Can Normative Accounts of Discrimination Be Guided by Anti-discrimination Law? Should They? A Critical Note on Sophia Moreau's Faces of Inequality : A Theory of Wrongful Discrimination

: In her recent book, Faces of Inequality (2020), Moreau aims at developing a normative account of discrimination that is guided by the main features of anti-discrimination law. The critical comment argues against this methodology, indicating that due to indeterminacy relative to their underlying normative principles, central anti-discrimination norms cannot fulfill this guiding role. Further, using the content of such norms to guide ethical discussions is likely to be misleading, as it reflects evi-dentiary considerations that are unique to the legal context. The critical comment ’s claims are developed based on a close examination of indirect discrimination (or disparate impact) norms, and, as such, have wider implications for ongoing moral and political debates that are heavily influenced by the content of these norms.

legally oriented discussions of the phenomenon.This is hardly surprising, considering that many of the attempts to confront discrimination have taken place in legal venues, where legal anti-discrimination norms provide a framework in which discrimination claims are mounted and evaluated.What is surprising to many, however, is that the normative underpinnings of anti-discrimination jurisprudence-including its underlying assumptions about the nature of discrimination and the reasons why it is wrongremain, to a significant extent, obscure and controversial.
Anti-discrimination legislation and constitutional provisions commonly express a general commitment to equality, but do not specify any more particular interpretation of this value that can guide judgments on whether given instances of group-based differential treatment constitute wrongful discrimination.Historically, then, anti-discrimination jurisprudence has evolved through judicial decisions determining whether particular instances of group-based differential treatment coming before courts violate these general provisions; on the basis of such concrete, case-based determinations, courts have then formulated rules and standards that designate certain classes of cases as instances of illegal discrimination (De Burca 2012).
Presumably, the inclusion of certain cases of group-based differential treatment within the purview of anti-discrimination norms reflects, at least in part, the judgment that they are instances of wrongful discrimination.However, in many of the accompanying judicial discussions, the grounds for such judgments have remained unspecified; nor has the continuous process of developing anti-discrimination law included any direct and systematic examination of normative questions about discrimination.And while the legal-theoretical literature contains several projects whose aim is to specify normative rationales or principles that could explain or justify existing anti-discrimination law,2 these have been confronted with the intractable nature of this now extensive body of jurisprudence.Ranging over a variety of diverse domains, including employment, healthcare, and the provision of goods, and comprising numerous, versatile legal sources and materials (Khaitan 2015, 2-3), existing anti-discrimination law is arguably too complex and internally inconsistent for such normative underpinnings to be specified.Furthermore, several components of anti-discrimination law-including some of its prominent features-are the subject of ongoing, heated legal and political controversies about their appropriate interpretation (De Burca 2012, 3-7;Khaitan 2015, 2-5;Collins and Khaitan 2018, 4-12).The philosophical-egalitarian literature, on the other hand, has remained largely isolated from these legallyoriented discussions of discrimination, and only recently started examining ethical questions associated with the phenomenon (Lippert-Rasmussen 2013, 4).
Against this background, Moreau's recent book, Faces of Inequality: A Theory of Wrongful Discrimination (2020), takes a fresh, integrative approach to the topic.Similar to many recent philosophical discussions, the book aims at developing a normative account of discrimination, that is, a set of claims about when and why instances of group-based differential treatment constitute wrongful discrimination. 3Diverging from both recent philosophical and legal-theoretical discussions, however, Moreau maintains that such a normative account should accommodate both moral intuitions about discrimination and the content of existing antidiscrimination jurisprudence.More precisely, while she permits the book's normative claims to ultimately diverge, to an extent, from those that can be gleaned from existing jurisprudence, Moreau maintains that ethical discussions of discrimination should at least start from, or be guided by "some of the basic ideas about discrimination given to us by the law", or the law's "widely shared features" (13-14).Further, she maintains that the resulting normative account must accord with these features, arguing that failure to do so amounts to a serious inadequacy flaw that provides a strong reason to reject such an account (13)(14)(27)(28).Based on this methodological approach, along with the general normative premise that the wrongness of discrimination is tied to a violation of the (abstractly defined) value of equality, Moreau develops a pluralistic account of discrimination: she argues that discrimination is wrong when and because it (1) unfairly subordinates people, or (2) interferes with deliberative freedoms to which they have a right, or (3) deprives the victims of some basic goods (11).
Arguably, difficulties accompanying attempts at specifying the normative underpinnings of anti-discrimination jurisprudence as a whole loom at such a project as well.Further, one may question the prospects of accounting for both the normative nature of discrimination and antidiscrimination law at the same time, considering the gaps between them.Particularly, anti-discrimination laws in liberal societies do not aim at regulating all instances of wrongful discrimination (as Moreau notes about discriminatory actions performed by individuals in their interpersonal interactions, 211-247), while presumably reflecting goals and considerations other than targeting wrongful discrimination as such.Hence, there are reasons to suppose that the content of anti-discrimination law does not neatly align with or reflect the normative principles making discrimination objectionable.
Since these difficulties have been noted in the literature (Khaitan 2015, 1-9;Moreau 2020, 12-23, 27), however, I focus here on an often overlooked difficulty with using the content of anti-discrimination law to guide ethical discussions of discrimination, or requiring that normative accounts of discrimination accord with the law's prominent features.I will argue (in section II) that even when ignoring the difficulties associated with gleaning clearly-specified normative principles from anti-discrimination law in its entirety, and focusing instead only on one prominent, clearly-defined anti-discrimination norm-namely, the norm pertaining to indirect discrimination (or disparate impact)-this does not result in any clear enough guidance for developing equality-related normative accounts of discrimination.This is because the norm's content is not determinate enough relative to this task.Further, I will argue (in section III) that the norm's content reflects, to a significant extent, considerations pertaining to the process of adjudicating the factual aspects of legal claims of discrimination.Hence, using the norm's content to guide ethical discussions, or requiring that normative accounts of discrimination match its content is likely to mislead us in developing accurate normative claims.
Before proceeding, some clarifications are due.I will assume here, with Moreau and prominent views in the literature and public discourse, that the wrongness of discrimination is tied with a violation of equality, abstractly defined (4-9).Normative accounts of discrimination should, then, specify more concrete equality-related reasons why discrimination is wrong, or when and why instances of group-based differential treatment constitute wrongful discrimination, over and above this general association with inequality (24).My argument will be that anti-discrimination law cannot provide significant and adequate guidance for this particular task.

II. INDIRECT DISCRIMINATION NORMS ARE INDETERMINATE RELATIVE TO THE EQUALITY-RELATED REASONS MAKING DISCRIMINATION OBJECTIONABLE
Indirect discrimination norms (IDN) are a central feature of anti-discrimination jurisprudence in many countries and jurisdictions (Moreau 2020, 13-18;Hepple 2006, 608-609).Their original formulation has been set forth in the landmark U.S. Supreme Court case Griggs v. Duke Power Co.
(401 U.S. 424 [1971]).There, a company that had previously openly discriminated against African-Americans changed its policy when a legal prohibition on discrimination took effect.Instead of openly proclaiming that African-Americans will not be considered for certain jobs, the company instated an eligibility requirement of passing a standardized competency test, whose implementation led to the rejection of a large number of African-American applicants (whereas White employees already occupying the relevant positions were not asked to take the test).Based on these background facts, the Court introduced a rule on which policies, laws, or practices that are 'facially neutral'-that is, that do not contain explicit reference to people's group identity, and do not openly proclaim any discriminatory aim such as excluding or disadvantaging members of a certain group-are to be considered unlawfully discriminatory if they satisfy two conditions: (1) they lead to a 'disproportionate disadvantageous' (or 'adverse') effect on members of groups protected by anti-discrimination norms; and (2) this outcome is not reasonably connected to a legitimate goal of the policy (IDN are often distinguished from direct discrimination norms, which prohibit the explicit designation of people belonging to certain groups for disadvantageous differential treatment).Influenced by the Griggs decision, many jurisdictions outside of the United States have introduced legal norms with similar formulations, where they are routinely used to adjudicate claims of discrimination in a variety of different domains. 4hich equality-related normative principle(s) can plausibly justify a determination that a given policy is illegally discriminatory (and presumably, wrongfully discriminatory), under this formulation of IDN? Before examining this question, it is important to address a preliminary hurdle relating to a longstanding controversy about the norms' appropriate interpretation.Some legal authorities and scholars maintain that IDN should be interpreted solely as an evidentiary tool, to be used in legal proceedings to 'smoke out' cases where discriminators motivated by inegalitarian attitudes use 'facially neutral' tools (such as standardized tests in employment contexts) to avoid legal liability for excluding or disadvantaging members of certain groups (as most likely occurred in Griggs).Advocates of this more restrictive interpretation maintain that IDN should only apply to cases where satisfying the mentioned conditions indicates an underlying 'discriminatory motivation/aim'-presumably, the presence of things such as group-based animosity, prejudice, or objectionable stereotypes that have motivated or influenced the policy's design or implementation.Conversely, those advocating for a more expansive interpretation of IDN maintain that, at least in some cases, showing that the outcome of a certain policy satisfies the said conditions should be enough for it to fall under these norms' purview, even in the absence of an indication of discriminatory attitudes (Collins and Khaitan 2018, 25-28;Primus 2003, 518-536).
These competing legal-interpretive approaches-often associated with broader political and ideological orientations-are commonly taken to align with or reflect competing views about what makes discrimination wrong.Thus, contrast is often drawn between 'motivation-oriented' or 'subjective' views of discrimination, vs. 'outcome-oriented' or 'objective' views (see, for example, Rutherglen 2006).What is commonly overlooked in these broader debates, however, is that both approaches are compatible with a wide range of views about what makes discrimination wrong.Thus, applying both interpretations to real-life examples of differential treatment would lead to designating a wide range of cases as unlawfully discriminatory, and those would plausibly be objectionable (when they are) for a variety of equality-related reasons.Further, when examining particular instances of plainly objectionable discriminatory policies that would normally be designated as unlawful under each approach, oftentimes this particular judgment could be grounded in several equality-related normative principles.Hence, even assuming that this controversy was settled and one interpretive approach would prevail as reflecting the normative nature of discrimination more accurately, this would only take us so far in restricting the range of normative accounts that are compatible with the content of IDN.
To see this, let us focus on the second ('outcome-oriented') interpretative approach (which appears more compatible with Moreau's substantive views, 9-10), in its narrow application within the context of standardized tests used to screen out job candidates.Suppose that using a certain standardized test that is not reasonably connected to a legitimate business aim screens out a disproportionate number of African-American candidates (say, relative to their number among the pool of candidates).Suppose also that inegalitarian attitudes do not underlie or influence the policy design or implementation in any way.Such a policy would be designated as unlawfully discriminatory based on its outcome, according to this interpretive approach.But what are the equality-related reasons why it may plausibly be regarded as wrongfully discriminatory?
One claim that plaintiffs or purported victims can raise in such a scenario is that the policy is not compatible with equality of opportunity: rejected candidates can plausibly complain, for instance, that the reliance on a written test denies members of groups that tend to underperform in such tests an equal chance of freely competing against other candidates (say, based on on-the-job performance).But another way of accounting for the moral objection to such an outcome is based on the claim that certain benefits-say, positions in certain governmental institutions, or highly competitive positions for artists and athletes-should be awarded based strictly on some appropriate distributive principles, such as merit, reward for previous achievements, or compensation for hard work.Yet in other domains where the same formulation is routinely used-for instance, in adjudicating claims of discrimination in health or housingsuch claims do not seem appropriate, whereas other equality-related claims do.Plaintiffs or purported victims may claim, for instance, that the distribution of healthcare resources or housing resulting from an identifiable policy does not conform to principles of need-based distribution, or does not give priority to the worse off; and that because it burdens members of a certain group, such a policy discriminates against them.
All of this illustrates, then, that many different normative commitments may underpin a legal judgment that a certain policy is unlawfully discriminatory, under this outcome-oriented interpretation of IDN.All of these egalitarian principles and claims may, presumably, be used as a basis for developing general accounts of discrimination-all of them compatible with IDN, that is, with one of the law's prominent features.Moreover, considering the versatility and complexity of contemporary anti-discrimination jurisprudence, the range of these accounts is likely to be even wider if the law in its entirety is to be considered.This indicates that existing law does not significantly constrain the range of equality-associated normative accounts available to the ethicist, and that requiring that such accounts accord with its prominent features (at least, where IDN are concerned)5 does not provide significant guidance as to their content; rather, much of the latter would have to be developed and settled based on reasoning that is largely independent of the law's content.
Notably, while Moreau's account is pluralistic in the sense that it appeals to the three different principles mentioned above, it does not accommodate this wide range of equality-related normative views compatible with IDN.At least, the book does not specify why its three suggested

III. ANTI-DISCRIMINATION NORMS REFLECT EVIDENTIARY CONSIDERATIONS
Even assuming that anti-discrimination norms were underpinned by one or several clearly identifiable equality-related principles, however, that would not necessarily mean that the set of cases they would typically designate as illegally discriminatory could accurately guide normative discussions of discrimination.This is because the content of anti-discrimination norms has been designed, and is consistently interpreted and implemented in the adjudication of particular cases, based on evidentiary considerations.These relate to the fact-finding process in legal proceedings, and not to the normative principles explaining why cases coming before courts are wrong.
To see this, let us again closely examine how IDN are interpreted and applied in employment discrimination lawsuits.As previously noted, some legal authorities and scholars maintain that IDN should be interpreted solely as an evidentiary tool, designed to uncover instances where discriminators have operated on inegalitarian attitudes.This ('subjective' or 'motivation-oriented') interpretive approach is often taken to reflect the normative view that what generates the wrongness of discriminatory actions is such attitudes.On this approach, then, the norms' function in legal proceedings involves assisting plaintiffs' efforts of supporting the factual aspects of their claim that wrongful discrimination has taken place in a given case: for instance, showing that an employer in a given employment discrimination lawsuit has in fact been motivated by discriminatory attitudes in designing their candidate screening policy.Applying the second condition of IDN's formulation to adjudicate this factual question serves, in a range of typical lawsuits, the role of showing that the employer had no other reason for adopting the policy, hence providing evidence that she operated on discriminatory attitudes.
In other words, the norms' role under this interpretation is not that of designating as unlawful all cases that would be objectionable under such an attitude-oriented normative view (while assuming that the relevant facts obtained).Rather, their content focuses on identifying those cases where the facts deemed relevant under this normative view actually obtain, bearing in mind evidentiary difficulties that are characteristic of typical discrimination lawsuits.This evidentiary role is, of course, important in legal proceedings, where both plaintiffs and defendants are interested in concealing facts that would assist the other side in supporting their legal claims; but no similar fact-finding problem exists in purely normative debates.
This indicates that using the content of IDN to guide ethical discussions, or requiring that normative accounts accord with their contentfor instance, by examining instances of differential treatment that typically fall within their purview, and specifying the reasons why discrimination is wrong based on their characteristics-is likely to be misleading.This is because the norms would not reliably identify the class of cases that are relevant for purely normative debates.Rather, their scope of application is likely to be skewed towards those cases where facts deemed relevant under a certain normative view are likely to obtain (in typical legal settings).This does not necessarily align with the class of cases that would be considered wrongfully discriminatory under the same normative approach, assuming the relevant facts obtained.For instance, applying IDN's rule would fail to designate policies that are underlain by discriminatory attitudes as unlawful, where operating on such attitudes also serves a legitimate business aim (for instance, if this attracts customers).This is even though such policies would, presumably, be considered objectionable under an attitude-oriented normative view of discrimination.
The injection of evidentiary considerations into the content of antidiscrimination norms in the manner just described does not seem to be restricted to this attitude-oriented interpretation of IDN.For instance, in employment discrimination lawsuits that are adjudicated based on an outcome-oriented interpretation, similar evidentiary considerations seem to be intertwined with more substantive considerations-that is, background normative convictions about which instances of differential treatment are, in principle, wrongfully discriminatory-in guiding the norms' application.In general, this can be learned from the usage of the term 'indirect discrimination' in such contexts.There, the norms' formulation is not understood as laying out the nature of a distinct normative phenomenon that the proceeding is concerned with, but rather as detailing what plaintiffs and defendants need to show to prove or refute a claim of (legal liability for) discrimination in such proceedings.The latter, presumably, includes both principled normative claims, and factual claims about the particular case being examined (Ricci, 557 U.S. at 18-19).
For instance, in a scenario where a standardized written test is used to screen out candidates for a job where the relevant skills are primarily physical or interpersonal, a legal claim that such a policy wrongfully discriminates against members of a certain group (based on an outcomeoriented approach) would involve a normative component (for instance, the claim that screening job candidates based on a test measuring irrelevant skills is wrongfully discriminatory, because it violates equality of opportunity or merit-based distribution) and a factual component (this particular written test does not measure skills relevant for the job).A 'bottom line' determination that the policy under consideration is unlawfully discriminatory rests on both these elements, and, importantly, the formulation of IDN is suitable for adjudicating both elements in typical employment discrimination proceedings (Ricci,(27)(28)(29)(30)(31)(32)J.,dissenting]).On this outcome-oriented interpretation too, then, IDN would not reliably identify cases that are representative of the phenomenon of wrongful discrimination, and requiring that normative accounts match their content is likely to mislead us in developing accurate claims about when and why instances of group-based differential treatment constitute wrongful discrimination.

IV. CONCLUSION
The discussion conducted throughout Moreau's book sometimes seems to suggest that there is much to learn from legal materials about how wrongful discrimination instantiates in real-life situations, especially considering the abstract character of much of the egalitarian-philosophical literature (29-30).This seems plausible: the variety of ways in which different policies and practices-with their everyday subtleties, peculiarities, and technical complexities-could fail to treat people belonging to certain groups as equals is not easily predicted or characterized from the philosophical 'armchair'.Discussions of discrimination would thus benefit from incorporating detailed and realistic descriptions of the institutional, societal, and interpersonal mechanisms and contexts in which the phenomenon is often embedded, and some of these are indeed documented in legal materials.Moreau's book makes an important contribution to highlighting and pursuing this non-ideal-theoretical approach that normative discussions of discrimination should arguably adopt.This is different, however, from maintaining that normative claims about discrimination-that is, claims specifying when and why such everyday practices are wrongful, by appealing to normative principlesshould be gleaned from or accord with existing anti-discrimination law.As I hope to have shown by closely examining one prominent, influential anti-discrimination norm, existing jurisprudence can serve this particular guiding role only to a very limited degree, and using its content to guide normative discussions of discrimination may be misleading.
Beyond this methodological point, the discussion here can be of assistance in beginning to clarify some contemporary political and moral debates about discrimination.Specifically, while interlocutors often appeal to the formulation of indirect discrimination norms in such debates-for instance, in claiming that a certain policy is wrongfully discriminatory because of its negative disparate or disproportionate impact

(
2022); andTobler (2005).For general descriptions of how IDN are implemented to adjudicate concrete claims of discrimination in the United States (here again, indicating their diverse and sometimes inconsistent and controversial usage) see Ricci v. DeStefano (557 U.S. 557[2009]) and Texas Department ofHousing and Community Affairs v. Inclusive Communities Project, Inc. (576 U.S. 519 [2015]).
principles better align with existing law, relative to many other equalityrelated principles.Specifically, it is not clear why the former should play a more central role in a normative account of discrimination, as compared to principles such as equality of opportunity and merit-based distribution, which figure more prominently in existing jurisprudence(Griggs, Ricci, 557 U.S. at 20, 25; Ricci, 557 U.S. at 13, 19, 28   [Ginsburg, J., dissenting];Moreau 2020, 5).6