II. The Alleged Irrelevance of Intention, Fault, and
Motivation to Permissibility 780
III. The Deception Accounts of Wrongful Discrimination 783
IV. Discrimination from Animus and Prejudice 787
V. An Objection 789
VI. Innocent Discrimination 790
VII. Disparate Impact 793
VIII. Suspect Classifications 795
A. Age 797
B. Sex 799
C. Sexual Orientation 801
D. Beauty and Ugliness 803
IX. Discrimination, Deontology, and Consequentialism 804
X. The Assimilationist Ideal 805
Legal prohibition of some types of discriminatory conduct may be morally acceptable even though the conduct being prohibited would not be immoral in the absence of legal prohibition. Consider Thomas Schelling’s analysis of patterns of racial segregation in residential housing.1 If one sees a sharply segregated housing segregation pattern (for example, African-Americans living next to African-Americans, whites living next to whites, and African-Americans living next to whites only at the neighborhoods’ edges) even though there is no legal requirement that forces this result, one might suppose that what explains the segregation is a strong desire of almost all members of one or both groups not to live in proximity to any members of the other group. Schelling presented a simple model of the dynamics of residential housing choice that showed that mild racial preferences could lead to strongly segregated outcomes.2 For example, if nobody wants to live in a neighborhood in which members of his racial group are a minority, and individuals occasionally move in and out of neighborhoods, eventually a strongly segregated pattern emerges.3 In other words, segregation can emerge even if no one is averse to living in proximity to members of another race.
Plausibly, the mild racial preferences that Schelling showed capable of inducing segregated housing patterns are morally innocent. It is not merely the case that they are not seriously morally wrong or not viciously racist. Arguably, they are not wrong at all.
Nonetheless, the segregated housing patterns induced by a Schelling mechanism might be the cause of serious social harms. To generate a simple example, imagine that children interact with other children who live nearby, that whites are wealthy and educated and African-Americans are poor and uneducated, and that interacting with children whose parents are wealthy and educated is a great boon if you are a child of poor and uneducated parents. Also assume that contact with children whose parents are richer and more educated than yours increases your expected lifetime wealth and education prospects, and to a far greater degree than interacting with children whose parents are poorer and less educated diminishes your lifetime prospects. Segregation in these imagined circumstances would do little, if anything, to help white children and would do a lot to hurt African-American children. Specifically, segregation would deprive African-American children of an important educational resource. In this setting, prevention of serious harm to African-American children might constitute sufficient grounds for passing a fair housing act that forbids discrimination on the basis of race in the rental housing and home sales markets.4
Even if discriminatory choices about where to live cause aggregate harm, none of the acts need be morally wrong according to nonconsequentialist moral principles.5 The acts might exhibit no more than morally permitted partiality toward those near and dear to us and violate no individual’s moral rights.
The Schelling analysis of residential housing segregation suggests the question, when an act of discrimination on the basis of race is morally wrong, what features of what is done determine that result? More broadly, what is morally wrongful discrimination? This essay seeks to answer this question.
Antidiscrimination norms single out particular categories of traits and forbid discrimination among persons in certain contexts on the basis of these traits. Neither theory nor practice tell us much about the principle for selecting these categories, if such there be, and the justification for this principle, if any. The principle of selection does not leap out and confront us when we inspect the particular categories of traits that are commonly agreed to be illegitimate determinants of choice. The idea that it is wrong to discriminate on the basis of race, creed, or color commands wide assent. The idea that it is wrong to discriminate on the basis of talent, virtue, citizenship, or friendship and family ties does not. Beyond that, the status of many classifications is uncertain, and the principle of selection looks elusive.
Discrimination that is wrong in one context may be acceptable in another. Inviting only old people to my sixtieth birthday party may be ill-advised, but would not be regarded as morally troubling in the way that refusing to extend an employment offer or university admission to an otherwise best-qualified applicant on the ground of old age is thought to be. In broad terms, home ground for an antidiscrimination norm is a public sphere of activity as distinguished from a private sphere. How to draw the relevant boundary line between public and private spheres for these purposes is not clear and is also contestable. An analytic account of the nature of wrongful discrimination should be able to explain why antidiscrimination norms apply only in certain contexts, and ideally should provide a principle of demarcation. On the other hand, the importance of this split between public and private should not be exaggerated. If I refuse to date an African-American woman or to befriend a Chinese-American man solely on the ground that I regard only people who are white-skinned as meritorious candidates for personal friendship, surely these prejudiced refusals are morally wrong if discrimination on the basis of race or skin color is ever morally wrong.6
The exploration of this topic will proceed within a given moral framework. This framework is assumed, and not even cursory attempts are made to defend it. The framework is not idiosyncratic.7 Arguably, it is part of the plain common sense of contemporary culture. At any rate, it is worth asking, what is the most sensible account of wrongful discrimination, given the framework? The framework is a deontological morality that holds, contrary to act consequentialism, that what it is morally right and wrong8 to do is fixed by constraints and options.9 There are moral constraints on conduct that restrict what it is permissible to do in pursuit of any of one’s goals.10 These moral constraints mainly take the form of moral rights of others that are correlative with moral obligations that one must not violate these rights.11 So long as one conforms one’s conduct to the moral constraints, one is not morally required to bring about the greatest good that one’s choice of act could achieve within these constraints.12 Thus, one has wide liberty to live one’s life as one chooses so long as one does not violate any moral constraints. In the leeway that constraints leave open, one has options.
Discrimination that is intrinsically morally wrong occurs when an agent treats a person identified as being of a certain type differently than she otherwise would have done because of unwarranted animus or prejudice against persons of that type.13 In other words, wrongful discrimination is a subcategory of defective discrimination. One person may fail to respond to another in the right way given the circumstances, or respond by treating the other in ways that fail to adequately respond to the reasons that dictate how the other ought to treated, without the failure amounting to wrongful discrimination. The extra bit that when added to generic defective discrimination constitutes wrongful discrimination is the fact that one is led to defective conduct toward the other by unjustified hostile attitudes toward people perceived to be of a certain kind or faulty beliefs about the characteristics of people of that type.14
II. The Alleged Irrelevance of Intention, Fault, and Motivation to Permissibility
At the outset, a serious objection applies to this proposal. The objection derives from work by Judith Thomson in clarifying the structure of a morality of moral constraints and moral options.15 Thomson’s idea is that we should separate cleanly the issues (1) whether or not an action by some agent would be morally permissible; (2) whether the agent would be at fault if he were to do the action; and (3) with what intention the agent would do the act if he were to do it.16 Issue (1) is claimed to be entirely independent of issues (2) and (3).17 Whether going to the store and buying bread right now is permissible depends on the features of that act, and according to Thomson, the crucial point is whether my doing this right now would violate anyone’s moral rights (that are not overridden by counterbalancing factors).18 The act can be completely innocent even if the actor intends something bad in doing the act, or would be (culpably) at fault in doing it. My hatred of Sally might induce me to repay a debt to her on time, because I know that she will be distressed that my right conduct in this instance, a rare phenomenon, deprives her of yet another opportunity to lament my failure to fulfill my obligations. My intention is bad but the bad intention does not taint the act, which remains the morally required thing to do. So, any proposal to characterize a type of wrong action in terms of the agent’s intention or the agent’s culpability of conduct is mixing up categories that for clarity’s sake had better be kept in separate bins.
Thomson illustrates her thesis by considering examples in which the distinction between what the agent is actually doing, in the thin sense, and what the agent intends to be doing is clear and sharp.19 For example, assume I intend to kill my wife in circumstances in which she plainly has a right not to be killed. I offer her what I take to be deadly poison and seek to persuade her to ingest it. In fact, the stuff I am giving to her is the medicine that she must take if she is to survive some threatening lethal disease. What I am doing is saving her life; what I intend to be doing is murdering her. Clearly what I am actually doing is morally right, despite my evil intent. I am morally at fault or culpable here in doing what is actually the right thing. The culpability arises from my evil intention.
The Thomson position does not officially deny that the permissibility status of proposed acts is independent of the motivation that would lead the agent to do the act if he were to do it, but the considerations that drive a wedge between the assessment of the act and its fault on one side and the agent’s intention on the other would appear to press motivation on the fault and intention side of the divide. I treat her position as striking down on these general grounds my proposed account of wrongful discrimination.
I do not wish to deny that one may characterize an agent’s act in given circumstances in abstraction from the motive, intention, or fault that might attach to the agent’s doing of that action. Call this the act thinly described. For example, suppose that from spiteful malice I decline to share my large ice cream cone with my little brother, who strongly desires some licks. My act can be thinly described without reference to the spiteful malice. One can raise the question whether the act thinly described is morally permissible or not. Let us suppose this question is settled in a deontological framework by inquiring whether any true thin description of an act characterizes it as a violation of someone’s moral right. If the act fully described violates someone’s right, then it would be morally wrong for the agent to do it, unless all the agent’s alternatives would violate more weighty rights, or unless failing to violate the moral right on this occasion would bring about excessively great losses for other people (affected non-rightholders). Here I am just assuming that a reasonable theory of moral rights and corresponding obligations can operate on acts thinly described and determine their moral status as permissible, forbidden, or required. This assumption might be open to challenge, but that discussion would take us too far afield.
Let us also grant the assumption that assessments of an agent’s intention in acting reflect on the quality of the agent’s exercise of agency in doing the act and maybe on the character of the agent, but not directly on the act itself (thinly described). Likewise, whether an agent would be at fault or culpable if she were to do an act does not bear on the different question whether the act itself, again thinly described, is permissible, impermissible, or required.
But all of this leaves open a further assessment of the act thinly described combined with the intention with which the agent acts. And we can assess the package of the act thinly described combined with the features of the agent’s doing of it that render her at fault or culpable in the doing. These combination assessments might issue in judgments to the effect that the combination is permissible, impermissible, or required. Given the above, I see no reason to suppose that combining assessments is impossible or illegitimate. Thus, such assessments will be made.
There is a purely verbal issue, whether to say that an assessment of an agent’s motivation, intention, or fault in action counts as part of an assessment of the “action itself.” The resolution of that issue depends on how one defines one’s terms. If there is a substantive issue, it is whether or not it can be morally wrong to perform bodily movements that constitute an unobjectionable act thinly described, if it is also the case that if one were to perform those movements, one would be acting with morally objectionable intentions, motivations, or be engaged in a doing that is faulty when viewed from a perspective that encompasses more than thin description.
Returning to the example of my refusal to share my ice cream cone, it might be the case that this act thinly described is morally permissible, doing it from spiteful malice is bad, and the combination of the act thinly described and doing it from spiteful malice is impermissible. It is not wrong to decline to share your ice cream, but declining to share it for that reason is wrong. These judgments do not conflict; they can be true together. We could say the act I propose to do thinly described is permissible, but this leaves entirely open the further issue, whether the act thickly described is permissible. Put another way, I suggest that what an agent proposes to do (thinly described) may not be wrong as such, yet it would be wrong to do the thinly described act with a certain intention, or from a certain motivation, or if one would be culpable if one were to do the act.
Thomson tends to treat examples in which there are very strong reasons to do the act thinly described.20 Although it would be best to do the right thing for the right reason, if the thing is important enough it is likely better to do the right thing for the wrong reason than not to do it at all. However, consider examples of acts that thinly described are merely permissible, not supported by a compelling weight of moral reasons. For such acts one might hold that it would be better not to do the act at all than to do the act with an evil intention or in a manner such that one’s performance of the act would render one culpable. For example, it might be permissible to hire a white male rather than a more qualified Hispanic female for a job, there being no right of the most qualified applicant to be offered the position, but this permission evaporates if what I would be doing is passing over a qualified candidate from racial animus or misogyny.
III. Thin Description Accounts of Wrongful Discrimination
Although the Thomson Irrelevance-of-Fault-and-Intention-(and Motivation) -to-Moral-Permissibility claims should probably be rejected, the considerations adduced against them do not amount to a decisive rebuttal. So it may be worthwhile to inquire how to conceive of wrongful discrimination on the assumption that what makes discrimination intrinsically wrong has to do with features which show up in the thin description of such acts.
One possibility is that in contexts such as selection among applications for employment, bank loans, and university admission, the Lockean right in question is the right of the most qualified applicant to be selected first and offered the position in question, then the second best applicant if the applicant chosen first declines the offer, then the third best on the same terms, and so on down to the worst ranked applicant.21 What it is to be best qualified varies from context to context. For example, the best qualified applicant for a university course of study is the one whose aptitude and background preparation plus the likelihood that she will work hard at it, taken together and appropriately weighted, render her reasonably expected level of success at that course of study higher than any other applicant’s level. The best qualified applicant for a bank loan is the one to whom giving the loan increases the bank’s reasonably expected profits the most. The best qualified applicant for employment is the one whose hiring would most advance the hiring enterprise’s morally innocent goals, appropriately weighted.22
A further qualification must be noted. An applicant for employment, a bank loan, or a position as a university student participates in a procedure. The procedure specifies a selection process. An applicant can be best qualified and yet fail to be selected owing to various causes: (a) the procedure is ill designed; or (b) the procedure is well designed, but even so is imperfect, so perfectly following the procedure does not guarantee selection of the best applicant; or (c) the procedure is well designed, but not correctly followed in this instance, and so on.
Given this clarification, the right of the best qualified applicant for selection might be understood as the right of the person who is really, from a God’s eye perspective, best qualified. The right might also be understood in a more down-to-earth sense, for example, as the right to have one’s application processed by a reasonable procedure and to be selected if that procedure, correctly followed, would single out one’s application as top ranked. The down-to-earth specifications of “best qualified” slide closer to a purely procedural understanding amounting to the right to a fair process. Here, the right is considered in its ideal sense and in a proceduralist interpretation after that.
On this view, what is morally wrong about wrongful discrimination is that it tends to produce the result that the meritocratic right of the best qualified to be selected is violated.
The difficulty in this approach is that the idea that there is any such moral meritocratic right is on its face not plausible. Consider whimsical hiring. I am hiring persons to work in a doughnut shop I own. There are several other doughnut shops in the neighborhood, so it will not be a great loss to any actual or potential customer if my doughnut shop is not run as well as it might be. I announce that I will respond to the applications according to my subjective mood and select an applicant to be hired by arbitrary whim. This does not seem to be in the ballpark of wrongful discrimination. Nor is it plausible that my whimsically hiring Fred to man the cash register violates the right of other, better qualified applicants to man cash registers instead.
In many hiring cases, whimsical hiring would violate fiduciary obligations to firm shareholders. The manager of the firm is contractually obligated to the shareholders to run the firm in a way that is profit-maximizing, and whimsical hiring fails this test. However, one can contract out of such duties. For example, hiring Arneson as manager, it is agreed in advance among all concerned parties that Arneson is bent on whimsical hiring. A similar point holds respecting obligations a business enterprise might be thought to have toward potential customers. Again, explicit announcements of one’s intentions would suffice to eliminate any such obligations. But, then it seems that in cases where such obligations are in force, it is explicit or implicit contracting—voluntary deals one makes when one could instead have made a deal on different terms—that is the source of any obligation to engage in meritocratic hiring that might plausibly be thought to exist. If there is a violated right, it is the right that contracts be fulfilled if one has done one’s contractually specified part. This has nothing per se to do with wrongful discrimination.
If there is no right for hiring selections in the order of the comparative merit, wrongful discrimination cannot be analyzed in these terms. What goes for hiring holds also for distributing bank loans and filling student slots in universities.
Turn now to the idea that the right that wrongful discrimination violates is the right, operative in certain familiar contexts, to submit an application and to have one’s application considered according to some process reasonably related to the goals that are supposed to be furthered by the posts for which applications have been solicited. To be fair, the process need not be perfect and need not be perfectly followed. Imperfect procedures and mistakes in the administration of the process can be morally innocent, not wrong at all. But one’s fair process right is violated if one is excluded from consideration or not given consideration comparable to other applicants on grounds that have no reasonable relation to the underlying goals the process supposedly serves.
The objection to the idea that one has a basic moral right that one’s applications be given due process is a diluted version of my objection to the supposed right of the best qualified to be selected. Against the latter, whimsical selection can be morally permissible on any reasonable deontological view and against the former whimsical application processing can equally well be morally permissible.
Return to the doughnut shop example. Advertising for job applicants, I announce that there is no implicit promise, in soliciting applications, to give any application any sort of consideration. I reserve the right to exclude batches of applications on arbitrary and subjective grounds and to short-circuit any application handling process by simply picking a random application and offering that applicant the job. Alternatively, I might single out a whimsical process for selecting applicants and follow that process rigorously, awarding the job to the candidate singled out by that procedure. Such an irritating announcement might reduce the volume of applications received, but the announcement itself does not plausibly violate any potential applicant’s standing right to due process.