,” but who began presenting herself as Black and
self-identifying as Black in every facet of her life sometime after she grad-
uated from graduate school in 2002.49 For example, before she garnered
media attention in June 2015, Dolezal was a former Africana Studies in-
structor at Eastern Washington University (EWU) and the former president
of Spokane’s chapter of the National Association for the Advancement of
Colored People (NAACP). Dolezal also filed complaints with Idaho police
48. Although Rachel Dolezal has legally changed her name to ‘Nkechi Diallo,’ she still uses her
former name for her public persona. As such, I will use her public persona name in this chapter.
49. This quote is from Dolezal’s November 2, 2015, interview on The Real, which I’ll talk about
in detail very soon.
27
about being a victim of anti-Black hate crimes. She graduated from Howard
University, which is a historically Black college or university (HBCU). Her
ex-husband is African American. She marked ‘Black/African American’ on
her job application to Spokane’s Office of Police Ombudsman Commission.
And the icing on the cake was that she curled her hair and suntanned enough
to look like a lightly pigmented African American!
Dolezal came to local media attention in Spokane after she filed anti-
Black hate crime complaints with Idaho police. However, Dolezal came to
national media attention when a local Spokane news reporter, Jeff Humphrey
of KXLY, interviewed her on camera about her hate crime complaints. In
the course of Humphrey’s interview, he asked her directly, “Are you African
American?” Dolezal was caught off guard and said, “I don’t understand the
question.” Then, Dolezal quickly walked off camera and ended the interview.
As you can imagine, Dolezal’s response drew immediate suspicion from other
news reporters. In particular, ABC news (the parent company of KXLY),
quickly found Dolezal’s birth parents and interviewed them. In that inter-
view, Dolezal’s parents said, “There seems to be some question of how Rachel
is representing her identity and ethnicity. . . . We are definitely her birth
parents. We are both of Caucasian and European descent—Czech, German
and a few other things” (Capehart 2015).
When Dolezal was invited to explain herself in several national television
interviews, she repeatedly said she was ‘Black’ or ‘African American,’ and,
sometimes, ‘not White.’ However, Dolezal eventually admitted on camera
that she was “biologically born white,” as I mentioned earlier. One fact about
the Dolezal case that’s relevant for philosophical race theory is how compli-
cated the national discussion was about Dolezal.
One debate was about whether Dolezal could accurately claim to be ra-
cially Black without possessing what was called Black ancestry in the con-
versation.50 Furthermore, this debate was at least partially motivated by a
genuine concern about whether Dolezal was taking away educational or em-
ployment opportunities that were intended for people with Black ancestry.
For example, during Dolezal’s interview on The Real, co-host Loni Love said
that she didn’t care about how Dolezal racially identified, but she did care
about whether Dolezal marked ‘Black’ on her college applications because
that act could have taken away scholarship money from a student with Black
50. For example, this was the term that YouGov used on its June 17–19, 2015, national survey
about Dolezal.
28
228 • W h a t I s  R a c e ?
ancestry.51 Interestingly, Dolezal said that Howard’s college application didn’t

ask about race, but she did say that she marked ‘Black’ on her job applica-
tion to Spokane’s Office of Police Ombudsman Commission. Furthermore,
Dolezal said she marked ‘Black’ because “we all have human origins in the
continent of Africa.”52
Some of the debate was about whether Dolezal could accurately claim to
be racially Black when underneath her artificially constructed Black appear-
ance (esp. her light brown skin color and curly hair), she looked like a typ-
ical Caucasian woman. Furthermore, this debate was motivated by a serious
concern that Dolezal was participating in blackface, which, if true, would
make Dolezal a participant in a long history of racist imagery. For example,
in a widely read op-ed on the Dolezal case for The Washington Post, Jonathan
Capehart said, “Blackface remains highly racist, no matter how down with
the cause a white person is” (Capehart 2015).
In addition, some of the debate was about whether Dolezal could accu-
rately claim to be racially Black without having gone through the so-called
Black experience. Furthermore, this debate was motivated by a genuine con-
cern that Dolezal was exercising White privilege by pretending to be Black
when it was convenient, but reverting back to identifying as White when that
was convenient. This became an increasing concern once the media found out
that Dolezal actually had sued Howard for racially discriminating against her
because she was White! An example of this debate can be found once again
during Dolezal’s interview on The Real. In that interview, co-host Tamar
Braxton expressed exactly this concern when she asked whether Dolezal
thought she had “walked the walk of a Black woman.” Interestingly, Dolezal
responded, “Absolutely,” and followed that up with, “the police mark ‘Black’
on my traffic tickets.”
51. The Real is a daytime talk show with the same format as The View and The Talk (all women
hosts who discuss the daily news), but with one big difference: all of the hosts on The Real
are racial or ethnic minorities in the US. One host ( Jeannie Mai) is Asian American, another
host (Adrienne Houghton) is Hispanic American, and the remaining three hosts (Loni Love,
Tamar Braxton, and Tamera Mowry) are all Black American. Actually, Braxton is no longer a
host on the show, but she was a host when Dolezal was interviewed.
52. I should clarify that it was actually Jeannie Mai, not Loni Love, who got Dolezal to reveal
which race(s) she marks on “applications.” Also, notice that we can use OMB race theory to
pinpoint exactly why Dolezal’s answer here is incorrect. Remember, according to OMB race
theory, OMB racial membership is about genomic ancestry, not ancestry simpliciter. So, while
it’s true that all living humans have African ancestry, it’s not true that all living humans have
African genomic ancestry.
29
And that’s just three debates Americans were having. I didn’t even discuss
the so-called transracial debate! But we can stop here because it’s clear to see
that the national discussion about Dolezal was complicated, and so compli-
cated that any attempt to simplify it to a discussion about a single meaning
of ‘race’ would not accurately capture what was going on. For example, is it
more plausible to say that Loni Love, Tamar Braxton, and Jonathan Capehart
were somehow, covertly, and unknowingly, using the same meaning of ‘race’?
Or, is it more plausible to say that Love was using whatever meaning of ‘race’
is frequently used on American college and job applications (which is the
OMB’s), Capehart was using a Glasgow-style visible phenotype meaning of
‘race,’ and Braxton was using a Haslanger-style racialized group meaning of
‘race’? I think the latter position is far more plausible than the former because
we can better make sense of the extreme complexity of the conversations that
Americans had about Dolezal. Nevertheless, here are two potential replies to
my empirical adequacy objection.
One reply can be extrapolated from something that Glasgow says in
Chapter 4. At the beginning of Chapter 4, Glasgow says:
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 (Glasgow, Chapter 4, 116).
Glasgow (Chapter 4, 117) uses the preceding fact about how to prevent “com-
munication breakdown” to motivate his search for “one overarching meaning
of ‘race.’ ” Also, Glasgow’s research isn’t entirely aspirational. Elsewhere he
has also said that he thinks it’s “implausible” that we Americans are “simply
babbling past one another when we talk about race” (Glasgow 2009, 75).
I think this is a clever reply. And while I agree that a shared meaning of
‘race’ must be held in order to have a disagreement about race, I don’t think
this reply assuages the empirical adequacy concern I’ve offered in the pre-
ceding. For one, it’s not implausible at all that, sometimes, Americans do “talk
past each other” when discussing race. For example, there was definitely some
cross-talk going on among Americans when discussing whether Dolezal was
Black. But also, there’s no need for there to be “one overarching meaning of
‘race’ ” in order to prevent communication breakdown in national discussions
about race. This is because Americans could know how to competently use
more than one meaning of ‘race’ and utilize whichever meaning is appropriate
320
230 • W h a t I s  R a c e ?
given the situation at hand.53 In short, just like a Taiwanese person can use con-
text to decipher whether her interlocutor is speaking Taiwanese or Mandarin,
Americans can use context to decipher which ‘race’ meaning her interlocutor
is using. In fact, Dolezal’s interview on The Real provides us with an example
of how we do this.
Even though Dolezal was nervous on her interview with the hosts of The
Real, she demonstrated an extraordinary skill in listening to her interviewers
and responding to their questions using the way of thinking about race that
was presupposed in the question that the interviewer asked. For example, re-
member that Dolezal responded to Braxton’s question (which presupposed
that races were Haslanger-like racialized groups) with evidence that she had
taken part in the so-called Black experience by being treated as a Black person
by the police. Also, remember that Dolezal responded to Love’s and Mai’s ques-
tion about whether she marks ‘Black’ on college and job applications (which
presupposes that races are ancestry groups) by explaining how she (and all
humans) possess Black ancestry. So, even if you didn’t like Dolezal’s answers,
we all can at least agree that Dolezal was disagreeing with her interviewers be-
cause she skillfully picked up on the appropriate race talk to use to answer each
question that was asked. So, given that it’s possible for Americans to disagree
about race without holding a single dominant meaning of ‘race,’ Glasgow’s
observation that Americans often disagree in their discussions about race
does not assuage the empirical adequacy concern that I’ve advanced.
However, another reply to this objection may stem from my assumption
that philosophical race theories (at least the ones advanced in this book) are
descriptions of what race is and whether race is real given a specific linguistic
context. But Haslanger and Jeffers could reply that their race theories are not
mere descriptions. For example, in Chapter 2, Jeffers (Chapter 2, 58) says,
he does not “draw a very sharp distinction between ethics and metaphysics,”
and as a consequence, he thinks it’s important to pay attention to “values and
ideals” when “thinking about the nature and reality of race.” In fact, in his
response to Paul Taylor’s (2013, 100–101) generic objection against cultural
constructionist views about race in the US context—which is that they’re
confusing a “prescriptive” race theory for what should be a “descriptive” race
theory—Jeffers (Chapter 2, 60) replies that his race theory, as well as that of
other cultural constructionists like DuBois, is committed to both prescriptive
and descriptive content.
53. I’m getting this response from Ann Morning’s (2009, 1186) “tool kit” quote.
213
In addition, Haslanger (Chapter 1, 6) explicitly says in Chapter 1 that the

“adequacy” of her race theory “is not to be judged simply by reference to ‘the
facts.’ . . .” Haslanger (Chapter 1, 6) even distances herself from scientists such
as “anthropologists” when clarifying what kind of race theory she’s engaging
in. Instead, Haslanger (Chapter 1, 8) views her race theory as an instance of
critical social theory, which, in her view, is not merely a descriptive, but is also
a prescriptive, exercise insofar as it recommends how to “improve” “ordinary
social practices.” So, in sum, both Jeffers and Haslanger might object to the
empirical adequacy objection I’ve offered because it seems to presuppose that
all a philosophical race theory is trying to do is accurately describe as opposed
to, also, helpfully prescribe.
This is a wonderful concern. However, notice that I never said that phil-
osophical race theories are completely descriptive theories. In fact, this would
be hypocritical of me since OMB race theory uses a definition of ‘biologi-
cally real’ that’s partly normatively justified.54 Rather, all I said was that philo-
sophical race theories are, in fact, descriptive theories about what race is and
whether race is real relative to a specific linguistic context. This statement is
compatible with philosophical race theories being more than just descriptive
theories. Nevertheless, insofar as any philosophical race theory is descriptive
at all, its adequacy is bound by the empirical adequacy standard just like any
other scientific theory.55 In other words, if you want to say that races are cul-
tural groups or racialized groups given the predominant way that American
English speakers use ‘race,’ then your theory needs to be better able than its
most serious rivals to explain or predict all of the relevant phenomena.
Also, upon closer inspection of what Jeffers and Haslanger have to say,
I think that they would agree with me that the adequacies of their race the-
ories are bound by the empirical adequacy standard. For instance, Jeffers
(Chapter 2, 61) explicitly says that his and other cultural constructionists’
“prescriptions” about race (e.g., DuBois’s prescriptions) have always been
54. For evidence, see Spencer (2012).

55. It’s also important to note that scientific theories are often not entirely descriptions as well,
and philosophers of science have known this fact since at least when Duhem (1906/1980, 208–
212) observed that natural scientists often use stipulative definitions for key terms in scien-
tific theories. For example, you might think that the claim “Homo sapiens immediately evolved
from the species Homo ergaster” is a purely descriptive claim. But it isn’t. It’s a descriptive claim
about an evolutionary relationship between two species given a preferred definition of ‘spe-
cies’—namely, the phylogenetic species concept (PSC). For example, Joseph LaPorte (2005)
has shown that if you switch the definition of ‘species’ from the PSC to the biological species
concept (BSC), then this evolutionary statement is false. Rather, under the BSC, the true state-
ment would be, “Homo sapiens immediately evolved from the species Homo erectus.”
32
232 • W h a t I s  R a c e ?
“inseparable from an understanding of the race being exhorted as already a

cultural group in some sense. . . .” Also, Haslanger (Chapter 1, 16) has the
following to say about her race theory: “. . . the goal is to provide an interpre-
tation of what has plausibly been at issue . . . ‘all along,’ as evidenced not only
by what we say, but what we do, such as the practices we engage in, the laws
we pass, and social scientific explanations of these.” So, since both Jeffers’s and
Haslanger’s race theories are, in large part, descriptions about what race is and
whether race is real given, at least, American English speakers’ usage of ‘race,’
both of their race theories are in just as much trouble as Glasgow’s when it
comes to being closer to empirical adequacy than a radically pluralist US race
theory that includes OMB race theory.
7.5. A Methodological Concern for Glasgow and Jeffers

While my primary concern with Haslanger’s race theory is that it’s simply
not as empirically accurate as a radically pluralist US race theory that
includes OMB race theory, I have an additional concern about Glasgow’s and
Jeffers’s race theories. It’s about methodology. Notice that both Glasgow and
Jeffers crucially rely on thought experiments whose experimental results are
drawn entirely from their own intuitions, or, at least, some small group of
philosophers’ intuitions. For ease of communication, I’ll call such thought
experiments intuition-based thought experiments. While I have no universal
worry about the reliability of intuition-based thought experiments in philos-
ophy, I do have a serious worry about their reliability in philosophical race
theory.56
In particular, I’m skeptical that intuition-based thought experiments
can yield any reliable information about what ‘race’ means when the lin-
guistic community is large and diverse, like, for example, American English
speakers, or, even more so, English speakers. If I’m right that an evidential
method that both Glasgow and Jeffers crucially rely on is unreliable under
the very circumstances in which it’s being used, then we should be worried
that both Glasgow’s and Jeffers’s race theories aren’t true, at least not given
the way these theories are currently formulated. But let’s explore this concern
in more detail.
56. However, for a criticism of using intuition-based thought experiments in philosophy at

all (esp. metaphysics and epistemology), see Machery (2011). For another criticism of using
intuition-based thought experiments in philosophical race theory, see Haslanger and Saul
(2006). What I am about to say is basically an extension of what I said in Spencer (2015, 52).
23
First, suppose we call any race that’s part of the correct US race theory
a US race. In that case, note that both Glasgow and Jeffers crucially rely on
intuition-based thought experiments in order to evidentially support their
claims about what is part of (or not part of ) the essence of a US race. Glasgow
does this directly with lots of intuition-based thought experiments. His Dalai
Lama thought experiment is supposed to show that “races must, by defini-
tion, be visibly distinct,” which, if true, is a problem for all theories of US
races as genealogical populations (Chapter 4, 122). Glasgow’s racial amnesia
thought experiment is supposed to show that “race persists even when the so-
cial facts change,” which, if true, is a problem for all social constructionist US
race theories (Chapter 4, 132). Glasgow (Chapter 4, 132) also cites his pre-
vious Utopia thought experiment from A Theory of Race, which is supposed
to show that “racial equality is not incoherent,” which is a problem for social
inequality–based US race theories.
To be clear, I’m not saying that Glasgow’s race theory is supported solely
by intuition-based thought experiments. He also uses relevant observations
to support his theory, such as his observation that biological populations pos-
sess the essential property of “migratability,” which is the property that one’s
members can immigrate into and emigrate from the group (Glasgow, Chapter
4, 121).57 Rather, what I’m pointing out is that the evidence for Glasgow’s race
theory crucially involves intuition-based thought experiments. For example,
the only evidence that Glasgow provides for his key claim that “races must,
by definition, be visibly distinct,” is the result from his Dalai Lama thought
experiment.
Similarly, Jeffers’s race theory crucially relies upon intuition- based
thought experiments for evidence. For one, remember that Jeffers (Chapter
2, 44) uses one thought experiment (the dark-skinned Bangladeshi British
woman) as a key source of evidence for rejecting “biological realism” as a “rea-
sonable . . . description of race.” But to get his result, all Jeffers (Chapter 2,
44) appeals to is his intuition that “it conflicts with common sense” and that
it’s less “illuminating” to describe this woman as “Caucasian” or “Caucasoid.”
Also, remember that Jeffers (Chapter 6, 192) adopts Michael Hardimon’s
“logical core” for what an ordinary race is as being a central element of his
race theory. However, Hardimon’s “logical core” for what an ordinary race
57. Notice that this constraint doesn’t apply to genealogical populations like monophyletic
groups or the human continental populations. Members of genealogical populations are born
into them and are not able to migrate in or out.
324
234 • W h a t I s  R a c e ?
is depends crucially—though not exclusively—on intuition-based thought

experiments for evidence.
For instance, in Hardimon’s (2017, 48) most recent defense of his common
ancestry requirement for what an ordinary race is, he relies crucially on what’s
called “the permuted world” thought experiment in order to reply to an objec-
tion from Glasgow. Without that thought experiment, the common ancestry
requirement of his “logical core” would be left wide open to Glasgow’s ob-
jection.58 So, now that I’ve shown that intuition-based thought experiments
are crucially relied upon for evidence by both Glasgow and Jeffers, I’ll move
on to what’s so problematic about relying upon intuition-based thought
experiments for evidence in philosophical race theory.
Well, in short, standard statistical theory advises against using a small and
non-random sample size in order to estimate the parameters of any statis-
tical population when both the variance in the sought-after parameter and
the statistical population size are large (Samuels and Witmer 2003, 57–6 4).
In our case, the target population is not the complete group of American
English speakers, English speakers, or so forth. Rather, it’s the group of ideas
about race that these latter groups of people hold. Also, we’re not looking at
an unchanging collection of ideas, but rather, a changing collection of ideas
stretching from the present day to as far back in the past as we need to go
to reach a semantic shift in what ‘race’ means in its dominant meaning in
the relevant linguistic community (if there is a single dominant meaning).
However, we can put a lower bound on this statistical population’s size in the
case of American English speakers.
During the 2010 US Census, the USCB counted 312,684,568 US residents
and citizens in the fifty US states, the District of Columbia, Puerto Rico,
Guam, and the Northern Mariana Islands (Humes et al. 2011, table 1; Guam
State Data Center 2012, table GU1).59 If we assume that each American
English speaker holds some idea about what race is, then the relevant sta-
tistical population size is at least 312 million. To think that we can sample
the thoughts of a single American English speaker—even if they are from in-
sightful people like Joshua Glasgow and Michael Hardimon—and arrive at
58. For Hardimon’s most recent discussion of Glasgow’s objection to his common ancestry
requirement for being an ordinary race, see Hardimon (2017, 48).
59. I include the citizens of Puerto Rico, Guam, and the Northern Mariana Islands because
they are part of the American English-speaking community given their histories as US ter-
ritories. Also, the Puerto Rico and Northern Mariana Islands counts are from the American
FactFinder, which is available at http://www.factfinder.census.gov.
235
any “core” semantic content in the widest shared meaning of ‘race’ among
American English speakers is extremely optimistic.
However, perhaps Jeffers and Glasgow will point out that my objection
is deficient in the following way. I said that a small and non-random sample
size is bound to lead to unreliable inductions about the parameters of a sta-
tistical population if both the variance in the sought-after parameters and the
statistical population’s size are large. But all I’ve done so far is shown that
the relevant statistical population’s size is large. So, it doesn’t follow yet that
there’s any problem with using intuition-based thought experiments in phil-
osophical race theory.
In fact, Peter Godfrey-Smith (2003, 583) has recently clarified that there
are at least two distinct types of “reliable inference” with respect to induction.
One must be based on a random and large sample of the statistical popu-
lation, but the other need not be. In the other type of reliable induction, a
sample of one might be okay! What’s relevant in the second case is that the
statistical population of interest has very low variance in its values for the rel-
evant parameter. Godfrey-Smith (2003, 585) points out that when the latter
event occurs, we are likely dealing with a “natural kind”—by which he means
a kind whose members do not vary that much in the properties they exem-
plify. The following is an example that illustrates Godfrey-Smith’s point.
Despite the fact that we know that our solar system makes up only a tiny
portion of the total mass-energy of the universe, all inductions in chemistry
about atoms are based on small and non-random samples of atoms from our
solar system. For instance, chemists infer that, at standard temperature and
pressure, all hydrogen ion pairs spontaneously react with electron pairs to
form hydrogen gas simply because that’s how hydrogen behaves in our solar
system (McMurray and Fay 1995, 123). However, according to Godfrey-
Smith, inductions about the chemical reactivity of hydrogen that are based on
the admittedly small and non-random sample of hydrogen in our solar system
are not unreliable because hydrogen atoms form a natural kind. That is, we
can expect any hydrogen atom in the universe to behave like the ones in our
solar system because hydrogen atoms themselves do not vary that much in the
properties they exemplify. So, much like chemists do, why can’t philosophers
of race appeal to a small and non-random sample of ideas about race (their
own) to make a reliable induction about what ‘race’ means for the commu-
nity of American English speakers based on the assumption that American
English speakers’ ideas about race form a natural kind?
While this is a possible way to justify intuition-based thought experiments in
philosophical race theory, I think you can guess what my response is to it. The
326
236 • W h a t I s  R a c e ?
Table 7.2 Glasgow et al.’s (2009) Results for Average Frequency of Using
Five Common Criteria for Racial Membership among 449 US Adults
Criterion Average Frequency of Use Scale Correction
One-Drop Rule 25.1% (2.01/8.00) +1

Ancestry 54.3% (2.17/4.00) 0
Social Relations 48.9% (1.96/4.00) +2
(i.e., Culture)
Ψ-Essence 64.5% (2.58/4.00) 0
Visible Phenotype 48% (0.48/1.00) 0
Notes:
The scale correction is the value added to the maximum and minimum value of the original
scale to mathematically translate the scale to have all positive values with a minimum value
of zero.
The scale minimum and maximum values are in parentheses.
presupposition that American English speakers’ ideas about race form a natural
kind is itself highly suspect and almost certainly false. In fact, we can quantify
how variant American English speakers’ ideas about race are using a recent ex-
periment by Joshua Glasgow et al. (2009).
Using a new psychological instrument called “the Racial Classification
Questionnaire” (RCQ) and a diverse sample of 449 US adults, Glasgow
et al. (2009) explored the average frequency with which ordinary Americans
use five often-discussed criteria for racial membership when actually clas-
sifying other people: the one-drop rule, ancestry, social relations (e.g., cul-
ture), ψ-essentialism, and visible phenotype.60 If one adds an appropriate
correction to make each scale of the RCQ composed of all positive values
with zero as the minimum value, then Glasgow et al.’s results can be summa-
rized in Table 7.2.
It’s important to remember that it’s incorrect to read Glasgow et al.’s results
as reporting the percentage of subjects who used each criterion.61 Rather, the
correct way to read the results is as reporting the average frequency with
60. Glasgow (2009, 66–67) clarifies that features are part of a ψ-essence just in case they are
“heritable, unchangeable racial features that are fixed no later than the moment one is born.”
61. The exception is the result for visible phenotype, since the scale for this criterion consisted
of a single item.
237
which each criterion was used across all subjects. With that said, one obvious
result from Glasgow et al.’s study is that Americans vary considerably in their
ideas about what race is. We can quantify how much by using the coefficient
of variation, which is a statistic that quantifies the dispersion of a frequency
distribution as the quotient of the sample standard deviation divided by the
sample mean (Samuels and Witmer 2003, 44).
If we look at Table 7.2, we can see that the coefficient of variation in
the average frequency with which each criterion was used is 0.30. In other
words, the standard deviation in US adults’ use of the five criteria for racial
membership that Glasgow et al. studied was 30% of its average use, which
is a very high variation in use! Furthermore, Glasgow et al.’s result is not a
fluke. Glasgow et al.–like results have been reproduced by Morning (2009,
2011), Compton et al. (2013), Guo et al. (2014), and Citrin et al. (2014), to
name a few.
What all of this implies is that, according to modern statistical theory
and the experimental studies that we’ve already done on American race-
thinking, we have enough evidence to say that using intuition-based thought
experiments is unreliable in philosophical race theory, at least when theorizing
on American English speakers’ dominant meaning(s) of ‘race.’ Furthermore,
by extension, intuition-based thought experiments are also unreliable when
theorizing on English speakers’ dominant meaning(s) of ‘race.’ In fact, when
we look at dominant race talk in English-speaking countries outside North
America, this point becomes obvious.
For example, in Singapore, the national government uses a fourfold ra-
cial classification that does not respect Hardimon’s requirement that ordinary
English races must be visibly distinct. In particular, the Singapore govern-
ment classifies people into Chinese, Indian, Malay, and Other (Kim 2010,
35). Other! That means Africans, non-Indian Caucasians, Native Americans,
and Pacific Islanders form a single race in a dominant race talk in an English-
speaking country. Even more amazing is that Niue’s government uses the fol-
lowing threefold racial classification: Niuean, Part-Niuean, and Non-Niuean
(Vaha 2012, 45, 119). Non-Niuean! That race is even more visibly diverse than
Singapore’s Other race. So, while I find Glasgow’s and Jeffers’s race theories
interesting and potentially accurate in highly contextualized race talks (e.g.,
races are probably visible-trait groupings in American law enforcement race
talk), their race theories are not based on a reliable evidential method when
applied to American English (or English) speakers as a whole. As a result,
we should be highly skeptical that these race theories are accurate given their
intended scope.
328
238 • W h a t I s  R a c e ?
7.6. Conclusion
To close, I will reiterate what I take myself to have done throughout
Chapters 3 and 7, and then I’ll share a few final remarks. What I take myself
to have done throughout Chapters 3 and 7 is to have developed and defended
a specific race theory for a specific race talk: OMB race theory for OMB race
talk. The theory was that OMB race talk houses an ordinary meaning of ‘race’
in American English whose meaning is its referent, and, is, in fact, a set of
five biological populations in the human species: Africans (or Blacks), East
Asians (or Asians), Caucasians (or Whites), Native Americans (or Americans
Indians), and Oceanians (or Pacific Islanders).
I used contemporary human population-g enetic studies and a partic-
ular meaning of ‘biologically real’ to defend the view that race, in OMB
race talk, is a biologically real division of people as well. I consider OMB
race theory to be a nuanced and contemporary way to defend biological
racial realism.
Nevertheless, I clarified in this chapter that OMB race theory is not a
stand-alone race theory. Rather, it is part of a larger radically pluralist race
theory about the dominant uses of ‘race’ in American English. So, unlike my
coauthors, I did not argue for a monistic US race theory because I don’t be-
lieve that the reliable empirical data we have (e.g., well-designed and well-
executed surveys but not intuition-based thought experiments) support a
monistic US race theory. Before I close, I should address one potentially con-
cerning aspect of OMB race theory. The concern is not about its truth, but
about its potentially negative impact on society.
As many philosophers of race have brought up in the past, one serious
concern that race scholars tend to have about any defense of biological racial
realism is that it possesses a potential to negatively impact society in unique
ways. For example, Lisa Gannett (2001, S489) has argued that “population
thinking”—such as thinking that OMB races are genealogical populations—
is not “inherently anti-racist,” and, in fact, such thinking makes it possible
for people to engage in a more sophisticated form of racist stereotyping that
happens to be statistical. Gannett (2001, S490) names this new type of racist
that population thinking allows “statistical racists.”
Also, Bernard Boxill (2004, 223–224) has argued that ancestry-based ra-
cial classifications tend to decrease our “compassion” for those who are not
members of our race, and he also suspects (though does not argue) that ancestry-
based racial classifications make us prone to “foolish rationalizations” and to
be “gullible enough” to believe these rationalizations. Next, Philip Kitcher
239
(2007, 314–316) has expressed a deep concern that any positive consequences
that result from recognizing a biological racial scheme in American society
(e.g., utility in “race-based medicine”) will be far outweighed by the nega-
tive consequences that follow in American society from that recognition
(e.g., reinforcing “racial stereotypes”). Finally, Charles Mills (2014, 91) has
expressed a sincere concern that modern versions of biological racial realism
might fuel new forms of “extrinsic racism,” and, as a result, add extra sup-
port to preexisting social hierarchies.62 Furthermore, empirical social science
findings have only supported these philosophers’ concerns.
For instance, we now know from various psychological studies that, for
some people, believing in the existence of biological human races is highly
and positively correlated with having racist attitudes (Morning 2009, 1169–
1170). We also know that, for some people, just reading genetic data in racial
terms (e.g., ‘African,’ ‘Caucasian,’ ‘White,’ ‘Black,’ etc.) makes one more likely
to develop not only racial bias, but slide into believing a racialist concept of
race (Donovan 2014).63
Of course, these philosophical concerns and empirical facts are worri-
some. However, I stand with Bernard Boxill (2004, 224) and Daniel Kelly
et al. (2010) in taking a scientifically informed approach to addressing these
concerns and facts. In short, it’s not obviously true that not publishing ac-
ademic defenses of biological racial realism will be an effective method of
preventing any statistically significant rise in American society’s statistical
and extrinsic racists, racial bias, racist attitudes, etc. So, instead, let’s do more
empirical social science to figure out enough about how these causal links ac-
tually work to be able to disrupt these causal links in a way that efficiently and
significantly reduces racial bias, racist attitudes, etc.
For instance, in Donovan’s study, the link between reading genetic data in
racial terms and developing a racialist concept of race turned out to be signif-
icantly negatively correlated with one’s comprehension of Mendelian genetics
(Donovan 2014, 481–482). In other words, the more you knew about genetics,
the less susceptible you were to developing a racialist concept of race just from
reading genetic data in racial terms. So, perhaps one morally respectable way
62. An individual is extrinsically racist when she treats people of a certain race differently based
on a belief that membership in that race is contingently correlated to morally relevant proper-
ties (Appiah 1990, 216).
63. I should say that neither of these correlations have been shown to hold for a nationally rep-
resentative sample of Americans. For example, Donovan’s sample consisted of 43 eighth grade
private school students in the San Francisco Bay Area (Donovan 2014, 469).
420
240 • W h a t I s  R a c e ?
to do philosophical race theory is not to suppress research on biological racial

realism, but rather, to improve the public’s understanding of genetics.
Just in case the latter is a morally respectable way to do philosophical race
theory, I will reiterate the disclaimer I made at the end of Chapter 3. If I’m
right that the OMB races are the human continental populations and that the
human continental populations form a biologically real human population
subdivision, then it’s not a metaphysically confused research project to search
for medically relevant alleles or any other phenotypic differences among
OMB races. Nevertheless, OMB race theory all by itself does not imply that
OMB races differ in any socially important traits (e.g., drug-metabolizing
enzymes, intelligence, beauty, moral character, etc.) or in any phenotypic
ways whatsoever. Furthermore, OMB race theory doesn’t imply that OMB
races don’t differ in any phenotypic ways either. Rather, determining whether
OMB races differ in any phenotypic ways requires a separate empirical inves-
tigation. Furthermore, I am not saying this out of political correctness. It’s
simply a fact that the DNA evidence that supports the existence of human
continental populations comes from non-protein-coding and non-functional
DNA in the human genome.
References
Appiah, A. 1985. “The Uncompleted Argument: Du Bois and the Illusion of Race.”
Critical Inquiry 12(1): 21–37.
Appiah, K. 1990. “Racisms.” In D. Goldberg (ed.), Anatomy of Racism.
Minneapolis: University of Minnesota Press, pp. 3–17.
Appiah, K. 1996. “Race, Culture, Identity, Misunderstood Connections.” In A.
Gutmann and K. Appiah (eds.), Color Conscious. Princeton, NJ: Princeton
Blum, L. 2002. I’m Not A Racist But . . . : The Moral Quandary of Race. Ithaca,
Boxill, B. 2004. “Why We Should Not Think of Ourselves as Divided by Race.” In M.
L. Pataki (ed.), Racism in Mind. Ithaca, NY: Cornell University Press, pp. 209–224.
Bryc, K., E. Durand, J. Macpherson, D. Reich, D., and J. Mountain. 2015. “The Genetic
Ancestry of African Americans, Latinos, and European Americans across the
United States.” The American Journal of Human Genetics 96(1): 37–53.
Capehart, J. 2015, June 12. “The Damage Rachel Dolezal Has Done.” The Washington
Post. Retrieved September 23, 2017, from https://www.washingtonpost.com/
blogs/post-partisan/wp/2015/06/12/the-damage-rachel-dolezal-has-done/?utm_
term=.ae19663876fd
Cavalli-Sforza, L.A., and Feldman, M.W. “The Application of Molecular Genetic
Approaches to the Study of Human Evolution.” Nature Genetics 33(Suppl.): 266–275.
214
Cavalli-Sforza, L., A. Piazza, P. Menozzi, and J. Mountain. 1988. “Reconstruction of

Human Evolution: Bringing Together Genetic, Archaeological, and Linguistic
Data.” Proceedings of the National Academy of Sciences 85: 6002–6006.
Chakraborty, B. 2017, August 2. “NY Times Story on Affirmative Action ‘Inaccurate,’
DOJ Says. Retrieved September 23, 2017, from FoxNews.com: http://www.
foxnews.com/politics/2017/08/02/justice-department-reportedly-g oing-after-
affirmative-action-programs.html
Chomicki, G., and S. Renner. 2015. “Watermelon Origin Solved with Molecular
Phylogenetics Including Linnaean Material: Another Example of Museomics.” New
Phytologist 205: 526–532.
Citrin, J., M. Levy, and R. Van Houweling. 2014. “Americans Fill Out President Obama’s
Census Form: What Is His Race?” Social Science Quarterly 95(4): 1121–1136.
Coalition of Asian-American Associations. 2015. Complaint against Harvard University
and the President and Fellows of Harvard College for Discriminating against Asian-
American Applicants in the College Admissions Process. Washington, DC: The
Chronicle of Higher Education. Retrieved September 23, 2017, from http://www.
chronicle.com/items/biz/pdf/Final%20Aisan%20Complaint%20Harvard%20
Document%2020150515.pdf
Donovan, B. 2014. “Playing with Fire? The Impact of the Hidden Curriculum in
School Genetics on Essentialist Conceptions of Race.” Journal of Research in Science
Teaching 51(4): 462–496.
Duhem, P. 1906/1981. The Aim and Structure of Physical Theory. New York: Atheneum.
Ennis, S., M. Ríos-Vargas, and N. Albert. 2011. The Hispanic Population, 2010. US
Census Bureau. Retrieved from http://www.census.gov/prod/cen2010/briefs/
c2010br-0 4.pdf
Epstein, B. 2015. The Ant Trap: Rebuilding the Foundations of the Social Sciences.
New York: Oxford University Press.
Feldman, M. 2010. “The Biology of Ancestry: DNA, Genomic Variation, and Race.”
In H. Markus and P. Moya (eds.), Doing Race: 21 Essays for the 21st Century.
New York: W. W. Norton, pp. 136–159.
Fine, A. 1999. “The Natural Ontological Attitude.” In R. Boyd, P. Gasper, and J. Trout
(eds.), The Philosophy of Science. Cambridge, MA: MIT Press, pp. 261–277.
Gannett, L. 2001. “Racism and Human Genome Diversity Research: The Ethical Limits
of ‘Population Thinking.’ ” Philosophy of Science 68(3): S479–S492.
Glasgow, J. 2009. A Theory of Race. New York: Routledge.
Godfrey-Smith, P. 2003. “Goodman’s Problem and Scientific Methodology.” The
Journal of Philosophy 100(11): 573–590.
42
242 • W h a t I s  R a c e ?
Guam State Data Center. 2012. Guam Demographic Profile Summary File. Hagåtña,
Guam: Bureau of Statistics and Plans.
Guo, G., Y. Fu, H. Lee, T. Cai, K. Harris, and Y. Li. 2014. “Genetic Bio-Ancestry and
Social Construction of Racial Classification in Social Surveys in the Contemporary
United States.” Demography 51(1): 141–172.
Hardimon, M. 2017. Rethinking Race: The Case for Deflationary Realism. Cambridge,
MA: Harvard University Press.
Hardimon, M. 2003. “The Ordinary Concept of Race.” The Journal of Philosophy
100(9): 437–455.
Harman, G. 2000. “Moral Relativism.” In G. Harman and J. Thomson (eds.), Moral
Relativism and Moral Objectivity. Oxford: Blackwell, pp. 3–19.
Haslanger, S., and J. Saul. 2006. “Philosophical Analysis and Social Kinds.” Proceedings
of the Aristotelian Society 106(1): 89–118.
Hatfield, G. 2003. “Objectivity and Subjectivity Revisited: Colour as a
Psychobiological Property.” In R. Mausfeld and D. Heyer (eds.), Colour
Perception: Mind and the Physical World. New York: Oxford University Press,
pp. 188–202.
Hixson, L., B. Hepler, and M. Kim. 2011. The White Population: 2010, 2010 Census
Briefs. Washington, DC: US Census Bureau.
Hixson, L., B. Hepler, and M. Kim. 2012. The Native Hawaiian and Other Pacific Islander
Population: 2010, 2010 Census Briefs. Washington, DC: US Census Bureau.
Hoeffel, E., S. Rastogi, M. Kim, and H. Shahid. 2012. The Asian Population: 2010, 2010
Census Briefs. Washington, DC: US Census Bureau.
Humes, K., N. Jones, and R. Ramirez. 2011. Overview of Race and Hispanic Origin
2010: 2010 Census Briefs. Washington, DC: US Census Bureau.
Kelly, D., L. Faucher, and E. Machery. 2010. “Getting Rid of Racism: Assessing
Three Proposals in Light of Psycholgoical Evidence.” Journal of Social Philosophy
41(3): 293–322.
Kim, S., C. Gignoux, J. Wall, A. Lum-Jones, H. Wang, C. Haiman, . . . I. Cheng. 2012.
“Population Genetic Structure and Origin of Native Hawaiians in the Multiethnic
Cohort Study.” PLoS One 7(11): e47881.1–10.
Kim, W. 2010. Census of Population 2010 Advance Census Release. Singapore: Republic
of Singapore Department of Statistics.
Kitcher, P. 2007. “Does ‘Race’ Have a Future?” Philosophy & Public Affairs
35(4): 293–317.
Kripke, S. A. 1980. Naming and Necessity. Cambridge, MA: Harvard University Press.
LaPorte, J. 2005. “Is There a Single, Objective Evolutionary Tree of Life?” The Journal
of Philosophy 102(7): 357–374.
Longino, H. 1990. Science as Social Knowledge. Princeton, NJ: Princeton University Press.
Machery, E. 2011. “Thought Experiments and Philosophical Knowledge.”
Metaphilosophy 42(3): 191–214.
243
Mallon, R. 2006. “‘Race’: Normative, Not Metaphysical or Semantic.” Ethics

116: 525–551.
Manichaikul, A., W. Palmas, C. Rodriguez, C. Peralta, J. Divers, . . . J. Mychaleckyj.
2012. “Population Structure of Hispanics in the United States: The Multi-Ethnic
Study of Atherosclerosis.” PLoS Genetics 8(4): e1002640.
Martínez-Cruz, B., R. Vitalis, L. Ségurel, F. Austerlitz, M. Georges, . . . E. Heyer. 2011.
“In the Heartland of Eurasia: The Multilocus Genetic Landscape of Central Asian
Populations.” European Journal of Human Genetics 19: 216–223.
McEvoy, B., J. Lind, . . . A. Wilton. 2010. “Whole-Genome Genetic Diversity in a
Sample of Australians with Deep Aboriginal Ancestry.” The American Journal of
Human Genetics 87: 297–305.
McMurray, J., and R. Fay. 1995. Chemistry. Englewood Cliffs, NJ: Prentice-Hall.
Mills, C. 2014. “Notes from the Resistance: Some Comments on Sally Haslanger’s
Resisting Reality.” Philosophical Studies 171(1): 85–97.
Moorjani, P., N. Patterson, J. Hirschhorn, A. Keinan, L. Hao, . . . D. Reich. 2011. “The
History of African Gene Flow into Southern Europeans, Levantines, and Jews.”
PLoS Genetics 7(4): 1–13.
Morgan, T. 1926. The Theory of the Gene. New Haven, CT: Yale University Press.
Morning, A. 2009. “Toward a Sociology of Racial Conceptualiztion for the 21st
Century.” Social Forces 87(3): 1167–1192.
Morning, A. 2011. The Nature of Race: How Scientists Think and Teach about Human
Difference. Berkeley: University of California Press.
OMB. 1995. “Standards for the Classification of Federal Data on Race and Ethnicity.”
Federal Registrar 60(166): 44674–4 4693.
OMB. 1997. “Document 97-28653: Revisions to the Standards for the Classification
of Federal Data on Race and Ethnicity.” Federal Register 62(210): 58782–58790.
OMB. 2000. Provisional Guidance on the Implementation of the 1997 Standards for
Federal Data on Race and Ethnicity. Washington, DC: Office of Management and
Budget.
Prewitt, K. 2013. What Is Your Race? Princeton, NJ: Princeton University Press.
Ramachandran, S., H. Tang, R. Gutenkunst, and C. Bustamante. 2010. “Genetics and
Genomics of Human Population Structure.” In M. Speicher, S. Antonarakis, and A.
Motulsky (eds.), Vogel and Motulsky’s Human Genetics: Problems and Approaches.
New York: Springer-Verlag Berlin Heidelberg, pp. 589–615.
Rastogi, S., T. Johnson, E. Hoeffel, and M. Drewery. 2011. The Black Population: 2010,
2010 Census Briefs. Washington, DC: US Census Bureau.
Risch, N., S. Choudhry, M. Via, M. Basu, R. Sebro, C. Eng, . . . E. Burchard. 2009.
“Ancestry-Related Assortative Mating in Latino Populations.” Genome Biology
10(11), R132: 1–16.
Rosenberg, N., J. Pritchard, . . . M. Feldman. 2002. “Genetic Structure of Human
Populations.” Science, 298(5602): 2381–2385.
42
244 • W h a t I s  R a c e ?
Samuels, M., and J. Witmer. 2003. Statistics for the Life Sciences, 3rd edition. Upper
Saddle River, NJ: Pearson Education.
Shin, H., and R. Kominski. 2010. Language Use in the United States: 2007. Washington,
DC: US Census Bureau.
Spencer, Q. 2012. “What ‘Biological Racial Realism’ Should Mean.” Philosophical
Studies 159(2): 181–204.
Spencer, Q. 2014. “A Radical Solution to the Race Problem.” Philosophy of Science
81(5): 1025–1038.
Spencer, Q. 2015. “Philosophy of Race Meets Population Genetics.” Studies in History
and Philosophy of Biological and Biomedical Sciences 52: 46–55.
Spencer, Q. 2016. “Do Humans Have Continental Populations?” Philosophy of Science
83(5): 791–802.
Tang, H., T. Quertermous, . . . N. Risch. 2005. “Genetic Structure, Self-Identified
Race/Ethnicity, and Confounding in Case-Control Association Studies.” American
Journal of Human Genetics 76(2): 268–275.
Taylor, P. 2013. Race: A Philosophical Introduction, 2nd edition. Cambridge: Polity Press.
Tighe, E., L. Saxe, R. Magidin de Kramer, and D. Parmer. 2013. American Jewish
Population Estimates: 2012. Waltham, MA: Brandeis University.
United Nations. 2017. World Population Prospects: The 2017 Revision, CD-ROM
Edition. New York: United Nations.
Vaha, K. 2012. Niue Census of Population and Households 2011. Alofi, Niue: Government
of Niue.
van Fraassen, B. 1980. The Scientific Image. Oxford: Clarendon Press.
Weatherson, B., and D. Marshall. (2017, Fall). “Intrinsic vs. Extrinsic Properties.” In E.
Zalta (ed.), The Stanford Encyclopedia of Philosophy. Retrieved from https://plato.
stanford.edu/archives/fall2017/entries/intrinsic-extrinsic/
Zack, N. 2002. Philosophy of Science and Race. New York: Routledge.
Zadeh, L. 1965. “Fuzzy Sets.” Information and Control 8: 338–353.
245
8 G L A S G O W ’ S R E P LY T O H A S L A N G E R ,
JEFFERS, AND SPENCER
In this book I’m concerned with the concept of race that has the
most currency in ordinary discourse. This concept lurks behind
our racial conflicts. We implicitly use it when we racially identify
ourselves and others. We deploy it when we worry about righting
racial wrongs and achieving a better future. While various people
no doubt use the term, ‘race,’ in various ways, we also have an on-
going, broad conversation about race and a vast set of practices that
implicate race, and that is the sense of ‘race’ I hope to be exploring.
With that as the backdrop, a good deal of common ground has
emerged between my coauthors and me. At the same time, some
disagreement remains. In what follows I first clear up three poten-
tial misconceptions about racial anti-realism (leaving basic racial
realism aside for the moment), in order to lay bare a broader frame-
work for agreement among the four theories advanced here. I then
take up some lingering issues of contention.
8.1. Preliminaries
8.1.1. Is Anti-Realism Always Based on the Idea That
Race-Thinking Is Inextricably Essentialist or
Racialist?
Chike Jeffers suggests that the argument for racial anti-realism rests
on the semantic premise that ordinary race-talk is committed to
the existence of some sort of biobehavioral racial essence, where
skin color is (mistakenly) supposed to reflect a suite of biologically
based traits like intelligence or virtue. Now we all agree—here’s our
first point of consensus—that if ordinary race-talk presupposed
that sort of classical racialism, then race in the ordinary sense
would be an illusion. But, as Haslanger and Jeffers rightly point out,
while old-school racialism was heavily featured in early waves of
426
246 • W h a t I s  R a c e ?
race-thinking, the word ‘race’ is not defined in terms of biobehavioral essences.

I agree here, as well—our second point of consensus. Some anti-realists have
argued in the way that Jeffers describes, but my argument proceeds differ-
ently. In fact, my analysis of ‘race’ is even more minimal than Jeffers’s anal-
ysis. On his view, race is conceptually intertwined with three features (and
more): visible traits, geography, and ancestry. On my analysis, visible traits are
the only element of that three-part formula that is actually implicated in the
conceptual core of race. So on my view, not only is it coherent to talk about
race without talking about biobehavioral essences; in addition, it is also co-
herent to talk about race without invoking geography or ancestry.
If we imagine earth being exactly duplicated in some freak four-dimensional
photocopying accident, we would judge that Hillary Clinton’s doppelgänger
is no less white than Hillary Clinton even though Twin Hillary has no racial
(or even human) ancestry. And it would not violate the definition of ‘race’ to
talk about Twin Hillary as white while knowing full well that she was created
from scratch in the same geographic location (Twin New York City) as lots
of non-white people. This shows that, because we can successfully talk about
races under the stipulation that those races have neither distinct ancestries
nor distinct geographic homes, neither ancestry nor geography is part of the
definition of ‘race.’ Of course, it is a contingent, empirical fact that we reg-
ular humans have inherited our racially relevant traits like skin color from our
ancestors, and that those traits developed independently in certain geograph-
ical regions due to local environmental pressures. Historical fact is one thing;
what’s part of the definition of the term ‘race’ is another.1
Once we strike these properties, along with biobehavioral essences, from
the definition of ‘race,’ we are left with the idea that races are defined in terms
of visible traits. The anti-realist argument only has to start with that point.
(And then, if we also find that race is by definition non-social, we’ll have ruled
out racial constructionism.) In this way, all of the views being examined in
this book can agree that race-talk does not necessarily presuppose classical
racialist essences.
8.1.2. Does Anti-Realism Reject Important Social Realities?

Although they acknowledge that anti-realists recognize racialized groups,
both Haslanger and Jeffers also portray anti-realism as out of step when it
1. For more on this argument, see Glasgow (2009, Ch. 2; 2010).

247
Glasgow’s Reply • 247
comes to explaining our social lives. Only constructionism, they say, can fully
capture the race-laden ways in which we have treated one another differently.
I believe this underestimates non-constructionist views. In fact, all of the
views discussed in this book can explain social reality in ways that are exactly
equally adequate. Our four views do disagree: we believe that different words
are more apt for capturing these explanations, and we have different views
about what there is in the world. But all four views can explain all elements of
our social lives in one fashion or another.
Consider Haslanger’s objection to anti-realism (or error theory): if race
is not real, then we cannot use race to explain group differences like cultural
practices or health outcomes, and we’d have to abandon seemingly sound
race-based reasons for making choices, such as trying to improve unequal ed-
ucational opportunities. After all, if race is not real, then it looks like race
can’t correctly explain anything or give any justifiable reason for acting.
However, as noted in Chapter 4, something race-related can explain and
be a reason for acting, namely racialized groups. These are real, even if race is
not. This is our next point of consensus, and it may be the most important
one: all sides in the race debate can agree that people have treated one an-
other differently based on the belief in race—in other words, that we have
racialized ourselves. We all agree that this treatment has impacted real lives
in enormous and morally significant ways. And we all agree that racialization
continues today and will likely continue for the foreseeable future. So I be-
lieve, with Haslanger, that there are racialized groups. I believe, with her, that
this system of racialization is implicated in massive injustices that require re-
pair. I believe, with her, that ignoring these facts would be both factually igno-
rant and ethically disastrous. Similarly, I agree with Jeffers that genuine social
and historical phenomena have led to both awful realities and meaningful
practical identities. Racialization explains different educational opportunities
and outcomes. Racialization explains disparities in housing. Both construc-
tionism and anti-realism agree that these disparities are best explained by how
we treat one another on the premise of race. Similarly, both views agree that
our reasons to remedy such disparities demand that we change how we treat
one another on the premise of race.
So we agree broadly on the social facts of racialization. Constructionists
then add the pivotal contested claim that racialization is sufficient for race: if
racialized groups are real (as they are), then, they claim, race is real. It is here,
not in explaining social life, where the debate lies. According to anti-realism,
races and racialized groups are different. But while we can and do disagree
about this, accepting a distinction between races and racialized groups does
248
248 • W h a t I s  R a c e ?
not hamper our ability to explain or correct racialized injustices, for that
explanation and action comes in the prior step where we (all) recognize
racialization.
Consider an analogy from K. A. Appiah (2007): to explain why some
people were (and still are, in places) persecuted for being witches and why
activists worked against that persecution, you don’t have to believe that
witchcraft is real. You just have to believe that people believed that witchcraft
is real. The fact that we ‘witchized’ people-that we treated them as if they
were witches—explains oppressive behavior and generates a reason to remedy
it. Similarly, racialization explains our race-related behavior, and so we must
recognize the reality of racialization. We live in a racialized world, and our
actions and theories should account for that fact. This is not a reason to be a
social constructionist about race, since anti-realists and biological realists can
(and usually do) recognize racialization as well.
8.1.3. How Does Anti-Realism Aid the Fight against Injustice?

Haslanger is also concerned that anti-realism weakens our ability to tackle
race-related injustice. “For example,” she writes, “the waning of racial essen-
tialism is not sufficient to undermine the legacy of economic deprivation”
surrounding race (p. 21 above). This gets us to our final preliminary ques-
tion: Even if all views can equally explain racialized injustice, how can the
views debated in this book help us repair racialized injustice? Is one of them
better than the others at tackling this all-important task?
I am frustrated that I have few ideas about what it will take to end racial in-
justice. But I do believe that anti-realism is neutral on this question. By itself,
racial anti-realism neither helps nor hinders that reparative project. (I agree
with Haslanger that denying that races are real is not enough to get people to
stop thinking and acting racially. Parts of western Europe have excluded race
from public statistics like censuses for decades, but that has hardly reduced
discrimination, inequality, or coalition-building based on race [Simon 2012].)
Importantly, the same is true of Haslanger’s SPR understanding of race: to de-
fine ‘race,’ as her SPR theory does, in terms of hierarchy and subordination by
itself neither helps nor hinders the project of undermining racial inequality.
For that reason, neither Haslanger’s SPR theory nor anti-realism is going to
steer us to a morally better place.2
2. That said, recall that there is some evidence that certain biological views of race may corre-
late with and even reinforce racist attitudes (Williams & Eberhard 2008; Glasgow, Shulman,
249
Undermining racial inequality requires two steps: recognizing inequality,

and fighting it. As noted earlier, all four theories discussed here can equally
call attention to racialized inequality. Rather than packing the injustice all the
way into the definition of ‘race,’ as SPR does, non-political theories of race can
just say that it is an (extra-semantic) fact about racialization that racialized
groups are unequal. Either way, though, all of these views can recognize in-
justice. But calling attention to racialized inequality—be it inside or outside
the definition of ‘race’—is only part of the reparative job. All the attention in
the world will not increase the urgency with which we confront and reduce
racialized inequality. After all, a white supremacist might look at Haslanger’s
SPR theory of race and celebrate rather than condemn the inequality it
captures. It would be an understatement to say that this is not the desired
outcome. The task of eliminating racial inequality therefore ultimately hangs
on us being motivated to end it, not just on recognizing it. No theory of what
race is, including the views discussed in this book, will itself contain such a
motivation to end racial inequality. That has to come from somewhere else.
The views being debated in this book are closer than they might appear
at first glance, as Ron Mallon (2006, 2009) has argued. We agree that it is
coherent to talk about race without talking about biobehavioral essences. We
agree that it is unlikely that biology will vindicate race based on visible traits
like skin color. We agree that racialization is real. We agree that racialization
has been paired with injustice, and we agree that this injustice must be resisted.
All four of our theories are compatible with that consensus. And I argue in
the Appendix to this chapter that we occupy a vast amount of shared meth-
odological ground as well—one that is vaster than is sometimes recognized.
With that in mind, let’s return to our central question to find where the re-
mainder of variance lies: What is race?
8.2. Constructionism’s Mismatch
Haslanger’s theory of race says that racialized groups—and so races, on her
analysis—must by definition be privileged or subordinated. While we all
agree that racialized groups are as a matter of fact slotted into unequal social
positions, Haslanger’s semantic position builds that inequality up into the
very essence of race: a group cannot be a race, according to her, if it is equal to
& Covarrubias 2009; Phelan et al. 2013). At the same time, not every study on this subject
points to this risk.
520
250 • W h a t I s  R a c e ?
another group. As we saw in Chapter 4, this would mean that racial equality
is an impossibility, a contradiction in terms, an incoherent ideal. I believe that
this account therefore violates one of our core commitments about race—in
fact, one of the most cherished and most central inputs for many people—
namely that if there are races, then racial equality is a valuable goal and so, by
extension, a coherent one. It is possible to talk, for example, about black lib-
eration, where that means that in principle a person could be both black and
fully liberated from racial subordination at the same time. Haslanger’s theory
of race makes this goal unattainable: on her view, if a person does not occupy
a subordinate social position, then by definition that person is not black. As
soon as one is liberated from anti-black oppression, one loses one’s blackness
as well. My view, in contrast, is that the dream of racial equality means that
racial groups are not defined by this fact of hierarchy—they could exist on
equal ground. Liberation need not mean the end of black racial identity in
particular or of race in general.
I won’t belabor the point any more. At the same time, I also argued in
Chapter 4 that this objection to political constructionism follows a formula
that works against every form of constructionism. So can it also apply to
Jeffers-style cultural constructionism?
Here again it is worth blanketing our disagreements in the warmth of
consensus. I appreciate the positive value of cultural difference that Jeffers
highlights. I agree that we must attend to cultural difference in designing
educational curricula or responding to stereotyping. And we agree that dis-
tinctive ways of life have historically been associated with different racialized
groups.
That said, when Jeffers claims that racial diversity should be preserved as
a mechanism for preserving distinctive cultural traditions, I pause. Race is
not required to preserve those distinctive ways of life. Theoretically, anyway,
it seems like we could preserve diverse cultural practices without tying them
to race. Instead we would have to (continue to) redistribute our cultural
practices in a way that is untethered to particular racialized groups. (In other
words, we’d have to pursue cultural integration without cultural homogeniza-
tion, cultural erasure, cultural imperialism, or unjust cultural appropriation.)
There is risk in this, of course. Humanity could easily fail at it. But it appears
to be possible.
That cosmopolitan possibility brings us back to the formula: What would
happen to race in such a world, on the ordinary concept of race? What if there
were no longer any correlations between cultural practices and the groups
commonly recognized as racial? Now, we’ll have to modify this question
215
to fully apply to Jeffers’s view, since he holds that cultural race can coexist
with political equality, and political race can coexist with cultural equality.
The modified question is therefore this: What would happen if all racialized
power hierarchies and all racialized cultural differences disappeared simul-
taneously? What if the only differences left were differences in visible traits?
Jeffers’s version of constructionism entails that without cultural or power
differences, race disappears. But I believe that on the ordinary concept of race,
race persists through such changes. Even if tomorrow all groups currently
recognized as racial had equal power and participated equally in eating the
world’s foods, dancing its forms of dance, playing its kinds of music, and so
on—even in such a world, I do not think we’d say that on the ordinary con-
cept of race Hillary Clinton somehow loses her whiteness or that Jeremy Lin
stops being Asian because of those points of equality.
Similarly, recall the babies-only world that we imagined in Chapter 4.
The babies—one who ordinarily would be recognized as Asian, another
as black, and so on—are sealed off from the racialized culture and power
struggles of their ancestors. The adults have all died and cleared every trace
of racialization from the limited resources they left with the babies. These last
humans share the only surviving culture and power equally. Does that mean
they lose their races?
To answer this, we must peel off a few related questions in our conceptual
centrifuge. Do the babies recognize race? No. Are the babies’ supposed racial
identities relevant to their lives? No. When they grow up, would it be useful
for the babies to one day re-institutionalize the notion of race for themselves?
Arguably not, but we can remain undecided on that. The decisive question
is this: Do the babies stop being Asian or black, simply because the previ-
ously uneven distribution of power and culture has been evened out? Not on
the ordinary concept of race, I believe: on our ordinary concept of race, an
Asian baby doesn’t stop being Asian at the exact moment when the last adult
dies. And that means that racial difference is neither cultural nor political
difference.
Jeffers recognizes the version of this argument as it applies to sociopo-
litical views of race, that is, he agrees that we could achieve power-equality
without that amounting to the end of race. I think we also must recognize
that cultural difference can be redistributed without that amounting to an
end to race. And if the broader analysis is sound, race won’t be found in any
social fact at all. This is how constructionism fails to match the ordinary con-
cept of race: ultimately the only essential ingredients for race, on the ordinary
concept, are certain distributions of visible traits. Power can change hands.
25
252 • W h a t I s  R a c e ?
Culture can come and go. Material well-being, health, education, and any
other socially determined possibility might fluctuate. In the face of all such
social transformations, if we continue to look the way we do now, then on
the ordinary concept of race, race remains throughout. And reasoning in the
opposite direction, if we became enduringly visibly uniform, with everyone
looking like the Dalai Lama, that is when race would disappear.
8.3. Biological Racial Realism’s Mismatch

In Chapter 3, Quayshawn Spencer uses the US Census as a sort of bridge
between the ordinary concept of race and the biological facts, thereby prom-
ising to avoid the Mismatch Objection: the OMB-derived Census categories
are an ordinary concept of race, and the OMB categories match real biological
populations, so race in one ordinary sense is biologically real. To evaluate this
theory, I focus first on the Mismatch Objection and then move on to a con-
cern about what Spencer’s theory means for cross-cultural communication.
In my view, we can have our Census or OMB or any other bureaucratic
categories fit the common-sense concept of race, or we can have them fit the
populations vindicated by biology, but they won’t fit both at the same time.3
Consider how categories like Middle Eastern and Latinx (or Hispanic) are ab-
sent from Spencer’s list of five races. Spencer notes that Hispanic Americans
can usually locate a place for themselves in the OMB system by self-identifying
as white or black in accordance with “their most prevalent ancestry.” Similarly,
3. Early in c hapter 3 above, Spencer points out that ordinary people can be mistaken about
the meanings of their terms and that there is a linguistic division of labor. Just for clarification,
I agree, as explained further in the Appendix (and see Glasgow 2010; forthcoming a; forth-
coming b). Spencer in this passage seems to suggest that my view is that we ask ordinary people
what they think ‘race’ means and that whatever they answer will constitute the meaning of race
in the ordinary sense. He writes: “Notice that limiting what an ordinary race talk is to what
‘ordinary people conceive about race’ implicitly assumes that ordinary people are the correct
people to consult to find out the meaning of the terms they are using” (p. 82 above). In one
sense, ordinary people have to be consulted—how else can we know what terms they are even
using? But in another sense, I would reject that interpretation, because how ordinary people
explicitly understand a term is not decisive in determining the meaning of that term: consulting
ordinary people does not mean that they always correctly account for the meanings or referents
of their terms. As I have suggested, we might even all be wrong about the meaning of some
term! The way I see it, such consultations only generate defeasible evidence as to what the con-
tent of the ordinary concept of race is. This includes, as Spencer himself tries to show, whether
ordinary people defer to experts on a term. In these cases what’s at issue is when the experts
have identified a meaning for our race-talk—race-talk in the sense relevant for our debate. To
figure this out, we have to look at how we use ordinary race-talk; there is no other option (see
also McPherson & Shelby 2004). Consequently, I think that our views on how to analyze racial
or any other terms are not that far apart.
253
Arab Americans can use the OMB system by classifying themselves as white.
But the ability to choose from a menu of options given by the OMB does
not mean that this menu corresponds to the menu that we use or otherwise
recognize in the rest of our lives. Spencer suggests that being able to use the
OMB’s categories on documents guided by the OMB (e.g., the Census, or a
college application) means that people intend to refer to the same object that
the OMB intends to refer to. While this may be evidence that we defer to the
OMB when using OMB-guided documents (and more on that momentarily),
it is not evidence about race as we live it outside of OMB-governed contexts.
It may well be that people just want to fill out the forms the way the form-
designers intended, and that they don’t think it really means anything about
race, in the ordinary sense. And to be sure, there is a gap between the OMB’s
1997 categories and common sense. This gap is so broad that the OMB has it-
self considered changes that would threaten Spencer’s argument. Citing some
of the same data Spencer cites, the OMB (2016) reports the following about
the 1997 Census standard:
Although many respondents report within the race and ethnicity

categories specified by the standard, recent censuses, surveys, and ex-
perimental tests have shown that its implementation is not well un-
derstood and/or is considered inadequate by some respondents. This
results in respondents’ inability and/or unwillingness to self-identify
as the standard intends. For a growing segment of respondents, this
situation arises because of the conceptual complexity that is rooted in
the standard’s definitional distinction of race from ethnicity. Nearly
half of Hispanic or Latino respondents do not identify within any of
the standard’s race categories. . . . With the projected steady growth
of the Hispanic or Latino population, the number of people who do
not identify with any of the standard’s race categories is expected to
increase.
This gap between self-identification and the Census options is why

the Census has considered using an expanded list of options that includes
Hispanic/Latino and MENA (Middle Eastern or North African). This is ev-
idence that the relevant government agencies defer to ordinary usage, not
the other way around.4 In any event, the broader lesson here is clear. If, on
4. The Census Bureau, and separately the OMB, might ask us the question one way and then
reclassify our answers according to their own standard. If this ends up happening, that will be
524
254 • W h a t I s  R a c e ?
the one hand, the OMB is aligned with biological fact, it fails to match the
ordinary concept of race, as others have observed (Compton et al. 2013;
Haney López 2005; Krogstad & Cohn 2014; Navarro 2003). On the other
hand, if the OMB chooses to align with (what it recognizes as) common
sense, then it will no longer represent something biologically real, since bi-
ology has not vindicated a set of seven races that includes both Latino and
MENA. The data from Rosenberg et al. (2002) yield as the sixth ‘race’ the
Kalash—a small population in Pakistan that has experienced reproductive
isolation from other groups. Once we go to seven categories, Li et al. (2008)
found this set: {African, Middle Eastern, European, Central/South Asian,
East Asian, Oceanian, and Indigenous American}. I have not seen any bio-
logical data of sufficient quality that demonstrate the existence of a Hispanic/
Latinx population that exists at the same level of partition as other purported
races. Whether common sense matches biology floats free of whatever is on
the Census.5
Now, as I wrote in Chapter 4, some might be willing to give up one or
another identity (such as Latinx/Hispanic) in order to preserve the biolog-
ical reality of race. But there are other complications on the different groups
front, not about which identities we invalidate but about what people be-
long to which groups.6 The biological populations revealed in the biological
data have wide bands of fuzzy boundaries. One consequence of this is that
many individuals will need to be slotted into a race not simply by their an-
cestry, but also by us choosing what fraction of which ancestry is sufficient
for which racial membership. Various sorting principles are open to us. For
example, if we were to say that having any of a population’s ancestry made one
eligible for membership in a population—as Spencer suggests regarding black
ancestry—then most people in Spencer’s data (his Figure 3.4 in Chapter 3)
evidence that the OMB is not deferring to ordinary usage; that would not tell us one way or
the other whether ordinary language users defer to it, though. I sometimes hear people say that
mismatch arguments won’t work for reasons having to do with semantic deference and direct
reference theory. For an explanation why this is not true, see the Appendix to this chapter and
Glasgow (2010; forthcoming a; forthcoming b).
5. Shortly before this book went to press, it was decided that the 2020 Census would not add
Hispanic/Latino or MENA. It remains an open question whether this decision will be revisited
in a decade. Ultimately, though, what the actual OMB or Census Bureau practice is doesn’t
really matter. The fact that we could conceivably change our Census categories in this way is
enough to highlight the key point: our government categories can map onto common sense, or
they can map onto biology, but either way what remains to be seen is whether common sense
and biology line up.
6. I am grateful to Shani Long Abdallah for helpful discussion on these points.
25
are going to be eligible for membership in more than one race (Atkin 2017,
146).7 If we instead used the criterion he says Hispanic Americans follow—go
with your most prevalent ancestry—then people will get classified differently.
These sorting criteria have different implications for Spencer’s theory of race.
Either the number of people in just one race is substantially smaller than we
currently recognize; or a large subset of people with mixed ancestry will be
sorted into races other than the one that we ordinarily assign them to and that
they ordinarily identify with; or a large number of people will have no race at
all—which is not to say that they have a race that’s hard for us to identify, but
that they literally are raceless.
Now one worry is that these options all flirt with the Mismatch Objection.
Ultimately, though, Spencer can insist that some revisions to common sense
are acceptable and do not rise to the level of violating the very definition of
‘race.’ (Basic racial realism essentially says this about the first option—that we
have memberships in many races, only some of which we pay attention to.)
How much, and which, revision we can allow until we depart from the ordi-
nary concept of race is a matter of interpretation, as this book has shown. But
my main concern at the moment is that it is not clear how Spencer’s theory
secures the biological credentials of race, if this account of race selects a sorting
principle in an extra-biological way. That is, whatever choice we make with
these criteria, the point is that it’s a choice, and it is hard to see where biology
is what is dictating the choice. If multiple ways of sorting people with mixed
ancestry are equally well credentialed by good biology (using Spencer’s cri-
teria for what makes something biologically real), then we’re choosing which
sorting principle to use in a way that is arbitrary from the perspective of the
biological sciences—in which case, where is the biological, scientific backing
for this?
Spencer might make a move here. He might say: well, all that I really care
about is that the choice plays some role in good biological science (or some-
thing like that—I’m blurring important details in his view in the hopes of
painting a broader picture). This courts questions all its own, though. If bi-
ology can only best use one sorting principle, then we need to know how this
will get all the right results—roughly matching both common-sense and bi-
ological assignments, via the OMB’s categories—for Hispanic people, black
people, people from across South and Southeast Asia, and so on. I believe
7. Spencer clarifies in the beginning of Chapter 7 that this is indeed the principle he adopts for
all races.
526
256 • W h a t I s  R a c e ?
that has not yet been demonstrated. If, instead, multiple principles are equally
usable for biology—most prevalent ancestry, any ancestry, etc.—the questions
multiply. We still need to see that match. And also we now need to know
what to do when more than one sorting principle applies to us. Which wins,
and why? Finally, isn’t it more sensible to say that, if multiple principles are
consistent with biology, then they are not biological principles? Consider
a politically disputed territory, like Northern Ireland or Alsace-Lorraine.
Biology doesn’t care which nation-state these lands belong to: any resolu-
tion to the political dispute is consistent with any biological facts. It seems
like having multiple biologically consistent principles for sorting people into
races, like sorting territories into countries, means that we’ve left the realm of
biological fact.8
Finally, again, there seems to be a distinctive gap between the concepts of
race and population. Most obviously, genealogical groups are not required
to stay visibly distinct, unlike races. Spencer objects to this argument on the
grounds that it misunderstands the relevant concept of race: two distinct
races, such as Pacific Islanders and black Africans, can share the same vis-
ible traits, he argues. But as he acknowledges elsewhere (Spencer 2015), the
relevant visible traits are not shared to the exact same degree. And all that
is required for racial distinctiveness on the ordinary concept of race is some
differentiation in visible traits, not necessarily a lot of differentiation (Alcoff
2006; Pierce 2015, 109). Spencer evidently disagrees with this judgment.
Readers must in the end determine whether they think that two groups that
are perfectly identical in visible traits could be racially distinct.9
Let’s now pivot from the Mismatch Objection to a separate worry. Recall
an observation from Chapter 4: semantic meaning sets a conversational
boundary, such that if we use one word with different meanings, we are
not truly dialoguing. We are not even disagreeing. We are just talking past
each other. As Mark Sainsbury (2014, 4) puts it, “Substantive disagreement
requires agreement in meaning. There needs to be some proposition that one
8. This is a variation on Philip Kitcher’s (2007, 304–306) point that in order to carve humanity
into races, we have to make an extra-biological choice about how many races to look for, such as
Spencer’s preferred K = 5 level. (Kitcher understands this choice to come down to pragmatics.)
I’m suggesting here that in addition to wondering how many races to look for, we also have
to figure out how to sort individuals into whatever number of racial groups we select. Both
choices appear to happen outside the realm of principled biological science.
9. And recall a related problem: if Roberta Millstein (2015) is right, we can move between bio-
logical populations in a way that we cannot move between races, namely by reproducing with
people in the new group.
257
party affirms and the other denies.” You instruct me to cut the deck of cards,
by which you mean that I am to take some cards off the top of the deck and
place them next to the bottom portion. But I don’t know that this is what
you mean, so interpreting you as best I can, I get out my scissors and start
cutting each card in half. If we were then to get into a heated discussion about
whether I cut the deck of cards, we don’t agree about how to cut a deck of
cards, but we don’t disagree, either. We simply miscommunicate. There is no
proposition that you affirm and I deny. We talk past each other. Quite unin-
tentionally, I simply mean something different by “cut the deck” than what
you meant. Our attempt at a conversation failed.
On Spencer’s view, the relevant meaning of ‘race’ is the set of five races
that the OMB recognized in 1997. This entails that when Americans op-
erate with this race-talk, they cannot communicate about race with people
who use other sets of categories. (And similarly for those many countries
that have no racial categories on their censuses.) Instead, on Spencer’s view,
Americans simply talk past people in those other countries when using the
language of race. To choose just one example,10 Canadians can select Arab,
Latin American, Filipino, and Chinese on their Census, whereas those in the
United States cannot. So if the meaning of the term ‘race’ were OMB-derived
Census racial categories, there would be no meaning of ‘racial’ that crosses the
US-Canadian border. In that case, we couldn’t even say that Canada and the
United States recognize different racial categories. Spencer’s semantic assess-
ment blocks us from comparing the two sets of categories at all—it removes
any words we might use to do so. There’s just race-in-the-US-sense and race-
in-the-Canadian-sense, and there’s no translating between the two senses of
‘race’ any more than in the ‘cut the deck’ case. Canadians and Americans lit-
erally would be unable to talk with each other about race, on this account. This
has a particularly peculiar implication for us. As a Canadian, when Chike
Jeffers talks about race, he would mean something different than Spencer,
Haslanger, or I (as US citizens) do, which would make our book akin to the
kind of miscommunication in the ‘cut the deck’ exchange. But this seems to
be the wrong diagnosis. Our book is not a gigantic exercise in miscommuni-
cation. Unlike the ‘cut the deck’ exchange, Jeffers and Spencer are not talking
past each other. They are genuinely dialoguing! To do so, they must share
some common meaning of ‘race.’ They might not know what that common
10. For more, see http://www.understandingrace.org/lived/global_census.html.

528
258 • W h a t I s  R a c e ?
meaning is, and they certainly disagree about it, but it must be one that is
shareable.11
Temporal shifts present a similar problem. Spencer’s view implies
that it is impossible for us today (when using OMB race-talk) to reflect
on and evaluate theories of race that were written prior to 1997. On this
view, Taylor (2000), Sundstrom (2002a; 2002b), Blum (2002), Glasgow
(2009a), and Haslanger (2012) are not disagreeing with Appiah (1985,
1996), Zack (1994), or Outlaw (1996), contrary to their own claims that
they are disagreeing. In fact, Spencer’s view means not only that those
race theorists are not in a dialogue with one another, but also that Appiah
(2007) and Zack (2002) are not even speaking continuously with their ear-
lier selves. This is a problem. We take these discussions to be continuous—
we think that post-1997 race theorists talk with pre-1997 race theorists. If
we take Spencer’s theory to be a theory of the ordinary concept of race, we
make this conversational continuity impossible. Similarly, if Americans add
Hispanic to the relevant lists of races in 2030, they will again be replacing
the current meaning of ‘race’ with a new one, according to Spencer’s anal-
ysis.12 In that case, in 2031 Americans will not able to react to 2029 writings
about race. This, too, is jarringly counterintuitive. In short, Spencer’s anal-
ysis renders us unable to unequivocally talk about a single thing across that
moment in time, not even to disagree.
These points suggest that we should not let any given set of racial
categories—Census or otherwise—constitute the very meaning of the term
‘race.’ Americans can react to what previous generations said about race, and
they can talk with Canadians about race, regardless of what particular racial
categories their censuses recognize. Spencer’s approach to the semantics of
race would mean that we end up equivocating on the word ‘race,’ if we try to
have a conversation about race across different racial classification schemes.
His view turns what is clearly a dialogue into a series of covertly isolated
monologues.13
11. To clarify, I’m still focused on the ordinary concept of race in the United States. I’m just
claiming that one feature of that concept is that it is supposed to be consistent with widespread,
though perhaps not universal, cross-cultural comparison and dialogue.
12. Or, for that matter, consider that “Mexican” was a US Census race category in 1930
(Gratton & Merchant 2016).
13. It might be tempting to deviate from Spencer’s view and say that ‘race’ just refers to his five
populations without ‘race’ actually meaning those five populations. That would avoid the mis-
communication problem, but it would revive the mismatch problem.
529
Although I don’t think that a specific set of racial categories constitutes the
very meaning of ‘race,’ I do think that we can use our racial categories as evi-
dence about what ‘race’ means. But even then we have to be careful about what
this evidence shows. ‘Race’ is supposed to be usable in both the United States
and Canada and indeed much of the world. It is supposed to be something
talked about both before and after 1997. These phenomena suggest that it is
defined in such a way as to be consistent with multiple systems of racial classi-
fication. It is supposed to apply to two siblings in Brazil, who get categorized
into different races because they look different, even though they have iden-
tical ancestries; and it is supposed to apply in the United States where ancestry
is given more weight; and we are supposed to be able to compare these sys-
tems of classification (Chen et al. 2018; Fish 2011). To make such cross-context
communication and comparison possible, rather than tying the meaning of
‘race’ to one highly localized set of racial categories, we need a definition of
‘race’ that holds across all of these contexts. If you’ve made it this far, you won’t
be surprised to see me say that one such solution is to define racial groups as
groups of humans that are supposed to be visibly distinctive in certain ways.
That said, I don’t doubt that people sometimes use ‘race’ in special
ways. The OMB may well have a specialized definition of ‘race’—though
given its writing and practice, I suspect that the OMB actually semantically
defers to ordinary usage, sometimes misdiagnosing it in the process.14 We
might also find specialized definitions at an Atlanta church, a forensics fa-
cility in Chicago, an immigration center in Los Angeles, or a conference of
anthropologists in Mexico City. And ordinary people in ordinary contexts
might well carry around their own “portfolios” of racial schemas, deploying
different schemas in different contexts (Roth 2012). Sometimes we just stip-
ulate, “What I mean by this word is . . .” In the process we change the sub-
ject from the mainstream conversation and put our little conversation into a
silo. But in addition to specialized, siloed understandings that have only local
range or authority, ordinary people also have many conversations about race
where they are not talking past one another or changing the subject or using a
cloistered language, where they can talk across a multitude of social contexts
and a variety of classification systems. This broad, social conversation uses the
most common ordinary concept of race.
14. Not only was the US Census considering adding Hispanic and MENA in response to ordi-
nary use, but also starting in 2000 the Census began allowing individuals to choose multiple
races to align with complex and mixed heritage, as a response to political activists highlighting
the complexity in everyday identities.
620
260 • W h a t I s  R a c e ?
This suggests a more ecumenical way to think about Spencer’s analysis. If

it is meant to capture racial discourse only when certain people use OMB-
directed documents (as Spencer intends it to be); and if, contrary to what
I have argued, the OMB is not interested in capturing our ordinary discourse
about race; then this view might be more insulated from the Mismatch
Objection—it’s just a specialized language with very restricted application. In
this case, though, it would not capture the more extensive concept of race that
stretches across cultures and moments in history and specific classification
systems. It would not capture the concept of race that most of us use most
of the time. In short, it would not be very relevant to our lives. (Except every
once in a while when you fill out a form, and only if you’re not just playing
the bureaucrat’s game, and only if the bureaucrats weren’t themselves trying
to capture something relevant to the rest of our social lives . . . and even then
it is unclear that it will be a truly biological concept at work.) To speak to
the broader cross-context use of ‘race,’ that is, to the discourse that has done
tremendous damage and that demands that we make sense of it and that
motivates my inquiry, at least, we must recognize a different concept of race,
one that is not defined by a specific and local set of categories but that instead
has spanned the centuries and the globe.
In my view, this concept of race has races being groups that have certain
visible traits to a disproportionate extent. As it happens, it’s hard to show that
this concept is biologically fruitful. But it is worth noting how flexible it is.
It can account for individuals who have atypical traits for their race—while
the racial group must have a distinctive pattern of traits to count as a race on
this analysis, it is not the case that every individual in the group needs to have
some specific set of visible traits (Glasgow 2009a: 79; Hardimon 2017: §2.3;
cf. Mills 1998). (This is what makes ‘passing’ possible.) It can explain why
we recognize different cultures’ varying sets of racial categories as racial.
And it can account for the fact that different people can have very different
representations of race while also having meaningful conversations about race
with one another. Capturing all of those facts at the same time is, in my view,
the main advantage that the visible-traits analysis has over Spencer’s OMB-
based semantics of ‘race.’
8.4. The Racial Categories with Which We Live

Soon after Justice Sutherland wrote that Bhagat Singh Thind was not ra-
cially white, even if he was biologically Caucasian, the United States forcibly
deported or “repatriated” over one million people to Mexico. Roughly half were
216
US citizens, and for many this racialized “repatriation” was their first time on
Mexican soil (Balderrama 2015). Three decades later, in 1954, riding the same
civil rights wave that crested with Brown v. Board of Education, the Supreme
Court took up the question of Mexican identity in Hernandez v. Texas. For years
Mexican defendants faced all-white juries in Jackson County, Texas. The state
defended this practice on the grounds that Mexicans were recognized as white.
Chief Justice Earl Warren responded by writing that Mexicans were treated as a
group apart and therefore were their own class, legally speaking:
The petitioner’s initial burden in substantiating his charge of group dis-

crimination was to prove that persons of Mexican descent constitute a
separate class in Jackson County, distinct from ‘whites.’ One method by
which this may be demonstrated is by showing the attitude of the com-
munity. Here the testimony of responsible officials and citizens contained
the admission that residents of the community distinguished between
‘white’ and ‘Mexican.’ The participation of persons of Mexican descent in
business and community groups was shown to be slight. Until very recent
times, children of Mexican descent were required to attend a segregated
school for the first four grades. At least one restaurant in town promi-
nently displayed a sign announcing ‘No Mexicans Served.’ On the court-
house grounds at the time of the hearing, there were two men’s toilets,
one unmarked, and the other marked ‘Colored Men’ and ‘Hombres
Aqui’ (‘Men Here’). No substantial evidence was offered to rebut the log-
ical inference to be drawn from these facts, and it must be concluded that
petitioner succeeded in his proof. . . .
Circumstances or chance may well dictate that no persons in a cer-
tain class will serve on a particular jury or during some particular pe-
riod. But it taxes our credulity to say that mere chance resulted in there
being no members of this class among the over six thousand jurors
called in the past 25 years. The result bespeaks discrimination, whether
or not it was a conscious decision on the part of any individual jury
commissioner.15
Another six decades later, in late 2016, the Supreme Court again heard a dis-
crimination case in Peña Rodriguez v. Colorado. One notable current running
underneath the oral arguments was that the Justices seemed to grant that it
was obviously a case of racial discrimination—“the best smoking-gun evidence
15. 347 U.S. 475 (74 S.Ct. 667, 98 L.Ed. 866).

26
262 • W h a t I s  R a c e ?
you’re ever going to see about race bias in the jury room,” as Justice Elena Kagan
put it (Barnes 2016). (The dispute was whether the significance of racial discrim-
ination overrides other legal principles at stake, not whether there was any racial
discrimination.)
These cases illustrate a fact repeatedly borne out not only by studies done
on the Census, but also by other sociological research (e.g., Roth 2012), by our
day-to-day practices, and by a mountain of anecdotal data: Latinx people are reg-
ularly treated and self-identify as a racially distinct group. We need a theory of
race that can account for this and the full range of racial practice. There is a sense
of ‘race’ that denies whiteness to Vaishno Das Bagai and Bhagat Singh Thind.
It also structures blackness differently than the biological population ‘African.’
It forbids people from changing races merely by visiting another society or by
reproducing with someone from a different group. It requires racial groups to
bear some distinctive portion of visible traits. If we’re trying to understand the
sense of ‘race’ with which we live, we must identify something that captures these
experiential realities and conceptual constraints.
It is these lived categories that invigorate our discussion. They are the cat-
egories where we hoard resources. They are where meaningful identities are
found. We discriminate. We repatriate. We go to war. We change lives, we
improve lives, we damage lives, and we end lives, all based on racialization.
The fences of racialization are meant to wall off biological traits like skin,
hair, and eye color. Because these traits are biological, our categories are going
to reflect our ancestors’ migrations, separations, and reproductions. But the
reflection has all the fidelity of a funhouse mirror. Ancestral patterns that
figure in biological explanation run through different valleys than those that
cradle the racial categories with which we live. We racialize humanity in ways
that float free of what biologists do. And in this lies a mistake in ordinary
race-thinking that cannot be repaired by simply paying better attention to
the science: we apparently believe that there is some fact about us that can be
read right off of our skin. In turn, that belief filters how we treat one another,
both across and within the fences of racialization, resulting in weighty and
disturbing effects on life chances.16 When disadvantage lasts the full course
of your existence; when stress and heart disease reflect not a bump on your
journey but rather an integral feature of the life-paths made available to you
by a white supremacist social architecture; when the prices paid are death for
16. Social scientists continue to document not only that we racialize but also that we treat
individuals differently within their racialized groups based on visible traits. For recent work
on colorism and ‘pigmentocracy’ see, e.g., Bailey, Saperstein, and Penner (2014); Monk (2014,
2015, 2016); Telles (2014).
236
many, a lower ceiling on reasonable hope for more, and unjust relations for
all; and when this whole moral mess stubbornly attaches to categories un-
broken by multiple generations of scientific correction—when that is what
needs explaining, the chasm between biology and social practice yawns wider
than our best bridges.
This is where we find the constructionist’s truth: we need to decode
the concept of race that we utilize in everyday life. As a separate matter,
constructionists also conclude that our ways of using and abusing this concept
can conjure race out of nothing: our practices, our identities, our power hier-
archies, and our cultures are said by constructionists to be the very building
blocks of race. Against this we also must take seriously the truth in biological
racial realism: those social facts are not what we are purporting to talk about
when we talk about race. What we are trying to do, a task that we seem to have
charged to the very concept of race, is to find something biological. And yet
that thing keeps eluding the biologists.
The traits that are supposed to distinguish the races—those visible traits
like skin color—are bequeathed to us by our ancestors and made significant
by social forces. But as we have seen, the sense of ‘race’ with which we live is
defined neither by genealogy nor by social ties. Which leaves us with two
choices. ‘Race’ in the sense with which we live either tries to latch onto bio-
logical kinds that end up being illusory, or onto a more basic, non-social, and
non-scientific kind of thing.
Appendix on Philosophical Method: Using Intuitions

about Fictional Cases
The arguments I have made use certain methods that I want to briefly de-
fend. In particular, I have repeatedly called on thought experiments to gen-
erate a partial ‘definition’ of ‘race,’ where a definition here is, roughly, a set
of properties that something must have to be what our term ‘race’ refers to.
These properties include, on my arguments: being a group of humans, being
organized by certain visible traits (as a group, not for each individual), and
being non-social. To get to this list of properties, I have imagined worlds
where there are no adults or where we all look like the Dalai Lama, socie-
ties that are fully culturally integrated, and situations of perfect equality and
mass amnesia. How we judge such possibilities—whether or not we say that
races would exist in these worlds—is pivotal in identifying what we mean by
‘race.’ I claim that our intuitions or judgments about these hypothetical cases
suggest that the ordinary concept of race requires both that race persists even
624
264 • W h a t I s  R a c e ?
when social facts change and that races are organized according to visible
traits rather than genealogy, calling into question the rival views presented
in this volume.
Using intuitive reactions to thought experiments is a widely accepted
method for analyzing concepts, and race theory is no exception (e.g., Appiah
1990; Mills 1998; cf. Glasgow et al. 2009). But as it happens, my three
coauthors have all taken issue with this method here and in other places.
And some readers may have similar hesitations. How relevant are outlandish
thought experiments—don’t we care about the real world rather than imag-
inary worlds? And why do we care about our intuitions, anyway—don’t the
facts of the world override whatever we might think our words refer to, like
when we mistakenly think that fool’s gold is real gold? And who am I—who
is any individual—to decide whether we would say that race exists in these
fictional worlds?
In a previous exchange, Jeffers (2013a, 2013b) objects to my use of fic-
tional thought experiments. These objections, along with some of Haslanger’s
comments in Chapter 5 here, suggest a question: How can a situation about
some possible world tell us anything helpful about race in this world?
The answer, in short, is that fictional scenarios can tell us about the limits
of our real concepts. These thought experiments do not speak to our biology,
of course. Instead, they reveal how words can and cannot attach to the world
around us. Sometimes we learn about conceptual possibilities through actual
experiments. We confirmed that it was conceptually possible for atoms to
be divided when Rutherford split the atom a century ago. We didn’t con-
tinue insisting that atoms were by definition indivisible, in which case he
must have split something else, the shmatom; instead, we called it an ‘atom,’
indicating that it was possible to coherently talk about a divisible atom. In
similar ways, real-world discoveries taught us, to our surprise, that whales are
not actually fish and that fool’s gold is not gold: real-world information told
us what our terms actually refer to. For the purposes of mapping conceptual
boundaries, there is no relevant difference between Rutherford’s actual ex-
periment and a fictional thought experiment: possible and actual cases allow
us to identify when a term can and cannot be successfully applied. To use
an example of an unrelated concept, many people believed that an agent is
morally responsible for some action she performed only if she had had the
ability to not perform that action. Then Harry Frankfurt (1969) presented
a fictional case of an agent who could not have done anything other than
what he did, but where his responsibility seems undiminished. We don’t
need to agree with Frankfurt here (and, to be sure, the ensuing discussion
265
has not been univocal). The point is merely that his method is legitimate: it
tests when we will and will not deploy the concept of responsibility. And
of course this example is just the tip of a very large iceberg. We use thought
experiments, often far removed from the actual world, to map the contours
of our concepts.
Because these cases test conceptual possibility, there is nothing wrong
with them being (very) unlikely. In fact, Jeffers deploys one of the fictional
scenarios that I use : the possibility—the goal—of social equality between
races means that the concept of race cannot be defined by power inequality
(Glasgow 2009a, 120). A world of ideal racial equality is no less fictional than
the other cases I’ve imagined—it is very far removed from the real world—
but nothing is wrong with using it to question whether political theories of
race are consistent with the ordinary concept of race. The same goes for any
other relevant thought experiments. These fictional stories are supposed to
tell us when we are, and are not, willing to use a term, which provides evi-
dence as to what the term could and could not refer to.
Once we open the door to thought experiments, how should we judge
them? We now know that our intuitions sometimes are not probative—they
can fail to be representative, and even if they are representative, what they rep-
resent may be the product of cognitive distortion (e.g., Machery et al. 2004).
In this spirit, Spencer (2015, 52) has objected that marshalling one’s own
intuitions about thought experiments is “merely clarifying an idiosyncratic
idea of race.” (He elaborates on this objection at the end of Chapter 7; and
in Chapter 5 Haslanger questions whether some interpretations of thought
experiments might be local to certain subgroups of the broader linguistic
community.)
I agree that we should consult data from other sources besides just
ourselves—ideally we would use a wide variety of tests to systematically gather
intuitions from a large and representative sample of people who competently
use the term ‘race’ to help establish what ‘race’ means (Glasgow 2009a, Ch.
3–4). But at the same time, we can reasonably have different degrees of con-
fidence about whether some of our own intuitions are representative; and
intuitions from one person can at least generate predictions that are starting
points for systematic experiments. After all, most theorists aren’t that iso-
lated from real life. If our intuitions turn out to be wildly inconsistent with
the reactions of other ordinary language users, we will need to explain the
discrepancy. Moreover, recall that we should not overestimate the value of
systematic and representative data: the best, largest, and most representative
study on how we use a term only delivers evidence, not conclusions, about
62
266 • W h a t I s  R a c e ?
the contents of our concepts. Even a representative and unanimous poll can
reveal mistaken results about the meanings of our words, as we would find if
we traveled back in time and asked whether whales are fish.17
As I see it, the tricky step is not in using your own intuitions, as long as
we are modest about what those intuitions might show. Every method uses
someone’s intuitions somewhere. Haslanger classifies Aung San Suu Kyi as
Asian; Jeffers judges that there can be politically equal races; Spencer makes
claims about how much visible-trait overlap is required to be included in a
race. These are intuitions, and they are unavoidable in philosophical anal-
ysis. The trick isn’t to avoid intuitions; rather, the trick is deciding what to
do when inconsistent intuitions result from various tests of a concept. (One
reason why we debate race is that we tap into inconsistent elements of ordi-
nary race-thinking.) Sometimes one individual’s intuition is idiosyncratic, as
Spencer notes. But sometimes it is the canary in the coal mine, indicating
a rift in our conceptual framework that requires eliminating a much more
widely held belief, such as, way back when, the belief that the earth is flat, or
that whales are fish, or that races have biobehavioral essences.
So how should we bring inconsistent intuitions into equilibrium?
There is no formula to apply in advance to decide which of our conceptual
commitments is stronger in any particular case. As I argued in Chapter 4,
all we have is conversation and analysis, and that starts with individual judg-
ment. There is no one person who represents the ordinary point of view or
represents all language users. As I also said at the outset of Chapter 4, an in-
credibly diverse set of people are part of the conversation about race, and our
job is to account for the concept that makes that conversation possible. No
one intuition occupies a privileged position in this conversation, but then no
intuition is irrelevant, either.18 Our task is to systematize these judgments.
17. Relatedly, for contested categories like race, intuitions might be unstable (Atkin 2017, 142).
This might mean either that there is no concept of race, or that we are working with multiple,
shifting concepts of race, or that the content of the concept is underdetermined.
18. Readers might wonder whether different self-identified racial groups come up with dif-
ferent understandings of race. The evidence does sometimes bear this out, such as one study
that found that whites scored higher than non-whites on a measure of whether individual
racial identity is socially determined (Glasgow et al. 2009). Again, though, the conclusion
from these differences is not—or at least not necessarily—that different folks have different
concepts of race. After all, another key data-point is that people across these different groups
seem to be having coherent conversations about race. What in combination this suggests is that
we need to dig deeper to identify the concept of race behind those conversations, the minimal
commitments required to make those exchanges possible.
267
For these reasons, I think we are warranted in rendering provisional, tenta-

tive, and guarded conclusions based on our (speculative) judgments about
what the intuitive thing to say is when confronted with a surprising thought
experiment.
Haslanger (2012) also questions the use of intuitions when addressing so-
cially pressing phenomena like race. Now here again there is a strong con-
sensus worth emphasizing. The field has benefited from Haslanger insisting
that good race theory must respect several key principles, including the
following:
• Semantic externalism: meaning or reference can be partly determined by

the physical world and social context; consequently, intuitions alone are
not always sufficient to determine what a term refers to.
• The division of linguistic labor: within certain constraints, we defer to some
experts to tell us what some of our terms refer to.
• Scientific essentialism: some objects have essences tracked by our words.
• Terminological opacity: we do not always know what our words mean or
refer to.
• Ambiguity and context: words often have multiple meanings, and which
one is operative depends on conversational context.
I believe that all sides in the race debate can, should, and (usually)
do agree with this analytical framework. Although I am sometimes
interpreted by others as disagreeing with one or another element of this
framework, I actually assume it throughout my arguments. The main ques-
tion is where the data point us within Haslanger’s framework, not what
framework to adopt.
Haslanger and Spencer emphasize that we can be wrong about what our
terms refer to, be it ‘water’ or ‘cancer,’ ‘atoms’ or ‘werewolves,’ or ‘race.’ This
is true, obviously. But that point leaves intact our main question: At what
point in revising our usage of the term ‘race’ have we changed the subject with
‘race,’ away from race in the ordinary sense and toward something else (per-
haps something better)? While sometimes we can fix our mistakes without
changing the subject—when our words are magnetically drawn to a surprising
referent—other times we do change the subject, and our words simply fail to
refer. Marathons aren’t races in the relevant sense, and talking about people
with rabies is not really talking about werewolves in the ordinary sense. This
is Haslanger’s point in requiring even the most surprising interpretations of
268
268 • W h a t I s  R a c e ?
‘race’ to converge on sameness of reference and shared linguistic and epistemic

practice. 19
What we are looking for here is the boundary of the tradition we are trying
to interpret, ordinary racial discourse. To identify this boundary, we can’t just
want or try to refer to the same thing, and we can’t just have an interest in
referring to the same thing. Marco Polo had an interest in trying to refer to
the same thing that a community of East Africans was referring to with the
term ‘Madagascar.’ He famously failed: while that community was using it to
refer to part of continental Africa, he mistakenly used the name to refer to
that big island off the coast. His mistake stuck, and so the referent of the term
eventually shifted; but in the short term, shared desires and interests were not
enough to help Polo and his interlocutors refer to the same thing. To succeed
in co-referring, we also have to have sufficiently overlapping usage. Otherwise
we talk past one another.
Even when we defer to the experts to tell us what our folk terms refer to,
the experts must leave a portion of semantic authority in ordinary people’s
hands: folk intentions about what their terms refer to limits the range of what
the experts can legitimately identify as the referent of those terms, if those
experts are to deliver the referent not just for a cloistered, specialized language
but for ordinary discourse. If experts start claiming that the ordinary term
‘table salt’ refers to H2O rather than NaCl, they will have changed the subject,
because ordinary usages of ‘table salt’ limit the range of what ‘table salt’ might
refer to—it has to be something like the white-ish, granular, salty-tasting stuff
that we season our food with, if we’re talking about table salt in the ordinary
sense (Glasgow 2010).
So when do we overlap? When do we stay on subject? What determines
when we succeed in sharing meanings and referents for our terms? How do
we know when someone is talking about race in the relevant sense, and when
they have started a cloistered conversation using the same old words but with
new meanings?
19. Herein lies my concern with Haslanger’s ‘person’ analogy in Chapter 1. The case she
describes is one in which we have false definitions of ‘person,’ and so rather than becoming
anti-realists about persons, we would just give up those false definitions and try to find a better
way of capturing our (opaque) definition instead. In my view the case of ‘race’ is different, and
the analogy buckles: it’s hard to give up defining ‘race’ in terms of visible traits (rather than
social facts) without it seeming like changing the subject. This is what pressures us to become
anti-realists about race but not personhood. In this respect, ‘race’ is more like ‘witch’ or ‘were-
wolf ’ than ‘person’—the apparently accurate definition of ‘race’ is inconsistent with reality.
269
The key factors are the ways we would and would not be willing to use
the terms in question—that is, how we would use the word when presented
with surprising (actual and counterfactual) cases (Glasgow forthcoming a).
The only indicator that we’ve made a mistake about what one of our words
means is some other semantic intuitions and intentions. The only way that
‘water’ or ‘gold’ or ‘cancer’ or ‘race’ turns out to refer to something surprising
or counterintuitive is if we are willing to call that surprising referent ‘water’
or ‘gold’ or ‘cancer’ or ‘race’—that is, when another intuition or intention says
that the referent is a surprise. One set of intentions and intuitions vetoes an-
other set, but whichever set prevails, either way our intentions and intuitions
fix what our terms can refer to. These semantic intentions just are among the
inputs about race that Haslanger asks us to consider. Haslanger’s input, “Aung
San Suu Kyi is Asian,” is an intuition about how those words can be used.
Whether that input is secure (or non-negotiable) depends in part on our
other commitments about how to use those words.
Elsewhere, Haslanger (2012, 305–307) holds that rather than appealing
to intuitions about various cases, we should instead identify the meaning
of ‘race’ by looking for paradigm cases and seeing what unifies those cases.
I believe this is a choice we don’t have to make. Really, we can’t make it (cf.
Chalmers 2011, 538–539; Glasgow 2009b). We cannot look to “what the
cases in fact have in common” (Haslanger 2012, 396) until we know which
cases are the relevant cases to consider, and identifying those cases is a job for
intuition. That is, to identify paradigm cases is to use intuitions, as we see in
the externalist tradition to which Haslanger belongs—see Burge (1979) on
‘arthritis,’ Kripke (1972) on ‘Gödel,’ Putnam (1975) on ‘water.’ We need to say
that a coffee cup does not count as Asian and that Aung San Suu Kyi does;
these are just intuitions.
Happily, I think the foregoing just expands the consensus: we can add to
Haslanger’s analytical framework that we must (somehow) consult intuitions
and intentions when identifying what our terms refer to and mean. At that
point the difference-making questions become which cases we consider; what
intuitions we have in those cases; whose intuitions count in rendering verdicts;
and how to systematize the intuitions when they are collectively inconsistent.
In this way, while we have some lingering disagreements about whether we
can and should analyze racial terms using descriptive definitions like I am
doing here (Haslanger 2010; Glasgow 2017, forthcoming b), I think we actu-
ally occupy a vast methodological common ground.
So as I see it, the choices that make a substantial difference to an-
swering this book’s question are relatively limited. In particular, the issues of
270
270 • W h a t I s  R a c e ?
dispute are usually not methodological. The live choice is not whether to use
intuitions or paradigms. Nor is it whether to use actual or merely possible
cases as evidence of what we mean by ‘race.’ Nor is it whether to go with one
person’s intuitions or a rigorous survey of many people’s intuitions. Nor is
it a choice between different theories of meaning or reference. We can and
must use all of the tools and evidence at our disposal. The difference-making
choice, as I see it, is how to reconcile the rival sets of clashing intuitions and
semantic intentions that we find ourselves with—different sets of judgments
that have been marshalled by each of the four authors in this book. When
faced with these incompatible judgments, the project is to identify which
ones we hold on to and which ones we compromise on, to find those core
commitments so that we can try to find out whether there is something in the
world that captures them. To determine what race is, the first question will
always be: What do we mean by ‘race’?
References
Alcoff, Linda Martín. 2006. Visible Identities: Race, Gender, and the Self. Oxford: Oxford
University Press.
Appiah, K. Anthony. 1985. “The Uncompleted Argument: Du Bois and the Illusion of
Race.” Critical Inquiry 12: 21–37.
Appiah, K. Anthony. 1990. “But Would That Still Be Me?” The Journal of Philosophy
87: 493–499.
K. A. Appiah and A. Gutmann (eds.), Color Conscious: The Political Morality of
Race (Princeton, NJ: Princeton University Press), pp. 30–105.
Appiah, K. Anthony. 2007. “Does Truth Matter to Identity?” In Jorge Gracia (ed.),
Race or Ethnicity: On Black and Latino Identity (Ithaca, NY: Cornell University
Press), pp. 19–4 4.
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.
Bailey, Stanley R., Aliya Saperstein, and Andrew M. Penner. 2014. “Race, Color, and
Income Inequality across the Americas.” Demographic Research 31: 735–756.
Balderrama, Francisco. 2015. Interview with Terry Gross, Fresh Air, “America’s Forgotten
History of Mexican-American ‘Repatriation.’ ” http://www.npr.org/2015/09/10/
439114563/americas-forgotten-history-of-mexican-american-repatriation
Barnes, Robert. 2016. “Supreme Court Hears Case Concerning Biased Comments in
Jury Room.” The Washington Post, October 11. https://www.washingtonpost.com/
politics/courts_law/supreme-court-hears-case-concerning-biased-comments-in-
jury-room/2016/10/11/f82de46c-8f2a-11e6-9c52-0b10449e33c4_story.html.
Accessed October 21, 2016.
217
Blum, Lawrence. 2002. “I’m Not a Racist, but . . .”: The Moral Quandary of Race. Ithaca,
Burge, Tyler. 1979. “Individualism and the Mental.” In P. A. French, T.E. Uehling, and
H. Wettstein (eds.), Midwest Studies in Philosophy, IV. Minneapolis: University of
Minnesota Press, pp. 73–121.
Chalmers, David J. 2011. “Verbal Disputes.” Philosophical Review 120: 515–566.
Chen, Jacqueline M., Maria Clara de Paula Couto, Airi M. Sacco, and Yarrow Dunham.
2018. “To Be or Not to Be (Black or Multiracial or White): Cultural Variation in
Racial Boundaries.” Social Psychological and Personality Science 9: 763–772.
Fish, Jefferson M. 2011. “What Does the Brazilian Census Tell Us about Race?”
Psychology Today, Dec. 6. https://www.psychologytoday.com/blog/looking-in-the-
cultural-mirror/201112/what-does-the-brazilian-census-tell-us-about-race
Frankfurt, Harry. 1969. “Alternate Possibilities and Moral Responsibility.” The Journal
of Philosophy 66: 829–839.
Glasgow, Joshua. 2009a. A Theory of Race. New York: Routledge.
Glasgow, Joshua. 2009b. “In Defense of a Four-Part Theory: Replies to Hardimon,
Haslanger, Mallon, & Zack.” Symposia on Race, Gender, and Philosophy 5(2): 1–18.
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.
Glasgow, Joshua. 2017. “A Metatheory of Race.” In Naomi Zack (ed.), The Oxford
Handbook of Philosophy and Race. Oxford: Oxford University Press, pp. 170–179.
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.
Gratton, Brian, and Emily Klancher Merchant. 2016. “La Raza: Mexicans in the United
States Census.” The Journal of Policy History 28: 4.
Haney López, Ian F. 2005. “Race on the 2010 Census: Hispanics and the Shrinking
White Majority.” Dædalus 134: 42–52.
Haslanger, Sally. 2010. “Language, Politics, and ‘the Folk’: Looking for ‘the Meaning’ of
‘Race.’ ” The Monist 93: 169–187.
Hardimon, Michael O. 2017. Rethinking Race. Cambridge, MA: Harvard University Press.
Jeffers, Chike. 2013a. Untitled comment. http://peasoup.typepad.com/peasoup/
2013/06/ethics-discussion-at-pea-soup-chike-jefferss-the-cultural-theory-of-race-
27
272 • W h a t I s  R a c e ?
yet-another-look-at-du-bois.html?cid=6a00d83452b89569e2019103efe4a3970
c#comment-6a00d83452b89569e2019103efe4a3970c
Jeffers, Chike. 2013b. Untitled comment. http://peasoup.typepad.com/peasoup/
2013/06/ethics-discussion-at-pea-soup-chike-jefferss-the-cultural-theory-of-race-
yet-another-look-at-du-bois.html?cid=6a00d83452b89569e20192ac22ccf7970
d#comment-6a00d83452b89569e20192ac22ccf7970d
Kitcher, Philip. 2007. “Does ‘Race’ Have a Future?” Philosophy & Public Affairs
35: 293–317.
Kripke, Saul. 1972. Naming and Necessity. Oxford: Blackwell.
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/
Li, Jun Z., Devin M. Absher, Hua Tang, Audrey M. Southwick, Amanda M. Casto,
Sohini Ramachandran, Howard M. Cann, Gregory S. Barsh, Marcus Feldman, Luigi
L. Cavalli-Sforza, and Richard M. Myers. 2008. “Worldwide Human Relationships
Inferred from Genome-Wide Patterns of Variation.” Science 22: 1100–1104.
Machery, E., R. Mallon, S. Nichols, and S. P. Stich. 2004. “Semantics, Cross-Cultural
Style.” Cognition 92: B1–B12.
116: 525–551.
Mallon, Ron. 2009. “Commentary on Joshua Glasgow’s A Theory of Race.” Symposia on
Race, Gender, and Philosophy 5(2): 1–8.
McPherson, Lionel K., and Tommie Shelby. 2004. “Blackness and Blood: Interpreting
African American Identity.” Philosophy and Public Affairs 32: 171–192.
Mills, Charles W. 1998. “‘But What Are You Really?’ The Metaphysics of Race.”
In Mills, 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
Monk, Ellis P. 2014. “Skin Tone Stratification among Black Americans, 2001–2003.”
Social Forces 92: 1313–1337.
Monk, Ellis P. 2015. “The Cost of Color: Skin Color, Discrimination, and Health
among African-Americans.” American Journal of Sociology 121: 396–4 44.
Monk, Ellis P. 2016. “The Consequences of ‘Race and Color’ in Brazil.” Social Problems
63: 413–430.
Navarro, Mireya. 2003. Going Beyond Black and White, Hispanics in Census Pick
‘Other’. New York Times. https://www.nytimes.com/2003/11/09/nyregion/going-
beyond-black-and-white-hispanics-in-census-pick-other.html
Office of Management and Budget (OMB). 2016. “Standards for Maintaining,
Collecting, and Presenting Federal Data on Race and Ethnicity.” https://www.
regulations.gov/document?D=OMB-2016-0002-0001
273
Outlaw, Lucius. 1996. On Race and Philosophy. New York: Routledge.

Phelan, Jo C., Bruce G. Link, and Naumi M. Feldman. 2013. “The Genomic Revolution
and Beliefs about Essential Racial Differences: A Backdoor to Eugenics?” American
Sociological Review 78: 167–191.
Pierce, Jeremy. 2015. A Realist Metaphysics of Race: A Context-Sensitive, Short-Term
Retentionist, Long-Term Revisionist Approach. Lanham, MD: Lexington Books.
Putnam, Hilary. 1975. “The Meaning of Meaning.” Philosophical Papers, Vol. II: Mind,
Language, and Reality. Cambridge: Cambridge University Press.
Rosenberg, N., J. Pritchard, . . . M. Feldman. 2002. Genetic Structure of Human
Populations. Science 298(5602): 2381–2385.
Roth, Wendy D. 2012. Race Migrations: Latinos and the Cultural Transformation of
Race. Stanford, CA: Stanford University Press.
Sainsbury, Mark. 2014. “Fishy Business.” Analysis 74: 3–5.
Simon, Patrick. 2012. “Collecting Ethnic Statistics in Europe: A Review.” Ethnic and
Racial Studies 35: 1366–1391.
Spencer, Quayshawn. 2015. “Philosophy of Race Meets Population Genetics.” Studies in
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. 2000. “Appiah’s Uncompleted Argument: W.E.B. Du Bois and the Reality
of Race.” Social Theory and Practice 26: 103–128.
Telles, Edward. 2014. Pigmentocracies: Ethnicity, Race, and Color in Latin America.
Chapel Hill: University of North Carolina Press.
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.
Zack, Naomi. 1994. Race and Mixed Race. Philadephia: Temple University Press.
247
275
INDEX
Aboriginal Australians, 91–92, 100, Dalai Lama thought experiment

101n61, 101, 207, 219–20 and, 121–22, 131–32, 184, 233,
affirmative action, 160–61, 224 251–52, 263–6 4
African Americans. See also Blackness elimination of racial categories
ancestry data and, 207–8 advocated in some, 41, 137–38
Du Bois on advancement of, 50–51, 60 intuition-based thought experiments
Office of Management and Budget and, 233–34, 235–36, 237, 238
definition of, 79t medical outcomes and, 76
as pan-ethnicity, 28–29 non-existence of race as core principle
Alcoff, Linda Martín, 200 in, 2, 117
Alien Land Act of 1913 (California), 112 post-equality disappearance of race
alleles, 75–76, 95, 96–98, 100–1, and, 131–33, 249–50
103–4, 240 racial injustice and, 137, 191,
Allen, Anita, 68–69 247, 248–49
American Indians. See Native Americans racialization’s social reality and, 137,
amniocentesis, 73–74, 75 188–89, 246–48
Andaman Islanders, 101–2 reconstructionist views of race and,
Andreasen, Robin, 42–43, 179–80 137–38, 188
anti-realist views of race social constructionist views of race
basic racial realism as alternative to, and, 45, 46, 190–91, 248
139–4 0, 141, 142–43, 144, 183, 255 Appiah, Kwame Anthony, 46, 64, 248
biological realism and, 42–4 4, 45 Arab Americans
critiques of, 45, 46, 86, 128–29, prejudice against, 85, 180–81
131–33, 137, 156, 160–61, US census data and, 122
183, 187–88 Whiteness and, 17, 84, 85–86, 87, 113,
cultural constructivist views of race 180–81, 214, 252–54
and, 191 Aristotle, 22, 191–92
267
276 • Index
Asians biologically real entities and,

Andreasen’s research on, 42–43 77–78, 94–96
aneuploidal fetuses among expectant classical racialism and, 41–42,
mothers and, 74–76 59, 245–4 6
as cultural group, 225 critiques of, 41–4 4, 86, 118–21, 127,
discrimination in United States 136, 140, 165, 182–84, 263
against, 111–12, 115 demes and, 96–98
genealogical ancestry and, 226 genealogical ancestry and, 120,
Harvard University affirmative action 153, 207
case and, 160–61, 224–25 genetic clusters and, 96–100,
Office of Management and Budget 182, 206
categories and, 79t, 81, 215–16, 225 human continental populations and,
as pan-ethnicity, 27–29, 99–100, 104, 204–5, 206, 217–18,
166–67, 194–96 220, 240
South Asians and, 207–8, 214–16, 217 medical outcomes and, 73–76, 77,
103–4, 119–20, 154
Bagai, Kala, 111 Mismatch Objections and, 120,
Bagai, Vaishno Das 180–82, 188, 213, 252–56, 260
immigration to United States by, non-essentialist forms of, 42, 45–4 6,
111, 143 86, 178, 187, 188
naturalization as US citizen of, 112 Office of Management and Budget
rescinding of US citizenship (1923) of, (OMB) race talk and, 94–96,
114, 131, 262 99, 100–4, 153, 180–81, 231,
suicide of, 115 238, 253–54
basic racial realism (Glasgow and prevalence until twentieth century
Woodward) of, 40–41
critiques of, 183, 184–85, 187 race talk and, 78, 82
definition of, 2, 139 social constructionist approaches to
genealogical ancestry and, 184 race and, 42–4 4, 45, 179–80, 182
Haslanger’s sympathy for, 156 Black Lives Matter movement, 54
mixed-race identities and, 142 Blackness. See also African Americans
operative concepts of race and, “American Africans” and, 90–92
142–43, 156–57 black cosmopolitanism (Pierre)
physical features and, 140, 143, and, 197–98
183, 184–85 Coates on, 129–30
Pierce and, 185–86 Dolezal and, 226–30
social constructionist view of race education emphasizing Black cultural
and, 186–87 heritage and, 66–68
spectrum view of race and, 142, genealogical ancestry and, 159, 207,
187, 255 227–28, 254–55
biological realist approaches to race in Latin America, 15
anti-realist views of race and, medical outcomes and, 119–20
42–4 4, 45 Obama and, 141–42, 186–87
72
Index • 277
Office of Management and Budget post-equality disappearance of race

definition of, 78–79, 79t, 90–91, and, 131–34
92, 100 racialization’s social reality and, 129,
“one-drop rule” and, 141–42 130–31, 133, 136, 210, 247–48
pride and, 62, 64 racialized hierarchies and, 130–31
skin bleaching and, 197 continental populations. See human
stereotyping and, 69–70 continental populations
Blauner, Robert, 27–28 critical theory, 7–8
Blum, Lawrence, 44, 46, 137, 162–63n17, cultural constructionist views of race
188–90, 191 anti-realist views of race and, 191
Blumenbach, Johann Friedrich, 113, cultures in contestation and, 55–56
217–20, 219t defining features of, 50, 57, 62
Boxill, Bernard, 238–39 diversity of experience within cultures
Braxton, Tamar, 228–29, 230 and, 65–66
Brown v. Board of Education, 260–61 DuBois and, 25n15, 49–51, 52–53, 60
Burchard, Esteban, 77 educational outcomes and, 66–68
ethnicity and, 61
Capehart, Jonathan, 228, 229 freedom to learn from the cultural
caste hierarchies, 169 practices of others and, 202
Cavalli-Sforza, Luigi Luca, 42–43, 179 historical patterns preceding racial
Central Asians, 101–2, 225n45 formation and, 63–6 4
Chinese Exclusion Act (United States, interracial marriage and, 68–69
1882), 111–12 pan-ethnicities and, 29
Citrullus lanatus, 216–17 personhood and, 55
Citrullus mucosospermus, 217 political constructionist views of race
Civil Rights Act of 1964, 224n43, 226 contrasted with, 51, 54–55, 56–58,
classical racialism (Taylor), 41–42, 62, 64–68, 192–94, 202
59, 245–4 6 racial difference after the end of racism
Coates, Ta-Nehisi, 129–30, 141 and, 58, 71, 170, 193–94, 199, 200–
Coleridge, Samuel Taylor, 2 1, 202, 250–52
Compton, Elizabeth, 86–87 racial pride and, 61–62
“The Conservation of Races” (Du Bois), social constructionist views of race
49–51, 52–53, 60 contrasted with, 24–25, 30
constructionist views of race. See also Taylor’s critiques of, 51–52,
cultural constructionist views of 58–62, 230
race; political constructionist views
of race; social constructionist views Davis, F. James, 15
of race demes, 96–98
“baby world” thought experiment and, descriptivism, 88–89
133, 142–43, 251 Dobzhansky, Theodosius, 178–79
critiques of, 128–29, 131–35, 136 Dolezal, Rachel, 158–59, 170, 226–30
Glasgow’s description of, 128 Dominican Americans, 84–85, 207–8
Mismatch Objections and, 128 Donovan, Brian, 239–4 0
287
278 • Index
Du Bois, W. E. B., 25n15, 49–51, 52–53, Hardimon, Michael

60, 130 minimal conception of race and,
Duhem, Pierre, 221, 231n55 19–20, 39–4 0, 77, 92–93, 165–66,
Dworkin, Ronald, 192 192–93, 207, 233–34
Office of Management and Budget
echogenic intracardiac focus race talk and, 93, 206–7
(EIF), 73–75 on race and common ancestry, 19,
empirical adequacy standard, 220–22, 39, 234
223, 229–30, 231–32 on race and geographic origins, 19,
Epstein, Brian, 205n5 40, 205n6
Espiritu, Yen Le, 27–29, 167, 194–95 on race and patterns of physical
Ethiopian Americans, 218 features, 19, 39, 49
ethnicity on race’s very limited explanatory
culture and, 61 value, 40
pan-ethnicities and, 27–29, on “racialist development,” 44n11, 44
167, 194–96 Harvard University affirmative action
pride and, 61 case, 160–61, 224, 226
Euthyphro (Plato), 150–51 Hawaii, 15, 79t, 101, 207–8
Hernandez v. Texas, 260–61
Feldman, Marcus, 204–5 Herzog, Don, 1–2
Fine, Arthur, 100n59 Hippocrates, 23, 191–92
First Fridays event (Accra, Hispanic Americans. See also Latinx
Ghana), 197–98 deportation in 1920s of, 260–61
Frankfurt, Harry, 264–65 genealogical ancestry and,
207–8, 254–55
Gannett, Lisa, 238 Hernandez v. Texas and, 260–61
gender, 134 race and, 85–86, 87, 214,
genetic clusters (Rosenberg), 96–100, 252–54, 260–62
182, 206 US census data regarding, 84–86
Ghana, 196–98 Whiteness and, 84–86
Godfrey-Smith, Peter, 235 Hmong ethnic group, 29, 195–96
Graves, Joseph, 118 human continental populations, 99–100,
Great Zimbabwe (medieval city in 103, 104, 204–5, 206, 217–18, 220,
southern Africa), 63–6 4 222, 240
Griffith, John Howard, 158–59 Humphrey, Jeff, 227
Guo, Guang, 102–3, 215 Immigration Act of 1917 (United
Gyekye, Kwame, 64 States), 111–12
Haiti, 141 interracial marriage, 68–69, 81,

Hall, George, 112 112–13, 140
Haney López, Ian, 130–31 intuition-based thought experiments,
Han people (China), 95 232–34, 235–36, 237, 238, 263–70
729
Index • 279
Jamaica, 197 Mendel, Gregor, 220–21

Jarawa people, 102 Mills, Charles, 55–56, 68n50,
Jews, 25n15, 86, 208–9, 222 196–97, 238–39
Jorgenson, Eric, 103–4 Millstein, Roberta, 121
Mismatch Objections
Kagan, Elena, 261–62 Arab Mismatch Objection, 214
Kalash Pakistanis, 216, 253–54 basic racial realism as response to, 140
Kelly, Daniel, 2–3, 239 biological realist approaches to
Kitcher, Philip, 178–79, 238–39 race and, 120, 180–82, 188, 213,
Kripke, Saul, 88–89 252–56, 260
constructivist views of race and, 128
Latinx. See also Hispanic Americans definition of, 120
ancestry data and, 207–8 Hispanic Mismatch Objection
as pan-ethnicity, 27–28, 166–67 and, 214
race and, 122, 123–25, 127, Office of Management and Budget
253–54, 262 (OMB) race talk and, 252–54, 260
Ling Sing, 112 scientifically provable genetic
Linnaeus, Carl, 216 populations and, 120–22
Loehman, Timothy, 121, 131 South Asian Mismatch Objection and,
Love, Loni, 227–28, 229 43–4 4, 180–81, 214–16, 217
Loving v. Virginia, 68 “way of life” view of race and, 167–68
Ludwig, David, 123n11 Mitochondrial Eve, 1
mixed-race identities
Madagascar, 101–2, 268 assimilation and, 15
Maglo, Koffi, 99 classification system’s varying
Malagasy people, 102 approaches to, 15, 113
Mallon, Ron cultural constructionism and, 202
on epistemic value of race talk, 9–10, in Latin America, 15
32, 177 Malagasy and, 102
on metaphysical agreement among Obama and, 141–42, 186–87
philosophers regarding race, 176, Office of Management and Budget
188–89, 249 race talk and, 209
Mismatch Objection and, 120 physical features and, 17
on normative debate about racial political constructionism and, 202
discourse, 177–78 South Asians and, 215–16
on race as a socio-psychological Morning, Ann, 211
grouping of people, 2–3 Moya, Paul, 130
semantic strategy regarding race and,
9–10, 150 Native Americans
Markus, Hazel Rose, 130 ancestry data and, 207–8
Maya people, 95 in Latin America, 15, 81
Melanesians, 93, 183–84, 219–20 legal discrimination against, 112
280
280 • Index
Native Americans (cont.) socially important traits and, 104

Office of Management and Budget Statistics Canada approach compared
categories and, 78–79, 79t, 81, 101 to, 181
as a pan-ethnicity, 27–28 truth-conditional approach to
stereotyping and, 69–70 language and, 87, 88–89, 94
Naturalization Act of 1790 (United “one-drop rule,” 130, 141–42,
States), 112, 113–14, 115 158–59, 236t
Niue, 237 Onge people, 102
“ordinary concept of race” (Glasgow),
Obama, Barack, 141–42, 186–87 115, 116–17, 122–23
Office of Management and Budget Outlaw, Lucius, 56, 190–91, 200
(OMB) race talk Ozawa, Takao, 113
Alternative Questionnaire Experiment
(Compton) and, 86–87 Pacific Islanders, 78–79, 79t, 81, 93,
biological realist approaches to race 101, 207
and, 94–96, 99, 100–4, 153, pan-ethnicities, 27–29, 167, 194–96
180–81, 231, 238, 253–54 Pemberton, Trevor, 98–99
Blumenbach’s racial taxonomy Peña Rodriguez v. Colorado, 261–62
and, 217–20 People v. Hall (California Supreme Court
census data and, 83–86, 155, 180–81, case), 112
222, 252–53 Pierce, Jeremy, 142n18, 185–87
civil rights laws and, 81, 210, 226 Pierre, Jemima, 196–98
college admissions and, 79, Plato, 150–51
80f, 222–25 Plessy v. Ferguson, 50
divisions of linguistic labor and, 82– political constructionist views of race
83, 86–87, 214 construction of race in everyday life
dominant meaning of “race” among and, 53
American English speakers and, cultural constructionism views of race
212–13, 222–23, 231, 234–35, 238 contrasted with, 51, 54–55, 56–58,
as “expert race talk,” 82, 154–55 62, 64–68, 192–94, 202
five races identified in, 78–79, 81, defining features of, 48, 56–57
153, 204 differential power relations
Haslanger’s critique of, 154–56 and, 56–57
identity thesis and, 100–3 educational outcomes and, 66–68
medical research and, 103–4, 153, 154 normative considerations and, 193
Mismatch Objection and, probabilistic forms of, 65–66
252–54, 260 racial difference generated in, 55
mixed-race identities and, 209 stereotyping and, 69
phenotypic differences and, 104 as variety of social constructivist views
“population thinking” as potential of race and, 48, 49–50, 51
negative effect of, 238–4 0 White supremacy doctrine and, 48,
referentialism and, 88, 94 53–54, 57
218
Index • 281
Polo, Marco, 268 “passing” and, 158–59, 169–70, 260

Prewitt, Kenneth, 217–18, 220 perceptual gestalts and, 17
Putnam, Hilary, 82, 83 physical features and, 2, 17, 18–19, 20,
25, 26–27, 39, 40, 41, 47, 48, 49,
Quine, W.V.O., 8 51–52, 87, 90, 117, 118–19, 126–27,
130, 140, 141, 142, 143, 157–58, 159,
race. See also anti-realist views of race; 160–62, 165, 171, 172, 182–85, 188,
cultural constructionist views of 194–95, 210–11, 245–4 6, 249, 256,
race; social constructionist views 260, 262–63
of race pluralist theories of, 212–13, 238
assimilation and, 15 practical roles as form of
biblical interpretations and, 22, 23–24 classifying, 17–18
colonialism and, 63–6 4 pride and, 61–62
definitions of, 4, 7, 19–20, 117 racial natures concept and, 18–19, 40–
epistemology and, 18 41, 120, 245–4 6, 249
explanatory roles as form of representational traditions and, 11–14
classifying, 18 semantic commitments and, 8, 33–34,
features-and-identities commitments 159–60, 162, 256–59
and, 124–25, 127 skin color and, 17, 18–19, 87, 90, 118,
genealogical ancestry and, 1, 19, 39, 126–27, 130, 143, 157–58, 160–61,
40, 41, 87, 120, 153–54, 159, 184, 162, 165, 171, 172, 182–83, 210–11,
188, 207–8, 226, 227–28, 228n52, 249, 262–63
234, 245–4 6, 254–56, 262–63 spectrum view of, 118–19, 141, 142
geographic origins and, 19, 40, 205n6, stereotyping and, 69–70
205, 245–4 6 as subdividing principle for species
heritability and, 24 other than humans, 178–79
hierarchical social structures and, symbolic roles as form of
22–23, 25–27, 29–30, 48, 49, 52, classifying, 18
130–31, 165–66, 248, 249–50 US citizenship and, 112–14, 115, 131
historical roles as means of way of life and, 167–68
classifying, 17 racial anti-realism. See anti-realist views
manifest concepts of, 156–57 of race
medical outcomes and, 73–76, 77, Racial Classification Questionnaire
103–4, 119–20, 153, 154, 247 (RCQ), 236–37
migratability and, 121, 122 racialization
normative dimensions of, 31 anti-realist views of race and, 137,
“one-drop rule” and, 130, 141–42, 188–89, 246–48
158–59, 236t constructionist views of race and, 129,
operative concepts of, 115, 130–31, 133, 136, 210, 247–48
126–27, 156–57 social constructionist views of race
particular instances as means of and, 25, 26–30, 31, 32–33
classifying, 16–17 representational traditions, 11–14
28
282 • Index
revelation theory of race, debunking accounts of race

124–25, 134–35 and, 32–33
Rice, Samaria, 121 genealogical ancestry and, 41,
Rice, Tamir, 121, 130–31, 136, 143 153–54, 188
Risch, Neil, 77 murder case analogy and, 46
Rosenberg, Noah, 96–99, 153–54, 180 pan-ethnicities and, 27–29
Rutherford, Ernest, 264–65 “passing” and, 169–70
physical features and, 2, 25, 26–27, 41,
Sainsbury, Mark, 256–57 47, 48, 49, 184–85, 188
Schroeter, Laura and François race talk and, 47
account of meaning offered by, 11–13, racial injustice and, 248–49
18–19, 192 revision of beliefs facilitated through,
race and the representational tradition 23, 31
for, 14–15 slavery and, 22
on the representational tradition and social hierarchies built on race and,
commitment to de jure sameness of 25–27, 29–30, 48, 49, 52, 165–66,
reference, 13–14 248, 249–50
water example of, 12, 13 Solomon Islands, 183–84
“Sean from Dayton” (National Public South Asians, 207–8, 214–16, 217
Radio caller), 160–62 stereotyping, 69–70, 85, 201
September 11, 2001 terrorist attacks, 85 Sullivan, Shannon, 75–76, 200
Shelby, Tommie, 7–8, 137 Sutherland, George, 113–14, 115,
Shipp, T. D., 74–76 130–31, 260–61
Shona people (southern Africa), 63–6 4
Sickle-cell trait, 119–20, 154 Taiwan, 212
Singapore, 237 Tajiks, 102
skin bleaching, 196–97 Tang, Hua, 215
slavery Taylor, Paul
Aristotle on, 22, 191–92 classical racialism and, 41–42, 59
Blackness and, 17, 47, 130–31, 158–59 on color-conscious endogamy, 68–69
Ghana and, 196–97 cultural constructionist view of race
social constructionist account of, 22 critiqued by, 51–52, 58–62, 230
social constructionist views of race on English roots of US conceptions of
anti-realist views of race and, 45, 46, race, 77
190–91, 248 on “modern racialism” and White
basic tenets of, 22 Supremacy doctrine, 51–52
biological realism and, 42–4 4, 45, probabilistic considerations in political
179–80, 182 construction of race and, 65–66
caste hierarchies and, 169 on race talk, 78
common-sense racial categorizations on the white supremacist
accepted in, 45–4 6 determination to link appearance
cultural constructionist view of race and ancestry with social
contrasted with, 24–25, 30 location, 48
238
Index • 283
Tay-Sachs disease, 119–20 Hispanic Americans and, 84–86

Thind, Bhagat Singh, 113–14, 122, in Latin America, 15
260–61, 262 Office of Management and Budget
Thunberg, Carl, 216 categories and, 78–79
Tran, S. H., 75–76 as pan-ethnicity, 198–99
Trump, Donald, 54, 224 “passing” and, 158–59
US citizenship and, 113–14, 115, 131
Uzbeks, 102 White pride and, 63–6 4, 198–200
White privilege and, 53–54, 137, 228
Van Fraassen, Bas, 220–21 White supremacy doctrine and, 2,
48, 52, 53–54, 55, 70, 130, 166,
Warren, Earl, 260–61 198–99, 262–63
watermelons, 216–17 Williams, Gregory Howard, 158–59
werewolf legend, 125 Woodward, Jonathan, 139, 183
Whiteness
Arab Americans and, 17, 84, 85–86, Yoruba people (Nigeria), 95
113, 180–81, 214, 252–53 Yudell, Michael, 75, 104
education systems privileging
of, 67–68 Zimbabwe, 63–6 4
284
258
286

Glasgow Et Al - What Is Race Four Philosophical Views (2019)

Uploaded by

Copyright:

Available Formats

Glasgow Et Al - What Is Race Four Philosophical Views (2019)

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Glasgow Et Al - What Is Race Four Philosophical Views (2019)

Uploaded by

Copyright:

Available Formats

i

Published in the United States of America by Oxford University Press

© Oxford University Press 2019

All rights reserved. No part of this publication may be reproduced, stored in

You must not circulate this work in any other form

Library of Congress Cataloging-​in-​Publication Data

Paperback printed by Sheridan Books, Inc., United States of America

and students in the philosophy departments at the following universities for

Historically, efforts to make sense of human diversity led to classi-

1 TRACING THE SOCIOPOLITICAL REALITY

Tracing the Sociopolitical Reality of Race • 5

Tracing the Sociopolitical Reality of Race • 7

to use existing terminology, but how to collectively orient ourselves toward

1.2. What Is the Question?

scientific thought, but they do so without taking a slavish attitude to-

1.3. The Semantic Strategy

Tracing the Sociopolitical Reality of Race • 9

Ron Mallon (2006, 527) considers this “semantic strategy,” widespread

While there is (or should be) a wide basis of metaphysical agreement on

I am sympathetic with Mallon’s (2006) suggestion that resolving the issue

Tracing the Sociopolitical Reality of Race • 11

1.4. Representational Traditions: ‘Water’ as an Example

our words and thoughts to represent genuinely interesting and important

. . . ideal epistemic methods for answering ‘what is x?’ questions hinge

Tracing the Sociopolitical Reality of Race • 13

stake in a representational tradition with ‘x’, and you should identify

A rationalizing interpretation, on their view, is not determined by reports of

From the deliberative perspective of a rational epistemic agent, the

known as ‘beer,’ I would be wrong. I would have failed to capture a reasonable

As rational epistemic agents, we normally take our words and

We postulate ambiguity or opt for an error theory only as a last resort.

1.5. Representational Traditions: ‘Race’

Tracing the Sociopolitical Reality of Race • 15

And finally, in some cases, there is a possibility of assimilation, so that after

So when I suggest we consider “our representational tradition,” what do

Tracing the Sociopolitical Reality of Race • 17

as a continent or US racial practices. There is disagreement about whether

athletics. It is also an important part of many people’s identity and sense of

Given these inputs (I don’t mean for these to be exhaustive—​this is just a

Tracing the Sociopolitical Reality of Race • 19

The method we have sketched precisely aims at determining what’s

(C1) that, as a group, is distinguished from other groups of human

Tracing the Sociopolitical Reality of Race • 21

of economic and political injustice, social segregation, and cultural stigma

• X is a person only if X has a soul.

There are compelling examples suggesting that accepting any of these

the concept. Communication in a context does not require a rigid represen-

Tracing the Sociopolitical Reality of Race • 23

I will consider (i) in this section and (ii) in the next.

Tracing the Sociopolitical Reality of Race • 25

that the “marking” is taken to be evidence of where, geographically, the group

Tracing the Sociopolitical Reality of Race • 27

in Europe and that is situated as privileged, as a result, has higher educational

populations. As a result of slavery, “the ‘Negro race’ emerged from the

The development of a pan-​ethnicity may emerge, for example, among

18. See also Alcoff (2000) on ethno-​race, and Gooding-​Williams (1998).

Tracing the Sociopolitical Reality of Race • 29

not a pan-​ethnicity, even if African American, or Diasporic African, is. This is

Library of Congress Cataloging-in-Publication Data

1.2. What Is the Question?

1.3. The Semantic Strategy

1.4. Representational Traditions: ‘Water’ as an Example

1.5. Representational Traditions: ‘Race’

Given these inputs (I don’t mean for these to be exhaustive—this is just a

The development of a pan-ethnicity may emerge, for example, among

18. See also Alcoff (2000) on ethno-race, and Gooding-Williams (1998).

not a pan-ethnicity, even if African American, or Diasporic African, is. This is

1.7. Going On: The Normative Dimension

SPR is to highlight—in the relevant cases—how our racializing practices and

2.1. Why Social Constructionism?

Despite the attractions of anti-realism, however, I reject it. When evalu-

2.2. Political versus Cultural Constructionism

14. Taylor, Race, 89–90.

2.3. Why Cultural Constructionism?

46. Taylor, Race, 82–83. Emphasis mine.

for our purposes, if it is effective with regard to non-white students, such a

3.2. OMB Race Talk as a US Race Talk

American Indian or Alaska Native—A person having origins in any of the

3.3. The Meanings of ‘Race’ and Race Terms in OMB

be so leads to several counterfactual conditionals with the wrong truth-value.

that by ‘recent’ we mean ‘before twenty-one generations ago.’36 Furthermore,

3.4. The Nature and Reality of Race and the Races