Discrimination: An Intriguing but Underexplored Issue in Ethics and Political Philosophy

1 Introduction

As I hope a reader of this brief introduction will agree (after reading it, if not before), discrimination raises a wide range of interesting philosophical questions. Given that these questions have received relatively limited attention compared to other questions in applied philosophy, it is a great pleasure to guest-edit this Moral Philosophy and Politics symposium. I hope that it will generate further philosophical interest in its topic and believe that the five papers of which it consists will take the discussion forward on a number of crucial points. Below I shall introduce three broad questions that philosophers need to address in relation to discrimination. In the second, and final, section I will present the five contributions to the symposium.

Complaints about discrimination consume extensive public discursive space. Almost everyone agrees that certain kinds of differential treatment are discriminatory and unjust. For instance, few people in Western countries would not regard the Saudi prohibition on female citizens leaving the country unaccompanied by a male relative discriminatory. However, other complaints about discrimination have a more uncertain status. Is it, for instance, discrimination against smokers, if lung clinics refuse to hire smokers? Is it discrimination against ordinarily looking people, if beautiful people are paid better wages than their not so beautiful colleagues? Or is it discrimination against obese people that seats on airplane are designed for average-sized people? Cases like these raise the question: What is discrimination? This conceptual question is the sort of questions that philosophers ask, and yet they have paid relatively little attention to it.

One important conceptual issue in connection with discrimination concerns who are possible discriminatees. Presumably, there is some form of intimate connection between discrimination and groups in that it is impossible to discriminate against an individual simply on account of her individuality. Suppose I need to hire a teaching assistant and refuse to hire Susan, despite the fact that she is the obviously best qualified applicant. Initially, the suspicion is that the reason that I treat Susan disadvantageously is that she belongs to a group, which commonly experiences discrimination in the relevant context, e.g., she is lesbian, black, or Muslim. Suppose, however, that she is not a member of any of these or similar groups. Or she is a member of some or all of them, but, on the basis of my previous behaviour, it is established that I am in no way disposed to treat members of such groups disadvantageously. Next, it is conjectured that I discriminate against groups that are not usually subjected to discrimination, e.g. heterosexuals, people who are fond of Dostojevsky, or people with green eyes and brown hair. Again it is shown that I simply have no such group-focused negative disposition. It seems that I have simply treated Susan disadvantageously qua the individual she is. In that case, I might have treated her unjust and unfairly – assuming, say, that the best qualified are entitled to be hired. But it also seems as though that I have not discriminated against her, since pure individualized discrimination conceptually impossible.

Elsewhere, I have argued that the group focus of discrimination is more specific in that it is tied to membership of a special kind of group – socially salient groups, i.e., groups which are such that perceived membership of them shapes social interactions across a wide range of different contexts (Lippert-Rasmussen 2013, pp. 30–36). Men and women each form socially salient groups, but people with green eyes and brown hair do not. Idiosyncratically differential treatment does not, I contend, amount to discrimination. However, as is apparent from the contributions to the present symposium not all theorists accept this view (cf. Malnes 2015, p. 249; Thomsen 2015, p. 315).

It is one thing to distinguish, conceptually, between those forms of differential treatment that amount to discrimination and those forms that do not. It is another thing to distinguish conceptually between different forms of discrimination. Most would distinguish between so-called direct and indirect discrimination, or, in terms more commonly used in the US, disparate treatment and disparate impact. There are various ways to draw a distinction between the two forms of discrimination.

On one influential account, direct discrimination takes place, where disadvantageous ‘acts or policies’ are aimed at certain socially salient groups. Indirect discrimination takes place where acts or policies are ‘not aimed – explicitly or surreptitiously, consciously or unconsciously – at persons for being members of a certain social[ly salient] group,’ but where these acts or policies ‘have the effect of disproportionately disadvantaging the members of a particular [socially salient] group’ (Altman 2011).

Historically speaking, the notion of indirect discrimination is intertwined with that of affirmative action. In the US context, courts found that it would often be difficult to show that, say, companies that after the 1964 Civil Rights Act continued to hire almost only whites for senior positions aimed to disadvantage African Americans. In the famous Griggs v. Duke Power Company (1971) case, the US Supreme Court ruled that certain employment policies of private businesses are indirectly discriminatory against African Americans and thus prohibited. In casu, the Supreme Court found that it was illegal for Duke Power Company to make a high school diploma a requirement for promotion, whereby due to past discrimination many fewer African Americans than European Americans fulfil this requirement and where this requirement is not suitably related to job performance. In response, many companies proactively initiated affirmative action programs intended to increase the percentage of underrepresented groups among their employees, despite the absence of any (demonstrable) direct discrimination.

The distinction between direct and indirect discrimination is not only important in a legal-political context; it is also important morally. For assuming that discrimination is morally wrong in virtue of the objectionable state of mind of the discriminator (cf. Alexander 1992) – e.g., its wrong-making feature is that the discriminator thinks the discriminatee qua membership of her group has a lower moral status – at least some kinds of discrimination are not necessarily wrongful. This follows from the definition above, since it implies that indirect discrimination does not involve a mental state of any particular kind. Alternatively, we must take the pluralist view that different forms of discrimination are morally wrong for different reasons and, thus, that it is not possible to give a unified account of the wrongness of discrimination. ^[1] This takes us to the second general topic in relation to discrimination.

Suppose we know what discrimination, as ordinarily understood, is. Or suppose that we clarify, stipulatively, what we mean by ‘discrimination’. We can now helpfully ask whether discrimination is unjust or morally impermissible. These are different questions, since an act might be just and nevertheless morally impermissible, and an act might be unjust and nevertheless morally permissible. Most believe that while justice is an important and not easily overridden moral value, it is only one value among others determining whether an action is morally permissible. For instance, it might be unjust to frame an innocent person, but if that were the only way to prevent the death of millions of people, doing so might nevertheless be morally permissible. Conversely, it might be just if you were to defend yourself against a culpable aggressor, who wants to steal your money, by knocking him unconscious in a way that causes amnesia, but nevertheless impermissible if he knows a cure to malaria, which he will remember and publicize saving the lives of millions of people only if you do not defend yourself. Similarly, justice and moral permissibility can come apart in cases of discrimination.

The morality of discrimination raises numerous intriguing issue. First, is discrimination always morally wrong in some way? Indeed, is it wrong by definition? In relation to the latter question, it is certainly true that, when labelling a case of differential treatment people often intend to convey their disapproval. Yet, the same is true of ‘act of terror’ and it is not true by definition that all acts of terror are in some way morally objectionable. Also, it involves no conceptual confusion to say, whether correctly or not: ‘It is age discrimination to give younger patients priority over older patients in the waiting-queue for live-saving organs. However, not doing so would be seriously morally objectionable.’

Even if discrimination is not by definition morally wrong, it might nevertheless always be morally wrong. Whether it is or not, depends on what makes discrimination wrong. Here, as in relation to all other comparable questions in applied ethics or applied political philosophy, philosophers give different answers. As indicated above, some appeal to the morally wrongful mental states in which discrimination is rooted, such as disrespectful beliefs about the inferior moral status of members of those groups against which one discriminates; others in the wrongful objective meaning of discriminatory acts (cf. Hellman 2008); and yet others locate the wrongness of discrimination in the harm causes to discriminatees (as well as third parties and, possibly even, the discriminators themselves) etc. (Lippert-Rasmussen 2013, pp. 153–189).

None of the parties to the debate have offered compelling arguments in favour of their position. In part for this reason, it is interesting how whichever of the three views mentioned that we endorse it follows, first, that it is possible for an act of discrimination not to be morally wrong. Or, at least, this follows assuming that it is possible – as seems to be the case – for an act of differential treatment neither to be rooted in any wrongful mental state, nor having any wrongful objective meaning or harmful effects. A case of statistical discrimination in which the police devote more resources to investigating tax fraud by members of traditionally privileged groups solely out of a concern to prevent crime and with the result that the state has a greater tax revenue, which it spends on public goods making everyone better off, would not be wrong according to any of the three accounts mentioned.

Second, on all of the three accounts mentioned discrimination is not morally wrong for a reason, which is distinctive of discrimination. For instance, non-discriminatory acts can be harmful – e.g. a serial killer who kills people at random does not manifest the vice of discrimination and yet he certainly harms people – and wrongfully demeaning – e.g. I treat all persons, myself included, as if our moral status were no different from that of a worm.

Suppose we know what discrimination is and what makes it morally wrong. In that case, it is natural to ask a third broad question about discrimination: What ought we to do, morally speaking, to eliminate, or counteract the effects of, discrimination? Knowing what makes discrimination wrong does not in itself answer this question. One reason this is so, is that relevantly effective measures might clash with other values. Religious groups that insist that heterosexual men, but not women or homosexual men, can serve as priests, imams, rabbis etc. engage in sex discrimination etc. ^[2] Yet many think that even if much religiously based sex discrimination can be eliminated through criminalizing sex discrimination in relation to the hiring of priests etc., doing so clashes with the freedom of religion and that this value takes priority over the value of non-discrimination. Such a view is controversial in itself, but it also raises the challenging question of why other values do not take priority over the value of non-discrimination. For instance, might the value of freedom of association take priority over non-discrimination such that racist, private country clubs ought to be legal?

2 Contributions to the symposium

Having introduced three broad questions philosophers must to ask in relation to discrimination I now briefly introduce the five articles comprising the present symposium. Four of them focus on conceptual issues in relation to discrimination, while one – by Xiaofei Liu – scrutinizes moral issues in relation to discrimination, to wit, whether ‘racial looksism’ is morally equivalent to ‘simple looksism’ (Liu 2015, p. 256). Two of the conceptual contributions – by Raino Malnes and Oscar Horta, respectively – discuss features of the general concept of discrimination – whether discrimination is a vague concept and whether the act of discriminating necessarily disadvantages the discriminatee. The other two conceptual contributions – by Ryan Cook and Frej Klem Thomsen – critically assesses the distinction between direct and indirect discrimination. In the following, I sketch the five contributions in the order that I have mentioned them here.

Xiaofei Liu opens with a real-life case of discrimination. In a recent article, activist Jimmy Nguyen complained about the fact that many online gay dating profiles contain the caveat ‘No Fats, Femmes, or Asians’. According to Nguyen, such caveat manifests an objectionable racial looksism. However, as Liu reports, commentators were quick to reply: ‘Mr. Nguyen, would you date a fat man?’ To put the implied challenge differently: if, as Mr. Nguyen thinks, racial looksism is morally wrong, is not simple looksism – ‘a preference that finds certain people aesthetically unappealing and thus sexually unattractive due to their having a certain physical appearance’ (Liu 2015, p. 256) – also morally wrong?

Different views are possible here. One is finding both forms of looksism to be wrong. Liu, however, defends a negative answer. He argues that, unlike simply looksism, racial looksism involves an overgeneralization and denies people a fair consideration for some important form of appreciation. The relevant overgeneralization consists of thinking that race is a reliable indicator of whether its individual members are aesthetically appealing. The overgeneralization is due to the fact that race ‘is not directly contributive to having a particular physical appearance’ and is thus irrelevant to assessing one’s aesthetic appeal (Liu 2015, p. 266).

Those disagreeing with Liu’s asymmetry thesis might part from his argument on various points: (1) They might deny that the ‘No Fats’-part of the relevant caveat does not involve a similar overgeneralization – people who write ‘No Fats’ in their personal adds might find some overweight people aesthetically very appealing, and indeed perhaps in some cases, find them appealing in part because of their being overweight; (2) They might deny that race could not be directly contributive to being aesthetically appealing, perhaps even for non-objectionable reasons; (3) Or they might deny that there is anything wrong about engaging in overgeneralizations in love life. Liu has things to say about all three challenges to his account and, in any case, it casts important light on an under-theorized form of discrimination, which, as Liu contends, is quite important for people’s well-being.

Turning now to the four contributions that primarily tackle conceptual issues in relation to discrimination, Raino Malnes explores our general concept of discrimination. While he thinks that labelling something ‘discrimination’ does not imply that one thinks it is morally wrong, doing so involves a ‘presumption of wrongdoing’ (Malnes 2015, p. 246). Hence, according to Malnes our concept of discrimination is moralized, albeit only in a very weak sense. Moreover, he argues that our concept of discrimination is vague. While some cases clearly amount to discrimination and others do not, between those extremes are a significant number of borderline cases in which we cannot definitively say whether they involve discrimination. One way in which Malnes thinks our concept is vague is that it is unclear, whether discrimination is restricted to the differential treatment of members of different socially salient groups, or whether it extends beyond that to idiosyncratic differential treatment (cf. Thomsen 2015, p. 304n5). The upshot of this point is that we should not set our hopes too high, when we define discrimination. Or to put his point differently: any lexical definition of discrimination must be imprecise, and any precise definition must, in part at least, be stipulative. As Malnes sees it, discrimination is best conceived as a graded property of action; that is, one that that actions have to a lesser or greater degree. Presumably, acts of idiosyncratic differential treatment have this property to a lesser degree than differential treatment based on membership of socially salient groups (cf. Malnes 2015, p. 254).

Oscar Horta challenges the view that ‘in order to be discriminated against one has to be treated or considered disadvantageously’ (Horta 2015, p. 277). Against this disadvantageous-treatment requirement he offers three counterexamples. First, he describes a case of epistemic discrimination, where the discriminator ascribes less credence to the view of a female TV-debater than those of her male interlocutor, even though the discriminator has epistemic reasons for making the opposite assessment. This act of discrimination takes place in a causally isolated context such that it does not negatively affect the woman (nor any other woman). Second, discrimination can take place against discriminatees who cannot be affected by the relevant acts of differential treatment; for example, because they are fictional characters or because they no longer exist. Third, there are possible cases of segregation – much different from real-life cases of segregation – where none of the segregated individuals are harmed – perhaps everyone even benefits relative to non-segregation. If these counterexamples are accepted, we cannot distinguish between discrimination against, on the one hand, and discrimination in favour of or discrimination between – e.g. possibly providing different sizes of clothes for men and women –, on the other hand, by saying that the former involves disadvantageous treatment (cf. Thomsen 2015, p. 304).

Constructively, Horta proposes substituting the requirement that discrimination against someone involves considering or treating them worse than others relative to a certain standard for the disadvantageous-treatment-requirement. This suggestion accommodates the counterexamples above, e.g. the epistemic discriminator treats the female debater worse than the male debater relative to a testimonial standard of credence. Also, the suggestion is pleasingly simple – unlike some of the other alternative substitutes for the disadvantageous-requirement condition that he considers (Horta 2015, p. 295). While Horta thinks this is the best way to go, he is open to the view that, for some, his worse-relative-to-some-standard requirement is too broad to be appealing. Arguably, this illustrates Malnes’s point about vagueness, and supports the view that the definition of discrimination will be stipulative, at least in part.

The two remaining contributions to the symposium both discuss the distinction between direct and indirect discrimination. Both contributors are critical of how the distinction is commonly drawn, although, as we shall see, for different reasons.

Ryan Cook argues that, contrary to what most theorists suppose, ‘direct and indirect discrimination are structurally similar’ (Cook 2015, p. 219). By this he means, first, that both forms of discrimination involve similar decision-making processes, and, second, that they ‘are about group membership similarly’ (Cook 2015, p. 219). Pace Cook’s first claim, it might be said that the common view is that direct discrimination involves some mental state to the effect that members of a particular group are treated more disadvantageously than others – or, bearing in mind Horta’s suggestion, worse relative to some standard. For instance, a sexist, directly discriminating employer has an intention to disadvantage, or at least, exclude women, unlike an employer who indirectly discriminates against women and may be unaware that the procedure he employs, when hiring new employees, disadvantages or excludes women.

Analysing two examples of discrimination – one involving disparate treatment and another involving disparate impact – Cook rejects this view. The former case is a US court case from 1857 – Dred Scott vs. Sandford. Dred Scott was a slave who had filed a lawsuit to obtain his freedom. The lawsuit was rejected, in part, on the ground that Scott was a ‘negro’ and that ‘negroes’ are morally inferior to whites such that unlike them, they are not self-owners. As Cook points out, to the extent that the judges reached their discriminatory verdict on this basis, they were not motivated solely by Scott’s group identity – being a ‘negro’ – but by the fact that a significant trait – not being a self-owner – is (believed falsely by the judge to be) ‘inextricably associated’ with the relevant group identity. Cook concludes: ‘it’s false to claim that disparate treatment discriminators are only motivated by the victim’s group membership. Rather, the discriminators are also motivated by what are perceived to be the victim’s relevant traits, much like in disparate impact’ (Cook 2015, pp. 225–226) and the two forms of discrimination are, thus, structurally similar in this way. ^[3]

Second, appealing to Iris Marion Young’s analysis of social groups Cook distinguishes between the constitutive and derivative traits of group members. The former are those traits, e.g. the having of certain beliefs or shared experiences, by virtue of the possession of sufficiently many of which one is a member of the relevant group. The latter are traits that members of the relevant group typically have ‘because of pre-existing constitutive traits, the person’s group membership itself, some social circumstance relevant to the group, or some combination thereof’ (Cook 2015, pp. 229–230). He proceeds to argue that, in cases of both direct and indirect discrimination, discriminators are ‘concerned with a trait that is both believed to be significant in some way and is (believed to be) a constitutive or derivative trait for some group’ (Cook 2015, p. 233–234). Hence, direct and indirect discrimination are about group membership similarly.

On the basis of his structural similarity claim, Cook proposes a complex criterion for when someone discriminates against another, which he then tests against various cases, including possible counterexamples (Cook 2015, pp. 235, 238—242). One putative counterexample involves a discriminator who treats members of a certain social group disadvantageously, not because of any trait – e.g., laziness – that he thinks is associated with the relevant social group membership, but simply on account of their membership in the relevant social group. Cook thinks that while such cases are conceptually, or least logically, possible, there are not actually any such cases. Moreover, since he believes that, simplifying somewhat, a criterion of discrimination should only be tested against actual cases – ‘Discrimination is an empirical event, and therefore must be conceptualized according to what actually happens’ (Cook 2015, p. 240) – it does not defeat his criterion.

If direct and indirect discrimination are structurally similar, does that mean that there is no underlying difference between the two? This would have important implications. For instance, we could no longer assume that direct and indirect discrimination are morally wrong for different reasons. Since Cook accepts that there is at least one (structural?) difference between direct and indirect discrimination, his argument should not be taken to imply this. More specifically, he thinks that, unlike in cases of indirect discrimination, ‘a disparate treatment discriminator… accounts of the victim’s group membership’ (Cook 2015, p. 242) and while Cook finds it intuitive to claim that this difference is ‘insignificant in the grand scheme of things’, he is also open to the view that this difference makes it ‘disingenuous to claim that [direct and indirect discrimination] are virtually identical to one another’ (Cook 2015, p. 243).

Like Cook’s contribution, Frej Klem Thomsen’s is about the distinction between direct and indirect discrimination. Unlike Cook, Thomsen’s main aim is not to argue that direct and indirect discrimination are structurally identical, or at least much more alike than is commonly assumed. Indeed, Thomsen thinks that one desideratum for a definition of indirect discrimination is that it allows us to classify all cases of discrimination as one, and only one, of the two forms of discrimination. Thomsen’s main aim is to canvass a competing account of how to draw the distinction. Specifically, he claims that indirect discrimination is best defined as equal treatment that is disadvantageous to the discriminatees because of their group membership. As Thomsen points out, however, this definition cannot satisfy the indicated desideratum. Thus, in Thomsen’s ingenious case of the inadvertent misogynist – a wealthy financier buys a factory, installs a new assembly belt that requires workers to be at least 175 cm tall, and accordingly orders all below-175 cm workers, all of whom happen to be women, fired – it cannot say that it is a case of indirect discrimination against women rather than a case of direct discrimination against those who are shorter than 175 cm (Thomsen 2015, p. 319).

Thomsen investigates a number of solutions, and argues that the only promising idea among them is to moralise the definition of discrimination, that is, to hold that in addition to other conditions an action constitutes discrimination only if it is objectively – not, as in Malnes’s piece, presumptively – morally wrongful. While this is an advantage of a moralized account, as Thomsen points out it comes with disadvantages, among them that it seems perfectly possible to talk about non-wrongful discrimination, such as statistical discrimination by police against young males, which suggests that while moral wrongness is a feature of common forms of discrimination, it is not a conceptual requirement (Thomsen 2015, p. 320). Indeed, friends of a non-moralized account might add that a similar pattern of reasoning underpins Thomsen’s rejection of the view that discrimination is conceptually tied to the notion of social salience (Thomsen 2015, p. 315).

In response, Thomsen might remind his reader that, first, while he wants to define ‘discrimination’ in a way that fits ordinary language this is not his only concern. Another concern is to define discrimination in such a manner that it fits interesting features of the phenomenon, to which the term refers, and this desideratum might relate differently to building moral wrongness and social salience into the definition. Second, in any case Thomson does not commit himself to a moralized definition. He leaves open whether a purely descriptive criterion is no less satisfying than the moralized one mentioned, all things considered, and explicitly notes that a moralized definition has’both strengths and weaknesses’ (Thomsen 2015, p. 319).

A moralized definition of discrimination leaves one with the task of explaining what makes discrimination morally wrong and here Thomson turns to the second of the three main questions about discrimination that I distinguished between in the first section. Thomsen takes a different approach from Liu. On his account indirect discrimination is wrong, not because it involves not treating people on the basis of who they are, but on the basis of how other people are (Liu 2015, p. 260). Rather, according to Thomsen direct and indirect discrimination are morally wrong because of how they harm people. ^[4]

As is probably apparent by now, there are various focal points where the five contributions differ, even disagree. However, the symposium also attests to the fact that discrimination, is not only an important real-life problem, but is also an intriguing topic in applied philosophy that presents challenging questions, the answers to which rest on and have implications for broader underlying philosophical questions – e.g., in relation to the question about the prospect for defining contested concepts and in relation to basic questions in normative ethics.