What is Discrimination?1
Relatively little philosophical work has been done on the nature of discrimination.2 But
discrimination raises a number of philosophically interesting issues. When we identify an action
or a rule as discriminatory, we are claiming not just that it draws a distinction between people or
results in some being treated differently from others, but also that this differentiation is in certain
respects unfair. What exactly are these respects: in what does the unfairness consist? And is it
ever justifiable for the state to impose legal sanctions on discriminatory actions, even though this
would appear to interfere substantially with citizens’ freedom to choose whom they contract with
and what sorts of reasons they act upon when they do so?
One reason why these questions have not been pursued in detail in philosophical
discussions may be that it is natural to treat intentional discrimination as the paradigmatic case of
discrimination, and there is a common view of intentional discrimination that provides answers
to the above questions that seem relatively straightforward and so philosophically uninteresting.3
Intentional discrimination –for instance, refusing someone entry to your bar because he is black
or denying someone a job because she is a woman– occurs when actions or rules are adopted
with the purpose of excluding some people because they possess a certain trait. Of course, this
general description also fits many innocuous acts of exclusion, such as denying a job to those
who lack the skills required by the job. According to a common view of intentional
discrimination, what distinguishes these innocuous exclusions from those that constitute
intentional discrimination is the motive for which the exclusion is undertaken: either the agent’s
sole motive is to cause harm to those excluded, or, although the agent is acting in furtherance of
some other purpose, he believes it acceptable to exclude these people in order to further this
purpose because he sees them as inherently inferior beings. So intentional discrimination, on this
common view, involves deliberately excluding someone on the basis of a certain trait, either with
the sole aim of harming her (we can call this “malice”) or in the belief that it is acceptable to
harm her in the service of some other goal because her possession of this trait renders her an
inferior person (we can call this “prejudice”). It is not particularly difficult to explain why such
actions are morally abhorrent. Most moral views that treat the content of a person’s motives as
relevant to the moral status of her actions imply that it is morally wrong –or in some other way
amounts to a moral failing– to act out of malice, or solely for the sake of causing harm to others.
And most also imply that it is similarly objectionable to act from the belief that others are
inherently inferior to oneself, that is, to act from what I have called prejudice. Moreover, it is
natural to suppose that these morally objectionable motives are also the source of our
justification for legally prohibiting discrimination. Most countries that prohibit discrimination
by one citizen against another do so only in certain special contexts, contexts where the
discriminator has chosen to offer certain goods to the public such as employment,
accommodation, or goods and services. It seems plausible to suggest that when we choose to
take on such roles in the public domain, then we must, like the state, treat everyone as though
they were of equal worth. And it can seem obvious that, whatever else this means, it must mean
at least that we cannot act with the sole purpose of causing harm to some of them, or act on the
belief that some are inherently inferior to others.
One of the subsidiary aims of this paper is to question this understanding of intentional
discrimination. I shall suggest that not all cases of intentional discrimination involve a belief in
the inferiority of others or a desire to cause harm to them. More importantly, I shall argue that
even in cases that do involve such objectionable mental states, the mental states themselves do
not play any role in justifying the legal prohibition of such discrimination; though, to the extent
that they cause greater suffering for the victim, they may justify requiring the agent to pay
additional compensatory damages to her.
The main aim of my paper, however, is to analyze a different form of discrimination.
This is what is known as “adverse effect” or “disparate impact” discrimination.4 This form of
discrimination can occur in the absence of a belief about the inferiority of those who are
excluded or a desire to harm them, and indeed, even in the absence of any exclusionary purpose
at all on the part of the discriminator. If an action or rule has the effect of excluding certain
people because of their possession of a certain trait and this trait is a recognised “ground” of
discrimination, then that action or rule may amount to adverse effect discrimination.5 For
instance, prior to the mid-1990’s, the Royal Canadian Mounted Police had a policy of requiring
the traditional “Stetson” hat to be worn on all ceremonial occasions. This meant that Sikh
officers were unable both to continue practicing their religion and to attend RCMP ceremonies.
It was not the purpose of the policy to exclude Sikh officers. But the policy had the effect of
excluding them from ceremonies because of their religion, in circumstances where making an
exemption for religious headwear would not have caused the RCMP any hardship. The policy
therefore came to be recognised as adverse effect discrimination.
So understood, adverse effect discrimination can seem puzzling in at least two respects.
First, if what makes adverse effect discrimination objectionable is not the motive or purpose with
which the act or rule is adopted but its undesirable effects, then it is hard to see how this form of
discrimination could involve any wrongdoing on the part of the discriminator. And, as I shall
explain in more detail below, we normally assume that such discrimination does involve
wrongdoing –and moreover, that it is the kind of wrongdoing that generates a personal claim on
behalf of the victim, as against the wrongdoer. Indeed, this assumption underlies the legal
institutions that most countries have established to deal with adverse effect discrimination in the
private sector. We allow victims to bring personal complaints against discriminators for redress,
and we treat these complaints not just as instruments through which the broader public good can
be promoted, but also, crucially, as attempts to rectify a personal wrong that the victim has
suffered. Second, and relatedly, if adverse effect discrimination is objectionable only because of
its unfortunate effects, then it has very little in common with intentional discrimination –or at
least with the common view of intentional discrimination that I sketched above, according to
which it is objectionable because of the agent’s motives. This disconnect seems odd, given that
we speak of these two forms of discrimination as though they were two forms of the same
phenomenon.
I shall try in what follows both to describe these two puzzling aspects of adverse effect
discrimination in more detail and to sketch out a solution to them in the form of my own theory
of adverse effect discrimination. I shall argue that, properly understood, adverse effect
discrimination does involve wrongdoing. But the wrongdoing has nothing to do with the agent’s
motives or purposes. Rather, it consists in interfering with another person’s rightful claim to
equal freedom. And I shall suggest that it is the same phenomenon as intentional discrimination
because that form of discrimination, likewise, is wrong not because of the agent’s motives or
purposes but because it amounts to an interference with the victim’s right to equal freedom.
I.
I stated above that if there is an action or rule that has the effect of excluding certain
people because of their possession of a certain trait, and if this trait is a recognised “ground” of
discrimination, then that action or rule may amount to adverse effect discrimination. I said
“may” rather than “must” because there is a qualification here, provided by a defence that all
countries make available to alleged discriminators in cases of discrimination. Different countries
have chosen to express this defence in slightly different language. For instance, most antidiscrimination
laws in the United States speak of “business necessity”, whereas most laws in
Canada speak more generally of “undue hardship”. But the broad purpose of the defence is the
same. I shall adopt the Canadian terminology of “undue hardship”, as it points most clearly to
the function of the defence. The defence of undue hardship allows an alleged discriminator to
claim that even if there was some reasonable way of accommodating those whom he has
excluded, such accommodation would have caused him “undue hardship” –that is, placed an
unacceptably large burden on him. If this claim succeeds, then according to the law, no
discrimination has actually taken place. Hence, the effect of the defence of undue hardship is to
narrow the cases in which discrimination occurs to those in which there is some reasonable
means of accommodating the excluded persons and the accommodation falls short of imposing
“undue hardship” on the agent. For this reason, it is important to think of the defence as an
integral part of our understanding of what discrimination is, rather than as an afterthought to it.
In order to understand the full structure of a claim of adverse effect discrimination and to
see more clearly how the defence of undue hardship works, it may help to consider three variants
on the following case.6 Suppose a blind customer with a guide dog tries to have lunch at a local
cafe. She sits down and waits for service. But the service never comes: no waiter approaches
her. On the first variant of this case, let us suppose this is because they simply do not see her:
perhaps there is a large tree in front of her table, whose gorgeous foliage obscures her. In this
case, she cannot complain of adverse effect discrimination because the fact that she was not
served did not result from any action or policy on the part of the waiting staff. So although they
may be faulted on other grounds –for instance, for having failed to be properly attentive to
potential clientele– they cannot be said to have performed an action that resulted in her
exclusion, much less her exclusion on the basis of a prohibited ground. On the second variant of
this case, let us suppose that they do see her. But the restaurant has a firm policy that no dog is
to be allowed on the premises because dogs lower the tone of the establishment. So they ignore
her, knowing that sooner or later she will take her dog elsewhere. This second variant does,
likely, constitute adverse effect discrimination: it seems unreasonable of the restaurant to refuse
to serve this client when doing so would not materially harm the premises or substantially
interfere with the running of the restaurant. It is true that, in this second variant, the waiting staff
have not deliberately set out to exclude the blind client: it is her dog that they want out of the
restaurant, and if she were willing to leave him by the entrance, they would happily serve her.
But because the guide dog’s constant presence is required by her, given her blindness, their
action has the effect of excluding her on the basis of a prohibited ground of discrimination.
Finally, consider a third variant on this case. Suppose that the restaurant staff have a religious
objection to approaching the dog: on their interpretation of their religion, a brush against a dog
would render them ritually impure and hence in need of cleansing themselves seven times.
Suppose it is a small restaurant –perhaps with only one waiter— so that for the waiter to have to
cleanse himself frequently as a result of encounters with a guide dog would pose substantial
difficulties for the operation of the restaurant. In this third variant on the case, we might think
that accommodating this client really would constitute an undue hardship for the restaurant. And
so we might conclude that this third variant is not, after all, a case of adverse effect
discrimination: there is, in these circumstances, no reasonable accommodation possible that falls
short of undue hardship because the only ways of accommodating this client are ones that
impose an unacceptably large burden on the restaurant.
Adverse effect discrimination, then, is any act or rule that has the effect of excluding
someone on the basis of a trait that is a prohibited ground of discrimination, in circumstances
where there is some reasonable way of accommodating this person that does not impose undue
hardship on the agent. So understood, adverse effect discrimination can seem philosophically
puzzling in at least two respects, respects that might lead us to question whether we can ever
justifiably use state coercion against someone for engaging in this form of discrimination.
The first respect concerns the apparent absence of any sort of wrongdoing on the part of
the agent.7 If adverse effect discrimination is just a matter of the unfortunate side-effects that
some actions and rules have on particular individuals because of characteristics such as race,
gender, or religion, then it is unclear how we could ever justifiably treat it as due to wrongdoing
on the part of the person who performed the action or adopted the rule. What did they do
wrong? There is, of course, an obvious but unhelpful sense in which any such agent acts
wrongly: when he violates legal prohibitions on adverse effect discrimination, he contravenes
positive law and in this sense commits a wrongdoing. But this is not the sense of “wrongdoing”
that I am concerned with. I am concerned with whether adverse effect discrimination involves
wrongdoing in some non-tautological sense –some sense that would justify the legal prohibition
on it. And it can seem as though it does not. Was it, for instance, really the fault of the RCMP
that Sikh officers were excluded from RCMP ceremonies? Was it, likewise, due to some
wrongdoing on the part of the waiters that, in variant 2 of my restaurant case, the blind client was
not served? All that the waiters did was notice a dog and decide that they did not want this dog
in their restaurant. It is true that it was this decision of theirs that caused the blind client to be
excluded in a way that others were not excluded, just as it was the RCMP’s policy that caused
the officers to face a burden that other officers did not face. But most actions and policies
impose an unequal burden on someone or other. Why should we treat these cases –cases in
which people are excluded because of some trait that we treat as a prohibited ground, even
though there are reasonable accommodations available—as ones where the burden amounts not
just to an unfortunate side-effect of the discriminator’s action but to wrongdoing on his part?
If we cannot explain why adverse effect discrimination involves wrongdoing on the part
of the discriminator, then we face at least two problems. Firstly, it becomes much harder to
justify the claim that the agent of adverse effect discrimination can be legally coerced into
shouldering the cost of whatever measures are required to accommodate those whom his acts
have excluded. And most regimes of private sector anti-discrimination law assume that this is
justifiable: when a case of discrimination is identified, it is the agent who is required to cover
the full costs of accommodating the victim. If the exclusion of the victim is not due to his
wrongdoing, however, then why should the costs be transferred onto his shoulders alone? Why
not instead have the costs covered by the public at large through a general taxation scheme, so
that the burden is spread across all of us? Secondly, if adverse effect discrimination does not
involve wrongdoing, then there is no sense in which the victim of adverse effect discrimination
can legitimately complain that she has personally been wronged by the agent of it. And this runs
counter to a further important fact about the way in which we conceptualize private sector
discrimination, and in particular the way in which we conceptualise the kind of wrongdoing that
it involves. Up until now, I have simply been speaking of “wrongdoing” on the part of the
discriminator. But we do not treat adverse effect discrimination as an instance of what Cardozo
might have called “wrongdoing in the air”.8 Rather, we think that it is the kind of wrongdoing
that generates a personal claim on the part of the victim. We believe that victims of
discrimination have a specific grievance against the discriminator: they have not just suffered a
loss as a result of his misconduct, but have rather been treated by him in a way that amounts to a
personal wrong against them. And this belief, too, is central to private sector discrimination law.
Most systems of private sector discrimination law use a “complaint model”: that is, they rely on
individual complainants to instigate legal proceedings against alleged discriminators and the
process is conducted throughout as an investigation into whether the complainant has herself
been wronged by the discriminator. Although carriage of the complainant’s case is sometimes
given to a legal body that is acting in the wider public interest, such as a human rights
commission, it is still the case that this body’s primary role in the proceedings is to bring forward
and resolve what is viewed as a personal dispute between the complainant and the alleged
discriminator. Moreover, many jurisdictions allow for special “dignitary” damages to be paid to
the victims of discrimination, over and above the costs of redressing whatever inconveniences or
material losses the victims have suffered. These special damages seem simply to be a symbolic
acknowledgment of the fact that these individuals have been wronged.
Hence, one puzzle raised by adverse effect discrimination is how it could be seen as
involving wrongdoing, in such a way as both to support our belief that those who discriminate
can legitimately be coerced into shouldering the cost of accommodating those whom they have
excluded, and to support our intuition that the kind of wrongdoing at issue involves a personal
wrong against the victims of discrimination. But there is also a second puzzling feature of this
form of discrimination. This is that it seems to have very little in common with intentional
discrimination. If intentional discrimination is legally objectionable because of the motives for
which it is done, whereas adverse effect discrimination is legally objectionable because of its
effects, then the legal prohibitions on the two sorts of discrimination have two completely
different sources. This seems odd, given that we speak of these two forms of discrimination as if
they were just that: two forms of the same phenomenon, which are objectionable for at least
some of the same reasons.
There are several responses to these puzzles that seem initially attractive, but that I think
we have reason to reject.9 One is to deny that there really is any objectionable form of
discrimination other than intentional discrimination. Adverse effect discrimination, insofar as it
merits the name “discrimination”, is just what New Zealand’s early Human Rights Commission
Act referred to as “discrimination by subterfuge”: that is, it is a malicious or prejudicial motive
hidden behind a facially neutral act or rule.10 So adverse effect discrimination really does
involve wrongdoing and really is the same phenomenon as intentional discrimination.
This response seems false to our experience of cases of adverse effect discrimination.
Most of them would be mischaracterized, and the genuinely difficult issues that they raise,
buried, if they were seen as cases of concealed malice or prejudice. To return to my earlier
example of the RCMP’s ceremonial uniform, it seems highly implausible to suggest that the
committee responsible for setting this uniform policy was really hostile to members of religions
that require certain headwear or assumed that such people were inferior and could legitimately be
excluded from ceremonies. The Stetson rule is old enough that, likely as not, there simply were
no Sikh officers (or officers practicing any other religion that required special headwear) at the
time the rule was adopted. When members of these religions did join the RCMP, it likely never
occurred to the authorities that these officers might have different needs. And when this was
finally brought to their attention, they probably felt it was regrettable but more important to
preserve tradition and uniformity of appearance. Many cases of adverse effect discrimination
have this pattern: initially, the act or rule is chosen not out of prejudice or a sense of superiority,
but simply without any awareness of how it may affect certain individuals. And when this is
brought to the agent’s attention, he regrets it but feels that other important reasons outweigh the
reasons for changing the policy. The question that adverse effect discrimination raises is
whether we can make sense of this phenomenon as fault-based and continuous with whatever is
objectionable about intentional discrimination; and it is no answer to this problem to suggest that
it does not exist.
But perhaps a more subtle version of this first response is in order. One might argue,
not that there is no such phenomenon as adverse effect discrimination, but that because it does
not appear to involve wrongdoing, we cannot legitimately prohibit it by law. So whatever we
may or may not say about adverse effect discrimination from a moral standpoint, we cannot say
that it is justifiably prohibited by law, in such a way as to require the agent cover the costs of
accommodating the victim. Insofar as we are to make sense of prohibitions on adverse effect
discrimination as legitimate laws, we must re-interpret them as prophylactic measures designed
to ferret out instances of covert intentional discrimination.
This view is popular among some American legal scholars.11 They argue that it can be
difficult for victims of discrimination to prove a malicious or prejudicial motive, especially in
cases where a rule or action is facially neutral. So by requiring proof only of burdensome effects
on the victim because of her possession of a trait marked out by a forbidden ground of
discrimination, we eliminate the evidentiary difficulties posed by intentional discrimination and
we cast a net that is wide enough to catch cases of intentional discrimination that might
otherwise be missed. But this net must not be mistaken for one that is designed to catch some
other form of discrimination. On this view, then, adverse effect discrimination functions as a
kind of res ipsa loquitur: just as that tort doctrine allows us to treat the defendant’s act in the
circumstances as “speaking for itself” and obviating the need for any further proof of negligence
on the defendant’s part, so this way of understanding adverse effect discrimination assumes that
certain sorts of objectionable consequences speak for themselves and render proof of ill motives
unnecessary.
It may be that this response offers us the only way of making sense of prohibitions on
adverse effect discrimination as legitimate law. If we accept this response, however, we will
have to abandon many of the current goals of anti-discrimination law. We look to antidiscrimination
law to combat oppression, to uncover and reform institutional structures that
disadvantage already stigmatized minority groups, and most broadly, as the Canadian Human
Rights Act states, to give effect to the principle “that all individuals should have an opportunity
equal to other individuals to make for themselves the lives that they are able and wish to have
and to have their needs accommodated.” This is not to say that we need (or should) take this
expansive statement literally. But it is to say that, at least in most countries, anti-discrimination
law has a more radical, reformist agenda than simply ferreting out exclusions motivated by
malice or prejudice. So before abandoning this agenda, we should try to see whether antidiscrimination
law can be understood in a manner that furthers this agenda while still
constituting a legitimate exercise of state power.
The last response I want to canvass before laying out my own view is one that questions
an assumption that has been underlying both my presentation of the problem and the two
responses to it that I have already examined. This is the assumption that in order for adverse
effect discrimination to be legitimately prohibited by law, we must make sense of it as involving
wrongdoing. One might argue that this is mistaken. John Gardner has suggested, for instance,
that those to whom private sector discrimination law applies –for instance, employers, serviceproviders,
accommodation-providers— are really just “agents of distributive justice”.12 Their
positions, he argues, give them access “to some of the most important opportunities that modern
life presents”. Consequently, they are well-placed to redistribute these valuable opportunities
from “the opportunity-advantaged” to “the opportunity-disadvantaged” in accordance with
whatever principles of distributive justice a society has chosen. Adverse effect discrimination is
any act or rule that fails to divide up the available opportunities in the manner that accords with
society’s chosen principles of distributive justice. So on this view, although adverse effect
discrimination is a legal wrong in the positivist sense that it is forbidden by actual laws (and
indeed Gardner repeatedly refers to it as “wrongdoing”), it does not depend on the perpetrator
having in any deeper sense committed wrongdoing or on his having personally wronged the
victim. All that it depends on is his having been a poor or incompetent agent of distributive
justice. He has, we might say, failed to aid people in circumstances where they need it, when
society’s chosen principles of distributive justice require that the state provide, or find someone
to provide, that aid. Let us call this the “instrumental view” of prohibitions on adverse effect
discrimination, since it treats these laws and those to whom they apply as instruments in
achieving the broader public purpose of redistributing opportunities to the disadvantaged.
One problem with the instrumental view is that it provides, at best, only a rather tortuous
explanation of the legal structures that most countries have in place to address adverse effect
discrimination in the private sector. The individual “complaint model” that I mentioned earlier is
most naturally interpreted as a way of enabling someone who has personally been wronged to
seek redress from the wrongdoer, in much the same way that the plaintiff in a private law dispute
seeks redress from a defendant for a personal wrong committed against him. And although
proponents of the instrumental view can tell an elaborate story about why this surface structure
of the complaint model actually tells us nothing conclusive about what the purpose of
discrimination law really is, this story lacks plausibility. The instrumentalist’s story must be that
the individual complaint model is in fact the most efficient way of achieving the broader
public purpose of redistributing opportunities to the disadvantaged: rather than entrusting to
legislatures the task of generating a series of specific prohibitions on discrimination in contexts
such as the workplace or schools or the provision of good and services, it is more efficient to
have a general prohibition on discrimination and to allow tribunals to develop specific rules if
and when complainants come forward with specific problems. So, for the instrumentalist, the
victim’s prospect of receiving redress is the incentive necessary to bring her before the tribunal
and alert the tribunal to a broader social problem. It does not represent the righting of a wrong
against her. The complaint model is therefore, on the instrumental view, a kind of convenient
legal fiction. It helps us achieve the outcome we want to achieve, but this is not because there
actually are any of the sorts of wrongs that it asks claimants to bring forward.
This story lacks plausibility, both for empirical reasons and for philosophical ones. On
the empirical side, it seems unlikely that the complaint model is the most efficient way of
achieving such redistributive goals. The extensive litigation that it requires is notoriously slow
and costly. And it is dependent on victims of discrimination coming forward to make their
complaints public, in circumstances where many of them have a deep mistrust of governmental
authorities and a great deal to lose if they publicly expose their colleagues or employers or
landlords. So, at least on the basis of experience, and assuming realistic rather than utopian
levels of government funding, it is hard to believe the empirical claim that the instrumentalist
must rely on if he wishes to defend the complaint model. Moreover, the instrumentalist’s
explanation of why we use a complaint model is much less economical than one that takes at
face value the idea that discrimination involves a personal wrong. So, assuming we can provide
a coherent account of anti-discrimination law as a means of redressing a personal wrong,
reasons of economy point toward adopting this account over the instrumental view.
A further problem with the instrumental view is that it admits that the second of our
puzzles –namely, how to make sense of adverse effect discrimination as the same sort of
phenomenon as intentional discrimination—is intractable. It implies that there are two radically
different forms of discrimination, which are prohibited for very different reasons. Gardner
himself explicitly acknowledges this but holds that it is not a problem: on his view,
discrimination can be of different types, each of which can be wrong for different reasons.
Although this may be right, it seems preferable to try first to work out a view of adverse effect
discrimination that accords with our intuition that the two forms of discrimination are versions of
the same phenomenon and amount to legal wrongs for some of the same reasons.
In the next two sections of the paper, I shall outline and defend an alternative view of
adverse effect discrimination. I shall argue that it does involve fault or wrongdoing. But the
fault has nothing to do with the agent’s motives or purposes. Rather, it consists in interfering
with another person’s rightful claim to equal freedom. The core idea is this. Most of us enjoy
certain freedoms of contract; and we believe, moreover, that everyone ought to have access to
these same freedoms because such access is necessary for full participation in society. These
freedoms include both freedoms to enter certain contracts and freedoms to deliberate about what
we wish to contract for in a manner that is insulated from pressures stemming from certain
extraneous facts about us, such as our race or religion. When someone else’s actions interfere
with your freedom, then the action can amount to adverse effect discrimination. Whether it does
will depend on whether the measures that this person would have to take in order to
accommodate you would constitute a greater intrusion into his freedom of contract. So the
aim of prohibitions on adverse effect discrimination, on this view, is to protect each individual’s
right to a certain set of important freedoms of contract. It is to help secure the conditions of
equal freedom. I shall argue that this is also the proper aim of legal prohibitions on intentional
discrimination: intentional discrimination, too, is a legal wrong not because of the objectionable
motives that accompany it, but because it amounts to an unjustifiable interference with another
person’s equal freedom. I hope to flesh out this view in what follows, and to show that it
provides a more coherent and attractive picture of discrimination than the ones we have so far
considered.
II.
Most of us enjoy certain freedoms in our contractual relations with others. Some are
freedoms to do certain things. We are free to accept jobs that we have been offered without
having to turn them down because our religion prevents us from working on those days of the
week that are required by the job. We are free to take public transit without having to get off one
stop early because the stop nearest our destination is not accessible to us. Other freedoms that
we normally enjoy are freedoms to have our deliberations about whom to contract with insulated
from certain pressures, pressures caused by the way others view certain facts about us.13 We are
free to consider which apartment to rent without having to factor in the contempt that the
management might show for us because of the gender of our partner; free to decide where to
shop without worrying about who will ask us to leave and breastfeed “somewhere more
discrete”; and free not to have to think at all about which sex’s washroom to use, because we
need not fear retaliation from people who see cross-dressing or sex-change surgery as
evidence of vice.
Most of us enjoy these freedoms because, when people enter into the provision of goods
and services, or employment, or accommodation, they tend to arrange their affairs in such a way
as to accommodate most people’s needs. The Sears Company schedules its workers’ shifts in
such a way as to accommodate those whose Sabbath falls on a Sunday; the transit system designs
its stations in such a way as to accommodate those who walk; and washrooms are assigned to
one or the other of two sexes in accommodation of conventional views about gender. In other
words, most of us enjoy these freedoms only because of constant accommodations which are
taken for granted and so remain invisible. But just because these freedoms are enjoyed by most
of us and the acts of accommodation that give rise to them are invisible, this does not mean that
there is no right to reasonable accommodation in play in these common cases. On the contrary,
of the freedoms that I listed above, we believe not just that many people happen in fact to have
these freedoms, but that each of us has a right to be accommodated in such a way as to have
access to them ourselves.
We can see this by contrasting these freedoms of contract with others, which we do not
think each person has a right to. There are many freedoms of contract that we do not think
people can rightfully demand: for instance, the freedom to take up a job regardless of your
qualifications for it or the freedom to decide which apartment to rent only on the basis of the
aesthetic qualities of the apartment, without having to factor in whether you can afford it or
whether it is close to your children’s schools. Such pressures of finance and family needs are
ones that we treat, to borrow a phrase from Lord Keith in the torts case of Baker v. Willoughby,
as “the ordinary vicissitudes of life”.14 When he wrote this, Lord Keith was not using
‘ordinary’ to refer to the frequency of these obstacles. That is, he was not making an empirical
claim, the claim that because they are so commonplace, each of us must live with them. Rather,
he was making a normative claim, the claim that there are some burdens that we believe ought to
lie where they fall, some freedoms that people do not have a right to. The same is true, I am
suggesting, of the freedoms we enjoy with respect to our contractual relations with others. There
are some that we think are not freedoms that anyone can rightfully demand; though some of us
may be privileged enough to enjoy them as a matter of fact. But there are other freedoms of
contract that most of us do enjoy, and that we believe each of us has an equal right to enjoy.
I want to suggest that laws prohibiting adverse effect discrimination are one way of
giving substance to these equal rights to certain freedoms in our contractual relations with others.
For one of the things that adverse effect discrimination does is to interfere with a person’s
freedom of contract –either by preventing him from taking a certain action or by interfering with
his freedom of deliberation by introducing extraneous pressures into his decision-making. Most
often it interferes in both of these ways. Laws prohibiting adverse effect discrimination work to
combat such interference. But significantly, they do not forbid every interference with another’s
freedom. All that they forbid are those interferences that satisfy two conditions: (i) they would
not have affected the victim in this way had it not been for her possession of a trait that
constitutes a prohibited “ground” of discrimination, and (ii) they cannot be eliminated by some
reasonable accommodation of the victim, without causing undue hardship to the agent. We can
see both of these conditions as ways of protecting equal freedom of contract for all.
Consider first the function of prohibited grounds of discrimination. Grounds of
discrimination seem to express our judgments about which sorts of freedoms of contract we can
rightfully claim of others, and which we cannot. For instance, every country that has antidiscrimination
laws recognises race, religion, and sex as prohibited grounds. This reflects our
shared belief in the importance of being free to contract without obstacles or pressures based on
these factors or on other people’s opinions about them. But interestingly, these grounds differ
widely in the type of trait that they involve. Some constitute chosen commitments, such as
religion. Others, such as sex, are traits over which we have no immediate control; and indeed, it
may seem important to protect us from the impact of others’ views about these traits precisely
because there is little we can do to alter them. Still other grounds, such as disability, involve
some traits that can become the basis for important commitments (think of the community of the
culturally deaf) and others that we are just saddled with (think of ADD). What matters for the
purposes of discrimination law, it seems, is not whether the trait is chosen or unchosen, an
important part of our life or something we would rather be rid of. It is that, whatever the nature
of the trait, the conflicts between the needs of those who possess this trait and the needs of
others, and the pressures stemming from others’ beliefs about this trait, are not ones whose
consequences the bearers of this trait can rightly be asked to bear. Their freedom of contract
cannot rightly be curtailed simply because they possess these traits.
One might at this point object that although I have been using the term “freedom of
contract” to refer to the different freedoms that the grounds of discrimination are intended to
demarcate and protect, this term is in fact inaccurate: this is not freedom of contract in the
normal sense of the term. Normally, we understand “freedom of contract” to mean “the freedom
to contract with whomever happens to want to contract with you”, and not “the freedom to
enter into those contracts that others can enter into, even in circumstances where the other
contracting party wishes to have nothing to do with you.” On the proper understanding of
freedom of contract, one might argue, victims of discrimination do not lack it at all. They may in
fact be unable to find anyone who is willing to contract with them, and so their freedom of
contract may not be of any value to them. But it is inaccurate to say that they lack freedom of
contract. They just lack the social conditions that enable them to make fruitful use of it. If this is
right, then it is misleading to suggest that anti-discrimination laws are about freedom of contract;
and misleading, also, to suggest that these laws attempt to limit the freedom of contract for some
in order that others’ rights to equal freedom of contract will not be infringed. For on this view of
freedom of contract, anti-discrimination laws do not give anybody freedom of contract. They
reduce the freedom of contract of discriminators by denying them the right to refuse to contract
with certain people on certain grounds; but although they do give something to victims of
discrimination, it is not freedom of contract. It is the social conditions that enable these people
to use the freedom of contract that they already have.
But is freedom of contract properly understood as “the freedom to contract with those
who happen to want to contract with you”? This is not how the law has historically understood
it. In quite a number of contexts, the law denies us the right to refuse to deal with others, either
on certain grounds or at all. Consider, for instance, the old common law restrictions on public
carriers, which required operators of such public facilities as ferries and taverns not to deny
admission to certain sectors of the public. Consider also the many rules within the law of labor
and employment that prohibit employers and employees from entering into, or backing out of,
certain kinds of agreements. If we take these laws as our basis, we may be inclined to define
“freedom of contract” precisely as “the freedom to enter into those contracts that others can enter
into, even in circumstances where the other contracting party wishes to have nothing to do with
you.” For this reason, I shall continue to use the terminology of “freedom of contract”
throughout the paper.
However, if one disagrees then I believe one could, without draining my account of
discrimination of any of its significance, replace my references to freedom of contract with
references to “the freedoms that give value to freedom of contract”. As I mentioned above, the
freedoms that I have been invoking as the basis for anti-discrimination law are all freedoms that
help give freedom of contract (in the narrower sense) its usefulness: without them, we are not
able to make much use of the freedom to contract with those who are willing to contract with us.
So my second response to the above objection is to suggest that those readers who prefer to use
“freedom of contract” in the narrow sense should substitute “freedoms that give freedom of
contract its value” for my appeals to “freedom of contract”. It is clear that this would not affect
most of my account; and I shall explain at the end of this section why it would not make a deep
difference to my response to Richard Epstein’s challenge to discrimination law.
Grounds of discrimination, I have said, reveal what we as a society take to be important
freedoms of contract at any given time –or, one might say, freedoms of contract that are so
important that access to them is necessary for full participation in society. If this is a correct
analysis of the function of grounds, then it helps to explain why so many human rights tribunals
and courts have faltered in attempting to locate a single criterion for something counting as a
prohibited ground. On the view I have been suggesting, whether some trait should be recognised
as a prohibited ground is a normative question whose answer depends on whether people have
a right to be free from the sorts of social pressures and obstacles that are encountered by those
with that trait, and whether this freedom is necessary for full participation in society. Because
there is no reason to suppose that all of the freedoms that are necessary for full participation in
society are valuable for the same reasons, there is likewise no reason to think that all grounds of
discrimination will meet some single criterion for constituting a ground. On the contrary, it
seems quite likely that the reasons will be diverse.15
I have argued that prohibited grounds show us which freedoms we feel individuals have a
right to claim in their dealings with others. I want now to turn to condition (ii), the requirement
that we must make reasonable accommodation of others, but only up to the point of undue
hardship. On the view that I am proposing, the function of this requirement is to ensure that the
freedoms of those who face discrimination are protected, but not at the expense of even greater
intrusions into the freedoms of the discriminators. In other words, the requirement ensures that
potential victims of discrimination are given access to the same freedoms that others commonly
have, but only insofar as this is compatible with the agent’s retaining his access to such
freedoms. If you are a full-time employee in my bookstore and I have a policy of requiring all
full-time employees to work one Saturday a month, I must make an exception for you if you are
a Seventh Day Adventist and need Saturdays for religious observance –at least, as long as my
store is large enough that other employees can cover for you. For requiring me to accommodate
you gives you the important freedom to work at a job that you are qualified for, and the freedom
to deliberate about the job without having to consider the impact of your religion on your work
(or the impact of your work on your religion). And the only freedom of mine that this interferes
with is a freedom which we do not think everyone has a right to, namely the freedom to dictate
every aspect of the running of my business myself. However, if you are my only employee and
my bookstore would be forced to close on Saturdays if you did not work that day, and if such
closure would substantially interfere with the viability of the bookstore, then I can reasonably
demand that you work on Saturdays. The law, then, requires reasonable accommodation, but
only to the point of undue hardship, because it is concerned not just with the freedoms of those
who face discrimination but also with the freedoms of those whose actions may result in it. It is
concerned to ensure that each of us is given no more and no less freedom of contract than the
majority of other people enjoy.
I have tried to explain how we might see prohibitions on adverse effect discrimination as
attempts to promote equal freedom of contract for all. But one might object that I have not yet
made clear how this picture of adverse effect discrimination makes sense of it as involving
wrongdoing, in such a way as to support both our intuition that the agent can legitimately be
coerced into shouldering the costs of reasonable accommodation and our intuition that the
victims of discrimination have been personally wronged by the agent. This objection might take
one of two forms.
First, one might be concerned that as I have described adverse effect discrimination, the
objectionable features of the action seem to consist solely in its effects, and not at all in the
nature of the action leading to these effects. Adverse effect discrimination, on the view that I
have articulated, is objectionable because it results in the denial of equal freedoms to some
people. But this seems to be just a matter of the effects of an action, and to have little or nothing
to do with what the agent was doing.
In responding to this version of the objection, it may help to compare adverse effect
discrimination to a tort that shares much the same form: the tort of defamation. This tort
involves publishing a statement that has the effect of damaging another person’s reputation, or
image in the community. Like adverse effect discrimination, defamation is a legal wrong
because of its effects, rather than because of the kind of action that the agent performed. Hence,
the plaintiff need not show that the defendant published the statement maliciously (nor even that
he published it intentionally; though unintentional publications can only be defamatory if the
publication was done negligently because the defendant could not otherwise be regarded as their
author). Nor does the wrong of defamation seem to lie in any facts about the way in which the
statement was published. Rather, it lies simply in the harmful effects on the individual’s
reputation. As one commentator has said, defamation is “defined by the character of the injury it
causes, rather than by the character of the insult which it produces”.16 The same could be said
about adverse effect discrimination.
But is this not problematic? In the case of defamation law, it might be. For defamation
law contains no other components that might help us locate wrongdoing on the part of the
defendant. Although there are a number of defences to defamation –such as the truth of the
statement published; or the “privilege” of publishing defamatory statements in the discharge of
certain private or public duties; or the need to allow for “fair comment” on matters of public
importance—these defences do not focus on legitimating features of the defendant’s actions. So
they do not work to suggest that, in the absence of these legitimating features, we can impute
wrongdoing to the defendant. Rather, the defences seem to articulate a series of policy-based
exceptions that focus on the importance of free speech and its ability, in certain circumstances,
to outweigh the harmful effects of the defamatory statement on the plaintiff’s reputation.
However, in the case of adverse effect discrimination, there is a further feature that helps
us explain how the agent’s actions could amount to wrongdoing. This is the requirement that the
agent must provide only reasonable accommodations, and only up to the point of undue hardship.
Because this requirement allows us to deny that discrimination has occurred at all in cases where
there is no way of reasonably accommodating others short of undue hardship, it implies that in
all cases where there is discrimination there must have been reasonable measures of
accommodation available to the agent and he must have failed to take them. Hence, it is not
really accurate to say that adverse effect discrimination consists solely in the creation of certain
effects. A better suggestion is that it consists in the agent’s failing to make reasonable
accommodations for others. It follows that those who engage in adverse effect discrimination
are acting unreasonably. We can therefore impute unreasonableness to agents who persist in
discriminatory actions. Their fault, then, can be understood as readily as the fault in any
negligence-related tort.
Or can it? One might object that although we can readily understand why most of the
actions that amount to negligence in the law of torts are unreasonable, and hence involve fault,
the same is not true of adverse effect discrimination. Why should we regard the failure to
accommodate those who are excluded by my actions on the basis of certain character traits as
unreasonableness on my part, as something that is my problem and falls on my shoulders to
rectify? Another way of putting this objection is to say that it is unclear how my own account of
adverse effect discrimination does not suffer from the same defects that, earlier, I found in the
instrumental view. On the instrumental view, as we saw, those who engage in adverse effect
discrimination are simply poor agents of distributive justice: all that they have failed at is
appropriately discharging the state’s obligation to redistribute resources and opportunities in
accordance with the preferred principles of distributive justice. I argued earlier that it is difficult
to see how this could be their fault or amount to wrongdoing in more than the positivist sense
that it is actually treated as a legal wrong. One might object that the same can be said of my
view. Why is it unreasonable, and hence my fault, if I choose not to accommodate you? Surely
all that I am refusing to do in such cases is come to your aid or benefit you, by making available
to you the freedoms that most people possess. But why is that my responsibility?
There is an important difference between the instrumental view and mine that I think can
help to answer this objection. On the instrumental view, adverse effect discrimination is
objectionable because it does not benefit certain people in circumstances where the state owes
them a benefit. Adverse effect discrimination amounts, on this view, to a failure to redistribute
resources and opportunities in the optimal way: hence, on this view, it is objectionable because
of the failure to benefit others. On my view, however, the objection to adverse effect
discrimination is not that it fails to benefit certain people. It is that it interferes with their
freedom; and this freedom is something to which they have an equal right. Each of us has a
claim on others to have the same important freedoms of contract that the majority of us enjoy.
And so I wrong you if my rules or actions prevent you from having one of these freedoms. It is
of course true that, when I interfere with your freedom, I normally harm you or at least fail to
benefit you. But what makes my action into a legal wrong is not the fact that I have harmed you
or failed to aid you, but rather the fact that I have interfered with a freedom that is rightfully
yours. On my view, then, the failure to accommodate others is unreasonable, and hence
amounts to wrongdoing, because it is an interference with their rightful freedom. And for the
same reason, it is a personal wrong against the victims of discrimination. It is not simply an
instance of misconduct that has the effect of harming them: rather, in interfering with someone
else’s rightful freedom, I commit a personal wrong against them.
I have tried to show that by construing adverse effect discrimination as an interference
with our equal right to certain important freedoms, we can make sense of it as involving
wrongdoing –and moreover, the kind of wrongdoing that involves a personal wrong toward the
victims of discrimination. We can therefore explain why it is that the victim of adverse effect
discrimination can rightly claim that she has been wronged by the discriminator, and why it is
just to require the discriminator to cover the costs of reasonable accommodation.
I now want to suggest that this is a plausible way of understanding why intentional
discrimination is a legal wrong, as well –and that, as a result, my account of adverse effect
discrimination helps us to make sense of these two kinds of discrimination as two forms of the
same phenomenon. Recall that intentional discrimination arises when actions or rules are
adopted with the purpose of excluding some people because they possess a certain trait, and this
trait is a prohibited ground. Hence, intentional discrimination differs from adverse effect
discrimination in that its purpose is to exclude, whereas the exclusionary effects of adverse effect
discrimination are not part of its purpose. On the common view of intentional discrimination
that we examined earlier, what makes this form of discrimination into a legal wrong is the
presence of one of two further features: either the agent’s sole motive is malicious, as she
excludes others in order to harm them, or it is prejudicial, involving a belief in the inherent
inferiority of these people. This view is problematic in at least two respects. On the one hand,
some exclusions are overt and purposive and seem objectionable, but they are not in any way
motivated by malice or prejudice. Consider, as an example, the recent case of a rape crisis centre
that refused to allow a male-to-female transsexual to volunteer as a counselor. 17 Because the
centre explicitly excluded her from volunteer work based on her gender identity, the case is one
of deliberate exclusion on the basis of a trait that is a prohibited ground of discrimination.
However, the managers of the crisis centre were not motivated by any animosity or belief in the
inferiority of transsexuals: their objection to her was largely that, having not been born or raised
as a woman, she lacked sufficient experience of the particular kind of discrimination that had
been faced by the women she would counsel. This reasoning may seem unimaginative and
insensitive; but it does not involve malice or prejudice.
The second, and more important problem with the common view of intentional
discrimination is that even where malice and prejudice are present, it is unclear that they can play
the type of role that the common view of intentional discrimination envisages them playing.
According to the common view, these motives turn intentional discrimination into a legal wrong.
But, at least in tort law, the presence of certain motives is rarely enough to turn an otherwise
lawful action into a legal wrong –unless, that is, the motive changes the character of what is done
and enables us to describe the action as a different action, an action that is wrongful. So, for
instance, in the case of intentional torts, certain intentions play a role in rendering an action
wrongful because they make a difference to the kind of action it is. If I inadvertently hit you
with my backpack while removing it, I have not assaulted you. But if I deliberately swat you
with it while removing it, I have. In this case, my intention to hit you makes a difference to the
legal status of my action because it turns what would otherwise have been an unintentional
touching into an action of a different kind: an unauthorized intentional touching, or an assault.
And an unauthorized intentional touching amounts to a legal wrong because it is a way of
drawing you into my purposes without your permission; whereas bumping you inadvertently
does not draw you into my purposes. But aside from cases in which the presence of a certain
intention changes the nature of the action performed, mental states are rarely treated in tort law
as sufficient to render an otherwise lawful action unlawful. Most courts, for instance, deny that
malice is ever, on its own, sufficient to convert lawful actions into legal wrongs.18 So the
common view of intentional discrimination sits uncomfortably with the normal role of motives in
tort law.
If we model an account of intentional discrimination on my explanation of what makes
adverse effect discrimination into a legal wrong, we avoid both of these difficulties. We can say
that intentional discrimination differs from adverse effect discrimination simply in the fact that
its purpose is to exclude, whereas the exclusionary effects of adverse effect discrimination are
not part of its purpose. Intentional discrimination may or may not, on this view, involve
objectionable motives. But what makes it into wrongdoing at law is never these motives. It is,
rather, wrongful in the same circumstances, and for the same reasons, that adverse effect
discrimination is wrongful: namely, because and to the extent that it constitutes an objectionable
interference with another person’s rightful freedom of contract. On this view, of course, the
distinction between the two forms of discrimination is not of very much legal importance. It
may be of minor importance: victims of intentional discrimination may suffer more when they
are purposively excluded, because such exclusions may constitute more of an insult to them (in
ways that I shall discuss further in section 3). So it may be that special damages should be
awarded in such cases, to reflect one or another of these additional harms. But these harms
should be regarded as further effects of discrimination for which the agent becomes responsible
because he is already responsible for the discrimination, in the same way that tortious agents
become responsible for the harms resulting from their wrongful actions. They should not, I am
suggesting, be treated as part of the reason why the discriminatory action was wrongful.
Before I go on to consider some objections to my view, I want to note one implication of
the view. It concerns a worry to which I alluded at several earlier points in the paper. This is
that anti-discrimination law constitutes an unjustifiable interference with freedom of contract and
association. Richard Epstein has eloquently articulated this worry in a number of places: he
laments the fact that our society has become “a world in which the antidiscrimination norm is
regarded as superior to the principle of freedom of association”.19 My account suggests that this
way of thinking about the relationship between anti-discrimination law and freedom of contract
or association. Anti-discrimination law does not simply function as a constraint upon freedom of
contract. Rather, it constrains some people’s freedom of contract in order that others, too, can
share in this same freedom. Or –if you prefer to use “freedom of contract” in the narrower way
that I discussed earlier in the paper— anti-discrimination law constrains some people’s freedom
of contract in order to provide other people with the conditions under which the purely nominal
freedom of contract that they already have becomes something that they can actually use, and
hence something that has value for them. Moreover, if I am right that the basic case is actually
one of accommodation –if I am right, that is, that those of us who do not have to launch
discrimination complaints have only avoided the need to do so because our rights have already
been respected by countless tacit acts of accommodation—then it only looks as though the
burdens of discrimination law fall unequally on those who end up having to accommodate
persons with traits marked out by prohibited grounds. Discrimination law only asks these
individuals to do what we all implicitly do most of the time: that is, make reasonable
accommodation for others, so that they may have access to the same important freedoms to
which others have access.
Of course, these considerations would likely not convince Epstein himself. On his view,
freedom of contract is important only instrumentally, insofar as it helps to maximize the overall
level of utility in society. Because Epstein values freedom only instrumentally, and because he
believes that the state’s role is to impose only those restrictions that will work to maximize
overall utility, he is not at all concerned with the way in which freedoms are distributed across a
society. So assuming, as Epstein claims, that the restrictions imposed on some people’s freedom
of contract by discrimination law do result in lesser overall utility, then Epstein would still regard
discrimination law as unjustifiable, even if could be shown that it vastly increases some people’s
freedom of contract (or, on the narrower view of “freedom of contract”, vastly increases some
people’s ability to use the freedom of contract that they already have). But for those of us who
take a different view of the value of freedom of contract and the purpose of the state’s attempts
to protect it, my account may help to quell worries about the extent to which discrimination law
interferes with it.
III.
One might at this point wonder what role my account gives to an important feature of
many discriminatory actions. This is the contempt or disdain for others that such actions convey
–or, as some scholars describe it, the attitudes that these actions “express”. Some scholars have
recently argued that this feature of discrimination is alone sufficient to justify the legal
prohibition of it.20 According to such expressivists, discriminatory actions should be prohibited
because of what they express: they express contempt and disdain for others, instead of equal
concern and respect. I shall suggest in what follows that there is something insightful in the
expressivists’ appeal to the attitudes that an action expresses, and in particular, in the idea that
the presence of these attitudes can aggravate the damage that is done by discrimination and
hence require more by way of compensation from the discriminator. But I shall try to show that
it is a mistake to appeal to the expressive dimension of discriminatory actions as the source of
the discriminator’s initial wrongdoing.
Why is this a mistake? Most importantly, because the same attitudes that are expressed
by actions that anti-discrimination law prohibits can equally well be expressed through other
actions, actions that are not regulated by anti-discrimination law and indeed are not subject to
any legal sanctions at all. Suppose I own a small pharmacy. I am not free to deny jobs at my
pharmacy to women on the grounds of their gender: this would constitute discrimination and
would be prohibited by law. But I am perfectly free to stage a peaceful demonstration against
hiring women pharmacists on the street outside and to wear a placard stating that women should
not be hired as pharmacists. This would constitute a legitimate exercise of free expression and
would not be prohibited by private sector discrimination law, since it is not something I do in the
context of being an employer. And this suggests that what is wrong with my denying jobs to
women as an employer is not reducible to what my action expresses. For surely most of this is
also expressed by my perfectly legal placard and my demonstration.
Moreover, it seems significant that although our constitutional entitlement to freedom of
expression protects many ways of expressing stigmatizing or exclusionary claims about others
(such as my marching up and down with my placard), no role at all is given in discrimination law
to the discriminator’s interest in freedom of expression. At no stage of the proceedings is the
alleged discriminator allowed to put forward the claim that his interest in freedom of expression
would be harmed if he were required to accommodate the victim; and there is no opportunity for
the adjudicating body to weigh this interest in freedom of expression against the victim’s interest
in being accommodated. And this too suggests that the focus of discrimination law is not on the
discriminator’s action qua expression.
I have argued that expressivism runs into difficulties as a theory of why discrimination
can justifiably be prohibited by law. But I want now to suggest that it does point us in the
direction of an important truth about many cases of intentional discrimination, and about the
ways in which the presence of malice or prejudice can aggravate the wrong that is done by
discriminatory actions. Expressivism starts from the plausible claim that many discriminatory
actions constitute expression of contempt for others. This is obviously true of intentional
discrimination, or more exactly, those acts of intentional discrimination that are motivated by
malice or prejudice. In such cases, the agent has not just overlooked a particular person and her
needs; rather, he has actively trampled on them, and this does express contempt and disdain for
her. This expression clearly causes additional anguish to the victims of discrimination; and it
may also increase the social stigma associated with possessing the trait in question. So in such
cases, the contemptuous nature of the action qua expression seems to aggravate the wrong that
is done. This may be why many jurisdictions award special additional damages in cases of
intentional discrimination, to compensate the victim for these further losses. So expressivism
helps us to see that when discrimination is done intentionally and is motivated by malice or
prejudice, the contempt that it expresses aggravates the wrong that is done to the victims. It is
important to note, however, that it is the particular way in which the action is done in such cases
that makes it an expression of contempt, and that this is not an essential feature of the action as a
discriminatory action. So the fact that a discriminatory action that is motivated by malice or
prejudice expresses contempt for the victims may aggravate the wrong or add a further
consequential loss which we think the discriminator can rightly be held accountable for. But it
does not constitute the wrong that is done in cases of intentional discrimination.
There remains the question of whether intentional discrimination that is not motivated by
malice or prejudice could ever similarly express objectionable attitudes and thereby exacerbate
the wrong that is done; and likewise, whether some cases of adverse effect discrimination might
also constitute objectionable expressions, leading to an exacerbation of the wrong in these cases
as well. It seems plausible that in some contexts knowingly excluding certain people, even for
quite benign reasons, can send the message that these people are not as worthy as others. And
likewise, completely failing to consider the impact of a rule or policy on a particular group of
people can also send the message that they are less worthy than others. When it does, it must
cause emotional suffering and social stigmatization quite similar to –though likely not as severe
as— that caused by discrimination that is motivated by prejudice or malice. So perhaps the
special additional damages that we now reserve for intentional discrimination should also be
available in certain other cases, though one would expect lesser amounts to be awarded in such
cases than in cases of intentional discrimination that involve malice or prejudice. This is a
question that requires further consideration and a more detailed analysis than I can give here of
the attitudes expressed by discrimination and the effects of these expressions on the victim.
In this paper, I have argued that we should understand both intentional and adverse effect
discrimination in terms of wrongful interference with another person’s equal freedom. If we do
so, we can make sense of discrimination as a single, coherent phenomenon. And we can explain
why those who discriminate seem to wrong those whom their actions exclude, and why they can
justifiably be coerced into covering the costs of reasonable accommodation. This account also
provides a clear explanation of why discrimination law is not an unjustifiable interference with
freedom of contract: it is not, because it limits some people’s freedoms only in order to enable
everyone fully to enjoy the same freedoms. As the last section of the paper suggests, there
remains work to be done in clarifying the precise role that should be given in discrimination law
to the effects on the victim of the attitudes that the discriminatory act expressed, and in particular
to the emotional suffering and stigma resulting from this. I have argued that these factors
aggravate the wrong, and I have suggested that although they are most clearly present in those
cases of intentional discrimination that involve malice or prejudice, they may also be present in
other cases of intentional discrimination and in cases of adverse effect discrimination as well
–though I am not certain that this is correct. But I hope that this paper has provided at least the
beginnings of a workable account of discrimination and has shown why it is a perplexing and
philosophically interesting phenomenon.
Notes
1For helpful comments on earlier drafts, I am particularly grateful to Arthur Ripstein, and also to
Ronald Dworkin, Evan Fox-Decent, Thomas Nagel, Denise Reaume, David Velleman, Jeremy
Waldron, and audiences at the McGill Legal Theory Workshop, the N.Y.U. Colloquium in
Legal, Political and Social Philosophy, and the Faculty Workshop at the University of Toronto.
2 For exceptions, see Matt Cavanagh, Against Equality of Opportunity (Oxford: Oxford
University Press, 2002) and Larry Alexander, “What Makes Wrongful Discrimination Wrong?”
University of Pennsylvania Law Review 141 (2002): 149-219. More philosophical work has
been done on affirmative action, which I take to raise somewhat different issues. See, for
instance, Thomas Nagel, “Equal Treatment and Compensatory Discrimination,” Philosophy and
Public Affairs 2 (1973): 348-63 ; Thomas E. Hill, Jr., “The Message of Affirmative Action,”
Social Philosophy and Policy 8.1 (1991): 108-129; and Elizabeth Anderson, “Integration,
Affirmative Action, and Strict Scrutiny,” New York University Law Review 77 (2002): 1195-
1271.
3 Those whose discussions presuppose what I am calling the “common view” of intentional
discrimination include Cavanagh and Alexander, supra note 1, and also John Gardner, “Liberals
and Unlawful Discrimination”, Oxford Journal of Legal Studies 9 (1989): 1-22; and Richard
Primus, “Equal Protection and Disparate Impact: Round Three,” Harvard Law Review 117.2
(2003): 493-587.
38
4 It is referred to as “adverse effect discrimination” in Canada and “disparate impact
discrimination” in the United States.
5 Subject to an important proviso, which I discuss in Section 1.
6 The case I discuss here is loosely based on Douglas Parisian v. Hermes Restaurant Ltd. (1987),
9 CHRR D/4756 (Manitoba Court of Queen’s Bench) and Commission des droits de la personne
et des droits de la jeunesse c. 9107-9194 Quebec inc. (Restaurant Jing Hua), [2005] J.T.D. P.Q.
no. 24, though neither of these cases had the religious dimension of my variant 3.
7 Note that I am using “fault” here, and throughout the paper, in the legal rather than the moral
sense. So, as I am using the term, it does not imply that the agent is morally blameworthy.
8 In Palsgraf v. Long Island Railroad 248 N.Y. 339, Cardozo spoke of negligence as a tort that
should be conceptualised not as “negligence in the air” but rather as negligence toward a
particular class of persons in respect of a certain action. I am suggesting that this is how we
normally conceptualise adverse effect discrimination.
9 My aim here is not to offer an exhaustive list of the alternatives to my view, but only to discuss
a few plausible ones. I have chosen these ones partly because each fails to capture some of the
features of adverse effect discrimination; so by seeing where each goes wrong, we can sharpen
our view of the phenomenon we are trying to account for.
10 Human Rights Commission Act, 1977, s.27 (N.Z). This legislation was subsequently replaced
by the Human Rights Act, 1993 (N.Z.), which eliminated the concept of “discrimination by
subterfuge” and recognised disparate impact discrimination as a distinctive form.
39
11 See, for instance, Christine Jolls, “Antidiscrimination and Accommodation,” (2001) 115
Harvard Law Review 643; A. Morris, “On the Normative Foundations of Indirect Discrimination
Law: Understanding the Competing Models of Discrimination as Aristotelian Forms of Justice,”
Oxford Journal of Legal Studies 15 (1995): 199-228; and George Rutherglen, “Disparate Impact
Under Title VII: An Objective Theory of Discrimination” Virginia Law Review 73 (1987):
1297-1345. See also Justice Stephens’ reasoning in Alexander v. Sandoval, 121 S. Ct. 1511,
1530 n.13 (2001).
12 Gardner, “Discrimination as Injustice” (1996) 16.3 Oxford Journal of Legal Studies 16.3
(1996): 353-67, p. 363; subsequent quotations are also from this page.
13 See Seana Shiffrin, “Egalitarianism, Choice-Sensitivity and Accommodation”, in Reason and
Value, ed. R. Jay Wallace et al. (Oxford: Clarendon Press, 2004): 270-302, for a detailed
discussion of the moral importance of such deliberative freedoms.
14 Baker v. Willoughby [1970] A.C. 467
15 One might object that unless something more general can be said of what ties together all of
the various freedoms that underlie the different prohibited grounds, we have no reason to think
that the victims of adverse effect discrimination have been wronged. For we have no single
principled explanation of that wrong. But in my view, we do have a principled explanation: it is
that victims of discrimination have been denied a freedom to which they have a right, a freedom
that is necessary for full participation in society. Which freedoms these are will vary depending
on the institutions that are central to a given society; and so, on the account that I have offered, it
is inappropriate to look for a more detailed explanation of the wrong that would hold true across
40
all societies. But I do not think this is a flaw in the view. It is a consequence of the fact that
some of these freedoms vary from society to society.
16 R. Browne, The Law of Defamation in Canada, 2nd ed. (Toronto: Carswell, 1999); my italics.
17 Vancouver Rape Relief Society v. Nixon, 2005 B.C.C.A. 601 (British Columbia Court of
Appeal).
18 As Denise Reaume has argued, it is consistent with this point to accept that malice can defeat a
privilege in certain contexts. For in such contexts, it is features of the action other than the
agent’s malice that make that action into a legal wrong. The role of malice is just to show that
the agent was not in fact acting in accordance with the rationale for the privilege. See Reaume,
“Harm and Fault in Discrimination Law: The Transition from Intentional to Adverse Effect
Discrimination,” Theoretical Inquiries in Law 2.1 (2001): 349-85, p. 360, note 19.
19 R. Epstein,“The Place of Caste under the Civil Rights Laws: From Jim Crow to Same-Sex
Marriages,” Michigan Law Review 92 (1994): 2456-478, p. 2470. See also Forbidden Grounds:
The Case Against Employment Discrimination Laws (Cambridge, MA: Harvard University Press,
1992) and “Standing Firm on Forbidden Grounds,” San Diego Law Review 3.1 (1994): 1-56.
20 E. Anderson and R. Pildes, “Expressive Theories of Law: A General Restatement,” University
of Pennsylvania Law Review 148 (2000): 1503-75; D. Hellman, “The Expressive Dimension of
Equal Protection,” Minnesota Law Review 85 (2000): 1-70.