This essay was the winner in the Undergraduate Category of the Oxford Uehiro Prize in Practical Ethics 2017
Written by University of Oxford student, Paul de Font-Reaulx
What makes discrimination wrong?
Most of us intuitively take discrimination based on gender or ethnicity to be impermissible because we have strong rights to be treated on the basis of merit and capacity rather than e.g. ethnicity or gender. I argue that, despite how this suggestion seems intuitive to most of us with a humanist perspective, it is indefensible. I show that well-informed discrimination can sometimes be permissible, and even morally required, meaning we cannot have absolute rights not to be discriminated against. In the last part I suggest an alternative account, arguing that acts of discrimination are wrong because they violate individuals’ weak right to be treated fairly and create negative externalities which – analogously to pollution – there is a collective responsibility to minimize. These results are counterintuitive, and require further attention.
1. What is Discrimination?
I take discrimination to be to treat someone very differently based on an irrelevant trait. A trait is relevant if and only if it by itself provides reasons for different treatment in some instance, such as constituting a difference in merit or capacity. Otherwise it is irrelevant. For example, limiting someone’s right to vote on the basis of age is not an instance of discrimination, because age is a relevant trait for participation in democracy. Doing the same on the basis of ethnicity is however, because ethnicity is not a relevant trait[1].
Discrimination based on bigotry such as racism is often indefensible simply because it rests on ungrounded beliefs about the relevance of traits such as ethnicity. Discrimination can be wrong even if it does not suffer from epistemic problems however. For this reason I consider what I call ‘epistemically grounded’ discrimination. By this I mean when different treatment is based on an irrelevant trait, but holding this trait makes it more likely that the same person holds another relevant trait. For example, ethnicity is not a relevant trait for performing a normal job. Criminality is however, and for various reasons (e.g. structural oppression and alienation) ethnicity can statistically correlate with criminality. In such cases different treatment based on ethnicity would constitute epistemically grounded discrimination.
2. Is Discrimination Impermissible?
I will now progress to consider what could make even epistemically grounded discrimination wrong. Here is a suggestion: We have the right to be judged based on individual merit and capacity rather than generalizations over traits we cannot control. Specifically, all individuals have an absolute right not to be judged or treated differently based on traits over which they have little control, unless those traits themselves constitute a relevant difference in capacity (e.g. the boxing case). Furthermore, it is impermissible to violate someone’s absolute rights. Discriminating based on e.g. ethnicity or gender is to violate someone’s rights in this way. For this reason it is impermissible. Let us call this the Strong Rights Account.
I believe the Strong Rights Account is highly intuitive, and corresponds well to a humanist conception of the ethics of discrimination which most of us share. This makes it worth studying in more detail. Let us test the suggestion by considering the following two examples:
Hiring: Harold is considering applicants for a position at his company. He knows that being a member of an ethnic minority strongly correlates with frequency of crime. Because of this he chooses not to consider applicants belonging to an ethnic minority.
Au Pair: Cassandra is considering hiring an au pair to take care of her children. She knows that being white and male strongly correlates with being inept at taking care of children. Because of this she chooses not to consider white male candidates.
These are both instances of epistemically grounded discrimination, and impermissible by the above suggestion. In their respective examples Harold and Cassandra both treat some group of individuals very differently based on an irrelevant trait over which they have little control. They do however have good reason to believe that these traits correlate with a relevant trait (criminality and child-caring respectively), and that therefore – ceteris paribus – someone with the irrelevant trait (e.g. being a white male) is more likely to possess the relevant trait (e.g. being inept with children) than someone lacking the irrelevant trait.
Most of us believe that Harold is acting impermissibly, and plausibly for the reasons above. He is treating individuals in a way which they have a right not to be treated by discriminating based on ethnicity. Many of us do not however find Cassandra’s action impermissible, and feel inclined to be more lenient in her case. In fact many ads for au pairs specify that they only consider females, and their owners are rarely considered to be violating absolute rights. This would suggest that our suggestion is too inclusive, making too many acts of discrimination impermissible. In order to still be able to consider Harold’s actions impermissible however, we need to find some further necessary condition which only rules out Hiring.
Here is a suggestion: Discrimination is only impermissible when the irrelevant trait (e.g. being a white male) constitutes the membership of an exposed group. By ‘exposed’ I mean a group which is generally considered to be at the receiving end of structural oppression, such as ethnic minorities or women. Cassandra’s discrimination is directed at white males, which are not an exposed group, while as Harold’s actions are directed at an ethnic minority, which is an exposed group. Therefore Harold’s discrimination is impermissible, but not Cassandra’s.
This condition captures a further intuition, but is insufficient. This is because sometimes it seems permissible, and even morally required, to discriminate even when the object of discrimination is a member of an exposed group. Consider the following example:
Alley: In order to get to the other side of town one needs to pass through either alley A or alley B. A tourist, who is internationally famous for spreading money around everywhere he goes, needs to get to the other side of town, and asks Pasha how to get there. Pasha knows only the following: (a) Passing through A is slightly quicker than B, (b) A is mainly populated with members of an ethnic minority and B by non-minorities and (c) being a member of the ethnic minority statistically correlates strongly with high propensity for violent crime, while there is no such correlation for non-minorities. The tourist will only understand a simple direction to A or B, and will then trust Pasha’s recommendation blindly.
In this case it seems to me that Pasha is not only permitted to direct the tourist to B despite A being the quicker path, but is also morally required to do so. The tourist is left at Pasha’s mercy, and directing him to A would likely put him at great risk. Doing so would however imply discriminating against the individuals in A which are members of an exposed group, because Pasha would treat them very differently by recommending the tourist to avoid them only on the basis of their ethnicity, thereby robbing them of the chance to receive some of his significant spending. If this is correct, it cannot be impermissible or even always morally wrong to discriminate on the basis of irrelevant traits which constitute membership of an exposed group. If correct this implies that the Strong Rights Account cannot be correct, and that discrimination based on e.g. ethnicity is sometimes permissible.
3. Fairness and Externalities
We might find it surprising that we cannot defend the suggestion that discrimination is impermissible. This does not mean that epistemically grounded discrimination is always permissible however; most often it is not. To determine when this is the case we need an account of what makes discrimination wrong in those cases when it is wrong. Below I sketch a short suggestion, arguing that the wrongness of discrimination comes from two independent and defeasible effects.
Here is my suggestion, which I elaborate on below:
Fairness: Discriminating against an individual is unfair to that individual, and individuals have a weak right to be treated fairly.
Negative Externalities: Discriminating against an individual produces negative externalities for individuals sharing the trait which has been the grounds for discrimination.
Fairness is being treated in proportion to one’s desert and capacity. When someone is not selected for a job because of their ethnicity, this is unfair to them, because they have ceteris paribus equal desert and capacity to perform the job to other applicants. For this reason both Harold and Cassandra are acting unfairly. To be treated fairly is however not an absolute right, and sometimes we are permitted to treat people unfairly when it is required for overriding reasons. An example of this is might be collective punishment of a platoon for the mistake of an individual, when it is required for discipline. When there are no such overriding considerations it is wrong to treat individuals unfairly. This effect varies only with desert and capacity, and not with e.g. whether the objects of discriminations are members of an exposed group.
When a group is discriminated against on the basis of being a member of that group, this can produce negative externalities for that group by reducing opportunities or solidifying alienation. For example, when Harold systematically disregards an exposed group, he is reducing its members’ chances for jobs relative to non-minorities, which constitutes a harm to members of that group. In other words, the harm of the act of discrimination can extend beyond the target of discrimination to others who share her traits. The harm constituting these negative externalities varies with the objects of discrimination, and is more significant for e.g. exposed groups than white males. This is why Harold’s actions are wrong while Cassandra’s are plausibly permissible, because discriminating against an ethnic minority produces more harm than discriminating against white males.
Why are Fairness and Negative Externalities the criteria that should be posited to capture the ethics of discrimination, rather than a more parsimonious suggestion? Firstly, there is a strong sense in which discrimination constitutes a violation of someone’s rights, and I believe unfairness is the most intuitive concept to capture this violation. Secondly the wrongness of discrimination seems to vary with its objects (e.g. it’s worse to discriminate against ethnic minorities than while males) and I find no other way to account for this variation but in the effects produced for members of the discriminated group. For these reasons a combination of (1) and (2) is the most parsimonious way to account for our intuitions regarding discrimination. There are however other implications:
Firstly, the nature of externalities is such that it affects large groups, and they are often produced collectively by many individuals (e.g. pollution). The responsibility not to produce such externalities is collective, meaning that the wrongness of discrimination is largely made up of a failure to live up to one’s part in the collective responsibility not to e.g. perpetuate alienating stereotypes. Specifically, any further wrongness to discriminating against an exposed group relative to white males is fully constituted by one’s marginal contribution to negative externalities for that group. This is counterintuitive, because discrimination is intuitively thought of wronging someone, not as – like driving an SUV – contributing to the production of negative effects for a group.
Secondly, like with pollution there can sometimes be overriding reasons to disregard this collective responsibility when the marginal harm produced is significantly outweighed by the harm of alternative actions. This is what happens in Alley, where there the risk to the tourist outweighs the harm caused to the individuals in alley A. Because both Fairness and Negative Externalities are then defeasible, this means that discrimination is permissible when counterweighed by independent considerations. For a counterintuitive example of this, it seems that Harold might be allowed to discriminate based on ethnicity if his business was extremely sensitive to criminality and the correlation between ethnicity and criminality sufficiently strong, which is highly counterintuitive.
I have argued that absolute rights are not sufficient to explain the wrongness of discrimination, and that it should instead be explained by unfairness and negative externalities. If correct then this means we should shift our view of the wrongness of discrimination as largely being a failure of a collective responsibility, and that we have to consider for each case whether there a sufficient independent considerations as to make discrimination permissible (e.g. how sensitive does Harold’s business need to be?). I believe this is counterintuitive, and indicates that the ethics of discrimination is more complicated than it might first appear and deserves further attention.
[1] We might use the word ’discrimination’ as simply referring to different treatment based on some trait. I will reserve the term for this ‘problematic’ sense, requiring selection based on an irrelevant trait. I also believe this definition corresponds closely to how most people use the word.
I think this is flawed from the start, by its very abstract definition of discrimination. That’s not the authors fault, since this definition as ‘selection on irrelevant traits’ is widely used, when considering the ethics of discrimination. However it is far removed from the reality of ‘discrimination’ and disadvantage, in fragmented and polarised western societies. Things have changed. We are not in the 1950’s anymore. It is no longer true that
Most academics perhaps, but in the real world discrimination is increasingly accepted as a normal reaction to changing societies. More and more people say openly that they discriminate, and see nothing wrong with that. In western Europe, we no longer live in societies where “ethnicity is not a relevant trait for performing a normal job”. On the contrary, ethnicity, religion, identity, tradition, and values have come to dominate politics, and to some extent broader social and economic interaction. In France the Front Nationale openly advocates a “préférence nationale” in employment, social benefits, and housing, and their views and support cannot be dismissed as marginal or irrelevant.
The model used in this essay is essentially that of the employer who has two equally qualified candidates for a job, and rejects one on the basis of a minor personal trait. This idea of ‘discrimination’ is closely related to the literal definition of ‘prejudice’ – snap judgements on an individual on the basis of visible traits, without any supporting evidence. Neither of these models correspond to the complex reality of polarised societies, with multiple structural disadvantages, and huge numbers of social transactions.
Obviously, if you start with model which is remote from reality, it is difficult to apply the conclusions to the real world. There are no direct policy recommendations in this essay, but the implication is that legal prohibition of discrimination has a weak foundation, and that suggests decriminalisation. I would certainly support decriminalisation, but on the much stronger grounds that the anti-discrimination legislation has been a total failure, offers no solution to the huge structural inequalities in western societies, nor a cure for xenophobia, nor an end to racism in its many current and historical forms. This essay does not use the language of racism and anti-racism, but in fact the legal prohibition of discrimination in the West is largely derived for international anti-racism treaties, in turn a post-war response to Nazi policy and practices. We can’t separate these issues (discrimination, inequality, xenophobia, sexism. racism and so on), and they are now unavoidable in the political arena. It is time for a general re-assessment of the ethics of discrimination, but this essay has little to offer.
The model of discrimination here can be categorised as liberal, and a common criticism of liberal ethics applies. It is often said that liberalism theorises from abstract ‘persons’ and their interactions, while ignoring the relevant characteristics of real human beings. And so too with this model of discrimination. It assumes a world of clones who have no identity, no culture, no tradition, no specific language, no religion, no ideology, no social class, no personality, and no personal history or background. It also assumes that each clone is entirely indifferent to the nature and actions of any other clone. They all have a function, which consists of simple execution of a task, and no more than a function. So they all get along, in the mythical conflict-free clone world, and prejudice and discrimination are inexplicable.
Real-world ‘societies’, which are also political communities, may claim shared values and a shared identity. In reality they are often bitterly divided, along multiple fault lines. This is where ‘discrimination’ comes from in the real world: it derives from enmity, hostility, and aversion between groups. These groups are divided by religion, nationality, ethnic origin, civilisation, identity, culture, gender, sexual orientation, class, ideology, language, and personality. What’s more, many of these groups consistently and systematically harm other groups, or they harm each other. There are many victims, but few are solely a victim.
Discrimination in work or housing, is simply an aspect of this polarisation and mutual hostility. It rarely takes the form of single-factor discrimination, as implied in the classic model. It is very unlikely that an employer will refuse a job applicant on the basis of ‘the colour of their skin’, for instance. In practice many factors play a role in selection: language, accent, social class, personality, education, experience, gender, life cycle stage, age, clothing, hairstyle, culture, sub-culture, and political preferences. Cumulative selection results in multiple cumulating social inequalities. There is never a simple function, for which all these factors are ‘irrelevant’. The employee has to work in an organisation alongside other people, who also have many different characteristics, inside a formal hierarchy with differing tasks and roles, and often with informal competition between various factions. A tenant never lives in an isolated house in the middle of the desert, but in proximity to neighbours who may have totally different values, religion, language, culture, and/or personality, in a local community which has a specific composition and historical background.
It is never as simple as the typical examples, used to consider the ethics of discrimination. For instance, the employer with two equally qualified candidates for a job, who selects one on some trivial irrelevant basis, such as hair colour. It that right or wrong? Politically and socially, it doesn’t matter. It is a mental puzzle for the ethicists, with no real-world application.
I agree that the ‘view from nowhere’ approach of contemporary ethics makes much of it irrelevant and possibly harmful.
Hi Paul! Thank you for your extensive comments. I will not be able to respond to them at the length they deserve, but wanted to give thoughts. As I understand your criticism it is directed against the abstractness of positing something like ‘discrimination’ under a general definition, and then trying to apply that to the world and prescribe actions, which requires some degree of historical understanding and consideration that such a methodology doesn’t allow. While I believe plausibly justified, I also think however that such criticism can be directed at most of the approach of practical ethics.
For example, you write:
I believe something similar could be argued to be the case for most moral concepts, and not just discrimination. Everything from ‘virtue’ (is Leonidas, Jesus or Bill Gates the embodiment of virtue?) to ‘impermissibility’ seem to be largely historically determined in how people view them, and perhaps it doesn’t even make sense to attempt to abstract away those circumstances to talk about some overarching form of the good. In the words of Sir Peter Strawson, they ‘neither call for, nor permit, an external ‘rational’ justification’.
While one might argue this, I believe it requires a deeper discussion of meta-ethics and about what ethics is supposed to do, in order to be settled. In this essay for example I have simply worked on the assumption that what we’re doing is exactly more or less trying to get at some form of the good (or sligtly less ‘queerly’ put maybe: find what we have external reasons to do). And so I commit myself to saying that concepts and moral claims both permit and call for external rational justification, and can be made true or false by external facts. Then I don’t think that positing some form of ideal person who doesn’t exist to be as strange.
Being a meta-ethical non-naturalist moral realist, this doesn’t bug me particularly. But I agree that if one were to hold another meta-ethical position, it’s not clear that the same kind of approach in practical ethics would make as much sense (though I’m not sure about this). On this assumption that we are seeking rational external justification for actions, I agree that ‘it is a mental puzzle for the ethicists’, but I’m not sure what kind of real-world application is demanded, and whether that is the business of moral philosophers. The suggestion seems similar to Bernard Williams’ brilliant comment on justification: ‘What does the professor’s justification do when they break down the door, smash his spectacles, take him away?’ The answer to both your comment and Williams’, I believe, is that it tells us and the pundits how they ought to act, nothing more.
I will add however that I agree with your comments that it’s not clear that we can talk about discrimination on the basis of a certain trait, as those are usually interwoven, and that is a complexity which needs to be pulled out if at all possible.
I disagree with the inconsistency with which the strong rights account is used in the essay. It’s ok, according to the author, to discriminate against some people on irrelevant traits, as long as they are not members of an exposed group. But this is an over-correction. If everyone follows this advice, then the group will become an exposed group, so you just create new injustices by applying the “exposed group” constraint. These injustices are at the group level, where you create new forms of discrimination on irrelevant, unchosen grounds, and at the individual level, where you treat people unfairly.
Also, the author treats as unproblematic the identification of exposed groups and non-exposed groups. Unemployed Bob from Toxteth might be surprised to find that it’s ok to discriminate him because Louis XIV and David Cameron enjoyed privilege, and they were also white men. (Then again, he actually won’t be surprised at all, which is why he votes as he does (but I’m sure modern academics have a plan to address that, too…))
My view is that discrimination on unchosen, irrelevant attributes is a genuinely dumb (and unjust) idea. If some form of diversity is relevant to the job, perhaps because you’re hiring a team and you’ve read Irving Janis, then it’s relevant and you should seek to hire a team which possesses the relevant attributes. And hiring team members because of how they fit into the team is reasonable, because there it’s the blend of talents, capabilities and perspectives that matter. Choosing an all-time Earth XI for football by individual football merit in isolation of other considerations would be a sub-optimal choice, since you’d have about 8 forwards, a couple of midfielders and Franz Beckenbauer. Choosing the Earth XI on the basis of their ability to form the best team would be a better strategy, even though this may result in the outcome that some excellent forwards (Cruijff and Puskas, say) would miss out, even though Cruijff and Puskas might be better pure footballers than the newly-promoted Phillip Lahm and Nilton Santos. In that case, there’s no injustice, because the selection is based on contributions to overall welfare. Discriminating on the basis of something *irrelevant* – like blue eyes or length of their index finger or whether or not they are now or have ever been affiliated with the Communist Party or FC Barcelona seems completely ridiculous. But you’d seem to admit these as possibilities as long as length of index finger or Barcelona affinities were not explicitly named “exposed groups”. I don’t like communists, Barca or stumpy fingers, either, but it seems poor form to discriminate on those grounds.
Hi Dave! Thank you for your comment. I actually believe you’ve misunderstood my claims, or I did not make them clearly enough.
Firstly, I did not claim that one is not allowed to discriminate against an exposed group. I set that up as a suggestion in the end of part two to judge whether it would be defensible, and then argue that it would not help, not considering any further problems that would arise from such a categorization.
Secondly, and relatedly, such a problem would – as you mention – be where exactly we would draw the line. When does somone qualify for membership in the exposed group? This would be a further problem if I took there to be a clear qualitative difference between it being permissible and not to discriminate against its members/non-members. Since I dismiss that suggestion however in the end of part two, I don’t believe it is a problem for me.
When I handle exposed groups in part three I end on the conclusion that there are cases where we would be allowed to discriminate against ‘exposed groups’ as well. The relevance of the distinction here is just shorthand for whether discrimination would cause large or small externalities for the group as a whole, where an exposed group is a group for whom it would cause large externalities. In this case we do not need any sharp cut-off, just approximations of how bad discrimination would be.
I’m not sure I followed your final comment, perhaps my response to Paul below might be a response to yours as well?
Let me give an example of how useless the ‘irrelevancy test’ is. I used to donate blood, but I stopped when I realised that my blood might save the life, or improve the health, of a right-wing person. I did ask formally, if it was possible to specify who gets my donated blood. The answer was no, and the tone of the reply was condescending, implying that donors nobly give blood for all humanity (not in fact the case).
It would be discriminatory in the simple literal sense, for me to donate blood to one person and refuse it to another, on the grounds of their politics. The Blood Bank will not however allow me to select recipients, and even if it did, my refusal would have no legal consequences, since blood donation is voluntary. I could not be prosecuted for discrimination in the legal sense. But is it discrimination in the ethical sense? Paul de Font-Reaulx defines discrimination as difference in treatment on the grounds of an irrelevant trait, which does seem to be the standard definition in ethics.
But what is ‘merit or capacity’ in the case of a blood transfusion, or an organ transplant? Capacity in the sense of ability to perform a task, is not applicable (the recipient is passive), and anyway every live human can receive a blood transfusion. What criteria would determine ‘merit’? If they were purely medical, then presumably only doctors would take the decisions, and donors would be unable to assess relevancy anyway. If the criteria are not medical, then what are they? Suppose an organ donor is allowed to choose who gets the donated organ, on the basis of ‘merit’. Which criteria determine the relative merit of potential recipients? If Paul de Font-Reaulx will not tell us what constitutes merit, then he should not insist that a decision be taken on that ground.
And that goes for all the others, who insist that selection must be on ‘relevant criteria’ What exactly are the relevant criteria? Who decides what the relevant criteria are? If we can’t find a consensus on relevancy, then we cannot even begin to discuss whether it is wrong to select on non-relevant criteria.
I had good reasons not to assist right-wing people. There are, by definition, no neutral and consequence-free right-wing positions. By definition, right-wingers seek a society in which certain others will be disadvantaged politically and socially. That includes me, again by definition, since I neither share right-wing goals, nor derive benefit from their implementation. In any case, right-wing people generally hate me, and some have threatened to kill me. One threatened to torture me to death. It is neither rational nor moral, for me to donate blood or organs to these people.
And that’s the general pattern for decisions which involve differential treatment. There are almost always conflicting interests, and a risk of harm to the person who is doing the selection. This is the social reality. So is at best detached, perhaps more accurately elitist, to suggest that there is some neutral set of ‘relevant criteria’ for all selective decisions. The person or organisation making the decision, is usually in the best position to judge what is relevant or not. In any case, we simply cannot expect consensus on what is relevant. Until we find a better definition of ‘discrimination’, then we cannot pronounce it right or wrong.
Here are some short comments on your comments on the matter of ‘irrelevance’:
You ask what the basis of ‘merit’ would be, or what constitutes it. This further seems to underlie the worry that we can’t determine irrelevance, being a function of merit and capacity (in my definition). Because I think that capacity is pretty clear, it seems we could alleviate this worry by making clear what merit is, if possible.
Personally I take the concept of merit to be pretty straight-forward, though I suspect your disagreement on this is derived from a suspicion of the kind of objective, timeless abstract conception of moral concepts which you mention in your comments above.
I take it that merit is close to desert, such that one can be treated differently if one has done something to deserve it. If for example I have killed someone, then the state is morally permitted to constrain my liberty (one might argue at least). And we don’t think that horrible people in general deserve help, as a further example. ‘What criteria determine the relevant criteria of merit[?]’. To me that just seems like a question of conceptual analysis which I think we can be spared as long as we have some intuitive notion of it as I suggest we do.
Your question might be more how we could determine who deserves to me treated differently unless we knew who deserved to be treated differently, and how do we know this if people judge merit differently? Well, I would just respond that disagreement doesn’t preclude the existence of an objective correct response. No-one ‘decides’ it, they just are, and with clear-headed thinking we can get a pretty good grasp of how it is.
In your question of right-wing people for example: We might argue that holding that political opinion constituted a difference in merit, which warrants differential treatment on your part. This doesn’t seem as implausible if we were imagining the same case for more extreme political opinions. Naturally one might plausibly doubt whether your average right-winger would in fact merit differential treatment, and furthermore we might question whether political views can ever justify differential treatment for medical considerations. That however, which I would agree with you on, does require an account of when a particular merit is relevant for a particular situation. I have not provided such an account, though I again don’t think that this constitutes any profound problem, and we might only need to appeal to intution to satisfy it.
I hope I’ve represented your views fairly, otherwise please let me know.
Hi Paul, Dave and Keith! Thank you for your comments. This is just to let you know that I will get back to respond to them as soon as I get the chance.
This is a complex issue, so I will split my comments. First, some historical background to the issues in the essay. English-language discussion of the ethics of discrimination, typically uses a model of discrimination which originates in the United States. It sees discrimination as a civil rights issue in a liberal society, and it is modelled on US experience in the 1950’s and 1960’s. There was a white majority with entrenched privilege, and a disadvantaged ‘Negro’ minority, so the resulting model is one-dimensional. It works best for formal binary selection procedures with one dimension of discrimination. An example is a job interview with only two candidates, one black and one white. No effects on third parties are considered, one person makes the decision, and the decision is binary (employ or not employ).
The same one-dimensional approach also influenced the American model of ‘racism’, which is still influential. We see its influence in terminology such as ‘people of colour’, which categorises every human as either white or not white. The American model proved inadequate when applied to nationalism and xenophobia in Europe, and especially in dealing with hostility to Muslim migrant minorities.
Anti-discrimination law in western states was also influenced by this American model. Much of this legislation is derived from a single source: the International Convention on the Elimination of All Forms of Racial Discrimination (ICERD). Article 2 commits signatories to “… pursue by all appropriate means … a policy of eliminating racial discrimination in all its forms”. It specifically requires them to legally prohibit racial discrimination by individuals and organisations.
Although this model does not explicitly predict future developments, it does favour the assumption that discrimination is controllable. Note that the ICERD is literally intended to eliminate ‘racial discrimination’ entirely. That is a grand ambition, but given the mood of the time (1966) it was probably considered feasible. We now know, or at least we ought to know by now, that discrimination is an intrinsic aspect of human society. It is here to stay. Equally we should recognise that ICERD’s later companion, the Convention on Elimination of All Forms of Discrimination against Women (1979), has failed for similar reasons.
For any serious consideration of the ethics of discrimination, we need to move beyond the simplistic models. The real world with seven billion people, is not a job interview with two candidates. Human life is more than a string of one-off binary choices. It is best to use real-world examples to look at the flaws in the ‘official’ model of discrimination.
Paul de Font-Reaulx says that there is an “objective correct response” to the question of who has merit, and that we can have an “intuitive notion” of what criteria determine merit. In the real world, however, merit or desert are just some of many factors in a decision. The question of whether they should be given priority is political, not ethical. Typically those who argue for selection by merit are opposing some form of positive discrimination, often a quota system. In current western politics, it is typically the right which argues for selection by ‘merit’, although meritocracy was plausibly radical and ‘left-wing’ 200 years ago. Typically also, the right use ‘merit’ to mean class-dependent factors such as educational achievement.
Let me give some more real-world examples, one by one, to illustrate the problems with ‘relevant criteria’, ‘merit’, ‘desert’, and other allegedly neutral norms. Some are contentious examples, and people do get angry about these issues, so there is a tendency to self-censorship. It is better to be open about them, however, even when this offends people.
Duindorp is an isolated neighbourhood in The Hague, inhabited by what would be called the ‘white working class’ in Britain. Unlike other low-income areas it has very few immigrants. It survived as a ‘white neighbourhood’, because its residents systematically intimidated migrant and minority households who moved into the area, usually in social housing. In the end, the housing associations gave in, and stopped allocating housing to non-Dutch households. In 2014, the media reported on this secret policy, resulting in predictable public condemnation from the authorities (who knew what was happening anyway). There are several similar neighbourhoods and villages in the Netherlands. They are easy to identify, because they stand out on election maps, as strongholds of Geert Wilders’ Freedom Party (PVV). The party got 45% of the votes in Duindorp at the last election, its highest score in an urban area.
The standard academic approach to discrimination, which is represented in the essay by Paul de Font-Reaulx, would condemn the ‘racist allocation of housing’. The academics will say: social housing is a scarce good, which should be fairly allocated. Each vacant house should be allocated by its owners (the housing associations) on the basis of ‘merit’ or ‘desert’ or ‘relevant criteria’. The ‘merit’ and ‘desert’ are qualities of the households applying for social housing, so the housing association must make a decision based solely on data about the applicant. The views of the neighbours don’t count. The academics would then say, that ethnicity is not a relevant criteria for allocating housing. (‘Capacity’ does not apply here, since anyone can live in a house). So they would conclude, that the housing association should allocate houses on the basis of objective relevant criteria, such as income, number of children, and possibly time on a waiting list. And if anyone interferes with the allocation, that’s a matter for the police, and when they sort it out, all will be well in Duindorp. The anti-discrimination law will make the discrimination disappear.
I used to think like that too, but I now recognise the political and social realities. The white residents of Duindorp are angry, and they would be even more angry if they did not see Geert Wilders as their imminent saviour. They typify the frustrated, bitter, and alienated white working class, which feels betrayed by the political and social elite. They are the core of the populist electorate, the Trump / Brexit / Le Pen voters. Their anger is here to stay as a political factor, until something really fundamental is done about it (way outside the boundary of mainstream politics).
What would happen if the authorities in The Hague took the advice of the academics seriously? They would allocate the social housing to the most urgent categories, for instance the recent wave of asylum seekers from Eritrea. But what would happen to a white working-class Dutch neighbourhood, if suddenly half its residents are Eritrean? It would disappear – gone, totally destroyed, ceased to exist. We know that, because it did in fact happen to hundreds of white working class communities all over western Europe. They were wiped out by mass immigration. That’s not to say they faced no other threats, but immigration is certainly a threat to such communities, which depend on ethnic and class cohesion.
This is why the white ethnic-Dutch residents of Duindorp don’t see themselves as the agressor. They see themselves as victims, and they will tell that to any journalist who asks. They feel threatened by the foreigners, the muslims, the mosques, the asylum seekers, the European Union which is filling their country up with immigrants and destroying their jobs, and the elite which never, never, listens to them. They see their neighbourhood as a last refuge, and they are well aware that it is an exception. So for them, if they suddenly find a Eritrean family living next door, that is a direct assault on everything they value. It is in their eyes an act of hostility by both ‘the elite’ (the housing association) and the Eritreans. It is hard to fault that logic, because they will in fact be harmed. They will suffer distress and fear, they will lose sleep, and if the same thing happens in the rest of the neighbourhood, their community is gone too. If they can’t take it, and they themselves ask to be rehoused, they will probably be moved out of the city entirely, possibly losing their job in the process.
Thats where the aggression comes from. If the authorities did in fact try to move several thousand Eritreans into Duindorp overnight, they would need to use the army. I think that is a realistic assessment. It would be different if was a gradual processes, but there have been historical instances of rioting as migrants moved into low-income ‘white neighbourhoods’.
Now, the standard academic approach to discrimination takes no account of all this. These complex historical and social factors are simply ignored. The case of Duindorp would be reduced to a simple bureaucratic procedure at the housing association. House vacant for family with two children? Open files of applicant families, filter for two children. Applicants qualify for social housing? Adults have income or benefit? Filter for any priority indication such as disabled child. Sort by date of application, send letter to first on list. If accepted, issue tenancy. End of matter. It is typical liberalism: reliance on blind formal procedure for ‘fairness’, while ignoring the complex political and social realities.
There is no easy ‘solution’ to such problems, but the state can certainly do something. However, the existing anti-discrimination laws are an obstacle for innovative policy responses. So are the international equality treaties, such as ICERD and CEDAW, from which those laws are derived. So is the whole culture of anti-racism and anti-discrimination, including the academic theorising on discrimination and justice. The whole idea of ‘equal treatment’ is outdated.
Instead of a futile quest for the ‘elimination’ of racism and discrimination, we should think about how to accept inter-group hostility, expressed in values and preferences. There is plainly a large xenophobic-populist minority in Europe, who will never feel comfortable in a multi-ethnic society. As individuals, they discriminate continuously. Even if it was feasible, there is no point in arresting them every week and putting them in jail repeatedly. It won’t make them any less xenophobic. And if the anti-discrimination laws are not enforced, what’s the point anyway?
I don’t want to single out Paul de Font-Reaulx and his essay here: it is simply an example of the failed mainstream thinking. Formally, my criticism here is that the alleged wrongness of discrimination is never established in a social and political context, but only in terms of isolated individual cases. In fact, most authors simply assume that discrimination is wrong, and proceed from there.
Let me give another example to illustrate the complexity of the issues, and to show why ‘capacity’ or ‘merit’ or ‘desert’ simply can’t help us, to assess the ethics of discrimination. There are no self-evidently relevant criteria, because all criteria can be relevant, or irrelevant. The choice of what is relevant might be political, social, cultural, economic, emotional, but in any case never self-evident, and rarely universally accepted.
In 2009, at the height of the crisis, there was a series of nationalist strikes in Britain, against the employment of foreign workers at the Lindsey Oil Refinery construction site. Just before the crisis, in 2007, Prime Minister Gordon Brown had promised a policy of préférence nationale in employment, similar to Front National policy in France, with the slogan British Jobs for British Workers. Like other mainstream politicians, he sought support from xenophobic populist voters, and like so many politicians, he failed to deliver on his promise. That became an acute political issue, when the labour market collapsed at the start of the crisis.
The motives of the strikers were xenophobic and nationalist. They were ‘racist’ in terms of the everyday usage of the word: not necessarily inspired by a theory of racial superiority, but simply anti-foreigner. The Lindsey strikers demanded that their employers, contractors building a new refinery, should dismiss Italian and Portugese workers at the site, and replace them by British workers. Strikers at other construction sites supported them, and more generally demanded that jobs in the UK be reserved for British workers. The trade unions, and the British left, supported the British strikers.
The contractors had brought in Italian and Portugese workers, under the European Union’s Posted Workers Directive 1996, which allows contractors to undercut local wages. It is a classic example of everything that frustrates the average populist, and in retrospect, the Lindsey strikes were a sign of the coming Brexit.
How does academic theory of discrimination approach this case? It simplifies the whole complex issue to a series of individual job interviews. The boss is hiring welders, and he has an application form on his desk. It is skilled work, so they must have the appropriate training. Beyond that, he would look for experience, which he can see on the applicants CV. And that’s it, more or less.
But that’s not how the company sees it. The contractor is trying to minimise labour costs, and if they don’t, they will go bankrupt, in this highly competitive sector. British workers cost too much, so they would prefer too avoid them entirely. If the EU would let them, the contractors would employ Vietnamese, much cheaper – or indeed North Korean prison labour. In practice, they used the cheapest realistic alternative, sub-contracting under the Posted Workers Directive to European Union firms.
The academic theory assumes that there is only one transaction: an employee is hired, and then goes to work, unproblematically. But that takes no account of the real world: workers have colleagues. The British workers at Lindsey were well aware that the contractors would replace them, if cheaper qualified workers were available in the EU. They had a direct interest in a British Jobs for British Workers policy. But that’s not all. They also knew that immigration has consequences for Britain as a whole. If every British worker was replaced by Italians and Portugese, then Britain would cease to exist in its current form. Its indigenous population would be reduced to poverty. If the Italians and Portugese bring their families, then they would form the majority of the population, and Italian and Portugese would displace English.
Immigration destroys the nation: all nationalists understand this. Logically, therefore, there was hostility to the foreign workers at Lindsey. The foreign workers harmed their British colleagues, by the mere fact of being there as foreign workers. The academic model of discrimination takes no account of such harm.
So in the end, the British workers at Lindsey demanded, quite logically, that their employer discriminate against foreigners. The employers were under huge pressure, not just from the strikers, but also from the media and public opinion, which took the side of the strikers. The academic model of discrimination ignores such factors. It takes no account of third parties at all. It does not even take account of the interests of the second party. Rejected job applicants apparently disappear into thin air, if they don’t get the job because the other candidate had more ‘merit’. Everything is reduced to one selection, made by one individual, applying to one individual, with no external factors, and no consequences post-selection.
In the case of the Lindsey strike, no consensus was possible on what constitutes ‘capacity’ or ‘merit’ or ‘desert’. There were no self-evident ‘relevant criteria’ for hiring employees. Everything was highly politicised and contentious. That is the real world. If ethics want to contribute something in such cases, then the abstract models used must be re-translatable to the real world. That’s not the case with the mainstream theory of discrimination: it has inherent flaws which prevent its application in real-world cases.
Here’s a final example of the complexities of discrimination. Recently Jewish student Izzy Lenga was elected as Vice President for Welfare of the British National Union of Students. The NUS is more established and more institutionalised than student organisations in other countries, and her election has implications for policy at British universities. Although her election was nominally about organising student facilities and provisions, Izzy Lenga is known primarily as a campaigner against antisemitism. She wants British universities to be safe spaces for Jewish students, and apparently she has the backing of the current NUS leadership. Her election is part of a wider controversy in Britain about antisemitism, especially inside the Labour Party.
In Britain, the Conservative Party, most of the Labour Party, the police, university administrations, and the media, treat antisemitism within a framework of racism and discrimination – the traditional framework which I criticised in my earlier comments. From that perspective, Jewish students such as Izzy Lenga want universities to suppress antisemitism among students and staff, which from their point of view is an obstacle to the completion of their studies. They see themselves as victims of discrimination.
However, in practice, a policy to suppress antisemitism can only be implemented by repression against antisemites, which means in a British university context primarily muslim and left-wing students. If the universities seriously implement a zero-tolerance policy for antisemitism, then these students will be expelled, or possibly refused admission in the first place.
This is a specific example of a general principle, that when a resource is allocated by selection, at least one party must be disadvantaged. If there is only one candidate for a job, then there can be no discrimination. If however there are more then one, which is usually the case in the real world, then both a discriminatory and an anti-discriminatory policy will result in objective disadvantage for some. Merit, capacity, desert, relevance, all the pseudo-neutral criteria which liberals propose to guide selection, can’t tell us anything about whether this advantage is unjust, because they simply ignore the disadvantaged parties.
If we apply the standard approach, which Paul de Font-Reaulx typifies, then we could compare for instance a Jewish student with a non-Jewish white male student, at the same British university. The Jewish student is confronted with antisemitism from muslim and left-wing students, and is too distressed to study effectively. The non-Jewish white male student is typically not confronted with antisemitic hostility, or indeed any racism or sexism, and can continue his studies undisturbed. He has been treated differently, given an advantage for which there is no justification, and that is classic discrimination.
So the classic approach says that the university must rectify the discrimination, and allow the Jewish student to study, in the same unstressed fashion as the white male student. However, in the real world of British universities, that might mean expelling hundreds of muslim and left-wing students for their antisemitism. How would they feel about the justice of that policy?
Once again, the mainstream ethics approach to discrimination is simply useless as a guiding principle in the real world, before it is far too remote from that real world to be translatable. There are no neat categories of relevant criteria, merit, desert, or capacity, in the messy real world. Discussion about them won’t produce any guideline or procedure, which can be applied to real-world discrimination issues.
In fact there is a simple solution to the issues faced by Jewish students in Britain, and in several other western countries: a separate Jewish university. It’s not immoral, it has historical precedents (religious universities), and it should not require extra finance since student numbers are the same. At the Jewish university (or universities), Jewish students would face no antisemitism, and enjoy a supportive cultural climate. There are other cases of discrimination where this principle can be applied – although not always, because resources are finite. Of course this will be abhorrent to liberals, but my point here is that the classic anti-discrimination principles can not, in themselves, show why such options are wrong.
Comments are closed.