Merit and Reaction Qualifications

I. MORAL MERITOCRACY

Most solutions to the puzzle of reaction qualifications argue that selectors cannot consider reaction qualifications if doing so would be immoral. I call this category of solutions Moral Meritocracy. This category of views encompasses different approaches. One such approach argues that selectors can consider reaction qualifications when choosing between applicants just in case the underlying preferences are not immoral or prejudiced. We can call such preferences antimeritocratic: they are rooted not in judgments about how well someone performs their job, but rather in prejudice or bias directed against membership in some social group.⁶ (To say a selection procedure is unmeritocratic, by contrast, is just to say that it deviates from selection by merit.) This solution thus draws “limits to what can be a qualification by reference to what constitutes a just reason for selecting or discounting a candidate.”⁷ Therefore, when a reaction qualification “is rooted in a particular kind of objectionable belief or behavior,” it should not be considered by selectors.⁸

Another variant of Moral Meritocracy holds that selectors may not consider a reaction qualification if doing so would unfairly disadvantage applicants. This solution begins by identifying a symmetry between the reasons we can permissibly appeal to when making consumer decisions and those that employers can permissibly appeal to when making hiring decisions. If “customers should not respond to services offered by employees on the basis of their [e.g.,] racial preferences,” then employers should not make hiring decisions on such grounds either.⁹ The thought is that the requirement not to base consumer decisions on antimeritocratic grounds is mirrored in a requirement on the part of selectors not to appeal to these same antimeritocratic grounds when making hiring decisions. The reason why making consumer or hiring decisions on antimeritocratic grounds is prohibited on this approach is because doing so tends to unfairly disadvantage applicants.¹⁰

Though the ground on which selectors are prevented from considering reaction qualifications on this approach is unfair disadvantage and not the immorality of the preferences that ground the reaction qualifications themselves, the unfair disadvantage approach nonetheless holds that because counting reaction qualifications rooted in antimeritocratic attitudes would unfairly disadvantage applicants, this “makes it morally wrong not to disregard such qualifications in standard cases.”¹¹ Therefore, it too is a form of Moral Meritocracy, for though it is framed in terms of unfair disadvantage, the reason why unfair disadvantage matters with respect to the consideration of reaction qualifications is because it is immoral to unfairly disadvantage applicants.

Recently, the two approaches just described have been united into a hybrid view developed by Andrew Mason.¹² The hybrid view holds that a reaction qualification should not be considered by selectors if any of the following hold:

1. it is grounded in the recipients’ morally culpable failure to attribute equal moral status to some of the potential applicants, or

2. it is grounded in preferences, feelings, or responses of recipients that express an objectionable meaning, even though recipients do not culpably fail to attribute equal moral status to any of the potential applicants, or

3. counting it would unfairly disadvantage some potential applicants in the selection process by giving insufficient weight to their interests, or

4. counting it would on balance exacerbate existing unfair disadvantages, or would contribute to or cause rights violations.¹³

On the hybrid view, conditions (1) and (2) reflect the view that reaction qualifications should not count if they are grounded in antimeritocratic attitudes. The thought is that when reaction attitudes are rooted in a failure to attribute equal moral status or an objectionable attitude, they should not be relied upon by selectors. Conditions (3) and (4) of the hybrid view incorporate unfair disadvantage variant of Moral Meritocracy, holding that selectors should not consider reaction qualifications if doing so will impose or exacerbate existing unfair disadvantage.

According to the hybrid view, when one of the above conditions holds, a reaction qualification is illegitimate, which means “there is a weighty moral reason not to count it.”¹⁴ Illegitimate reaction qualifications should not be considered by selectors unless certain exceptions hold, in which case counting an illegitimate reaction qualification may be “nonetheless morally justified all things considered.”¹⁵ Thus, both the test for what renders a reaction qualification ineligible for consideration and whether there are overriding reasons to consider it anyway both appeal to moral considerations. Given the centrality accorded to the moral consequences of weighing reaction qualifications, therefore, the hybrid view represents a form of Moral Meritocracy. In particular, the hybrid view holds that reaction qualifications should not be considered if doing so would mean treating applicants immorally. It is immoral to deny the moral equality of all people, express objectionable attitudes, unfairly disadvantage someone, or exacerbate existing unfair disadvantage. The thought behind the hybrid view is that selectors should not act immorally by counting reaction qualifications when one of these conditions holds.

Moral Meritocracy has intuitive appeal: part of why it strikes us as impermissible to count reaction qualifications in Race is that it is immoral to hire based on racist attitudes. Such hiring practices are undoubtedly morally wrong. They are also violations of procedural fairness, and this fact is overlooked on views like Moral Meritocracy, which elide the distinction between fairness within selection procedures and considerations of morality that bear on the outcomes of procedures. The distinction between the morality of the outcomes of a procedure and a procedure’s fairness is important because the two can come apart, as I will now argue. The judgment rendered by Moral Meritocracy about the following case reveals a deeper problem with these approaches.

Looks: Executives at a record label are considering who to sign. They consider two candidates who are equally technically talented (i.e., equally good at singing and performing). Nonetheless, one of the contenders is considerably more conventionally attractive than the other. The executives believe audiences will respond favorably to the more attractive contender because of her good looks. As a result, she will make more money for the record label. They choose to sign the more attractive contender.

The more conventionally attractive candidate has better reaction qualifications: she is more likely to elicit positive responses from audiences and, therefore, more likely to be more commercially successful.¹⁶ Moral Meritocracy would deem these reaction qualifications illegitimate, potentially on two distinct grounds. First, the hybrid view might hold that Looks violates condition (3), the requirement that a reaction qualification should not be weighed if doing so would unfairly disadvantage applicants. We can look to Mason’s discussion of a structurally similar case, which he calls Good Looks, wherein selectors consider the reaction qualification of good looks when hiring sales associates.¹⁷ If the preferences of customers to interact with good-looking salespeople have only to do with “the pleasure they receive from interacting with him or her [and] are rooted in feelings they do not possess any reasons for having,” hiring based on reaction qualifications rooted in these preferences would unfairly disadvantage applicants and so should not be permitted.¹⁸ We can apply a similar analysis to Looks: if the preferences of audiences in favor of conventionally attractive musical entertainers are not grounded in any particular reason and instead have to do with the pleasure they derive from watching a good-looking person sing and dance, then counting reaction qualifications in Looks would presumably be disallowed on the same grounds as in Mason’s case.

Weighing reaction qualifications in Looks would also likely violate condition (4) of the hybrid view, the requirement that a reaction qualification should not be considered if doing so would exacerbate existing unfair disadvantage. We can again turn to Mason’s analysis of Good Looks: “those who lack good looks suffer from a general bias against them simply because interviewers favour attractive candidates or are influenced by stereotypes concerning the attractive or unattractive.”¹⁹ Moreover, there may be other considerations that make violation of (4) all the more likely: norms concerning conventional attractiveness are often gender-specific and “place greater burdens on women than men,” which is likely to exacerbate “existing unfair disadvantages.”²⁰ As a result, Mason suggests that weighing reaction qualifications tied to attractiveness in selection procedures for sales associates would be prohibited by the hybrid view. For similar reasons, the hybrid view likely would deem it impermissible for selectors to weigh reaction qualifications when selecting between candidates for the record deal in Looks.

However, I argue that it is not procedurally unfair for selectors to hire based on reaction qualifications tied to conventional attractiveness in Looks. In saying so, I do not mean to claim that allowing consideration of reaction qualifications in Looks would result in a perfectly just hiring procedure or outcome. Mason is right to point out that norms of conventional attractiveness are burdensome and carry differential impacts for applicants depending on what social groups they belong to. Weighing reaction qualifications in Looks would not be above reproach or moral criticism; however, nor would it constitute procedural unfairness in part because adherence to conventional norms of attractiveness is plausibly part of the job of a musical entertainer and, thus, part of one’s merit.

Considerations that speak to merit concern the job or position in question. Unlike in Looks, where adherence to norms of conventional attractiveness is standardly taken to be part of the job of a musical entertainer, race is not plausibly part of a salesperson’s job. As a result, weighing reaction qualifications in Race would not only be discriminatory—it would also amount to judging applicants on characteristics irrelevant to the job of a salesperson that do not confer merit, which violates procedural fairness. The injustice of the hiring procedure in Looks has to do with appeal to unjust standards of assessment, but not with departure from meritocratic criteria. The hiring procedure in Race, on the other hand, is unjust in both ways. The first, and of course graver, injustice involved in the shoe store’s hiring procedure is that it allows discriminatory attitudes to determine that the white applicant is hired. However, a second thing also goes wrong in the hiring procedure: the Black applicant is denied the job because he lacks reaction qualifications for reasons that have nothing to do with the job or his ability to perform it.

One might argue that it is not true that attractiveness is salient to a musical entertainer’s job in a way that race is not salient to a salesperson’s job. In either case, what matters is the applicant’s ability to contribute to the ultimate aim of selling records and shoes, respectively. In reply, I note that it is important to keep in mind the distinction between different ways in which one might advance their employer’s ends: one can do so by effectively performing their job, but one can also advance their employer’s ends in ways unrelated to job performance. For instance, a skilled blackmailer might make many sales, but this does not mean they merit the job of salesperson more than non-blackmailers. Often, the promotion of an employer’s ends will overlap with merit-conferring criteria (since one way of promoting an employer’s ends is to perform one’s job effectively). However, these can come apart, as in the case of the salesperson’s successful blackmail. The difference between Race and Looks can be understood in these terms: the job of salesperson does not involve one’s race (just like it does not involve being skilled at blackmail), even if, because of the attitudes of customers, this fact bears on how many shoes will be sold and so is relevant to how effectively one can advance the store owner’s end of growing profits.

There is thus an important distinction between unjust standards of assessment and violations of procedural fairness that Moral Meritocracy cannot capture. It does not follow from the fact that the standards of assessment are unjust that the hiring procedure violates selection by merit. The injustice of the grounds on which applicants are assessed is distinct from the further question of whether the applicants have been treated fairly with respect to these unjust standards. While both Race and Looks involve unjust standards of assessment, the selectors in Race do not treat applicants on their merits and accordingly violate procedural fairness—something which is not true of Looks. Moral Meritocracy rules out consideration of reaction qualifications in all cases where doing so involves appealing to disrespectful attitudes or disadvantaging applicants, even if doing so does not involve procedural unfairness. As a result, Moral Meritocracy is committed to the claim that Race and Looks are impermissible in the same way, namely that in both cases, selectors count reaction qualifications that there are moral reasons not to count. In collapsing the distinction between unjust standards of assessment and violations of procedural fairness, Moral Meritocracy is premised on an idealized conception of what it means to merit a job: possession of qualities that, were society morally enlightened, would be treated as qualifications and nothing else. This deprives us of the ability to assess the fairness of selection procedures where the standards of assessment are themselves unjust.

The solution I develop to the puzzle of reaction qualifications understands the permissibility of considering a reaction qualification in terms of whether doing so violates procedural fairness by rendering a selection procedure non-transparent or discriminatory. I will introduce and defend this approach throughout the following three sections, but first, I wish to offer a preliminary reply to a worry that might arise: if a solution to the puzzle of reaction qualifications is rooted in actual attitudes about what jobs consist in, does it amount to a relativist appeal to societal conceptions that reflect morally problematic views? The short answer: it does not, though this is a serious concern, which I address in Section III. My goal is to thread the needle between Moral Meritocracy—which places too little emphasis on procedural fairness, understood partly in terms of hiring in predictable ways that conform to societal conceptions of what jobs involve—and complete relativism—wherein selectors can hire based on societal conceptions of jobs without regard for how discriminatory these conceptions may be. I will argue that we can strike a balance between these two approaches, such that we can make sense of assessing applicant qualifications in cases where standards of assessment are unjust. We should want our solution to the puzzle of reaction qualifications to both (a) speak to hiring practices that take place in less-than-perfectly just conditions (so as to be of use to selectors navigating actual hiring procedures) and (b) do so without abandoning all normative constraints on selectors.

II. PROCEDURAL FAIRNESS AND TRANSPARENCY

One important way in which my view departs from Moral Meritocracy is in focusing on procedural fairness. In the sense I have in mind, a procedure will be fair just in case it can be justified to those who are party to it. Applicants’ interests in the procedural fairness of the selection procedures they are party to is, on my view, a fundamental interest applicants have with respect to how they are evaluated within hiring procedures. A procedure will be justifiable to those who are party to it when no applicant is disadvantaged relative to others without good reason. This is an important sense in which fair procedures constitute a level playing field. Considerations of fairness, framed in terms of the metaphor of a level playing field, provide a strong reason in favor of selection by merit. In meritocratic hiring procedures, the best-qualified applicant gets the job. This means that all applicants have an equal chance at the job in the sense that they are all judged on the same considerations. Selection by merit, then, is valuable in large part because it facilitates procedural fairness in the sense that it gives applicants the ability to compete on a level playing field: everyone is judged on the same criteria and no one is held to a higher or lower standard than any other applicant.

However, the plausible thought that selection by merit is prima facie fair seems to get us into trouble in cases like Race, where the prima facie fairness of considering merit-conferring reaction qualifications is in tension with the discriminatory consequences that follow from doing so. Note that this setup of the problem assumes that the reaction qualifications in Race confer merit. But we can interrogate this assumption. How? One might start by asking what skills or abilities are required for the job of salesperson. The trouble is that while we can fairly easily identify some core skills and abilities required to perform a job (a pilot is plausibly required the have the ability to fly a plane, for instance), our intuitions are much less clear about ancillary or penumbral skills. As such, defining the precise parameters of what skills a job requires through reliance on a priori reflection is unlikely to be a successful strategy.²¹

I propose we take an alternative approach to determining whether reaction qualifications confer merit. Given the close relation between procedural fairness and selection by merit, we have reason to think that a reaction qualification will confer merit only if relying on it is procedurally fair. I will argue that there are two steps to determining whether some reaction qualification confers merit, each pertaining to a feature of the meritocratic principle that enables selection by merit to promote procedural fairness. In particular, selection by merit ensures that selection procedures are conducted in ways that are (1) transparent and (2) non-discriminatory. When consideration of a reaction qualification would make a selection procedure non-transparent or discriminatory, this provides reason to think the reaction qualification does not confer merit and, consequently, should not be considered by selectors.

The first way in which selection by merit promotes procedural fairness is by protecting what I will call transparency. A selection procedure is transparent when applicants know how they will be evaluated prior to a selection procedure and can prepare accordingly by acquiring education, training, and experience that enable them to compete for the position. Generally, transparency will speak in favor of a procedure’s fairness. All else equal, applicants have strong objections to non-transparent selection procedures. This is because people have a strong interest in being able to control how their lives go and pursue their chosen career. In order to enact their life plan, one must be able to apply their efforts in ways that give them a chance to achieve their goals. Only fair procedures, which take up transparent means of assessment, can offer this.

This is why selection by merit is often instrumentally valuable in achieving procedural fairness: one of the main reasons we favor meritocratic selection is that knowing how selectors will evaluate applicants in a selection procedure allows one to plan and prepare accordingly. We value the ability to do so because this enables us to pursue our chosen path in life. Selection by merit can ground reasonable expectations in applicants about the considerations they will be judged on, and thereby enables them to plan how to prepare for such selection procedures, as well as how to present themselves and their achievements within the selection procedure itself. Selection by merit thus promotes the transparency of selection procedures. One way to treat applicants unfairly is to subvert their expectations and subject them to judgment along metrics they could not have foreseen.²²

Thus, the first step in determining whether a reaction qualification confers merit (and, therefore, whether it would be prima facie procedurally fair to consider) is to look at whether it would be transparent to assess applicants on its basis. It is not transparent to assess an applicant on some consideration if doing so would frustrate their reasonable expectations about how they will be assessed. Applicants tend to expect to be judged on grounds that speak to how well they can perform the job. As a result, assessing applicants based on some reaction qualification will promote transparency when the reaction qualification pertains to what the job involves; when this is the case, applicants may reasonably expect selectors to consider their reaction qualifications as a part of meritocratic selection.

But determining whether some reaction qualification pertains to what a job involves, in turn, relies on our ability to determine what tasks or abilities various jobs involve. Some ways of characterizing what various jobs involve will undermine the transparency of selection procedures that appeal to them, such as appealing to Platonic forms (how are applicants to know what the form of an accountant is?) or obscure and idiosyncratic institutional aims (how are applicants to know that a certain firm is after an accountant who can juggle?). Instead, for present purposes, we should assess what jobs involve by looking at widely held societal conceptions. Appealing to widely held societal conceptions of what various jobs involve promotes transparent selection procedures because this information is accessible to all applicants. For example, it would be unfair to judge applicants for a nursing job on whether they can code in JavaScript or flawlessly perform Chopin’s Nocturne in E-flat major. This method of evaluation is non-transparent because widely held conceptions of what nursing involves do not include these abilities. As a result, applicants to nursing positions do not expect to be assessed along these metrics.

When there is a societal understanding that certain reaction qualifications bear on one’s ability to perform some job, and this understanding engenders expectations on the part of applicants that selectors will judge them based on these reaction qualifications, it will, at least prima facie, be procedurally fair to consider these reaction qualifications. Thus, when a reaction qualification pertains to what the job involves based on widely held societal conceptions, assessing applicants based on that reaction qualification will be justifiable insofar as it will be transparent. When this is true of some reaction qualification, we have good reason to think that the reaction qualification speaks to merit.

Of course, grounds of assessment can be transparent yet unmeritocratic. For instance, if a selector announced they would select applicants by lottery, this would preserve transparency despite being unmeritocratic. Moreover, a selector might announce that they will assess applicants on grounds that do not have to do with what the job involves understood in terms of widely held societal conceptions. This, too, would preserve a kind of transparency.²³ A final way in which a selection procedure may be transparent but unmeritocratic (and so fail to promote procedural fairness) is if the societal conception of what the job involves that selectors hire on the basis of is itself antimeritocratic. Such cases present a challenge for my view and answering them will be the focus of part of Section III.

To review: when a reaction qualification does not pertain to what the job involves based on widespread societal conceptions, evaluating applicants on its basis will not be transparent and will frustrate their reasonable expectations. This gives us good reason to think that the reaction qualification in question does not speak to merit, since non-transparent assessment is unfair and selection by merit is fair in large part due to its promotion of transparency. Thus, we have good reason to conclude that assessing applicants on reaction qualifications that are not included in widespread societal conceptions of jobs is procedurally unfair on grounds of non-transparency and is therefore a violation of selection by merit.

We can now apply this transparency analysis to our central cases. In Looks, applicants will likely expect to be judged on reaction qualifications concerning conventional attractiveness. This expectation is engendered by a societal understanding of what the job involves. Although in a more just world conventional attractiveness would not be a qualification for being a musical entertainer, in our nonideal world, it is; being attractive is simply part of what the job involves in actual fact.²⁴ Considering these reaction qualifications is admittedly criticizable insofar as they are grounded in lookism, “the widespread practice of commenting upon and judging the appearance of others.”²⁵ However, given that such assessment accords with applicant expectations grounded in widely held conceptions of the job, considering reaction qualifications in Looks does not undermine transparency by assessing applicants in ways they could not have predicted or prepared for. This provides us with reason to think that the reaction qualifications in Looks confer merit and that to consider them would not violate procedural fairness.

In Race, by contrast, widely held societal conceptions of the job of salesperson do not involve race. Rather, in this case, there are business incentives to hire in accordance with reaction qualifications that do not have to do with what the job involves because of localized prejudice. As a result, applicants do not expect to be judged on reaction qualifications grounded in their race. This is because our societal conception of a salesperson’s job does not include one’s race. Therefore, considering reaction qualifications in this case subverts applicant expectations and violates procedural fairness. This gives us reason to think that the reaction qualifications in Race do not confer merit, and so would be procedurally unfair to consider.²⁶

I have argued that procedural fairness requires transparency and that when considering a reaction qualification would be non-transparent, this gives us reason to think it does not confer merit. But transparency alone is no guarantee of procedural fairness; societal conceptions of what jobs involve can themselves reflect discriminatory (that is, antimeritocratic) attitudes. Under these conditions, an applicant might, despite recognizing such attitudes are unjust, not be surprised to learn that selectors appeal to them. Thus, considering reaction qualifications rooted in these attitudes may not undermine transparency. Nonetheless, as mentioned earlier, subverting applicant expectations is not the only way to violate procedural fairness. In the next section, I will turn to the second step of my proposed test for when a reaction qualification can be considered, which has to do with non-discrimination. For now, I have argued that although the satisfaction of reasonable expectations rooted in societal conceptions of what jobs involve does not guarantee that a hiring procedure will be fair, it does eliminate one source of unfairness, namely, the unfairness of undermining transparency by subverting applicant expectations about how they will be assessed.

III. PROCEDURAL FAIRNESS AND NON-DISCRIMINATION

In the last section, I argued that we should not count reaction qualifications where doing so would undermine transparency. I will now argue that we cannot rely on transparency alone to determine whether it would be procedurally fair to consider a reaction qualification; this means that appeal only to transparency will underdetermine whether a reaction qualification confers merit. This is because selection by merit promotes fairness not only by means of transparency, but also because it guards against assessment of applicants on discriminatory grounds.²⁷ This is often taken to be analytically true: part of what it means to assess applicants based on merit is to refrain from assessing them on discriminatory grounds. As a result, if considering a reaction qualification would be discriminatory, that gives us good reason to think it does not confer merit and so should be disregarded, for to allow it into selector deliberation would violate procedural fairness.²⁸

Much of the time, if considering a reaction qualification would undermine transparency, it would also be discriminatory, meaning that considerations of transparency and non-discrimination will point in the same direction. However, this is not always true. In some cases, the two values come into conflict: prohibiting reliance on a reaction qualification may preserve non-discrimination at the cost of transparency. This is why the test for when to allow consideration of a reaction qualification must incorporate an assessment of whether doing so is discriminatory in addition to whether it is non-transparent. In such “conflict cases,” procedural fairness cannot be preserved by appeal to considerations of transparency alone.

Conflict cases arise when societal conceptions of jobs are themselves discriminatory, for instance, by endorsing the thought that certain jobs are only fit for members of certain of social groups. Under such conditions, appealing to societal conceptions of what a job involves can be unfair because doing so will make the selection procedure discriminatory. Thus, appealing only to transparency cannot explain why some reaction qualifications should be disregarded. To do so, we must also appeal to the value of non-discrimination. Because procedural fairness also involves a prohibition on discrimination, we have good reason to think that reaction qualifications do not confer merit in conflict cases where counting them is discriminatory (even if doing so does not violate transparency).

I raise conflict cases not because they are the only instances where appealing to the value of non-discrimination in the context of reaction qualifications is relevant but because they show why appealing to transparency alone is not enough to ensure procedural fairness. The upshot of conflict cases is that considering a reaction qualification is only procedurally fair if it preserves both transparency and non-discrimination. I will first outline how to assess whether a reaction qualification undermines the value of non-discrimination in standard cases before turning to conflict cases. I will then apply both forms of the non-discrimination analysis to our central cases, Looks and Race.

In light of procedural fairness’s prohibition on discrimination, selectors should not consider reaction qualifications when doing so would be discriminatory. When is it discriminatory to consider some reaction qualification? The answer, on my view, is that considering a reaction qualification renders a selection procedure discriminatory when it violates what I call community anti-discrimination norms.²⁹ In the sense I have in mind, community anti-discrimination norms are widely (though not necessarily universally) accepted norms against disadvantaging or wronging individuals on grounds of their membership in a social group, which are paradigmatically enshrined in anti-discrimination law. We should care about community anti-discrimination norms not (or not only) because they typically represent legal doctrines that govern particular selection procedures; rather, we should care about them, in the present context, because they express foundational commitments that are widely shared across a society. If consideration of a reaction qualification violates these norms, considering it in selection procedures is antimeritocratic insofar as selection by merit is incompatible with discriminatory hiring. It is procedurally unfair to consider such a reaction qualification; accordingly, selectors must disregard it.

Consideration of a reaction qualification can violate community anti-discrimination norms by incorporating facially neutral criteria into the selection procedure that act as proxies for protected grounds that selectors are otherwise prohibited from considering in virtue of community anti-discrimination norms.³⁰ A protected ground, for my purposes, is a characteristic picked out by community anti-discrimination norms as something that selectors are prohibited from (de)selecting applicants on the basis of. Reaction qualifications can function as proxies because they are themselves facially neutral: they concern the ability to elicit positive responses, a consideration that does not itself violate anti-discrimination norms. Nonetheless, reliance on them can, in some cases, de facto violate such norms.³¹

If a reaction qualification is rooted in discriminatory conceptions of what the job involves, whether an applicant possesses reaction qualifications is no longer an innocent consideration. If the reasons for an applicant’s failure to possess reaction qualifications derive from discriminatory attitudes, then considering applicants’ reaction qualifications will allow these discriminatory attitudes into the selection procedure. Under such conditions, the reaction qualification acts as a proxy, allowing selectors to de facto (de)select applicants on protected grounds that they are prohibited from considering in virtue of anti-discrimination norms. This brings to light a way that selection procedures can be unfair for reason of being discriminatory: selectors must not only avoid selecting directly based on protected grounds like race or gender; they must also pay mind to whether considering facially neutral criteria like reaction qualifications would amount to selection based on protected grounds.

This line of reasoning suggests a plausible claim about procedural fairness: if there is a non-accidental connection between the way a procedure is run and the unjust effects it produces, this can amount to a kind of procedural unfairness itself.³² When a hiring procedure non-accidentally and predictably violates anti-discrimination norms, it thereby violates procedural fairness. Though this claim appeals to outcomes, it does so differently than Moral Meritocracy. On the account I am arguing for, outcomes are relevant to the assessment of whether selectors can consider a reaction qualification not in virtue of its all-things-considered (im)morality but rather because non-accidental connections between procedures and outcomes can themselves constitute unfairness within the procedure. For a procedure to be fair, selectors must ensure that their method of evaluation will not produce certain outcomes not because they are independently unjust but because doing so undermines the fairness of the procedure. This way of appealing to outcomes is itself rooted in considerations of procedural fairness.

We thus have a preliminary case for the claim that selectors should not consider reaction qualifications when doing so would be discriminatory. But what are we to do in cases where setting aside discriminatory reaction qualifications comes at the cost of transparency? This can occur in cases where widely held societal conceptions of jobs are themselves discriminatory. As argued in Section II, considering reaction qualifications rooted in widely held societal conceptions of jobs is transparent since these conceptions inform how applicants expect to be evaluated. Nonetheless, if a societal conception is discriminatory, appealing to reaction qualifications rooted in it is procedurally unfair. Discriminatory hiring procedures are unfair (in addition to being independently unjust) because they deviate from selection on grounds of merit. Thus, if considering a reaction qualification would undermine procedural fairness by making it discriminatory, this gives us good reason to think it does not confer merit, just as violations of transparency give us good reason to think reaction qualifications do not confer merit. The trouble is that if the discriminatory societal conception is widespread and applicants (while realizing it is unjust) expect selectors to appeal to these discriminatory conceptions when deciding whom to hire, refraining from considering the reaction qualification rooted in these conceptions will make the procedure less transparent.

Be that as it may, transparency and non-discrimination are jointly necessary for procedural fairness. When consideration of some reaction qualification is either non-transparent or discriminatory (even if not both), considering it will be procedurally unfair. To illustrate, consider a case where the selection procedure is non-discriminatory though non-transparent: expecting applicants for a job in nursing to code in JavaScript. Though this mode of evaluation is non-discriminatory, it is also non-transparent and, therefore, unfair. Now, consider a case where the selection procedure is transparent though discriminatory: selectors appeal to a widely held though discriminatory conception of a job, and as a result, members of some social groups lack a real chance to compete for the job. If these conceptions are widespread, applicants may not be surprised that selectors appeal to them. Accordingly, the procedure may be transparent. Despite this, it will be discriminatory and, therefore, unfair. Thus, conflict cases show that if a reaction qualification undermines either transparency or non-discrimination, we have good reason to think it does not speak to merit and should not be considered.

We can now apply the non-discrimination analysis to our central cases, first looking at versions of the cases that engage a standard non-discrimination analysis, then modifying them so that they take the form of conflict cases and seeing how appeal to non-discrimination impacts the permissibility of considering the reaction qualifications in question.

We can first turn to Looks and ask whether considering reaction qualifications rooted in lookist preferences violates community anti-discrimination norms. I suggest it does not. We do not have widely accepted norms (such as legal doctrines) forbidding individuals from disadvantaging or wronging others based on their looks. Though lookism is often identified (plausibly, we might note) as a prejudice, we do not tend to treat (un)attractiveness as a protected ground. Therefore, when societal conceptions of the job musical entertainer hold that reaction qualifications rooted in conventional attractiveness form part of what a job involves, it may be permissible for selectors to consider them in a selection procedure without violating the non-discrimination element of procedural fairness.

This does not mean that the hiring procedure in Looks is fully just, all things considered; since the reaction qualifications in Looks are grounded in lookism, a selection procedure that takes them into account is, in a broad sense, criticizable. While such a hiring procedure would not be praiseworthy, it nonetheless would not fly in the face of community anti-discrimination norms so as to render the procedure antimeritocratic and, therefore, procedurally unfair. The selection procedure in Looks, therefore, does not violate procedural fairness for the reason of being discriminatory, though the outcome of such a hiring procedure may be worthy of criticism. My claim is that it is procedural fairness, rather than moral permissibility more broadly, that we should look to in determining whether a reaction qualification speaks to an applicant’s merit. Thus, appeals to values of transparency and non-discrimination in Looks both yield the same result: it is not a violation of procedural fairness to consider the reaction qualification.

However, suppose we modify Looks to involve a conflict between transparency and non-discrimination. Suppose the societal conception of the job of musical entertainer engages a racialized conception of conventional attractiveness, according to which whiteness is considered a central component of attractiveness, such that contenders will lack reaction qualifications if they are not white. Those wishing to secure record deals may be aware that racialized beauty norms inform society’s conception of the job of musical entertainer, so considering reaction qualifications rooted in these conceptions may not violate transparency. Nonetheless, if these reaction qualifications are allowed into the selection procedure, they will act as proxies, enabling selectors to de facto select for the protected ground of race, which is prohibited by community anti-discrimination norms. Considering these reaction qualifications would, therefore, be discriminatory, which gives us good reason to think they do not confer merit and that their consideration would be procedurally unfair.

We can next turn to Race. Recall that in the original version of the case, while the storeowner has business incentives to hire in accordance with localized prejudice, it is not the case that widely held societal conceptions of the job of salesperson involve race. Clearly, considering reaction qualifications rooted in customer prejudice violates community anti-discrimination norms. Our community anti-discrimination norms include a strong norm against discrimination on the basis of race. This means that the reaction qualifications in Race do not confer merit: anti-discrimination norms establish that one cannot fail to merit a job simply because they are a member of a race that others are prejudiced against. Reaction qualifications grounded in racial prejudice, therefore, cannot affect merit.

Again, in this version of Race, transparency and non-discrimination analyses point in the same direction: the reaction qualification does not confer merit and would be unfair to consider. However, in a conflict case version of Race, where societal conceptions of the job of salesperson do involve race and yet are out of step with community anti-discrimination norms that forbid race-based discrimination, although appealing to reaction qualifications might preserve transparency, it would nonetheless be unfair to consider them, because to consider them would be to discriminate based on race, just as it would be to consider reaction qualifications in the modified version of Looks, where conceptions of attractiveness were racialized. Therefore, selectors would be prohibited from considering reaction qualifications in both the original and the conflict case versions of Race.³³

In this section, I have argued that if considering a reaction qualification will render a selection procedure discriminatory, this gives us good reason to think that the reaction qualification does not confer merit. This is because discriminatory selection procedures are unfair (in addition, of course, to being unjust more broadly). Selection by merit, all else equal, promotes the fairness of selection procedures. This gives us good reason to think that reaction qualifications fail to confer merit when considering them undermines either transparency or non-discrimination.³⁴

IV. FAIRNESS-SENSITIVE MERITOCRACY

Selection by merit promotes procedural fairness in large part because it promotes transparency and non-discrimination. A selection procedure can be unfair by being non-transparent or by being discriminatory. Violations of either transparency or non-discrimination alone provide sufficient reason to disallow consideration of a reaction qualification since the fact that considering some reaction qualification undermines either of these values provides strong evidence that the reaction qualification does not confer merit. Though violation of either value is sufficient reason to disallow consideration of a reaction qualification, preservation of either value is not sufficient reason to permit consideration of it because of conflict cases. This is why the test for when to allow reaction qualifications must ensure that consideration of a reaction qualification violates neither transparency nor non-discrimination. On my view, then, we can determine whether a reaction qualification confers merit (and so whether it would be fair for selectors to consider it) by appealing to the following principle:

Fairness-Sensitive Meritocracy: If relying on a reaction qualification would violate either transparency or non-discrimination, the reaction qualification does not speak to merit. Accordingly, selectors must disregard it, for to consider it would violate procedural fairness by making the procedure unjustifiable to those who are party to it.

The principle of Fairness-Sensitive Meritocracy explains why relying on reaction qualifications in Race violates procedural fairness and why relying on reaction qualifications in Looks does not. It also explains why, if we modify the cases so that the societal conceptions of the relevant jobs are themselves discriminatory, thereby producing a tension between transparency and non-discrimination, it would violate procedural fairness to rely on reaction qualifications in either case.

Of course, if selectors were to rely on reaction qualifications in either the original or conflict case version of Race, doing so would be not only procedurally unfair but also independently unjust. To suggest that we rule out consideration of reaction qualifications for reasons of procedural fairness rather than independent considerations of justice is not to deny the latter’s importance. Rather, it is to say that when we should allow selectors to consider reaction qualifications is a procedural question, one that we can assess independently from considerations of all-things-considered justice. I suggest that developing such a framework is necessary to distinguish between the kinds of questions at issue when considering whether to permit reaction qualifications versus whether the outcome of a selection procedure is all-things-considered just. If this distinction is lost, our test for when to permit consideration of a reaction qualification will return results like those delivered by Moral Meritocracy in Looks, which implies that consideration of the reaction qualifications in Looks and Race should be disallowed for similar reasons.

It is worth considering a final objection to my solution to the puzzle of reaction qualifications. In particular, one might worry that my view would yield implausible verdicts in cases where no community anti-discrimination norms are in place. In such cases, my view might seem to imply that considering reaction qualifications grounded in discriminatory conceptions of jobs would be permissible. First, it is worth noting that in nearly any contemporary case involving reaction qualifications, some form of anti-discrimination norms will be in place in the form of constitutional guarantees of equal rights or anti-discrimination legislation. The enactment of such laws reflects the norms operative within a community, even if the community contains some intolerant citizens who reject these norms. To be operative, community norms to do not have to be universally accepted. Liberal and pluralist societies will be marked by disagreement; this does not mean that there are no norms or values that characterize such a society’s commitments.³⁵ These norms can block appeal to reaction qualifications even if societal conceptions of jobs are discriminatory, like in the modified conflict case versions of Looks and Race.

Societies with no anti-discrimination norms, such as wholly illiberal societies or societies in the distant past, are so fundamentally unlike contemporary liberal democracies that we should not expect a test for reaction qualifications or for procedural fairness more broadly to have traction. The test I propose here is not designed to evaluate selection procedures in distant history, where the equality of all persons was not recognized and no anti-discrimination norms were in place. This is not to say that my view endorses the hiring practices of these societies or implies that they are not deeply unjust.

Instead, these sorts of cases expose the limits of appealing to procedural fairness as a way to secure justice in hiring. As I have said throughout, procedural fairness does not guarantee perfectly just selection procedures. Nonetheless, it is a tool that we can appeal to in many nonideal contexts that enables us to identify why some selection procedures are not only unjust but also unmeritocratic and, therefore, unfair. I do not regard this as a shortcoming of the approach I develop here. Rather, I take it to reveal how we should think about solutions to the puzzle of reaction qualifications and the methods we use in thinking about justice in selection procedures: they respond to a particular set of conditions, allowing us to leverage criticism of hiring practices as procedurally unfair. They are tools rather than articulations of timeless moral truths. They can be relied on to characterize and guard against procedural unfairness, but we should not expect more from them than they can feasibly deliver. I take myself to have provided a case for why appeal to actual conceptions of what jobs involve is crucial to selection by merit and why solutions to the puzzle of reaction qualifications that ignore transparency render implausible verdicts in some important cases. As a test that appeals to actual, nonideal attitudes, we cannot expect the test for permissible reaction qualifications to produce results that more accurately carve morality at its joints than the data it appeals to. Still, the view I have developed guides the assessment of the permissibility of reliance on reaction qualifications in contemporary nonideal contexts while avoiding the twin dangers of over-idealization and total relativism.

V. CONCLUSION

Fairness-Sensitive Meritocracy provides a novel solution to the puzzle of reaction qualifications. It explains why relying on reaction qualifications is permissible in some cases, like Looks, but not in others, like Race. It also explains why relying on reaction qualifications in the conflict case versions of both cases would be impermissible. Fairness-Sensitive Meritocracy provides an account of the relation between transparency, non-discrimination, and merit as they figure in procedural fairness. Whether a reaction qualification speaks to merit is ultimately sensitive to whether reliance on it would undermine fairness understood in terms of transparency and non-discrimination. As a result, Fairness-Sensitive Meritocracy sheds further light on the nature of fairness in selection procedures, providing an analytical framework compatible with a range of substantive accounts of justice.

ACKNOWLEDGEMENTS

I would like to thank Steve Wall, Connie Rosati, Allen Buchanan, Guido Pincione, Travis McKenna, Travis Quigley, and Joe Metz for their helpful feedback and discussions. I also extend my appreciation to two reviewers for this journal whose insightful comments helped strengthen this manuscript. Special thanks go to Tom Christiano and Aaron Segal for their thoughtful written and oral feedback on successive drafts. Additionally, I am grateful to the organizers and participants of the Faculty Workshop series at the Georgetown Institute for the Study of Markets and Ethics, particularly Thomas Mulligan, for their generous and perceptive feedback.

COMPETING INTERESTS

The author declares that she has no competing interests.