Discrimination Without Traits: From Social Construction to the Politics of Discrimination

Abstract

Theories of discrimination typically select properties such as race, gender, sexuality, or ethnicity as being of special concern. These properties, which are customarily identified as genuine grounds of discrimination, are also at the forefront of constructionist efforts to understand reality as a product of social interactions. Theories of discrimination have so far neglected the important question of how understanding the nature of these properties impacts our theoretical views of the kind of phenomenon discrimination is. This article outlines some pitfalls of assuming away complexities regarding the ontology of the underlying properties, and systematically develops a constructionist account of discrimination, which I call Discrimination without Traits. I argue pursuing a constructionist view of grounds reveals discrimination to be not a discrete process involving a discriminator and a victim, but an ongoing process of (re)negotiating social reality that is fundamentally political. This uncovers neglected avenues for designing political remedies to discrimination.

INTRODUCTION

A stubborn fact about our understanding of the social world is that it proceeds as if social categories were not primarily products of construction, but ways of carving nature at its joints. We tend to think Danes, Dalits, and dyslexics are types of people – much like dahlias, dandelions, and daisies are types of flowers. This naturalist presumption persists despite a long and well-received body of literature attempting to debunk the reification of categories like gender, race, ethnicity, disability, or culture (Haslanger 2012; Brubaker 2004; Oliver and Barnes 2012; Phillips 2007). Even if acknowledging the social construction of such categories has become “the epitome of academic respectability, even orthodoxy” (Brubaker 2004, 3), the tendency to think of the social world as a world of substances remains irresistible.

How does this tension between a tendency to think in naturalist terms, and a considered constructionist judgment, affect our understanding of wrongful discrimination on grounds of race, gender, or ethnicity? The question is relevant since such properties or traits are central to legal and theoretical approaches to discrimination. Discrimination theorists frequently emphasize such “extraneous traits” (Moreau 2010, 147), “traits [imbued in a] history of mistreatment or current social disadvantage” (Hellman 2011, 21–2), “the property of belonging to a socially salient group” (Lippert-Rasmussen 2014, 34), or more generally “certain attributes or characteristics that persons have” (Khaitan 2015, 29). How to conceive these properties, traits, or characteristics is at the forefront of constructionist efforts to reevaluate them as products of collective processes of constructing the social world. Yet their nature remains under-scrutinized by discrimination theorists, with some even claiming the ontology of relevant properties does not “ha[ve] much significance” for how we understand discrimination (Shin 2018, 199).

The question is particularly pressing since when instances of discrimination are identified, natural-seeming traits are never too far behind. Paradigmatic cases of discrimination are those where underlying properties are likely to strike us as naturally demarcating types of persons. We are more likely to identify genuine grounds with properties that seem to separate individuals into contrasting groups (Khaitan 2015, 30), or when “it is evident whether or not someone is [sic] member of a certain group” (Lippert-Rasmussen 2014, 34). Indeed, Cass Sunstein takes the visibility of underlying traits to be defining for discrimination, saying “when the characteristic is not highly visible, we cannot have a [discriminatory] system as I understand it here” (Sunstein 1994, 2433). Those we typically identify as victims of discrimination tend to meet (and thus risk confirming) comfortable naturalistic intuitions about traits as given properties that naturally demarcate human beings.

What would a constructionist approach to discrimination look like, and would consistently pursuing it change the way we understand and combat discrimination? This article develops an account of discrimination which takes as its explicit starting point the constructed nature of properties of concern such as race, gender, or disability. It also argues that, contrary to thinking that how we conceive underlying properties is of little significance for our understanding of discrimination, consistently pursuing constructionist ideas uncovers distinct possibilities for viewing the grounds of discrimination, interpreting claims that someone was discriminated against because of a trait, conceptualizing discriminatory treatment, and designing remedies.

The article’s argumentative strategy is to gradually progress from a constructionist notion most would find unobjectionable (traitification), to increasingly distinctive commitments of a constructionist approach to wrongful discrimination. I refer to this account as discrimination without traits (DwT), and show it comprises four conditions:

(Traitification) There is an outcome of traitification T in the social context S of the discriminatory act Δ such that A perceives B as different from C when exhibiting T.

(Grounds) B and not C is an occupant of a distinct social position marked by T in the social context S of the discriminatory act Δ.

(Because) The fact that B (and not C) is perceived by A as occupying a distinct social position marked by T is the best explanation that can be inferred as to what rendered B a victim of discrimination through act Δ based on the behavior of A on the one hand, and the social position marked by T on the other.

(Treatment) A treats B differently than C in ways that express and/or entrench an inferior social position occupied by B and not C in the social context S of the discriminatory act Δ.

While these conditions are not meant to offer a comprehensive diagnosis of what makes discrimination morally wrong, DwT nonetheless reveals possibilities of understanding wrongful discrimination neglected by other theorists which otherwise embrace constructionist ideas, such as Kasper Lippert-Rasmussen, Deborah Hellman, and Sophia Moreau. These overlooked aspects ultimately reveal an irreducibly political dimension of discrimination (and anti-discrimination) struggles, which I discuss in the final section. It is this political nature of discrimination that gives the measure of what is lost when folk assumptions occasionally overshadow considered constructionist judgments.

TRAITS AND TRAITIFICATION

An enduring lesson about the social construction of reality is that the more successful processes of construction are, the more likely we are to forget they are in fact outcomes of social processes and mistake them for starting points in our analyses. Naturalist presumptions are deeply embedded in everyday vocabulary, where it is a matter of course to refer to social categories without acknowledging their socially contingent nature. Discrimination theorists are not immune to relying on this way of speaking, even as it rarely characterizes their considered way of thinking. For example, when discussing the challenge of setting apart genuine from idiosyncratic grounds of discrimination, the trait of having large earlobes is sometimes mentioned as a testbed for whether theories are too lax in allowing idiosyncratic grounds to count as genuine (Arneson 2006, 796; Koppelman 2006, 812; Lippert-Rasmussen 2017, 451). The issue is formulated as if something about earlobe size itself renders it a ridiculous candidate for grounding genuine cases of discrimination, even by theorists whose interest in the conditional nature of the grounds of discrimination strongly suggests they accept their transient character.

This way of speaking obscures the insight that in themselves, traits carry little indication regarding their potential to systematically structure social reality. As the case of the Cagots in early modern France and Spain shows, it is optimistic to assume something as innocuous as earlobe size could be safe from patterning genuine discrimination. Described by one doctor who inspected them as having ears that were “round and gristly, without the lobe of flesh into which the ear-ring is inserted” (Gaskell [1855] 2013, 21), the shape of their ears marked Cagots as a European under-caste. They were segregated in separate settlements, were forbidden from touching water from public fountains or using the central part of the roads, were denied access to certain professions, and had a separate entrance in church (Mentzer 1996, 23; Robb 2007, Chapter 4; Hawkins 2014). Speaking as if earlobe size itself holds any clues regarding its suitability to ground genuine discrimination or not obscures this bizarre but telling case.

A constructionist approach to discrimination would therefore be right to insist at the outset on highlighting the socially contingent nature of traits. Constructionism focuses on our “virtually unrestricted capacity… to carve up the external world into named categories, and then arrange the categories to suit our social convenience” (Leach [1976] 2012, 35–6). Similar to the terms ethnicization or racialization, I introduce the term traitification to refer to mechanisms through which apparently natural properties are brought into existence even in the absence of any naturalistic hook. Anthony Appiah describes such a mechanism of trait emergence in the formation of the groups of “Eagles” and “Rattlers” among boys from homogenous backgrounds, who manufactured numerous differentiating traits within just four days (Appiah 2005, 63–4). This “nascent trait divergence” between the “prayerful, pious, and clean-living” Rattlers and the “boisterous, tough, and scrappy” Eagles (Appiah 2005, 64) is rightly considered by Appiah not an exception but an insight into how quickly identities can be conjured into being.

Traitification gives an account of how properties which can be used to describe persons gain intersubjective purchase. Economically powerless groups often became “traitified” as polluted undercasts (Stuart 2000), or “ethicized” as minority groups like the Cagots (Robb 2007, Chapter 3). Some of the most grievous forms of injustice involve the invention of traits because—not despite the fact that—there are no natural markers to single out those we have decided to frame as less than human. For instance, it is a feature of dehumanization to postulate “natural” but invisible differences between perpetrators and victims (Smith 2014). Once such differences become traitified, they are perceived as naturally ordering the social world in ways that become part of societies’ common knowledge. They are no longer random and ridiculous classifications like dividing animals into “embalmed,” “tame,” and “fabulous,” as in Borges’s imagined encyclopedia (see Foucault [1970] 2002, xv), but ways of carving up the world which appear inevitable.

We can therefore formulate a first component of DwT as:

(Traitification) There is an outcome of traitification T in the social context S of the discriminatory act Δ such that A perceives B as different from C when exhibiting T.

It is important to note that traitification does not claim traits do not exist. The bracketing of traits—like Brubaker’s bracketing of groups (Brubaker 2004) and Phillips’s bracketing of culture (Phillips 2007)—should not be understood as asking to cease speaking about traits or assigning them explanatory power. What constructionism does demand is moving from a framework of (naturalistic) traits to one of socially constructed traits. Acknowledging traitification means acknowledging the primacy of human aims and interests in (re)creating such categories on the one hand, and the mirage of seeing them as natural on the other. While discrimination theorists would likely not contest the former, they would, for reasons explored in the reminder of this article, benefit from being reminded of the latter.

GROUNDS AS SOCIAL POSITIONS

Traits range from relatively innocuous or benign ones like being scrappy, to morally relevant ones such as being violent, to ones which constitute genuine grounds of discrimination, such as sex or skin color. Constructionism insists the difference lies not in any inherent aspect of traits themselves but in the kinds of processes and relations they pattern. In particular, constructionist approaches to the properties usually recognized as genuine grounds of discrimination—such as race, gender, ethnicity, or religion—have focused on whether these traits mark people out as occupants of social positions within hierarchical social relations.

While there may be multiple constructionist understandings of the grounds of discrimination, approaching properties as social positions is perhaps the most salient one, and is traceable to Rawls. John Rawls defined social positions as locations within the basic structure of society with attached opportunities, expectations, and life chances that shape the situation of occupants in pervasive ways. Social positions are shaped by wider structural and societal systems, like the workings of social institutions pertaining to “the political system as well as … economic and social circumstances” (Rawls 1999, 7).

Social positions are particularly well suited to capture constructionist intuitions given their focus on underlying social relations. Iris Young pays foremost attention to social relations when defining social positions in terms of the “relations in which [occupants] stand to other persons” (Young 2011, 57), especially hierarchical “[r]elations of privilege and disadvantage” shaped by wider social systems (Young 2011, 59). Social positions not only emerge from patterns in social interactions but also in turn shape interactions between occupants and others, as there are “different rules [that] apply to people in the different positions,” regulating patterns of interaction and behavior (Young 2011, 60). Identifying social positions is not a matter of identifying natural objects but of adopting “a certain way of looking at the whole society, one that sees patterns in relations among people and the positions they occupy relative to one another”s (Young 2011, 70, my emphasis). To identify a social position, we must perceive patterns as systemic, behaviors as non-accidental, and effects on multiple aspects of the lives of occupants as connected.

Understanding the grounds of discrimination as social positions is supported by how constructionist thinkers have approached categories like race, gender, or ethnicity. As Charles Mills puts it, echoing Young’s emphasis on hierarchical relations, race represents “a politically constructed categorization” which marks “locations of privilege and disadvantage in a set of power relationships” (Mills 2015, 76–7). Race is not a property individual persons possess but a marker of positions within relations between racialized individuals and others, created and sustained through social institutions and practices such as law, the economy, science, religion, and even philosophy (Lott 1999, 8; Heng 2018, Chapter 4). Similarly, being a woman according to Sally Haslanger is a matter of “occupy[ing] a particular kind of social position, one of sexually marked subordinate” (Haslanger 2012, 239). Gender is not reducible to biological sex, but defined by the differential opportunities, expectations, rules, and life chances associated with occupying a certain location in a hierarchy patterned by gender. Likewise, discrimination based on religion, language, or ethnicity is not, properly speaking, discrimination because of the trait of religious membership but because of how traits assumed to relate to religion or ethnicity map onto a hierarchical “unequal ‘us-them’ relationship” (Modood [2007] 2013, 35).

Viewing grounds as social positions allows differentiating idiosyncratic from non-idiosyncratic (i.e., genuine) grounds of discrimination. Properties which are merely products of traitification but do not demarcate social positions are idiosyncratic. However, because social positions are defined by systematic patterns of human action and the functioning of large-scale institutions, being treated in disadvantageous or demeaning ways because one occupies a distinct social position is different from idiosyncratically withstanding the “petty likes and dislikes of our fellow citizens” (Scanlon 2008, 73). While the former are idiosyncratic grounds and the latter are genuine grounds, whether a particular trait is idiosyncratic or genuine is not decided permanently. Earlobe size does not demarcate a distinct social position today, but historically did so in France and Spain. Nevertheless, when a certain property does demarcate a distinct social position, its status might be perceived as permanent by those marked by it in the relevant context.

Focusing on social positions hence has an additional advantage of capturing the intuitive notion that discriminatory treatment is inescapable for those marked out by the properties in question. Individuals are placed in positions where they are repeatedly subjected to disadvantageous or demeaning treatment by structures, expectations, institutions, and practices that conspire to mark them as different from others in ways they have little control over. While the idea of social construction might summon an optimistic image of arrangements being changeable, the pervasive ways in which social positions shape occupants’ lives show there is, from an individual’s perspective, a socially produced inescapability to being marked by relevant traits. Those branded as witches could not more escape being placed in a distinct social position than any person with an immutable characteristic could—even if no material substratum served as its basis. Social positions capture the thought that whether one likes it or not, systemic processes locate persons in positions marked out for certain forms of treatment.

A question that arises is whether social positions are reducible to social groups. This possibility is suggested by the fact that racial or ethnic groups are paradigmatic examples of social positions, as discussed above. Moreover, the grounds of discrimination are often conceptualized in terms of social group membership (e.g., Lippert-Rasmussen 2014, 26; Khaitan 2015, 30).

Yet social positions differ from groups in at least two ways. Firstly, whereas for groups the main relation between individuals and traits is group membership, social positions focus on how individuals are marked by traits. Traits pattern institutionalized behaviors, forms of regard, and structured interactions in ways that create a field of force in which individuals are placed along with others. These others are usually group members, but not necessarily. For example, relatives of caretakers of disabled individuals occupy a social position marked by disability, without necessarily being disabled themselves. As books like Renate Welsh’s Dragon’s Wings (1989) or series like There She Goes (2020) illustrate, family members of disabled persons are also marked for differential forms of treatment and regard in pervasive ways, from being excluded from leisure facilities and socializing opportunities, being gawked at or stigmatized for maintaining private positive attitudes toward a disabled loved one in public, or encountering difficulties forming social identities other than their relation to disabled persons. These represent enduring constellations of institutionalized interactions and behaviors patterned by disability, and affect nondisabled persons because of how their association with disabled persons positions them relative to others. Relatives and caretakers of disabled individuals are occupants of a social position marked by disability, without being group members.

An advantage of understanding the grounds of discrimination as social positions is capturing the unfairly neglected case of discrimination by association. The term refers to discriminatory treatment suffered by victims because of protected grounds, although the victims themselves are not members of the relevant group, but merely associated with group members. For example, Coleman vs Attridge Law in the EU (2008) or the Equality Act 2010 in the UK recognize caretakers of disabled persons as victims of direct discrimination on grounds of disability, despite not belonging to the group of disabled persons themselves. Understanding the grounds of discrimination as social positions allows capturing such cases by virtue of highlighting how the underlying properties pattern relations of association. As discussed above, disability patterns social relations in ways that mark not just disabled people but also others associated with them, for different and inferior forms of treatment.

Admittedly the group criterion can partially capture this dynamic since the source of the problem can be construed as making structural provisions or accommodations for the needs of the dominant group, which appear “normal and natural” as a result (Moreau 2019, 132). Yet the relative inferiority of those with nondominant needs and experiences needn’t fall along group lines. While “normal” needs indeed differentiate nondisabled from disabled people, “normal” individuals are simultaneously constructed as persons who do not care for or are related to disabled people. The latter become excluded by structurally accommodating normal needs, but do not form a separate group. Understanding grounds as social positions allows capturing widespread experiences of caretakers, relatives, and allies of members of relevant groups as discrimination.

A second reason for focusing on social positions instead of groups when identifying the grounds of discrimination is that grounds sometimes cut across the group criterion. On the one hand, grounds can create hierarchies within social groups, as when colorism differentiates between lighter- and darker-skinned members of racial minorities (Jones 2000). On the other, grounds can cut across the group criterion by transcending the social groups they partially overlap with. As Kimberlé Crenshaw argued in her discussion of DeGraffenreid vs General Motors (1976), adopting a strict group criterion in which the plight of Black women was analyzed in terms of Blackness or gender taken in turn blinded the court to the institutionalized practices and asymmetrical privileges that produced a clear configuration of disadvantage affecting the plaintiffs as Black women (Crenshaw 1989, 141–3). Persons occupying a social position at the intersection of race and gender might, moreover, find it difficult to coalesce into a cohesive group since tensions between racialized and gendered attitudes, expectations, norms, and behaviors might give rise to “conflicting political agendas” (Crenshaw 1991, 1252). Focusing on social positions accounts for these grounds by viewing them as products of distinct configurations of attitudes, life chances, and opportunities. They emerge if we adopt “a certain way of looking at the whole society” (Young 2011, 70, my emphasis), which reveals them as demarcating distinct positions within racialized, gendered, etc. relations.

We can then formulate the second component of DwT as:

(Grounds) B and not C is an occupant of a distinct social position marked by T in the social context S of the discriminatory act Δ.

One might object that in attempting to be less restrictive than the group criterion, DwT risks including even morally relevant properties in the protected list. For example, in a society where income distribution marks people as occupants of distinct social positions, but income tracks morally relevant criteria, DwT would recommend treating income distribution as a genuine ground of discrimination—a problem which we would not encounter on the group criterion. DwT then risks offering a protected status to properties that do not deserve it.

The first thing to note in response is that issues judging novel or borderline cases are often encountered when the judgment relies on identifying patterns in the underlying social relations. The group criterion is not immune to such challenges, especially when issues of judging novel cases arise. Bourdieu’s argument for recognizing rural farmers, large families, or the unemployed is that their interactions with mainstream institutions are patterned in similar ways to those of politically recognized groups (Bourdieu and Ferguson 1999, 627). Hence difficulties with novel or borderline cases are not specific to social positions, but also plague the group criterion.

Furthermore, a more careful look at what is needed to establish a trait delineates a social position reveals resources for overcoming the issue. In the case of income, the relevant evidence for constituting a distinct social position is not the size of people’s pay-checks but whether people on low incomes are customarily suspected of moral or intellectual inferiority, whether accents or clothing choices associated with the trait limit people’s access to desirable jobs, whether it is easy to escape, and whether it continues to mark people—or, as in the movie Stella Dallas, even their children—well after income differences disappear. The evidence a person occupies a distinct and inferior social position is evidence of opportunities and attitudes that are systematically structured, and conspire to keep occupants on an inferior rung of the social ladder. These parameters are those we should consider when deciding whether we can rest assured a property tracks moral relevance. If they are met, there is reason to suspect the occupants are the ones subjected to morally arbitrary treatment in ways that are masked by legitimizing myths, stereotypes, and rationalizations. It is an enduring lesson of constructionism that presumed moral failings are manufactured for marking targeted populations as deservingly subordinated—from accusations of heresy, leprosy, or homosexuality being widely used to justify the persecution of minorities in medieval Europe (Moore [1987] 2007), to the construction of gender or race in ways that “justify and motivate” their bearers occupying inferior social positions (Haslanger 2012, 232). This possibility justifies viewing discriminatory treatment because one occupies a social position as pro tanto morally wrong, although as discussed below, this can be rebutted in light of other features of the account.

THE BECAUSE CONDITION

Constructionism’s concern for how the underlying properties should be conceived also impacts our understanding of other defining features of discrimination. One of these is the because condition, ¹ regarding how to understand statements that someone is discriminated against on account of their race, gender, ethnicity, or sexuality.

A straightforward proposal is to take statements that a victim suffered discriminatory treatment because of a ground to mean the act was causally prompted by the property. Lippert-Rasmussen proposes that we understand claims that A discriminates against B because of ground P as meaning either (a) the thought that B (and not a third party C) has P “is part of [A]‘s direct, motivating reason” for the discriminatory act, or (b) the fact that B (and not C) has P “causally explains” the act “and this is in turn causally explained by the fact that people with P are often treated worse than those without P” (Lippert-Rasmussen 2014, 37–8, my emphases). The former covers cases of direct discrimination, where the property prompts the act through the discriminator’s direct motive for acting. The latter covers cases where the property plays a causal role in producing the discriminatory act without directly motivating agents—and includes instances of indirect discrimination. For example, a policy of not hiring applicants likely to request childbirth-related leave indirectly discriminates against women, since gender causally explains why the employer does not hire female applicants, and women are in this disadvantaged position because of how they are often treated regarding, for example, disproportionate care work (Lippert-Rasmussen 2014, 37).

Against this causal understanding, DwT denies properties themselves are prime movers in the order of explanation. In the case of direct discrimination, the causal account claims the property explains the act by being (part of) what directly motivates it. But focusing on what causally prompts the act means focusing on the act’s mechanics instead of on what is truly needed to explain it: the way the ground functions. While, for example, race-related motivations may causally prompt acts, this does not mean they are primary in the order of explanation too since, as Donald Davidson puts it, “we never do more than move our bodies” (Davidson 1980, 59). Viewing properties as outcomes of social processes means acknowledging that without references to the underlying differential norms, attitudes, and institutionalized forms of treatment and regard which render skin color and not eye color as demarcating a distinct social position in our society, it would be as difficult to establish race as a motive behind directly discriminatory acts as it presently is to establish one was discriminated against because of eye color. In themselves, properties do not establish discriminatory acts happened because of them.

Accounting for the distinction between causation and explanation requires, in turn, acknowledging the role of background knowledge in explaining discriminators’ observable actions. On a constructionist reading, properties such as race, gender, or ethnicity explain social processes through “the tacit, taken-for-granted background knowledge, embodied in persons and embedded in institutionalized routines and practices, through which people recognize and experience objects, places, persons, actions, or situations as ethnically, racially, or nationally marked or meaningful” (Brubaker 2004, 17). Claims that someone was (directly) discriminated against because of their race, gender, or disability are claims that the act is best understood in terms of the associated rules, routines, institutions, and attitudes which delineate persons as occupants of racialized, gendered, or ethnicized social positions. DwT hence recommends understanding the “because” condition as (a) aiming to make sense of the act rather than to pin down its causal details and (b) going beyond what is causally entailed by the act itself to relevant background knowledge in doing so.

Explanations that meet these two criteria are called inferences to the best explanation, or IBE (Douven 2021). Unlike deductive inferences, which seek to derive the actual mechanics behind acts or events from given premises, IBE seeks to make sense, or explain, acts or events. In a classic example, when Sherlock Holmes infers Moriarty is the culprit based on motive, blood stains, and fingerprints, these clues do not amount to causal evidence for Moriarty’s actions. Moriarty’s guilt merely “provide[s] a better explanation of the evidence than would anyone else’s” (Lipton 2000, 185). Nor does IBE attempt to identify a prime mover. As a “self-evidencing” type of explanation, “[t]he phenomenon that is explained in turn provides an essential part of the reason for believing that the explanation is correct” (Lipton 2000, 185). In the case of discrimination, the prevalence of discrimination based on gender, race, or religion as a social phenomenon lends plausibility to the explanation that a given act happened because of gender, race, or religion. Secondly, IBE requires going beyond the mechanics of acts and appealing to external criteria for weighing the relative strengths of alternative explanations. While the exact criteria for providing the “best” explanation are hard to pin down, there is widespread agreement that better explanations are those that can “explain more types of phenomena, explain them with greater precision, provide more information about underlying mechanisms, unify apparently disparate phenomena, or simplify our overall picture of the world” (Lipton 2000, 187; see also Weintraub 2013, 204). Identifying the best explanation is, consequently, a matter of inferring which possible explanation best renders our overall understanding more coherent.

As applied to cases of direct discrimination, IBE does not focus on singling out particular properties as direct motivating causes (as the causal approach does), but on widespread patterns of behavior and interaction which render explaining the act in terms of the property convincing. For example, to establish an employer’s act of not hiring a female applicant constitutes direct discrimination, the causal account seeks to show that the applicant was treated disadvantageously “simply because she is a woman” (Lippert-Rasmussen 2014, 146–7, my emphasis), that is, that being a woman is the direct motivating cause all things being equal. By contrast, individuating the ground on the DwT account requires referring to ways in which things are not equal, that is, ways in which the present case is connected to wider social processes, differential rules, and institutionalized behaviors that construct being a woman as a distinct social position subordinate to men. In short, identifying a directly discriminatory act as gender-based discrimination requires referring to the way gender is socially constructed and understood. The “because” condition is satisfied by gender not in light of it being a clean cause for the act (as “simply” suggests) but in light of the messy backdrop of how women are habitually regarded and treated. IBE moves away from engaging with discriminators’ direct motivating reasons and toward more behavioral approaches like the one adopted in the United Kingdom, where evidence of direct discrimination is whether the act or policy excludes 100% of those marked by the ground, regardless of the perpetrator’s motive.

What about indirect discrimination? Here, the case for a casual approach seems stronger, since demonstrating indirect discrimination has taken place often requires establishing “a causal connection between the defendant’s facially neutral act and some adversity suffered by a set of persons” (Khaitan and Steel 2018, 202, my emphasis).

Yet looking more carefully at what demonstrating a “causal connection” entails, and the requirement to assess whether the disparate impact that was evidenced is justified, lends further credence to IBE. Firstly, to demonstrate differential disadvantage has taken place, the law “no longer requires the claimant to prove statistical significance of the differential impact [but] to make a plausible prima facie case that the difference is unlikely to be random” (Khaitan and Steel 2018, 203). The importance of establishing the act was due to the grounds rather than due to chance is captured by IBE’s requirement to demonstrate the ground provides a better explanation for the act than salient alternatives, in this case, than chance. In Lippert-Rasmussen’s example of the facially neutral policy of not hiring employees likely to request childbirth-related leave, IBE more naturally captures the demand to prove gender is a more plausible explanation than chance compared to the causal view. IBE focuses specifically on what makes gender a more plausible explanation for the policy, highlighting the fact that gender provides “more information about underlying mechanisms” (Lipton 2000, 187) than the alternative that discrepancies are explained by chance. Explaining the differential impact as due to gender can moreover “simplify our overall picture of the world” by “unify[ing] apparently disparate phenomena” (idem) that delineate gender as a social position in the first place. Gendered restraints on working schedules connect the case with other ways in which society is designed to accommodate males such that “[men’s] needs define auto and health insurance coverage, their socially designed biographies define workplace expectations and successful career patterns,” and so on (MacKinnon 1987, 11). In contrast, attributing the differential impact to chance does not increase our understanding beyond the details of the particular act. Hence, just like in cases of direct discrimination, establishing an indirectly discriminatory act happened because of gender, race, or ethnicity requires making a statement about how to interpret social reality.

Secondly, IBE better captures the significance of having to establish if the disparate impact once evidenced is justified or not. When a plaintiff makes a prima facie case, there is a “burden to the other party to offer a justification for the negative impact” (Hellman 2018, 106, footnote 8) and the defendant has the “opportunity to justify her conduct” (Khaitan and Steel 2018, 198). Compelling justifications can prevent disparate impact from constituting discrimination. For example, the fact that more female than male applicants were rejected from their desired graduate program at the University of California, Berkeley, in 1973, is prima facie evidence of discrimination because of gender. Yet the disparate impact was ultimately due to women disproportionately applying to more selective programs than men (Khaitan and Steel 2018, 205), and was not, therefore, discriminatory. IBE can account for the possibility of rebutting a prima facie case by showing the disparate impact is nonetheless justified, since as the available information changes, the best explanation we can infer based on the evidence also changes. The additional information about women applying to more selective programs tips the scales against the explanation the disparate impact happened because of gender. Until such information is provided, IBE places the burden of proof on defendants once a plausible prima facie case is made, in line with the practice initiated by the ECJ and adopted by the Equality Act 2010 136(2). The practice is sometimes defended as a way of minimizing the impact of information asymmetries between defendants and victims (Fredman 2011, 286–7); IBE reveals it is also recommended by constructionist approaches to discrimination.

DwT’s third condition can be formulated as:

(Because) The fact that B (and not C) is perceived by A as occupying a distinct social position marked by T is the best explanation that can be inferred as to what rendered B a victim of discrimination through act Δ based on the behavior of A on the one hand, and the social position marked by T on the other.

One might object that DwT is nonetheless ill-equipped to account for cases of indirect discrimination since its insistence on interpreting the social meaning of traits in a given context renders it a type of expressivist account of discrimination. DwT’s emphasis on the social meaning of acting because of protected characteristics might strike readers as similar to Hellman’s view that determining a classification is demeaning (and thereby discriminatory) as it requires scanning the social context and making “an interpretive claim about our culture” (Hellman 2011, 32). But this focus on what acts express is widely regarded as a weakness in capturing instances of indirect discrimination where the disparate impact need not (and usually does not) express a problematic meaning.

DwT indeed shares the expressivist accounts’ focus on understanding the social context. But instead of viewing grounds as carrying overt meanings they transparently express like expressivist accounts, DwT regards our understanding of the social meaning of grounds as dependent on our knowledge of the social world—in particular, on how accurately we can trace the corresponding social positions. For example, to assess whether a certain act constitutes discrimination because of sexual preference, DwT requires assessing not the act’s objective demeaning meaning but whether the act is best interpreted by cis-het norms, assumptions, or institutionalized practices which shape LGBTQIA+ identities. The latter cannot be immediately gleaned from the context of an act, but only become visible if we adopt “a certain way of looking at the whole society” (Young 2011, 70)—one that reveals nonstandard sexuality and/or gender identity as a social position. As indicated by the “+” sign, which “holds space for the expanding and new understanding of different parts of the very diverse gender and sexual identities” (Princeton University N.d.), we sometimes do not know (and know that we do not know) the exact ways in which LGBTQIA+ identity is constructed as distinct from cis-het norms. The more we understand the underlying social processes which shape LGBTQIA+ identity, the more determinate the subgroups currently covered by the “+” sign become. The meaning of the grounds of discrimination depends on our understanding of underlying social processes and relations, instead of being an objective feature of grounds. Allowing for the possibility of discriminatory acts which are not transparently demeaning permits DwT to capture cases like that of indirect discrimination, where the treatment is explained by relevant grounds, but without featuring objectively demeaning meanings.

TREATMENT

The discussion of the Because condition showed that claiming a discriminatory act happened because of a protected characteristic means making a statement about how social positions pattern discriminatory acts. But DwT is also sensitive to how discriminatory acts in turn shape the social positions marked by relevant properties through discriminatory treatment.

When seeking to define discriminatory treatment, it is tempting to assume the analysis would benefit from narrowing down our focus. For example, Lippert-Rasmussen argues we should only focus on particular goods and interactions, rather than on the treatment’s wider consequences, leading him to differentiate between disadvantage and harm. He illustrates the distinction through the example of a Jewish student who is denied a place at the University of Vienna in 1938 by a Nazi official, which causes the student to go into life-saving exile (Lippert-Rasmussen 2014, 18). Although the official’s act does not harm, and indeed benefits the student, he is nonetheless guilty of discrimination for treating the Jewish student disadvantageously compared to others with respect to “the particular goods the distribution of which the official is in charge of”—namely university admissions (Lippert-Rasmussen 2014, 18). Isolating the specific wrong-making feature of discriminatory treatment hence requires focusing on the particular goods discriminators can divvy up (instead of on its wider implications).

However, excluding the Jewish applicant, together with similar acts by other admissions officers, employers, and bureaucrats in 1930s Vienna, contributed to (re)defining the norms, opportunities, and interactions that shaped being Jewish as a social position in 1930s Austrian society. Just as each legal order eroding the status of Jewish persons was “not haphazardly demanding just any crimes unconnected with each other, but building up with utter consistency and care the so-called new order,” as Hannah Arendt puts it (Arendt [1964] 2003, 41), so too individual acts of differential disadvantageous treatment against Jewish persons together with other acts of excluding Jewish people from public life through, for example, the Aryan paragraph, produced and reproduced being Jewish as a subordinate social position. Together, they rendered Jewish identity socially salient, normalized differential treatment against Jewish persons, and “built up with utter consistency” a social order without Jewish people in positions of power or privilege—and ultimately a social order without Jewish persons altogether. Focusing on how discrete goods like university admissions were distributed in localized interactions overlooks how discriminatory acts impact the wider patterns in the distribution of goods.

Accounting for this dynamic requires adopting the constructionist insight that the grounds of discrimination and the treatment premised on them are co-constructing. Discriminatory acts exclude individual victims from the benefits, opportunities, or norms of respect enjoyed by others, but this does not exhaust their mark upon the social world. Human acts do not only accomplish the states of affairs they aim to bring about, but also often produce and reproduce the structured rules, institutions, and opportunities they rely on for accomplishing intended states of affairs (Giddens 1979). As acts by human agents, discriminatory acts simultaneously impact individual victims, and create or reproduce the structures, differential opportunities, attitudes, and behaviors which define the social positions occupied by victims alongside others. While the effects of individual acts might be difficult to trace, DwT nonetheless shares the assumption that it is at least “highly likely” for individual acts to have such effects (Moreau 2019, 138), and effects on groups can be traced “in most, if not all, cases” (Khaitan and Steel 2018, 206).

By focusing on how discriminatory treatment shapes the grounds of discrimination, DwT traces its wrong-making feature not in the effects on individual discriminatees, but in furthering inegalitarian “[r]elations of privilege and disadvantage” (Young 2011, 59) that underline relevant social positions. A crucial wrong-making feature of discriminatory treatment is reproducing a distinct and inferior social position for victims as well as others marked by the relevant ground. This reverse dynamic whereby widespread differential treatment in turn contributes to the social construction of the underlying grounds is what makes discrimination without traits possible. DwT thus maintains that discrimination is morally wrong at least partially because it contributes to social subordination (Fiss 1976; Colker 1986; Sunstein 1994; Balkin and Reva 2003; Moreau 2019)—in particular, the subordination of persons occupying social positions that constitute genuine grounds of discrimination. However, subordination is not taken by DwT to exhaust the wrong-making features of discrimination. DwT takes a pluralist approach to what makes discriminatory treatment wrongful (Khaitan 2015, 9–13; Moreau 2019, 146) and does not deny the significance of the harm, denigration, or exclusion which frequently accompany it. Rather, DwT regards social subordination as a defining wrong of discriminatory treatment whatever else its effects may be, since in addition to its impacts on individual victims, discriminatory treatment also impacts them as occupants of subordinate social positions.

Discriminatory treatment furthers subordination not just by entrenching a subordinate position but also by virtue of what such acts occasionally express. Discriminatory acts can further social subordination by expressing demeaning meanings toward occupants of hierarchically inferior social positions. The demeaning meaning expressed by discriminatory acts is indeed considered defining for discrimination by Hellman (2011, 21–2). However, she focuses on what acts express about individual victims rather than on their subordinating effects for social positions. For example, spitting on a homeless person is said to “constitute demeaning [treatment] because (a) spitting is a conventional way of showing disrespect and (b) the relative disparity in status between a homeless person and myself allows my expression of disrespect to put him down” (Hellman 2011, 35). Hence, Hellman regards discriminatory treatment as something that affects individuals, not groups (Hellman 2011, 23), and is aggravated by the superior status of perpetrators.

Nevertheless, focusing solely on what is expressed toward individuals overlooks what discriminatory treatment expresses about the corresponding social positions. In particular, it overlooks the intriguing possibility that acts expressing positive meanings toward individuals may discriminate by virtue of the negative meanings expressed toward them as occupants of social positions. From back-handed compliments like “you are not like other girls,” “you are one of the good Muslims” or “I would not have guessed you were disabled,” to Nazi officials sometimes expressing views such as “’the others are pigs, but this particular Jew is first rate’” (Arendt [1963]1994, 133), occupants of inferior social positions can be praised for exceptionally exhibiting traits associated with privileged positions. Sometimes these result in egalitarian treatment for the individuals concerned; even Hitler, Arendt tells us, “is said to have known three hundred and forty ‘first-rate Jews,’ whom he had either altogether assimilated to the status of Germans or granted the privileges of half-Jews” (idem). In such cases it might appear the individuals are treated as the equals of more privileged others. Hellman is right to insist the social meanings expressed by seemingly egalitarian treatment should be examined in light of background injustice. Yet it is relevant that individuals’ identities are framed as a shortcoming that requires a commendable effort to overcome. Individuals are praised for not exhibiting properties such as their gender, religion, or disability, which lifts them to the status of relative equality at the expense of lowering their social position. DwT captures this dynamic as demeaning by focusing on what such statements express about social positions instead of individuals.

Accepting that discriminatory treatment can both entrench a subordinate position (as in the example of the Jewish student) and express the view that the victim occupies an inferior social position (as discussed above) means that accounting for discriminatory treatment requires a two-pronged approach. This sets DwT apart from accounts that focus on the subordinating aspect of discriminatory treatment but only focus on one aspect of subordination, be it differential status-related honor or deference (Fiss 1976, 157) ² or as “systematic social disadvantages” in education, income, wealth, or health (Sunstein 1994, 2429). Instead, DwT reveals the status and disadvantage components of subordination usually coexist, as social positions are shaped through both attitudinal and distributive components. Hence, DwT agrees with accounts which take a pluralist view to the wrongs of social subordination, such as Moreau’s understanding of discriminatory treatment as behavior that can either “play an important causal role in sustaining … subordination” or “constitute an expression of censure [or] lack of deference toward” victims (Moreau 2019, 140, my emphases). Indeed, these components do not just coexist, but can be mutually dependent. As Catharine MacKinnon puts it in her discussion of defamation and discrimination, differential segregation which causally (re)entrenches exclusion “cannot happen without someone saying ‘get out’ or ‘you don’t belong here’ at some point” (MacKinnon 1993, 13), that is, without authoritatively expressing the view that certain types of people are to be excluded.

The fourth condition of DwT therefore becomes:

(Treatment) A treats B differently than C in ways that express and/or entrench an inferior social position occupied by B and not C in the context of the discriminatory act Δ.

This fourth condition completes the DwT account; yet a critic might object that the four conditions do not represent the only way we can construe a constructionist understanding of wrongful discrimination. Even if we accept traitification as necessary, it is not inevitable that we end up with social positions, inference to the best explanation, and subordination as defining components.

In response, it is important to remind the reader that DwT does not claim to offer a unique answer to all necessary and sufficient conditions of wrongful discriminatory acts. DwT sets a more modest goal of showing that a particularly plausible constructionist understanding of the grounds of discrimination opens new possibilities for defining discrimination that break away from established views of the grounds of discrimination as given, the process through which they prompt discriminatory acts as causal, and their relation to discriminatory treatment as unidirectional instead of co-constitutive. But while DwT is not the only possible constructionist approach to discrimination, it is particularly well suited for highlighting how adopting a constructionist outlook requires departing from widespread views about the nature of discrimination. DwT delivers on the promise of providing plausible alternatives to understanding the grounds of special concern, saying an act happened because of grounds, and what is distinctive of discriminatory treatment. It also uncovers new areas where discrimination is (re)produced new avenues for combating it, to which I now turn.

THE POLITICS OF (ANTI)DISCRIMINATION

So far, this article focused on definitional consequences of embracing a constructionist approach. Yet DwT also impacts how remedies to discrimination should be organized. In particular, DwT recommends going beyond the culprit and plaintiff model of remedying discrimination. The latter view, that remedies to discrimination require compensation for costs endured by plaintiffs, forms the backdrop of legal approaches. Moreau has offered a forceful normative account of such a straightforward legal approach. For Moreau, discrimination consists in “a personal dispute between the complainant and the alleged discriminator” (Moreau 2010, 146). Consequently, remedies to discrimination should neutralize the costs placed on discriminatees for their traits, and restore what Moreau calls the deliberative freedom of victims, defined as freedom from factoring in protected traits as costs in deliberation, such that victims are not forced to “live one’s life with these traits always before one’s eyes” (Moreau 2020, 84).

Despite the conceptual clarity and wide applicability of this approach, viewing discrimination as a personal dispute between complainant and discriminator whose resolution is compensation for costs endured nevertheless overlooks three political aspects of discrimination. These are as follows: (1) the politics of formulating claims; (2) the political responsibility of third parties in perpetuating discrimination; and (3) reconfiguring the political imagination to envisage remedies. I take each in turn and show they amount to overlooking the politics of discrimination—and the necessity of political remedies.

Firstly, legal approaches take the moment complaints are raised as starting points in efforts to redress the wrongs of discrimination. The (alleged) victims of discrimination are equated with their role as plaintiffs, especially in private law approaches to discrimination. This assumes away efforts, tensions, or achievements behind raising claims in the first place. In some cases, even identifying behaviors as discriminatory requires contesting normalized inegalitarian relations patterned by traits in nontrivial ways. Raising a claim is not always a matter of simply registering an injustice that was easy to identify; in some cases, it can be a creative act of framing and of perceiving the behavior in question as an injustice to be addressed, and not just a misfortune one should put up with. Where to draw the line between injustice and misfortune is “not a simple rule that can be taken as a given,” but “a political choice” (Shklar 1990, 5). Raising a complaint hence presupposes a prior step of realizing that behavior which is socially compliant is nonetheless morally arbitrary.

The legal account overlooks that moving from victim to plaintiff is a political achievement. To raise a claim, victims must first understand themselves as unfairly affected by norms, policies, or practices, and view their plight as worthy of public scrutiny instead of internalizing costs. In later work, Moreau herself acknowledges such difficulties in drawing the line between “’injustice’ …and a mere …unfortunate state of affairs” (Moreau 2018, 126). Yet she overlooks that raising a complaint that, for example, offices of the Canadian police should be open to all regardless of religious restrictions on wearing headgear then becomes an act of political dissent. It might appear the emergence of such dissent is inevitable, since even when the nature of injustice obscures its morally arbitrary nature, “[t]here are always some people and after a time there are a great many, who think the seizure is not justice but usurpation” (Walzer 1983, 12). Yet assuming dissent is inevitable blinds us to the most severe forms of discrimination. As Susan Okin puts it, “the closer the social system is to a caste system in which social meanings ‘overlap and cohere,’ the less likely is the appearance or development of … dissent” (Okin 1987, 59). While easier to highlight in the case of private law approaches to discrimination, this oversight also affects public law approaches that do not require identifying individual victims for addressing discrimination, insofar as the political line between injustice and misfortune is overlooked in the latter case too. In both cases, the more pervasive and overlapping the norms, meanings, and justifications that conspire to paint victims as acceptable targets of differential treatment, the more difficult it becomes to capture the phenomenon as unjust.

The second simplifying assumption of the legal approach is that in framing discrimination as “a personal dispute” between plaintiff and (alleged) discriminator (Moreau 2010, 146), perpetrators only feature as accused parties, who are either guilty or not guilty. In contrast, DwT uncovers that those marked by a relevant property are not aberrantly placed into social positions we innocently happen upon but continuously framed as persons of a certain kind through the expectations, attitudes, and behaviors of others who are not legally liable for discrimination. Third parties play a role in perpetrating discriminatory acts by shaping the grounds of discrimination as social positions, weaving hierarchical relations and differential norms and expectations into the fabric of social reality through seemingly unrelated combinations of actions and interactions. As co-citizens, individuals influence the beliefs, attitudes, norms, and expectations that enable unjust social processes, often simply by “minding their own business and acting within accepted norms or rules” (Young 2011, 106).

Assuming individual plaintiffs are the only responsible parties overlooks the fact that others who are not guilty of discrimination nonetheless bear political responsibility for their contribution to social processes that shape discriminatory outcomes. Unlike guilt which pertains to individuals for their own actions, political responsibility pertains to individuals as (co)participants in society, who should not be indifferent to what is done in their name (Young 2011, Chapter 3). For example, while Nazi officials are the ones guilty of perpetrating crimes against Jewish persons, others also carried political responsibility to the extent that they “accepted the language of self-deception that authorities promulgated” and stood by “as sites of political participation were shut down” (Young 2011, 86). In the case of discrimination, while, for example, individual employers are the ones who are in positions to determine “some of society’s most important distributions” acting “like a government” (Gardner 1989, 11) in distributing jobs, employers also speak in the name of others—as customers or consumers—when claiming that, for example, hiring a Black person as a shop assistant would decrease sales due to consumer preferences. Ordinary citizens contribute to the social construction of racialized or gendered social positions as inferior in the labor market, which makes them politically responsible (but not personally liable) for expressing or acquiescing in publicly perceived preferences and expectations that employers disseminate and act upon.

The third simplifying assumption of the legal model is that framing the costs of discrimination as potentially offset by compensation overlooks the more encompassing nature of required remedies. Compensation comes after plaintiffs navigate their circumstances as persons marked out for differential treatment and regard. Victims of discrimination (and others sharing the same social positions as them who have not yet been discriminated against) inhabit a social world in which disadvantageous or demeaning treatment is commonplace. The possibility of being treated differently because of gender, race, or disability—and the steps needed to avoid such treatment—will feature prominently among their possibilities of interaction. This renders compensation particularly ill-suited for Moreau’s proposed remedy of preventing traits from featuring as costs in victims’ deliberations. If persons navigate their social environment while being marked out for certain forms of treatment, they live their lives “with the traits before their eyes,” whether they fully realize it or not.

Framing the remedy for discrimination as compensation paid by individual culprits overlooks the fact that being truly free from discrimination requires reforming structures, processes, and institutions to prevent them from constructing relevant properties as grounds of discrimination. Citizens will be free from factoring traits as costs in their deliberations when race, gender, or disability are no longer intersubjectively understandable as reasons that explain discriminatory treatment. Remedying discrimination requires reconfiguring norms, structures, and institutions such that treating people differently on account of their sex or skin color becomes as idiosyncratic as treating people differently because of eye color is today. This in turn demands redrawing a community’s political imagination, with its “institutional arrangements and imaginative preconceptions” that frame relations among citizens (Unger 1987, 10), toward a more egalitarian vision of how society should be organized.

The above diagnosis paints discrimination as pertaining to the political as a continuous project of (re)shaping collective life (Wolin 1960; Arendt [1966] 2005; Mansbridge 1996, 49). This understanding in turn points to corresponding political remedies to discrimination. Firstly, potential struggles in identifying claims reveal the need to enable occupants of inferior social positions to formulate complaints. Andrew Dobson has argued that dominant groups have obligations to actively “listen out for” the undefined claims of people at the margins of political discourse, not just “listening to” demands raised by those privileged enough to voice them (Dobson 2010, 851; 2012). “Listening out for” unstructured grievances is necessary in cases of complex injustice. In such cases, processes of obfuscation, erasure, and indifference toward marginalized modes of being might render some identities continually invisible and unknowable (Dotson and Gilbert 2014, 874). The emergence of the term sexual harassment from discussions and surveys in the 1970s can be seen as an instance of successfully “listening out for” women’s claims regarding workplace discrimination (Farley 1978). “Listening out for” undefined grievances could both uncover previously neglected grounds (such as, potentially, class) and reveal previously neglected forms of discrimination against already recognized grounds (as the case of sexual harassment illustrates).

Secondly, since individuals who are not guilty of discrimination nonetheless contribute to the social processes which underline it, DwT recommends that third parties take political responsibility when discriminatory acts, practices, policies, and institutions are done in their (collective) name. This requires repoliticizing areas of social life previously framed as inevitable, by examining how our attitudes, behaviors, or expectations might render those marked out by relevant traits more vulnerable when combined with the attitudes, behaviors, and expectations of others. For example, citizens should take political responsibility for reaction qualifications, where an employee’s “characteristics which contribute to job effectiveness [serve] as the basis of appropriate reactions in the recipient” toward the goods or services offered (Wertheimer 1983, 100). While debates regarding reaction qualifications highlight the employer’s responsibility, it is the consumers or customers who are meant to actually react to the race, ethnicity, or gender identity of service providers or sellers. Customers and consumers hold a collective power to challenge what are perceived to be their collective attitudes and expectations in ways that incentivize employers to pursue more egalitarian practices. They have a responsibility to engage others in public, political action (Young 2011, 89). DwT hence opens avenues for researching what our obligations are for enjoining others in collective action aimed at challenging allegedly neutral market behaviors of transmitting preferences and their combined effects.

Thirdly, DwT revealed that the aim of anti-discrimination is to stop traits from functioning as intersubjectively understandable reasons for differential treatment. An important part of this comprehensive effort is to de-reify the grounds of discrimination and the differential norms, institutionalized interactions, and patterns of regard they supervene on. While this aim is forward-looking, a salient way of recovering the contingent nature of the way the grounds of discrimination function is to engage in a backward-looking effort of exploring the politics of the unjust past (Nuti 2019). This can inform not just the search for remedies by highlighting, for example, the role of the unjust past in practices of racial profiling (Mogensen 2019), but also the way remedies once identified are delivered. For example, affirmative action policies that seek to increase the participation of racial minorities in higher education usually identify beneficiaries in terms of their racial or ethnic origin rather than our structural responsibility for an unjust past. Instead, university questionnaires could inquire whether applicants have “a history of discrimination, connection with underserved communities or a likelihood of having encountered racism to their detriment” (Ford 2002, 7). Making the legacy of the unjust past salient in this way would help combat critiques that affirmative action policies place on beneficiaries a “demeaning burden of presumed inferiority” (Cohen and Sterba 2003, 110). It would do so, moreover, by showing the meaning of affirmative action policies is not an inherent aspect of the benefit provided but is up for political (re)negotiation. Institutions that adopt affirmative action policies thus have an unacknowledged reason-giving function and are a site where the grounds of discrimination can be reified or remade. As such, especially approaches to affirmative action which view it as a forward-looking effort of achieving a desegregated society (Anderson 2010, 148–54) would benefit from specifying intended beneficiaries by asking the prior question: how did we get here?

While the above analysis has sought to reveal political aspects of (anti)discrimination which take us beyond legally demanded compensation, this is not to deny the law contributes to the political by its very nature. ³ Firstly, while the law often does not precede but follows the social articulation of political interests, legal protections are political achievements. Legally recognizing a trait as a protected ground bestows political recognition to the social struggles of activists and litigants, and contributes to their cause by attempting to shift attitudes that persons marked by protected traits in some sense deserve being treated as inferior. In this sense, the law has a transformative role in taking anti-discrimination efforts from the social to the political. Secondly, legal protections also signal where individual and collective efforts to combat discrimination should be focused. The law has an agenda-setting power in prioritizing which practices and their effects should be examined, which meanings should be contested, and which costs should be shouldered.

CONCLUSION

At the outset, I asked how embracing a constructionist approach to the grounds of discrimination might impact the way we conceptualize the nature of wrongful discrimination. I showed that taking seriously the primacy of processes over properties leads to a novel understanding of the defining characteristics of discrimination and a distinctly political approach to its remedies. Doing justice to the insight that properties which ground discrimination are primarily outputs and not just inputs of social processes requires rethinking anti-discrimination struggles. This reconceptualization shapes not just empirical efforts to combat discrimination but also normative appraisals of whether and to what extent reaction qualifications or racial profiling should be resisted in the name of combating discrimination.

The initial proposal of viewing the grounds of discrimination as social positions also has the more far-reaching implication that discrimination is not primarily a matter of individual ethics or even of legal responsibility, but of social justice. While Iris Young criticized discrimination for being an “agent-oriented, fault-oriented” concept, blind to the collective and continuous nature of group-based injustice (Young 1990, 195), DwT shows discrimination can also be viewed as a systemic and continuous process instead of simply an individual “fault.” This possibility opens fruitful conversations about the relationship between discrimination and oppression. Such far-reaching implications of DwT prove that fully embracing constructionism is not about simply acknowledging properties touch the world of natural kinds only at the margins; it is about actively examining the processes through which they come to be, their relations with mainstream institutions and structures, and the avenues of political resistance needed to uproot them.