Skip to main content

Abstract

Feminist philosophers have long worried that intersectionality undermines the viability of the concept and category of woman, thereby undermining feminist theory and politics. Some have responded to this problem by abandoning intersectionality; others have attempted to find some suitably inclusive way of reconceptualizing woman. I provide a novel solution that focuses on conceptualizing oppression in light of intersectionality, rather than trying to provide an account of what it is to be a woman. By enabling us to understand feminism as responding to gender oppression, this account shows that intersectionality does not conceptually undermine and fragment feminism. Feminism should be intersectional.

I.  Introduction

A common accusation against intersectionality is that it is divisive. This is an accusation that we see on social media and that we hear from political pundits, but it is limited neither to political conservatives nor to mainstream political discourse. Feminist philosophers have also worried about the seeming divisiveness of intersectionality.

The significance of the worry, for feminists, comes from the centrality of woman for feminism—as Alcoff puts it, “the concept and category of woman is the necessary point of departure for any feminist theory and feminist politics.”1 This is because feminist theory is broadly conceived of as theorizing about and for the sake of women, and feminist politics is conceived of as mobilizing around women in order to liberate them from their oppression.

Intersectionality has, however, brought into question the very viability of the concept and category of woman. Intersectionality theorists note that attempts to find common characteristics among women and their experiences have done so by ignoring the experiences of Black and other marginalized women. Further, they highlight how women’s characteristics and experiences differ along lines of race, class, sexual orientation, ability, and similar social divisions, such that there appears to be nothing universal either to women or to their experiences of oppression that would justify conceptualizing women as a unified category. This seems to lead, in the first instance, to a fragmentation of the concept and category of woman—the group woman becomes splintered into finer and finer groups like American, able-bodied, middle-class, queer white woman—and ultimately threatens a dissolution of social groups altogether as the distinctions illuminated by intersectional considerations become so fine-grained as to leave us only with a sea of individuals.

In undermining the viability of woman as a coherent, unified category, intersectionality seems to spell existential trouble for feminist projects: conceptualizing the oppression of women, making moral or legal claims on behalf of women, theorizing about women and their social conditions, and mobilizing politically around women all seem to require being able to conceptualize women as a coherent, unified group. Woman, again, seems to be the very starting point for feminist theory and politics.

Stated generally, then, the worry is that intersectionality commits us to the ontological fragmentation of woman, and that in so doing it politically fragments feminism by undermining the very possibility of feminist projects. Intersectionality thus presents a challenge for feminists. On one hand, intersectionality seems to provide crucial insights about the complexity and diversity of women’s experiences. On the other hand, accepting these insights seems to undermine feminist theory and politics.

Some have responded to this problem by rejecting intersectional commitments. Many have strived to find some way of reconceptualizing woman so as to salvage feminism. Others have reversed the common assumption that feminist politics starts with women and instead take women to be a group constructed through feminist political coalition. These approaches have focused on rehabilitating the concept and category of woman, because these approaches have retained some commitment to the standard view that feminism centrally starts with or revolves around women.

In this article I take a different approach to the fragmentation challenge. Rather than working on giving some unified account of woman that can serve as a focal point for feminism, I shift the focus from identity and group kinds to oppression. That is, rather than trying to find some suitably unifying basis for groups like women, I show how it is that we can (1) conceptualize collective forms of oppression while retaining intersectional commitments and (2) conceive of responding to oppression as the rational basis for feminism and similar political movements. By focusing on oppression rather than identity, I show that intersectionality does not have the politically fragmenting implications that many feminists have feared—and, further, that we do not, as has been standardly thought, need to find some way of rehabilitating woman as a unified kind in order for feminist theory and politics to be coherent.

By focusing first and foremost on oppression, my view is in many ways aligned with Black feminist approaches that have identified shared struggle, rather than some shared, preexisting identity, as the impetus for entering into political coalition.2 However, my view departs from these in allowing for a more expansive pluralism about identity and group kinds, rather than strictly defining or fixing these in terms of political coalitions. Thus, another benefit of my account is that it allows for complicated and heterogeneous relationships between experiences of oppression, political coalitions, identity, and group kinds.

Again, my approach will be to provide an account of oppression that accommodates the intersectional commitments that have been thought to spell trouble for feminism and to show that we can use this account of oppression to provide a theoretical and political basis for feminism. On my account, oppression is understood in terms of collections of patterns of injustice that are importantly related to certain ideological conceptions of the social world. On this picture, there can simultaneously be patterns of injustice that are shared by a collection of individuals and subpatterns of those injustices that represent how particular subsets of that collection (like white women, Black women, trans women, etc.) more specifically tend to experience that kind of injustice.3 Further, on this account there can be oppression-constituting experiences of injustice that are not shared by all members of the collection, and there need be no experience of injustice that all members of the collection share. This allows the account to reflect the intersectional idea that experiences of a particular kind of collective oppression are heterogeneous, and every individual’s experiences of that oppression are shaped by their overall social positioning. By providing a way of conceptualizing oppression that incorporates key intersectional insights, the account allows us to conceive of feminism and other specific, radical political movements as organizing around specific, collective oppressive kinds.4

While the worry about fragmentation is ultimately political in nature—that intersectionality undermines and thereby fragments feminist and similar political movements—it is predicated on the thought that intersectionality fragments ontological kinds, and particularly the group women. The article will therefore proceed as follows. In Section II, I will present the fragmentation challenge in more detail: I will review three core intersectional notions and show why these are thought to lead to the ontological fragmentation of groups like women, as well as why this has been thought to result in political fragmentation. In Section III, I will present my account of oppression and show that it is compatible with the core intersectional notions thought to lead to fragmentation. In Section IV, I will show how this account can ameliorate the worries about political fragmentation. Finally, in Section V, I will contrast my solution with alternative approaches in the literature to highlight some further benefits of my view.

II.  Understanding the Fragmentation Challenge

My aim in this article is to resist the idea that intersectionality necessarily leads to political fragmentation because it undermines the rational basis for radical social movements like feminism. The first step is to understand why philosophers have thought that intersectionality has this consequence.

The worry about politically damaging fragmentation arises from widespread (though, as with any philosophical area, not universal) commitments to core notions of simultaneity, inseparability, and mutual constitution that characterize intersectionality.5 Intersectionality pushes us beyond recognizing multiplicity to considering race, class, gender, ability, and so on, as being importantly intertwined. To understand why this has been thought to have fragmenting implications, it is helpful to first give an overview of these core intersectional notions.

A.  Core Notions: Simultaneity, Inseparability, and Mutual Constitution

The notions of simultaneity, inseparability, and mutual constitution all get at the idea that race, class, gender, ability, and so on, are importantly interconnected. First, simultaneity emphasizes a kind of holism by emphasizing that one has multiple identities, belongs to multiple social groups, and experiences multiple forms of oppression at the same time. Importantly, this claim should not be read as simply indicating multiplicity. It’s not just that one has these multiple different identities, all at the same time, but that these different aspects of one’s identity are integrated, or fused.6 As Anzaldua expresses it, “What am I? A third world lesbian feminist with Marxist and mystic leanings. They would chop me up into little fragments and tag each piece with a label. … Not so. Only your labels split me.”7 In this way, simultaneity evokes not just multiplicity but also a notion of integration.

Similarly—turning to the notion of inseparability—there is a widespread rejection of the idea that we can cleanly separate out individuals’ identities, or different forms of privilege and oppression, into “pure” parts (though, as we will see, there is also significant debate around this idea).8 Inseparability (also discussed under the labels of anti-essentialism and nonadditivity) challenges the idea that there is some universal characteristic or experience shared by members of an identity group that captures what is true of them qua group members—like universal “pop-beads” that represent the essence of an identity and that can be mixed and matched to generate any individual’s particular combination of identities.9 Similarly, with respect to oppressions, inseparability challenges the idea that “Black men and Black women experience the same racism and white women and Black women experience the same sexism.”10 Again, the idea being rejected is that there is something that is universally shared that characterizes the particular kind of oppression in question, such that a comprehensive analysis of oppression could simply add together an analysis of “pure” racism, “pure” sexism, “pure” classism, and so on (and similarly for “pure” forms of privilege). The illusory appeal of separability is typically attributed to a tendency to take the features and experiences of more privileged members of a group (e.g., white women in the case of women) as definitive.11

To see why intersectionality theorists reject separability, consider the following example from the Combahee River Collective: “We also often find it difficult to separate race from class from sex oppression because in our lives they are most often experienced simultaneously. We know that there is such a thing as racial-sexual oppression which is neither solely racial nor solely sexual, e.g., the history of rape of Black women by white men as a weapon of political repression.”12 The example that they provide here—the rape of Black women by white men as a weapon of political repression—cannot be divided into some purely gendered part and some purely racial part. Rather, the phenomenon is inseparably racial and gendered. This makes it impossible to partition experiences of oppression into mutually exclusive (or “pure”) categories, because there are experiences, like this one, that span across divides.

A different kind of failure for additive analyses can be seen in Crenshaw’s example of a law that required individuals who immigrated to the United States for marital purposes to remain married for at least two years before applying for permanent residency.13 One consequence of this law was that many immigrant women in abusive relationships felt unable to leave their partners for fear of deportation. While this law might seem to have nothing to do with gender, by making these women more dependent on their partners, it exacerbated the effects of domestic violence. This shows that we cannot separate out racism, classism, sexism, xenophobia, and so on; analyze these pieces in isolation; and then combine these separate analyses to get an adequate analysis of oppression. Such an additive approach would overlook the ways in which different factors interact with and exacerbate each other to create instances of oppression that are simultaneously racial, classed, gendered, and so on.

Going a step further, many have taken the central notion of intersectionality to be the idea that race, class, gender, sexuality, and so on, are mutually constituting.14 Shields, for instance, takes it that “a fundamental assumption in every influential theoretical formulation of intersectionality is that intersectional identities are defined in relation to one another.”15 She spells this out by saying that instead of being discrete, separable identities, intersectional identities are mutually constituting, where “by mutually constitute I mean that one category of identity, such as gender, takes its meaning as a category in relation to another category.”16 Collins and Bilge also speak in these terms, describing both identity categories and oppressions as “gaining meaning” in relation to each other.17 We can understand “meaning” here as referring roughly to social meaning, which encompasses how one is treated and regarded, what is expected of one, and what kinds of experiences one has. For example, suppose that Liz is a straight, middle-class, white woman. Applying the notion of mutual constitution to Liz, what it means for her to be a woman, or her experience of being a woman, will be shaped by her status as straight, middle-class, and white. And likewise, what it means for her to be a white person—how she experiences the world as a white person—will be shaped by the fact that she is straight, middle-class, and a woman. And so on for other parts of her identity. This example highlights an important point emphasized by Carbado’s notions of color-blind and gender-blind intersectionality: these core intersectional notions don’t just apply to oppressions or individuals with multiple marginalized identities.18 Rather, they apply across the board. Everyone has mutually constituting intersectional identities, and everyone’s experiences of different kinds of oppression, as well as privilege, are mutually constituting.

B.  Ontological Fragmentation

Having reviewed these core intersectional notions, we can now ask, why has intersectionality been thought to be politically fragmenting?

It is useful to split this question into two parts by distinguishing between ontological fragmentation (i.e., the fragmentation of kinds like women) and political fragmentation. This is because the concern about political fragmentation comes from the conjunction of two claims: (1) that intersectionality leads to ontological fragmentation (in the sense that accepting intersectional commitments ends up committing one to a view on which kinds like women no longer exist), and (2) that ontological fragmentation leads to political fragmentation (because it undermines, in some important way, the basis for political projects).19 More specifically, feminist philosophers have worried that intersectionality fragments the group women and that the fragmentation of women undermines the coherency of feminist theory and politics. Hence, the conclusion is that intersectionality is politically fragmenting because it fragments ontological kinds. Given that one finds separate arguments for (1) and (2), it is useful to distinguish between them. I will start by reviewing why people have thought that intersectionality leads to ontological fragmentation, and then I will address the political implications this is thought to have in the next subsection.

There are two main arguments from the core notions of intersectionality to the claim that intersectionality leads to ontological fragmentation: one starts from the notion of inseparability, and the other from the notion of mutual constitution.

Starting with the former, there is disagreement over whether accepting inseparability simply equates to a wholesale rejection of social categories.20 In particular, some equate inseparability with a rejection of essentialism—that is, they equate inseparability with the rejection of the claim that “members of a particular group all possess some set of defining characteristics, universal among group members, and not dependent upon context.”21 This strong, anti-essentialist version of separability is typically taken to coincide with a postmodern rejection of social categories.22 On this view, intersectionality straightforwardly leads to ontological fragmentation because accepting inseparability amounts to rejecting social categories.

Many reject this straightforwardly anticategorical stance, however. Instead, they think that what accepting inseparability demands is a recognition of the complexity within categories, rather than a rejection of categories.23 They take categories like race, racism, gender, and gender oppression to point to real, important social phenomena, albeit phenomena that are not fixed, universal, ahistorical, or acontextual.24 Others take a more pragmatic view, on which categories are a kind of useful fiction. Harris, for instance, suggests that we must recognize the artificiality of categories and “make our categories explicitly tentative, relational, and unstable,” while Anthias argues that, though the categories may not be truly separable, we can and should treat them as separable for analytic purposes.25 Thus, while some take a commitment to inseparability to constitute an anticategorical stance, many (if not most) intersectionality theorists reject an outright anticategorical stance.

The second path to ontological fragmentation takes the form of a regress argument. Many have argued that the commitment to mutual constitution sets off a regress that ultimately results in the fragmentation, or indeed outright dissolution, of important social kinds. Zack, for instance, worries that intersectionality splinters the group women by replacing it with more specific groups (e.g., straight, working-class, able-bodied, Latinas) that correspond to “the reification of intersections as incommensurable identities.”26 Others, like Young and Ehrenreich, take this worry a step further, claiming that intersectionality leads to the dissolution of social groups altogether, leaving us only with individuals.27 More recently, Gasdaglis and Madva have pushed this worry even further, arguing that the regress would not even stop at the level of individuals but would leave us with space-time points. Gasdaglis and Madva also apply their argument to oppressive kinds, not just group kinds. They argue that if one understands intersectionality—which they understand in terms of mutual constitution—to be a thesis about oppression, then racial oppression, gender oppression, class oppression, and so on, cannot be genuine kinds.

Because Gasdaglis and Madva provide the most detail in developing this ontological-fragmentation-through-regress worry, I will focus on their argument, which they claim will apply to any domain that one takes mutual constitution to be a thesis about.28 Focusing first on oppressive kinds, they point out that if oppressions are mutually constituting, then “what it means to be oppressed in virtue of blackness differs for black men and black women,” and they conclude from this that Black oppression is not a genuine kind.29 One might then want to say that we have two genuine kinds where we previously thought we had one: instead of Black oppression, we have Black women’s oppression and Black men’s oppression. But they argue that this cannot work: “By the same token, however, ‘black women’s oppression’ isn’t a genuine kind either, because gender, race, and class intersect: what it means to be oppressed in virtue of black-womanhood differs for rich and poor black women. The same goes for sexuality, ability, religion, and a host of other significant social categories, potentially ad infinitum.”30 Again, they indicate that the same kind of regress argument applies to identities or group kinds. But whereas others have assumed that the regress stops at the level of the individual, they note that an individual’s experiences and interests are also going to fluctuate as their age, levels of ability, employment status, education, social context, and so on, change. They thereby conclude that if we take identities to be mutually constituting, then we seem to end up with nothing more than “isolated, atomic space-time points.”31

In order to highlight the role that mutual constitution is playing here, as well as to make explicit what I take to be important background assumptions, I will recast this argument in more formal terms. First, there is the thesis of mutual constitution, which says that, for any G, what it means for g to be G, or for g to experience G-oppression, is going to depend on the other identity categories that g is part of. This means that, for any F and for any G, the experiences and attributes of members of G are going to differ between those who are also members of F and those who aren’t. But then this means that there are not going to be any attributes or experiences that are shared by all members of G and that distinguish either G-ness or G-oppression. But if, for all G, there is no attribute or experience that is shared by all members of G that can distinguish either G-ness or G-oppression, then, for all G, neither G nor G-oppression is a genuine kind. Note that this step of the argument seems to rest on an implicit background assumption that there must be some universal feature that unifies members of a kind in order for there to be a kind.32 From there it is concluded that mutual constitution entails the dissolution of all social categories and undermines all claims of group oppression—that is, ontological fragmentation.

C.  Political Fragmentation

Having presented why many take intersectionality to lead to ontological fragmentation, I will now turn to the primary concern: ontological fragmentation, the thought goes, leads to political fragmentation.33

We can identify four major strains of the worry that ontological fragmentation is politically damaging. The first and dominant strain is based on the conception that feminism organizes around women.34 If feminist politics is aimed at combatting the systematic injustices that women face, then if the group women is dissolved, feminist politics also dissolves. As Bach puts it, “if there is no real group ‘women,’ then it is incoherent to make moral claims and advance political policies on behalf of women.”35 The second and third strains of the worry are presented by Young. Young argues that “without conceptualizing women as a group in some sense, it is not possible to conceptualize oppression as a systematic, structured, institutional process.”36 The thought is that by leaving us with an individualistic perspective, ontological fragmentation impedes our ability to identify and politically mobilize around women’s oppression, or any kind of oppression for that matter. Further, Young worries that if we are unable to identify women as a group, then we are unable to identify feminist politics as a specific, distinctive movement; at best we are left only with a general radical politics aimed at establishing social justice for all people. This connects back to the first strain of the worry: there can be no specific, feminist politics without recovering the group women, since what distinguishes feminist politics is that it organizes around women in particular.37 Lastly, Zack sees groups as carving out domains of assistance and cooperation. Members of the same group, by virtue of being relevantly similar, are entitled to equal treatment and assistance from their fellow group members. By denying that more and less privileged women are part of the same group, women, ontological fragmentation undermines this moral entitlement and corrodes the motivation that privileged women would have to assist other, less privileged women.38 All of these strains see the ontological fragmentation of women as undermining and thereby fragmenting feminist politics.

This leads me to the central challenge of this article: to find a conceptual basis for feminism and similar social movements that retains the key intersectional commitments. My intervention is going to be on (1): that intersectionality leads to ontological fragmentation. But while most of the literature has focused on group kinds (like women) when worrying about fragmentation, I will instead focus on oppressive kinds. This is largely because, like others, what I take to be most worrisome about fragmentation is the idea that it corrodes our ability to identify and mobilize around systemic injustice. What seems most directly pressing, then, is whether we are able to conceptualize, identify, and ameliorate oppression. I will argue that we can have an account of oppression that is compatible with intersectional commitments and that can serve as the basis for political movements focused on particular forms of systemic injustice—thereby ameliorating the worry about politically damaging fragmentation.

To be clear, I think there is also room to push on (2): that ontological fragmentation leads to political fragmentation—or roughly conversely, that political cohesion requires ontological cohesion. I do not think that effective political movements necessarily require shared struggle, and I certainly want to leave room for solidarity.39 However, a key question here is whether there can be a basis for political movements that organize around specific forms of oppression, and indeed whether we can retain a concept of oppression at all. In response, I argue that we can hold on to intersectional commitments while retaining both claims of oppression and a nonarbitrary basis for specific political movements like feminism.

III.  Resisting Ontological Fragmentation

I now turn to the central aim of the article, which is to provide an account of oppression that is compatible with core intersectional notions. In so doing, I show that intersectionality does not, at least, lead to the ontological fragmentation of oppressions—and this, I argue, is sufficient for resisting political fragmentation. I will focus on the ontological portion in this section, and I will show how this ameliorates the different strains of the political fragmentation worry in the next section.

A.  An Account of Oppression

The account of oppression that I offer distinguishes between three levels of phenomena: events, patterns of events, and collections of patterns of events. By events I mean particular occurrences, or something that happens at a particular time and place, like Nayeli being catcalled by a truck driver at 2:00 p.m. last Tuesday. Patterns of events, like a pattern of women being catcalled on the street, are not particular occurrences, but are at least partially constituted by particular occurrences.40 Following Dennett, a pattern exists—really exists—when it constitutes an efficient description of some data.41 So there is a real pattern of women being catcalled in the street when this is an efficient way of describing a set of events. This pattern would include Nayeli being catcalled by the truck driver at 2:00 p.m. last Tuesday, but also other instances of catcalling experienced by Nayeli and other women. On a still higher level, we can also talk about collections of patterns of events. For example, one collection of patterns of events could consist in a pattern of women being catcalled on the street, a pattern of women being paid less than their male counterparts, and a pattern of women being “mansplained” to about their areas of expertise.

One important feature of the Dennettian account of patterns is that there can be multiple real patterns that exist in the same data simultaneously. This is because there can be multiple efficient descriptions of the same data. For instance, consider a pattern of Black women being treated as inappropriately angry. This same set of events could also be described as a pattern of women being treated as overly emotional. Both of these patterns can be real patterns that describe the same set of events, because both serve as an efficient description of the same data, just at different levels of specificity.42 Further, not only can there be multiple real patterns that capture the same set of events, but there can also simultaneously be real patterns that describe subsets of those data—I will use subpatterns to refer to real patterns that describe more specific subsets of a broader pattern. For instance, a real pattern of women being treated as overly emotional could also have subpatterns of Black women being treated as inappropriately angry, Latinas being regarded as “spicy,” and white women being treated like they could burst into tears at any moment. All of these patterns and subpatterns can be real patterns that exist simultaneously, giving rise to a picture on which there are multiple, layered, and overlapping patterns. Which patterns we choose to emphasize or talk about may vary across contexts and depending on our purposes, but they are all real.

Moving forward with an overview of the account, I will understand oppression in terms of collections of patterns of injustice that have certain properties. Part of this picture is that these patterns affect individuals who possess features that are connected in an important way to an ideological conception of what the social order is or should be like. This gives rise to collections of patterns of injustice that cluster around those individuals. Due to the subpatterning discussed above, these patterns, as well as the clusters that they help constitute, can be more or less coarse-grained, giving rise to nesting and overlapping oppressive kinds. With that overview in place, I will now get into the details of the account.

Properties of oppression.—On my account, oppression consists in a collection of patterns of injustice that have certain properties.43 In particular, I will say that the relevant patterns must be persistent and Sort based and that the collection of patterns as a whole must be domain-crossing, interlinking, and robust.44

By persistent I mean that the patterns are not short-lived but long-lasting and enduring.

By domain-crossing I mean that there are many patterns of injustice in this collection and that they span across many, or virtually all, domains of life. For example, the patterns of injustice in the collection affect individuals across financial, medical, educational, legal, religious, environmental, and interpersonal domains.

By interlinking I mean that different patterns in the collection “interlink” in such a way that the overall effect of these patterns of injustice is greater than the sum of the effects of these patterns considered independently. For example, it would be bad enough if poor women of color in the United States often lacked access to health insurance, or were frequently unable to take time off work to see a doctor, or frequently had their reports of pain dismissed by medical providers. But it is even worse when all of these patterns operate simultaneously, because managing to surmount one barrier is still insufficient to access care. The notion of being interlinking captures the ways in which the effects of some patterns exacerbate the effects of other patterns in the collection.

Lastly, by robust I mean that the overall conditions established by this collection of patterns of injustice are quite stable. Attempts to change these conditions often result in the reemergence of these patterns in a different form.45 For example, one might have thought that getting more women into the workforce would lead to a more equitable division of household labor, but instead sociologists have described a “second shift” phenomenon where working women often end up taking on the bulk of household duties after they’ve completed their paid workday.46

There is one more important property—the patterns in the collection must be Sort basedbut this requires more extensive discussion.

Sort based.—Another important feature of this account is that the patterns must be what I call Sort based.47 This is analogous to the widespread idea that oppression is group based, meaning that there must be some causal connection between membership in some social group and having the relevant experiences of oppression.48 But I am introducing the term ‘Sort’ to correspond, roughly, to an ideological conception of a group or kind—which is importantly different from a social group or identity.49

A Sort is a conception of a group or kind that is constituted by widely shared, conceptually interconnected schemas (e.g., beliefs, concepts, attitudes, dispositions) about that group.50 As Sorts are conceptions of a group, the group in question need not be real. Consider the Sort <vampire>, for instance.51 Due to a body of lore about vampires, there is a general conception of what vampires are, what they are like, how one should interact with them, and so on: they are characteristically undead creatures who drink blood, are generally malevolent, can be warded off using holy water, and can (and should) be destroyed by driving a stake through their heart. There are conceptual relations between the different schemas that help constitute <vampire>: the idea that they have a demonic or unholy nature, for instance, is connected to the idea that they can be kept at bay with holy water. While it is now a common part of the conception of a vampire that they are mythological creatures, this has not always been the case—vampires have historically been blamed for incidences of plague, for instance.52 Again, the conception of a vampire is cultural, or ideological in the nonpejorative sense—it is made up of culturally shared notions of a certain type of creature. There are many beliefs, concepts, and so on, that help constitute <vampire>, so as to give rise to a fairly rich conception of vampires.

Likewise, we can consider the Sort <women>. <Women> refers to a conception of a particular kind of social group. As with <vampire>, <women> is made up of various conceptually interconnected schemas that together give rise to a rich conception of women. For instance, while nowadays <vampire> contains the belief that vampires are fictional, it is part of the ideology constituting <women> that there is such a group called “women.” Further, there are various schemas that specify the features that regulate membership in this group, what its members tend to be like, what its members should be like, how others should interact with those individuals, and so on.

While <women> is a matter of ideology and is therefore abstract and immaterial, it has real effects because of how ideological schemas coordinate our behavior. To illustrate, consider another Sort: <breakfast>. As with vampires and women, we have culturally specific conceptions of breakfast—what it consists of, what does and does not count as breakfast, when one should eat breakfast, how breakfast relates to other meals, who prepares breakfast, and so on. This conception of breakfast leads us to think about and interact with objects in ways that are guided by this conception—for example, to cook eggs; to crave pancakes for brunch; to identify cinnamon rolls, but not salad, as breakfast food; to urge someone to eat breakfast because it’s the most important meal of the day. Similarly, the widespread internalization of the schemas constituting <women> results in there being people who, for example, are thought of as women, who think of themselves as women, who are treated in the ways that the ideology says women should be treated, and who feel responsive to the norms meant to regulate women’s behavior—although it should be emphasized that different people will fit into these different categories (e.g., not everyone who identifies as a woman is identified by others as a woman).53 A Sort, then, is a conception of a group or kind that is constituted by conceptually interconnected ideological schemas and that often shapes how we categorize, conceptualize, act, and interact with the world in our everyday lives. A Sort thereby has connections to identity—it can lead individuals to identify themselves and others in certain ways—but it does not correspond neatly to identity. Likewise, a Sort, being a conception of a social group, is importantly different from a social group—even when it leads to there being collections of individuals who fit, in differing ways and to varying degrees, the descriptions in the ideology.54

To proceed with developing my account of oppression, the patterns of injustice must be Sort based. In saying that a pattern of injustice is Sort based, I mean that there are one or more structural mechanisms that underlie the pattern in question and that track features associated with a Sort. To first spell out what I mean by these features, on this picture the features could take a variety of forms. They could include physical features, like height, body shape, or skin color. But they could also include more complex social and historical features, like having a particular job title, being thought of in certain ways (e.g., being identified as a woman), or financially relying on subprime lending sources. A feature is directly associated with a Sort if it is taken to be a criterion for Sort membership (e.g., having a uterus, for <women>), or if the ideology generates an expectation that members will have this feature (e.g., being nurturing, for <women>). A feature is recursively associated with a Sort if the individual’s having this feature is explanatorily connected to their having features that are directly associated with the Sort. For instance, historically, individuals identified as people of color in the United States, and particularly Black Americans, have been shut out of prime lending sources and thereby have been forced to rely on subprime lending sources.55 Thus, there is an explanatory connection between having the feature of financially relying on subprime lending sources and having features that are directly associated with marginalized racial Sorts. Thus, we can say that the feature of financially relying on subprime lending sources is recursively associated with these Sorts.

Again, in saying that a pattern is Sort based, I mean that there are one or more structural mechanisms that underlie the pattern in question and that track features that are associated with some Sort. The structural mechanisms can take two main forms: the form of the very same ideological schemas that help constitute the Sort, or the form of laws and institutional policies (e.g., think about the policies that govern a university, or the regulations that a bank follows).56 I will give two examples to illustrate how structural mechanisms can track these features. First, consider schemas that depict women’s bodies as being available for men in various ways. Plausibly, for those who have internalized them, such schemas generate a sense of entitlement to comment on the bodies of individuals with features that are directly associated with <women>. This in turn helps to explain the pattern wherein individuals with such features are catcalled—thus, this pattern is Sort based. Second, consider institutional policies that rely on credit scores to make decisions (e.g., about whether to approve a mortgage). Because reliance on subprime lending sources leads to a lower credit score (by virtue of how these scores are calculated), institutional policies that rely on credit scores to make decisions are going to disadvantage individuals who rely on these sources. Thus, these policies track reliance on subprime lending sources, which, as discussed, is recursively associated with marginalized racial Sorts in the United States. In these examples, then, there is a structural mechanism (the relevant law, policy, or schema) that tracks a feature that is associated with some Sort such as to give rise to a Sort-based pattern of injustice.

Before moving on, I would like to make a few notes about this account. First, Sorts may be implicit to varying degrees. By this I mean that while there may be various schemas that constitute a Sort, there may be no term associated with the Sort or even an explicit awareness that such a Sort has been carved out.57 For instance, consider the mid-twentieth-century Afro-Brazilian women’s movement.58 This movement emerged in a context where there was not yet a conception of Afro-Brazilian women as such. Nevertheless, through consciousness raising and political organizing, members of this movement came to realize that they shared common experiences of systemic injustice that were distinct from the experiences of injustice faced by non-Black Brazilian women and Black Brazilian men. Accordingly, while there was not yet a particular term or identity associated with these individuals, there were schemas and policies in place that systematically tracked them—the ideology had implicitly carved out a Sort, though explicit awareness of this was lacking. Through consciousness raising they were able to recognize that they were experiencing a distinct form of oppression that tracked this Sort, and it was by mobilizing around this shared oppression that they carved out a robust political identity of Black womanhood in Brazil.59 The lessons here are, first, that a Sort can be implicit and, second, that both Sorts and corresponding oppressive kinds can be uncovered. Relatedly, I want to flag that in speaking of “experiences” of oppression, I do not mean that the person is necessarily aware that what they are experiencing is oppression (or some particular kind of oppression), or that there is some particular phenomenology associated with those experiences.

Second, Sorts can be nested and overlapping and need not be totally coherent. For instance, consider the Sort <white women>. This Sort is partly constituted by schemas with contents specifically about “white women” (e.g., schemas that depict white women as needing protection, or that cast white women as the paradigm of womanhood). There are also schemas that connect <white women> to <women>: for example, the ideology takes white women to be women. Consequently, <white women> can be considered a compound Sort that generally functions as a sub-Sort of <women>, and various schemas that help make up <women> also apply to <white women>. It is important to note, however, that schemas aren’t always straightforwardly inherited by sub-Sorts in this way. Other features associated with the sub-Sort may mask or block the application of a schema. For instance, schemas sexualizing women often do not straightforwardly apply to individuals identified as disabled, plausibly due to conflicts between schemas desexualizing people with certain disabilities and schemas sexualizing women. This illustrates some of the ways in which Sorts (or more broadly, ideologies) are not always coherent: the ideology may say that Xs are Y and Ys are Z yet deny in various ways that Xs are Z. This is reflected in Sojourner Truth’s refrain “Ain’t I a Woman?,” which captures the tension wherein Black women are simultaneously identified as women and overlooked as women, in part due to ways in which the ideology positions white women as the normative baseline for womanhood.

Lastly, the account leaves open questions about whether or how to conceptualize social groups like women. For instance, each of the following would be compatible with the account: that women are those who identify as women, that women are those with features that <women> depicts as constitutive of membership in the group women, that women are those who experience <women>-tracking patterns of injustice, that women are those who mobilize around the oppressive kind that tracks <women>, or that there is no coherent social group corresponding to “women.”60 Thus, the account does not attempt to define women, nor does it depend on resolving the ontological fragmentation of social groups. This has two important upshots: (1) it does not commit us to exclusive views of women, and (2) feminism does not first need to define women as a social group in order to conceptualize gender-based oppression.

To put these pieces together, I will characterize oppression as follows:

Collective Oppression: 

A collective oppressive kind O𝛴 is constituted by a robust collection of persistent, domain-crossing, and interlinking patterns of injustice that track the Sort 𝛴.

Individual Oppression: 

An individual S is oppressed when there exists a collection of persistent patterns of injustice such that

afor each pattern I in the collection, I tracks features that S has, such that S either has had or is liable to have experiences of this kind; and

bthe collection of patterns is domain-crossing, interlinking, and robust.

B.  Compatibility with Mutual Constitution

With this conceptual infrastructure in place, I will now show that this account of oppression is compatible with core intersectional notions.

Inseparability.—Recall that the notion of inseparability rejects the idea that we can cleanly separate out oppressions into “pure,” universal parts. As reflected in the regress argument, many have thought that we cannot have oppressive kinds unless we can identify some distinctive, universal experience that can be used to unify and individuate the kind. My account, however, does not require identifying any such universal experiences. Rather, my account allows that (1) not all individuals associated with Sort 𝛴 need to have some given experience of injustice in order for that experience of injustice to help constitute 𝛴-oppression, and (2) there need be no experience of injustice that all individuals associated with 𝛴 have in order for there to be 𝛴-oppression. On a general level, this is because what unifies a collective oppressive kind on my account is not any universal experience of injustice, but rather the clustering of injustices in relation to an ideological Sort.61 There needn’t be any universally shared experience because individuals’ heterogeneous experiences may still bear the right relation to the Sort. This helps us to see where the infinite regress argument goes wrong—universality is not required in the way that the argument assumes.

Mutual constitution.—Next, the account is also able to reflect the notion of mutual constitution. Recall that mutual constitution says that identities and oppressions gain meaning in relation to each other. We can see this reflected on two levels—on the level of individual experience, and at the level of Sorts and the schemas that constitute them.

On a more individual level, we can think of mutual constitution as saying that what it means for an individual to experience 𝛴-oppression is going to depend on all of the Sorts that they are associated with. To see how my account reflects this, I want to show that (1) individuals associated with sub-Sorts of 𝛴 may experience distinctive subpatterns of the patterns that help constitute 𝛴-oppression (such that an Asian woman’s experience of catcalling, e.g., reflects racial dynamics) and (2) individuals associated with sub-Sorts of 𝛴 may have unique experiences of injustice that are not shared by other individuals associated with 𝛴 (e.g., Black women may have unique experiences that other women do not have), but which still help constitute 𝛴-oppression.

First, the account allows that different sub-Sorts of 𝛴 may experience distinctive subpatterns of the patterns of injustice that track 𝛴. As previously discussed, there can subpatterns of any pattern of injustice. In particular, there can be subpatterns that reflect how association with other Sorts affects the experience of a certain kind of injustice. For example, there can simultaneously be a general pattern of catcalling that tracks <women> and various subpatterns that reflect how association with other Sorts (e.g., being identified as queer, trans, white, or Asian) modulates that experience. This could happen through a combination of mechanisms that track different Sorts—Crenshaw’s example of the immigration law that led many women to feel trapped in abusive relationships owing to fear of deportation seems to be like this: the pattern seems to result from laws tracking immigrant status interacting with schemas surrounding domestic violence. Or this could also result from sub-Sort-specific versions of a mechanism that tracks the Sort—for example, schemas that specifically objectify Asian women in particularly exoticizing ways, as compared to more general schemas that objectify women.62 Thus, how one individual experiences 𝛴-oppression may differ from how others experience 𝛴-oppression in ways that depend on the other Sorts that they are associated with.

Second and relatedly, the account also allows that different sub-Sorts of 𝛴 may have unique experiences of oppression that are not shared by other individuals associated with 𝛴, but which still help constitute 𝛴-oppression. For instance, consider the pattern, which tracks <Black women>, of touching someone’s hair without permission.63 There are two points to be made here. First, the account allows that this pattern could be distinctive to <Black women> and still help constitute the oppressive kind that tracks <women>, insofar as it is persistent and tracks features associated with <women>. Second, the account allows that this pattern may simultaneously help constitute multiple oppressive kinds—including one that tracks <women> and one that tracks <Black women>. This shows that the account accommodates mutual constitution both by allowing that an individual’s experiences of 𝛴-oppression will be a function of the different Sorts they are associated with and by allowing that individuals associated with sub-Sorts of 𝛴 may simultaneously experience forms of collective oppression that are sub-Sort based.64 Thus, combined with the first point about subpatterns of injustice, the account allows us to capture Crenshaw’s insights about the variety of ways in which Black women’s experiences can be similar to and also distinct from white women’s and Black men’s experiences: “Black women sometimes experience discrimination in ways similar to white women’s experiences; sometimes they share very similar experiences with Black men. Yet often they experience double-discrimination—the combined effects of practices which discriminate on the basis of race, and on the basis of sex. And sometimes, they experience discrimination as Black women—not the sum of race and sex discrimination, but as Black women.”65 We can thus see how the account again accounts for inseparability—there is no “pure” racism or “pure” sexism, but rather 𝛴-oppression is partly constituted by patterns that result from the interaction of mechanisms that track 𝛴 with mechanisms that track other Sorts, as well as by patterns that track specific sub-Sorts of 𝛴. Thus, racial oppression will include patterns of injustice that are inseparably raced, gendered, and so on. Moreover, we see that what it means to experience 𝛴-oppression is dependent on, and so relationally defined in terms of, the other Sorts one is associated with.

So far this looks at mutual constitution from a more individual perspective, but we can also think about mutual constitution on a higher level—not just the mutual constitution of individuals’ experiences of race, class, gender, and so on, but also the mutual constitution of race, class, gender, and so on, itself. The ways in which an individual’s experience of, for example, race is modulated by the other Sorts that they are associated with reflect the ways in which Sorts are relationally defined—the norms and roles of one Sort interact with those of other Sorts. For instance, gender norms that help constitute <women> are raced and classed in various ways: sometimes different norms apply to specific sub-Sorts of <women>, or there are different variations of the norm for different sub-Sorts, or different individuals are afforded different degrees of leeway when being held accountable to the norms. Relatedly, compound Sorts often carve out distinctive roles, creating class roles, for instance, that are raced and gendered. Collins illustrates this with the example of the typical American university, where working-class men of color typically perform maintenance roles, working-class women of color clean and feed people, wealthy white men occupy high-level positions of power, and middle-class white men and women fill faculty and staff positions.66 These effects result from the ways in which schemas that constitute gender Sorts also constitute racial Sorts, class Sorts, and so on, such that Sorts are in this sense co-constituting. But further, the schemas that make up these co-constituting Sorts reflect a broader ideological view of what the social order should be like that often assigns individuals to specific roles in relation to compound Sorts. Thus, gender divisions interact with racial divisions, class divisions, and so on (and vice versa).

In these ways, the account of oppression I have offered is compatible with mutual constitution. It accommodates the ways in which individuals’ experiences of any particular kind of oppression will be affected by their simultaneous association with multiple Sorts, which in turn reflects the broader ways in which these Sorts mutually constitute each other to create gendered racism, raced classism, and so on.

Simultaneity.—Lastly, the account also reflects the notion of simultaneity. In particular, on this account an individual’s experience of oppression will be a cluster of persistent, domain-crossing, and interlinking injustices that reflect the overall complexity of the individual’s social position, without chopping up either the individual or the oppression they experience into a combination of “pure” parts.67 In this way, the account is able to reflect the fact that individuals are integrated wholes who simultaneously have multiple identities and experience multiple forms of oppression.

IV.  Resisting Political Fragmentation

At this point I have provided an account of oppression and argued that it is compatible with core intersectional notions, thereby allowing us to reject the claim that intersectionality leads to the ontological fragmentation of oppressive kinds. Given that the concern about fragmentation is ultimately political in nature, as well as that political fragmentation has been thought to result from the ontological fragmentation of social groups (because of the primacy such groups have been accorded in conceptualizing radical political movements), I will now show that my account can ameliorate the worries about political fragmentation.

Recall that there are four major strains of the political fragmentation worry. The first and most dominant strain conceptualizes feminism as organizing around women, such that if the group women is fragmented or dissolved, then so too is feminist politics. The second strain worries that we cannot conceptualize oppression, and particularly women’s oppression, as a systematic form of injustice without conceptualizing women as a group. The third strain, closely related to the first, worries that there can be no specific, feminist politics without the group women, since what distinguishes feminist politics is that it organizes around women in particular—and likewise for other specific, radical political movements. Lastly, the fourth strain worries that without group kinds, and specifically the group women, to carve out domains of assistance and cooperation, there is no reason for more privileged women to assist other, less privileged women in achieving their political aims.

To start with the second strain, which worries about our ability to conceptualize oppression, giving an account of oppression that is compatible with core intersectional notions directly alleviates this worry. To briefly consider what we can learn from this, while it is correct that an individualistic approach inhibits our ability to identify oppression, we don’t need social groups per se.68 Rather, what we need is some “connective tissue” that will enable us to identify patterns of injustice. Social groups could provide this—unjust events can be bundled together according to whether they affect members of a particular group—but they are not the only thing that can. On my account, Sorts play this unifying role. On one level, the schemas that help constitute a Sort, as well as the institutional structures that they help shape, give rise to patterns of injustice that affect individuals associated with that Sort. In this way, Sorts unify the events that make up a particular pattern of injustice. On a higher level, the interconnected nature of those Sort-constituting schemas also helps to unify and interlink different patterns of injustice. In this way, Sorts provide the connective tissue that allows us to bundle together patterns of injustice into clusters that individuate oppressive kinds. It is true that Sorts “conceptualiz[e] women as a group in some sense,” but as we have seen, Sorts are not social groups, nor do they straightforwardly give rise to some coherent, unified account of social groups. This is not to say that we cannot figure out a way of providing an account of women as a social group, nor that we should necessarily abandon the notion of social groups like women. It is just to say that we do not, as many have thought, need to first resolve debates in the metaphysics of gender and recover women as a coherent ontological kind in order to conceptualize gendered oppression.

Shifting briefly from the metaphysics of collective oppression to the epistemology of coming to recognize oneself as sharing experiences of oppression with others, it seems right to say that this will be facilitated by some recognition of similarity and difference. Plausibly, one needs some sense that “we” share experiences that “they” do not in order to conclude that there is some shared form of collective oppression.69 But as illustrated by the case of the Afro-Brazilian women’s movement, this sense of similarity need not take the form of some fully fledged conception of shared membership in a social group. Being willing and able to compare experiences with someone does not require there to be a coherent, unified account of a group that you share membership in. Further, it seems possible that the sense of similarity and difference at play could be sharpened and made more nuanced as one continues to make comparisons and explore the political terrain. In this way, one again does not need social groups per se to get the necessary connective tissue for conceptualizing oppression, though some sense of similarity and difference will be epistemically useful.

Turning now to the remaining strains of the political fragmentation worry, my account allows us to think about feminism as organizing around a particular oppressive kind, as opposed to a particular group kind. In particular, we can conceive of feminism as organizing around a collection of persistent, domain-crossing, interlocking injustices that track <women>. This provides a specific and nonarbitrary basis for feminist politics that can distinguish it from other radical political movements (thereby responding to the first and third strains of the worry). Further, we can also see oppressive kinds as carving out domains of assistance and cooperation—there is reason to work with and aid those who share experiences of systematic injustice with you (thereby responding to the fourth strain).

But I want to note that this account affords us choices about how we conceive of feminist politics, as well as radical politics more generally. One option, as just outlined, is to see feminist politics as organizing around a particular collective oppressive kind. Put differently, since collective oppressive kinds are here individuated by Sorts, this Sort-based approach would take specific radical political movements (like feminism) to organize around injustices that track particular Sorts. This seems like the closest match to the way in which feminism has typically been conceived, except that instead of organizing around the group women, it takes feminism to organize around injustices that track <women>. But another option would be to take specific radical political movements to organize around particular ideological orders. For instance, we can see the ideological schemas that construct <women> as helping constitute an ideological vision of what the social order is and should be like that is organized around gender. So feminism could be thought of as a movement that aims to challenge this oppressive gender order. Yet another option is to see feminism and similar movements as organizing around changing underlying social structures. Recall that part of the account of oppression is that there are structural mechanisms that track the features associated with a Sort so as to give rise to patterns of injustice that constitute a particular collective oppressive kind.70 The structure-based approach would focus on transforming or getting rid of these oppressive structures. Depending on whether we go for a Sort-based, order-based, or structure-based approach, feminist politics may be narrower or broader in scope. It can look much like the kind of feminism I assume Young imagined, but it can also be broader than traditionally conceived, including the struggles of, for example, queer men who are also oppressed as a result of structures that uphold an unjust gender order. The important point is that there are coherent grounds—and indeed, a choice of grounds—for feminist politics.71

What this shows is that my account provides the resources to ameliorate the different strains of the political fragmentation worry and to recover a basis for feminist politics roughly as it has been envisioned by feminist philosophers and activists. But my account also offers the resources for a more expansive feminist and radical politics that is unified by its commitment to transforming or eradicating oppressive social structures. Further, it makes room for solidarity—one can join a political movement that focuses on changing an unjust social order because one recognizes it as unjust, and not merely because one is directly affected by it.

To be sure, this account does not guarantee political smooth sailing. Hard work will be needed to ensure that those who enter into political coalition actually listen to each other, support each other, and adequately balance their respective material interests. But the account does ameliorate the worry that intersectionality undermines the very coherency of these kinds of coalitions, and it also serves as a reminder to be attentive to the important differences that will exist within any coalition.72

V.  Alternative Views

I have shown that we can retain an understanding of oppression while accepting intersectionality and that we can conceive of feminist and other radical politics as organizing in response to oppression. My approach has the benefit of allowing us to retain intersectional commitments while focusing on what makes the prospect of fragmentation most troubling—namely, the worry that it will impede our ability to identify and respond to systematic injustice. Further, by focusing on oppression rather than identity or group kinds, my approach also leaves open the possibility of adopting a pluralism about gender kinds and identity. To highlight these further benefits of my account, I will briefly review the three main approaches to the fragmentation problem that have been taken in the literature: first, to abandon the intersectional commitments that are thought to give rise to it; second, to attempt to provide an account of gender that incorporates intersectional ideas; and third, to reverse the assumption that feminist politics mobilizes around women and instead conceive of women as being constructed by feminist politics.

The first kind of response to the fragmentation worry has been to reject the commitments of intersectionality. This is Zack’s response to intersectionality, though she also attempts to provide an account of gender that is meant to be inclusive in a way that accommodates at least some insights of intersectionality.73 More recently, although they certainly remain committed to the spirit of intersectionality, Gasdaglis and Madva have also adopted a version of this position, in that they reject mutual constitution as a thesis and instead provide a methodological interpretation of intersectionality.74 They suggest that we should understand intersectionality as a guiding methodological principle rather than take it to be literally true that identity categories or oppressions are mutually constituting in order to avoid an infinite regress. More specifically, they propose that we understand mutual constitution as a maxim directing us to “(a) treat social classification schemes, groupings, or categories as if they are indefinitely mutually informing, (b) with the aim of revealing and resisting inequality and injustice.”75 Adopting this maxim urges us to consider and seek out ways in which these schemes may be mutually informing as we work in a number of domains, thereby facilitating the correctional work that intersectionality was introduced to do. While I agree that we should adopt this methodological principle, on my account we do not have to merely act as if mutual constitution were true; we can also take it to be true.

The second kind of response to the fragmentation worry has been to “fix” gender by providing some unifying, metaphysical account of women that doesn’t require women to share some universal feature. Within this camp, strategies have included providing disjunctive accounts,76 family resemblance accounts,77 and accounts that understand gender on the basis of social relations such as systems of heterosexuality and the sexual division of labor.78 However, these accounts have struggled to meet their goal of providing an inclusive metaphysics of gender, and in particular to provide an account that does not marginalize or exclude trans women, or inappropriately include trans men and nonbinary people.79 Moreover, this strategy does not directly address what seems to be most troubling about fragmentation—namely, the suggestion that it impedes our ability to identify and address systematic injustice. Rather, it has been assumed that to identify and organize in response to gender oppression we first must solve the metaphysical problem of gender. But, as I have shown, this is not the case.80

The third kind of response to the fragmentation worry has been to conceive of the construction of identity and group kinds as itself a coalitional political project, rather than to see politics as organizing around fixed, preexisting identity groups.81 Crenshaw, for instance, suggests that we should recognize identity groups as actual or potential coalitions, and Carastathis argues that we should take this suggestion literally, as an account of what identities are.82 Fuss describes the view as follows: “Many anti-essentialists fear that positing a political coalition of women risks presuming that there must first be a natural class of women; but this belief only masks the fact that it is coalition politics which constructs the category of women (and men) in the first place.”83 Some, like Young, have worried that the coalitional view seems arbitrary—yes, individuals can choose to come together to form a political movement and from there forge a group identity, but why would they choose to come together in the first place?84 However, this worry does not quite do justice to the view, which sees coalition, and hence identity, as resulting from choices to organize together around a shared struggle.85

My view shares the core idea underlying the coalitional view: that shared struggle provides a nonarbitrary basis for political coalition. My view can be seen as helping to develop the response to Young’s worry by spelling out in more detail how, in light of intersectional insights, we can see individuals as sharing a struggle. My view also elucidates the fact that individuals have choices as to which of the nesting and overlapping oppressive kinds they mobilize around, and thus who they enter into coalition with. It is in this sense that shared struggle is not, as some emphasize, a road map for politics.

Where I depart from the traditional coalitional view is with respect to the focus on identity, and particularly the tying of identity to political coalitions. Unlike the coalitional view, I do not think that we should take political coalitions to fully define identity groups like women. To be clear, I do not want to deny that this kind of political organizing can give rise to strategic political identities—it seems to have done so in the case of the Afro-Brazilian women’s movement. I also think that Carastathis and Crenshaw are right in saying that there is a substantial benefit to recognizing that any identity group will be a coalition of heterogeneous individuals. What I want to deny is that identities or group kinds must be defined or fully fixed by a political movement organizing around an oppressive kind.86

I want to hold off on fixing identities and group kinds in terms of political coalitions for various related reasons. First, I want to remain open to a pluralistic view of gender kinds like the one defended by Dembroff, which takes gender kinds to be social kinds and allows us to recognize dominant gender kinds as being ontologically oppressive precisely because of the ways in which they marginalize trans women and exclude nonbinary people.87 A pluralistic picture allows us to identify this kind of ontological oppression while also leaving room for inclusive gender kinds and, more broadly, the development of strategic identities for pragmatic, political purposes.

Second and relatedly, I want to allow that individuals may be affected by an oppressive kind and, for this reason, choose to be part of a coalition that responds to that oppressive kind, without it being the case that an identity tied to that coalition (or the corresponding Sort) is authentic for them. I am here influenced by Bell’s view of social authenticity, which recognizes authenticity and identity as “a project of constructing an intelligible self out of available materials from one’s social context.”88 Different individuals who share a struggle and who organize together in response to that struggle can differ significantly in terms of what is authentic for them and therefore in their identities. There could be trans men and nonbinary individuals, for instance, who choose to join in coalition against the oppressive kind that tracks <women>—and indeed, part of their motivation may be that they face certain injustices because the ideology attempts to impose an identity on them (viz. woman) that is not authentic for them. Further, individuals may choose to join a political coalition out of solidarity with the oppressed, even if they are not directly affected by the oppressive kind that the coalition centers on. Defining “women” in terms of a feminist coalition would thereby get things wrong. This can be true even if for other individuals such a coalition does provide the basis for an identity that is authentic for them.

More broadly, I want to leave room for complicated relationships between oppression, coalition, identity, and social groups. I want to allow that for some individuals the best move is to reject an identity like woman as an oppressive imposition, while for others it makes sense to accept that identity because it captures certain experiences they take to be formative, and for still others the best move may be to embrace a counterhegemonic conception of the identity that emerges through coalition. Different moves and different combinations of moves may feel authentic for different individuals who share a struggle against an oppressive kind.

VI.  Conclusion

I have offered an account that understands oppression in terms of collections of patterns of injustice with certain high-level properties, rather than in terms of any particular kind or set of injustices, and that individuates collective oppressive kinds in relation to Sorts, rather than social groups. This account allows that oppressions are inseparable and mutually constituting and that individuals are integrated wholes who simultaneously have multiple identities and experience multiple forms of oppression. The account is thereby compatible with the core intersectional notions thought to lead to ontological fragmentation. Further, and most importantly, I have shown that this account allows us to conceive of feminist politics and other radical political movements as organizing around collective oppressive kinds, oppressive social orders, or oppressive social structures—thereby resisting political fragmentation.

By shifting the focus from identity to oppression, I have joined others in calling into question the assumption that feminist politics must start with the metaphysics of gender, suggesting instead that what it requires is a recognition of systematic injustice.89 By letting go of identity as a necessary starting point for feminism, my account allows for pluralism about gender kinds and complicated relationships between oppression and identity. My account leaves room for strategic choices about how we organize, as well as which identities we adopt to live authentically.

However we choose to organize, this account shows that intersectionality does not undermine the possibility of specific, radical political movements. Rather, as others have emphasized, leaning into intersectional insights can help us to form bridges with others and strengthen our political coalitions.90 Intersectionality is not a threat, and both our theorizing and our politics should embrace the important insights that intersectionality provides.

Notes

* For helpful discussion and feedback on earlier versions of this article, I would like to thank Eric Bayruns García, Caroline Bowman, Daniel Brinkerhoff Young, Yarran Hominh, Laura Martin, Ariel Melamedoff, and Philip Yaure. I would also like to thank audiences at the 2022 Central APA session of the Society for Analytical Feminism and at the Centre College Colloquium, where I presented this article. Finally, I would like to thank anonymous referees and editors at Ethics for their comments and suggestions.

1.  Linda Alcoff, “Cultural Feminism versus Post-structuralism: The Identity Crisis in Feminist Theory,” Signs: Journal of Women in Culture and Society 13 (1988): 405–36, 405.

2.  See, e.g., Anna Carastathis, “Identity Categories as Potential Coalitions,” Signs: Journal of Women in Culture and Society 38 (2013): 941–65; Patricia Hill Collins and Sirma Bilge, Intersectionality (Hoboken, NJ: Wiley, 2020); Combahee River Collective, “The Combahee River Collective Statement,” in How We Get Free: Black Feminism and the Combahee River Collective, ed. Keeanga-Yamahtta Taylor (Chicago: Haymarket, 2017), 15–27; Kimberle Crenshaw, “Mapping the Margins: Intersectionality, Identity Politics, and Violence against Women of Color,” Stanford Law Review 43 (1990): 1241–99, 1241; Mari J. Matsuda, “Beside My Sister, Facing the Enemy: Legal Theory out of Coalition,” Stanford Law Review 43 (1990): 1183–92; Sharon Parker, “Understanding Coalition,” Stanford Law Review 43 (1990): 1193–96.

3.  As discussed later, I am here aligning with Carbado in thinking that everyone has an intersectional identity, not simply those who are multiply oppressed. Thank you to an anonymous referee for pressing me to clarify this. See Devon W. Carbado, “Colorblind Intersectionality,” Signs: Journal of Women in Culture and Society 38 (2013): 811–45.

4.  Though, as I will discuss, this is not the only option for how to conceive of feminism on this account.

5.  Collins discusses some of the challenges of defining intersectionality. Patricia Hill Collins, “Intersectionality’s Definitional Dilemmas,” Annual Review of Sociology 41 (2015): 1–20.

6.  See, e.g., Gloria Anzaldua, “La Prieta,” in This Bridge Called My Back: Writings by Radical Women of Color, ed. Cherrie L. Moraga and Gloria E. Anzaldua (Berkeley: Third Woman, 2002), 220–33; Kathryn Sophia Belle, “Interlocking, Intersecting, and Intermeshing: Critical Engagements with Black and Latina Feminist Paradigms of Identity and Oppression,” Critical Philosophy of Race 8 (2020): 165–98; Maria Lugones, “Purity, Impurity, and Separation,” Signs: Journal of Women in Culture and Society 19 (1994): 458–79; Elizabeth V. Spelman, Inessential Woman: Problems of Exclusion in Feminist Thought (Boston: Beacon, 1988).

7.  Anzaldua, “La Prieta,” 228.

8.  See, e.g., Alison Bailey, “On Intersectionality, Empathy, and Feminist Solidarity: A Reply to Naomi Zack,” Journal for Peace and Justice Studies 19 (2009): 14–36; Anna Carastathis, “The Concept of Intersectionality in Feminist Theory,” Philosophy Compass 9 (2014): 304–14; Patricia Hill Collins, “Toward a New Vision: Race, Class, and Gender as Categories of Analysis and Connection,” Race, Sex and Class (1993): 25–45; Collins, “Intersectionality’s Definitional Dilemmas”; Crenshaw, “Mapping the Margins”; Ann Garry, “Intersectionality, Metaphors, and the Multiplicity of Gender,” Hypatia 26 (2011): 826–50; Katherine Gasdaglis and Alex Madva, “Intersectionality as a Regulative Ideal,” Ergo, an Open Access Journal of Philosophy 6 (2020): 1287–1330; Spelman, Inessential Woman; Iris Marion Young, “Gender as Seriality: Thinking about Women as a Social Collective,” Signs: Journal of Women in Culture and Society 19 (1994): 713–38.

9.  Spelman, Inessential Woman.

10.  This includes rejection of the idea that Black women merely experience sexism and racism to a greater degree than white women and Black men, respectively. Devon W. Carbado and Cheryl I. Harris, “Intersectionality at 30: Mapping the Margins of Anti-essentialism, Intersectionality, and Dominance Theory,” Harvard Law Review 132 (2018): 2193–2239, 2221; Angela P. Harris, “Race and Essentialism in Feminist Legal Theory,” Stanford Law Review 42 (1989): 581–616.

11.  See, e.g., Carbado and Harris, “Intersectionality at 30”; Kimberle Crenshaw, “Demarginalizing the Intersection of Race and Sex: A Black Feminist Critique of Antidiscrimination Doctrine, Feminist Theory and Antiracist Politics,” University of Chicago Legal Forum 1989 (1989): 139–67; Harris, “Race and Essentialism”; Lugones, “Purity, Impurity, and Separation.”

12.  Combahee River Collective, “Combahee River Collective Statement,” 19.

13.  Crenshaw, “Mapping the Margins,” 1247.

14.  See, e.g., Bailey, “On Intersectionality”; Collins, “Intersectionality’s Definitional Dilemmas”; Collins and Bilge, Intersectionality; Garry, “Intersectionality, Metaphors”; Gasdaglis and Madva, “Intersectionality as a Regulative Ideal”; Stephanie A. Shields, “Gender: An Intersectionality Perspective,” Sex Roles 59 (2008): 301–11; Spelman, Inessential Woman; Young, “Gender as Seriality”; Naomi Zack, Inclusive Feminism: A Third Wave Theory of Women’s Commonality (Lanham, MD: Rowman & Littlefield, 2005).

15.  Shields, “Gender,” 303.

16.  Ibid.

17.  Collins and Bilge, Intersectionality, 53, 58, 226.

18.  Carbado, “Colorblind Intersectionality.”

19.  I will discuss different, specific ways in which it is thought to be undermining in the next section.

20.  And, for this reason, as we shall see shortly, there are also varying levels and kinds of commitment to the notion of inseparability.

21.  Kimberlé W. Crenshaw, “Close Encounters of Three Kinds: On Teaching Dominance Feminism and Intersectionality,” Tulsa Law Review 46 (2010): 151–89, 153 n. 156. Note, however, that characterizations of essentialism also differ; cf. Harris, “Race and Essentialism.”

22.  Cf. Judith Butler, Gender Trouble (New York: Routledge, 1990). See Alcoff, “Cultural Feminism versus Post-structuralism”; Floya Anthias, “Rethinking Social Divisions: Some Notes towards a Theoretical Framework,” Sociological Review 46 (1998): 505–35; Lugones, “Purity, Impurity, and Separation”; Mari Mikkola, “Gender Sceptics and Feminist Politics,” Res Publica 13 (2007): 361–80; Mari Mikkola, The Wrong of Injustice: Dehumanization and Its Role in Feminist Philosophy (Oxford: Oxford University Press, 2016); Young, “Gender as Seriality.”

23.  See Carastathis, “Identity Categories as Potential Coalitions”; Leslie McCall, “The Complexity of Intersectionality,” Signs: Journal of Women in Culture and Society 30 (2005): 1771–1800.

24.  Carbado and Harris, e.g., characterize this as an important commitment of critical race theory, and Crenshaw also takes this kind of stance. Carbado and Harris, “Intersectionality at 30,” 2214; Crenshaw, “Close Encounters of Three Kinds,” 152.

25.  Anthias, “Rethinking Social Divisions”; Harris, “Race and Essentialism,” 586. Spivak’s notion of strategic essentialism also captures this kind of stance. See Gayatri Chakravorty Spivak, “Subaltern Studies: Deconstructing Historiography,” in In Other Worlds: Essays in Cultural Politics (New York: Routledge, 2012), 197–221.

26.  Zack, Inclusive Feminism, 18.

27.  Nancy Ehrenreich, “Subordination and Symbiosis: Mechanisms of Mutual Support between Subordinating Systems,” UMKC Law Review 71 (2002): 251; Young, “Gender as Seriality.”

28.  To clarify and do some foreshadowing, I ultimately want to claim that there are oppressive kinds, and so I do not think that this argument ultimately works.

29.  Gasdaglis and Madva, “Intersectionality as a Regulative Ideal,” 1304.

30.  Ibid.

31.  Ibid., 1305.

32.  As will be discussed later, my account of oppression does not require this. Thus, I think we can reject this premise and thereby block the conclusion.

33.  Note that, for much the same reason, there is also a concern that it impedes feminist theorizing. Here, too, however, the worry is ultimately political in nature. See, e.g., Alcoff, “Cultural Feminism versus Post-structuralism.”

34.  Ibid.; Theodore Bach, “Gender Is a Natural Kind with a Historical Essence,” Ethics 122 (2012): 231–72; Ann Garry, “Intersections, Social Change, and ‘Engaged’ Theories: Implications of North American Feminism,” Pacific and American Studies 8 (2008): 99–111; Gasdaglis and Madva, “Intersectionality as a Regulative Ideal”; Sally Haslanger, “Gender and Race: (What) Are They? (What) Do We Want Them To Be?,” Noûs 34 (2000): 31–55; Mikkola, “Gender Sceptics and Feminist Politics”; Natalie Stoljar, “Essence, Identity, and the Concept of Woman,” Philosophical Topics 23 (1995): 261–93; Young, “Gender as Seriality.”

35.  Bach, “Gender Is a Natural Kind,” 234.

36.  Young, “Gender as Seriality,” 718.

37.  Ibid., 714–19.

38.  Zack, Inclusive Feminism, 9–11,18.

39.  Thank you to an anonymous referee for pressing me on this point.

40.  I am describing the pattern in terms of “women” here as a convenient, if not quite accurate, shorthand before getting into the details of my account.

41.  Daniel C. Dennett, “Real Patterns,” Journal of Philosophy 88 (1991): 27–51.

42.  Ibid., 34–37.

43.  Note that in referring to “patterns of injustice” it need not be the case that the events that compose these patterns, or even the patterns themselves, are intrinsically unjust. Rather, they may constitute an injustice by virtue of their relations to other events or patterns.

44.  To briefly motivate this account, what it does is seize on persistent, systematic injustice as the central notion of oppression and tease out different important ways in which oppression is persistent and systematic. While this account draws on existing accounts of oppression, and particularly Frye’s account of oppression, it is distinctive in its formulation of oppression in terms of patterns of injustice—which, as we shall see, plays a key role in allowing the account to accommodate key intersectional notions. See Ann E. Cudd, Analyzing Oppression (New York: Oxford University Press, 2006); G. M. Eller, “On Fat Oppression,” Kennedy Institute of Ethics Journal 24 (2014): 219–45; Miranda Fricker, Epistemic Injustice: Power and the Ethics of Knowing (Oxford: Oxford University Press, 2007); Marilyn Frye, The Politics of Reality: Essays in Feminist Theory (Trumansburg, NY: Crossing, 1983); Sally Haslanger, “Oppressions: Racial and Other,” in Racism in Mind, ed. Michael P. Levine and Tamas Pataki (Ithaca, NY: Cornell University Press, 2004), 97–124; Iris Marion Young, Justice and the Politics of Difference (Princeton, NJ: Princeton University Press, 2011).

45.  Haslanger also discusses this feature, though she describes it in terms of there being a dynamic homeostasis at the level of the system. Sally Haslanger, “Racism, Ideology, and Social Movements,” Res Philosophica 94 (2017): 1–22, 17.

46.  Arlie Hochschild and Anne Machung, The Second Shift: Working Families and the Revolution at Home (London: Penguin, 2012).

47.  While Sorts might end up being conceptually related to sortals in important ways, I do not mean for them to be the same thing. Sortals centrally concern our ability to count objects, with some taking sortals to provide identity and persistence criteria for certain kinds of objects, or to specify the essence of a kind of thing. Sorts, in contrast, are not tied to essences, and they also do not centrally concern our ability to count things (there can be Sorts for things that aren’t or cannot be counted, like <marriage>). For an overview of sortals, see Richard E. Grandy and Max A. Freund, “Sortals,” in Stanford Encyclopedia of Philosophy, ed. Edward N. Zalta and Uri Nodelman (Stanford, CA: Stanford University, 2023), https://plato.stanford.edu/archives/sum2023/entries/sortals/.

48.  See, e.g., Haslanger, “Oppressions,” 111–15.

49.  While it can be challenging to find a good term for what I have in mind, I have landed on ‘Sort’ to play off the double meaning of (1) a sort as a type of thing and (2) sorting as an activity that we undertake. This is to emphasize that I am interested in conceptions of kinds of things that we create through our social practices. Note that Sorts can be ideological in the nonpejorative sense.

50.  I am mostly following Haslanger’s conception of ideology here. See Sally Haslanger, Resisting Reality: Social Construction and Social Critique (New York: Oxford University Press, 2012); Sally Haslanger, “What Is a (Social) Structural Explanation?,” Philosophical Studies 173 (2016): 113–30; Sally Haslanger, “I—Culture and Critique,” Supplement to the Proceedings of the Aristotelian Society 91 (2017): 149–73; Sally Haslanger, “What Is a Social Practice?,” Royal Institute of Philosophy Supplements 82 (2018): 231–47.

51.  I will use angle brackets to denote a Sort.

52.  Matthew Beresford, From Demons to Dracula: The Creation of the Modern Vampire Myth (London: Reaktion, 2008).

53.  See Rowan Bell, “The Role of a Lifetime: Trans Experience and Gender Norms” (unpublished manuscript); Katharine Jenkins, “Amelioration and Inclusion: Gender Identity and the Concept of Woman,” Ethics 126 (2016): 394–421.

54.  One could think that a social group may be produced as a result of a Sort (e.g., women may be produced as a result of oppression that emerges in connection to <women>). I am open to this possibility in general, although I do not want to accept it about women at this stage, given that such accounts have struggled to be suitably trans*-inclusive. I discuss this more later.

55.  See Lisa Rice and Deidre Swesnik, “Discriminatory Effects of Credit Scoring on Communities of Color,” Suffolk University Law Review 46 (2013): 935.

56.  If these laws and policies track a Sort, there is a good chance that they are influenced by past or present ideology. Nonetheless, it is useful to separate laws and institutional policies from ideological schemas because they are distinct mechanisms, and because the former may reflect the influence of ideological schemas that are no longer widely internalized by individuals in the milieu.

57.  I also suspect that it may be possible for a Sort to be implicit in a more thoroughgoing way. For instance, perhaps there could be a case where there are no schemas whose contents are about a particular collection of individuals, and yet there are implications for individuals with particular features. I am here influenced by Johnson’s idea that there can be biases that are nowhere represented in an individual’s cognitive repertoire. See Gabbrielle M. Johnson, “The Structure of Bias,” Mind 129 (2020): 1193–1236.

58.  I am here following Collins and Bilge’s discussion in Intersectionality, 25–31.

59.  To be clear, I don’t think that a political identity must map onto a Sort or a shared oppression.

60.  To be clear, in saying that my account is compatible with each of these views is not to say that I am neutral between them. This will be discussed further later.

61.  To see why (1) is the case in more detail: On my account, a pattern of injustice may help constitute 𝛴-oppression if it is persistent and tracks 𝛴. This requires that the mechanism(s) underlying the pattern track features correlated with 𝛴, but not that all individuals associated with 𝛴 have these particular features, nor (therefore) that they all experience the relevant kind of injustice. For example, there are some individuals associated with <women> who know that they are unable to get pregnant and do not have their bodily autonomy directly restricted by abortion bans. Nevertheless, insofar as being able to get pregnant is a feature that is directly associated with <women>, the restrictions to bodily autonomy that directly result from such bans may still form part of a collection of patterns of injustice that track <women>.

62.  I take Bernstein’s account of intersectional categories as being explanatorily prior to their constituents to be capturing this kind of case. See Sara Bernstein, “The Metaphysics of Intersectionality,” Philosophical Studies 177 (2020): 321–35.

63.  One might think that this pattern could also be described as a subpattern of a more general pattern of being touched without consent. If so, then one could replace this with another example that better fits this category. I take Crenshaw’s idea that Black women sometimes experience discrimination as Black women to imply that there will be such an example. If not, then all such cases will be covered by the subpattern case. Thank you to an anonymous referee for pressing me on this point.

64.  Further, it shows how the account gives rise to a picture of nesting and overlapping oppressive kinds.

65.  Crenshaw, “Demarginalizing the Intersection,” 149.

66.  Collins, “Toward a New Vision.”

67.  At the same time, the account allows us to recognize the injustices that individuals experience as helping constitute various kinds of collective oppression and as resulting from structures that uphold those forms of oppression.

68.  In this way my view aligns with that of Mikkola, who argues that we do not need to resolve the question of who exactly counts as a woman in order to retain feminist politics. Mikkola’s approach differs from mine in that she focuses on dehumanization as the central and fundamental kind of injustice that feminists should be concerned about. My approach, in contrast, does not identify any particular kind of injustice as central, universal, or fundamental. See Mikkola, “Gender Sceptics and Feminist Politics”; Mikkola, Wrong of Injustice.

69.  Thank you to an anonymous referee for highlighting this and pressing me on this point.

70.  Dembroff argues that we should think of intersectionality in terms of different systems of injustice being co-constituted by the same causal structure. I think that Dembroff is right that there will be this kind of co-constitution, such that this structure-based approach opens up the way to a broad, radical politics, although I think even here that there might be choices about how broadly or narrowly we construe the underlying structures (e.g., this particular policy vs. this institution), which can make a difference to the scope of feminist politics. Robin Dembroff, “The Metaphysics of Injustice,” in New Conversations in Philosophy, Law, and Politics, ed. Ruth Chang and Amia Srinivasan (New York: Oxford University Press, forthcoming).

71.  One question is whether the label “feminism” will still make sense for all these options. Perhaps not—feminism may need to look substantially different than it has in the past. Some may be unhappy with this, but I do not think it would necessarily be an unacceptable result. It may be in line with intersectional critique.

72.  For some helpful discussion of how to think about the hardships but also potential associated with coalition building, see Matsuda, “Beside My Sister”; Carastathis, “Identity Categories as Potential Coalitions.”

73.  Zack, Inclusive Feminism.

74.  Gasdaglis and Madva, “Intersectionality as a Regulative Ideal.”

75.  Ibid., 1314.

76.  See, e.g., Stoljar, “Essence, Identity”; Zack, Inclusive Feminism.

77.  Garry, “Intersections, Social Change”; Garry, “Intersectionality, Metaphors.”

78.  Young, “Gender as Seriality”; Haslanger, “Gender and Race.”

79.  Jenkins, “Amelioration and Inclusion”; Stephanie Julia Kapusta, “Misgendering and Its Moral Contestability,” Hypatia 31 (2016): 502–19.

80.  Mikkola has also argued that we do not need to first find an inclusive theory of gender in order to have effective feminist politics. Mikkola, “Gender Sceptics and Feminist Politics”; Mikkola, Wrong of Injustice.

81.  Alcoff, “Cultural Feminism versus Post-structuralism”; Anthias, “Rethinking Social Divisions”; Collins and Bilge, Intersectionality; Combahee River Collective, “Combahee River Collective Statement”; Crenshaw, “Mapping the Margins.”

82.  Carastathis, “Identity Categories as Potential Coalitions”; Crenshaw, “Mapping the Margins.”

83.  Diana Fuss, Essentially Speaking (New York: Routledge, 1989), 36.

84.  Young, “Gender as Seriality,” 722.

85.  See, e.g., Alcoff, “Cultural Feminism versus Post-structuralism”; Carastathis, “Identity Categories as Potential Coalitions,” 945, 955; Collins and Bilge, Intersectionality, 186–87, 233; Carmen Vázquez, “The Land That Never Has Been Yet: Dreams of a Gay Latina in the United States,” in The Third Pink Book: A Global View of Lesbian and Gay Liberation and Oppression, ed. Aart Hendriks, Rob Tielman, and Evert van der Veen (Buffalo, NY: Prometheus, 1993), 217–24, 222.

86.  Or, likewise, by the corresponding Sort or oppressive kind.

87.  Robin Dembroff, “Real Talk on the Metaphysics of Gender,” Philosophical Topics 46 (2018): 21–50. I think that my view of oppression can complement Dembroff’s view of ontological oppression, insofar as we can see oppression as being importantly tied to the construction of oppressive kinds. On the other hand, we might think that introducing Sorts allows us to capture the marginalization that Dembroff is interested in without needing to posit the oppressive group kinds. In any case, I think there is an advantage to leaving room for pluralism here.

88.  Rowan Bell, “Being Your Best Self: Authenticity, Morality, and Gender Norms,” Hypatia (forthcoming). Part of Bell’s view, which is also providing important motivation for me, is that the social processes that lead an individual to internalize social norms, and particularly gender norms, can give rise to cases where an individual is responsive to norms that do not match the category that others associate them with. This norm-responsiveness forms part of the social context that one responds to and is working with in constructing an intelligible self. See also Bell, “Role of a Lifetime.”

89.  Mikkola, “Gender Sceptics and Feminist Politics”; Mikkola, Wrong of Injustice.

90.  Carastathis, “Identity Categories as Potential Coalitions.”