There’s a little semantic game that’s being played a lot these days, which seems to me worthy of analysis. (And since philosophers are so often of accusing to getting hung up on “semantic questions,” who better to comment on it?) It has become quite standard in many quarters to condemn Canadian society, along with all of its institutions, as being thoroughly and systematically racist. There is however an important ambiguity in the way that the term “racist” is being used, with critics often shifting back and forth between two quite different meanings of the term, in a way that vitiates the force of their criticism.
When most people hear the word “racism,” the way that they understand it is in terms that would have been familiar to civil rights activists of the 1960s. This type of racism was interpreted first and foremost as a derogatory attitude certain individuals have, that leads them to engage in discriminatory behaviour – treating some people better than others based on their racial characteristics. These attitudes could be conscious or unconscious, and the racism could be overt, subtle, or covert. The point is that racism, on this traditional conception, occurs through derogatory attitudes, which then have a systematic effect on interpersonal interactions.
On this traditional conception, the injunction to avoid being racist can be satisfied by purging oneself of these attitudes, or if that proves impossible, by putting in place checks to ensure that they do not influence one’s behaviour. In many cases, instituting “colour-blind” procedures is a way of ensuring this. For instance, it is common in many countries for job applicants to submit a photo along with their CV. In North America we do not do this, in part because it just seems like an invitation for employers to engage in discrimination (of various forms, including racial). Similarly, a professor might ask students to submit assignments with just their student number on it, not their name, in order to avoid the possibility of bringing any discriminatory prejudice to bear upon the evaluation.
With this type of racism, if one stands accused of having been “racist” in one’s interactions with another person, it is a legitimate defence to show that one was not aware of their race at the time of the interaction. (For instance, I once spent an unpleasant couple of days sitting around waiting to testify at the Ontario Human Rights Commission proceeding, responding to a complaint from a student who was accusing the university of racial discrimination in its grading practices. As associate chair of undergraduate studies in my department I had handled a series of petitions she had made to have her course work regraded. The university’s principal defence against the charge of racial bias was that my interactions with the student were all conducted by email, and the rereading of her course work was done blind, and since there was nothing in her name that gave it away, no one was aware of what her race was. There were some other complicating factors, but the core defence was, I assume, successful, since the university won the case.)
One final thing about this type of racism: in our society it is generally regarded as a very serious character flaw. Thus if you accuse the average person of being “racist,” and they interpret this accusation in the “discriminatory behavior” sense, they will typically find the accusation upsetting and seek to defend themselves against it. At the same time, they will also understand that, if it is true, then there is a serious problem with their behaviour and they should do what they can to stop acting this way. In other ways, saying “that’s racist,” in the discriminatory behaviour sense, leads immediately to the normative judgement “you should stop doing that.”
There is, however, a different use of the term “racism,” which is more common among academics, and is generally unfamiliar to the average person on the street. This is often described as “institutional racism,” or increasingly, as “systemic racism.” The idea, stated roughly, is that there could be a “system” in place that is discriminatory in its effects, but where no single individual is being “racist” in the attitudinal sense. This is easiest to explain using the language of the U.S. Supreme Court, which distinguishes in civil rights law between “disparate treatment” of individuals and “disparate impact.” Old-fashioned racism, with respect to job applicants for instance, involves disparate treatment of individuals based on their race. It is however possible to impose employment criteria that do not single out individuals for differential treatment in accordance with their race, but nevertheless have a differential impact upon members of different racial groups. The U.S. Supreme Court has ruled that, if a criterion does have such an impact, it must be shown to constitute a bone fide occupational requirement in order to be permissible.
To take a concrete example, the Toronto Police used to have a minimum height requirement for officers. This obviously involved disparate treatment of short people, but at the time I suppose this was considered justifiable, on the grounds that officers should have an authoritative presence. In any case, as the population become more diverse through immigration, the rule turned out to have the untoward effect of excluding most East Asians from the service – and thus had disparate impact upon a particular racial group. This raised the question whether it could be considered a bone fide occupational requirement, and the conclusion that it wasn’t (or that whatever gain in efficacy the police force enjoyed by being tall and imposing was outweighed by the disadvantage of having practically no Asian officers). So the rule was abolished.
What’s going on in these cases? Because racial groups are different in a variety of ways, not all of which are constitutive of racial identity, certain neutral traits (e.g. height) may be correlated with racial traits (e.g. Chinese ancestry), and as a result, a policy that involves disparate treatment based on the neutral trait can have disparate impact with respect to individuals bearing the racial trait. As a result, it is not an adequate response to the charge of discrimination to point of the neutrality of the trait that serves as the basis for disparate treatment. There are a variety of reasons for this, but the most obvious is that this would make it too easy to engage in covert discrimination by gerrymandering a set of rules that are racially discriminatory without actually mentioning race. In order to defend the rule, you have to show that it is independently justifiable.
There is an important normative difference, however, between this sort of case and the more familiar one involving disparate treatment. Merely observing that the rules being applied by a particular institution are having disparate impact on racial groups does not amount to a condemnation of that institution or its rules. This is why it is somewhat tendentious to call this systemic or institutional “racism.” With people, we move automatically from saying “that’s racist” to “you should stop that.” But with institutions, saying “that’s racist” (in the systemic sense) does not automatically entail the “you should stop that” judgment. In order to get to the moral condemnation, you must show that there is something else wrong with the rules being applied, beyond the mere fact that they have disparate racial impact. It must be shown that they are somehow not justifiable. (For instance, an employer could have a particular level of education as a job requirement, and if different racial groups have different rates of educational attainment on average, this will have disparate racial impact, and yet the requirement might be perfectly justifiable.)
Unfortunately, many people seem to think that, in order to make the case for the reform of a particular institution (or to condemn it as “racist”), all that one needs to do is identify some respect in which it is having a disparate negative impact on a disadvantaged racial group. This does not follow. Typically, the argument works by exploiting the ambiguity between racism in the attitudinal sense, which is always wrong but hard to prove, and racism in the systemic sense, which is easier to prove but not always wrong. Another way of putting it is to say that, if you want to preserve the judgement that racism is always wrong, then you should not refer to institutions with disparate racial impact as “systemically racist,” because systemic racism in this sense of the term is not necessarily wrong.
So far I’ve been speaking rather abstractly about how these arguments go, but let me give a concrete example of how people play fast and loose with this distinction. On February 26, journalist Marci Ien published an op-ed in the Globe and Mail entitled “The Double Standard of Driving While Black – In Canada” recounting the details of being pulled over by the police for failing to stop at a stop sign. As the headline suggests, she alleged that various details of the encounter, including the fact that she had been pulled over in the first place, were due to police racism. As she put it
But there is no walking away from the truth. The stop signal at my daughter’s school is half a kilometre away; why wasn’t I pulled over there? Why did he follow me home? Why, after seeing the address on my driver’s licence, did he still ask if I lived at my home?
Who you are doesn’t matter; it’s what you are. If you are black in Canada, you are subject to a different standard and, often, seemingly, different laws.
The allegation being made here is obviously one of disparate treatment. In fact, saying that, if you are black, “you are subject to a different standard” is almost the definition of disparate treatment. Furthermore, Ien was clearly alleging that this disparate treatment was motivated by racism in the attitudinal sense, since later on in the piece she says “So how do we fix this? There are no easy answers, but one solution would be to start with our kids. We know that children are not born with prejudice. Racism is learned.” The implication is clearly that the officer was prejudiced.
The Toronto Police disputed Ien’s allegations, releasing video from the incident that showed the infraction (failing to stop at a stop sign), and pointing out that Ien’s race was not visible to the officer until after she had been pulled over and stepped out of the car. This is the standard “colour-blind” defence against charges of disparate treatment. The officer was, in effect, saying “I couldn’t have been treating you differently due to your race, because I didn’t know what your race was.”
Things got more interesting when Shree Paradkar, at the Toronto Star, wrote a column defending Ien (“Toronto Police reaction to Marci Ien shows woeful ignorance of racism basics.”) On her view, the police made an elementary error in thinking that the officer’s ignorance of Ien’s race constituted a defence. Paradkar claims that “Nowhere in Ien’s piece is the allegation that the man who stopped her was racist” (which I suppose could be true, if you read the piece in the most literal way possible). She goes on to suggest that the Toronto Police misunderstood the meaning of the term “racism” as Ien was using it: “Racism isn’t just about intent. It’s also about outcomes. Racism can occur without anyone having to be a racist — or without someone being actively prejudiced against a person of colour.”
What Paradkar is doing is here is shifting from an allegation of attitudinal racism to one of systemic racism (or from disparate treatment to disparate impact). Indeed, she goes on to cite some statistics showing the black drivers in Ottawa (weirdly, not Toronto), are pulled over more often than others. The problem with this argument though, as I indicated, is that this sort of racism is not necessarily wrong. If the rule being applied is justified (as most traffic regulations are), and it is being enforced in a way that does not involve disparate treatment of different racial groups, then the fact that it has disparate impact on different racial groups is neither here nor there. More importantly, it is neither the mandate nor the responsibility of the police to avoid disparate racial impact in their enforcement of the law. (For instance, the people that they arrest for murder are overwhelmingly male, and yet the police have no obligation to eliminate disparate gender impact in their enforcement efforts.) The police have to take society as it is given to them, then enforce the law in as neutral a way as possible. Since the social determinants of crime (such as various forms of social disadvantage) are not distributed in a racially neutral manner, there are going to be differences among racial groups in the rate of different crimes committed, and so it is natural that enforcement of the law will have disparate racial impact.
If this is the case, then one might wonder why people bother to make accusations of systemic racism against the police. One possibility is that they are simply confused – they think that just demonstrating disparate impact is sufficient to condemn the behaviour of the police. There may be some of this going on, but the more likely explanation is that they consider the difference in outcome to be evidence of attitudinal racism. In other words, they think that the most plausible explanation for the disparate impact of the laws is that police are engaged in disparate treatment. In fact, Paradkar basically says as much, later in the column, when she quotes with approval Ijeoma Oluo saying, “Those who demand the smoking gun of a racial slur or swastika or burning cross before they believe that an individual encounter with the police might be about race are ignoring what we know and what the numbers are bearing out …We are being targeted.” Obviously, the claim that “we are being targeted” is an allegation of disparate treatment, not impact.
One can see here the semantic shell game that is being played. An allegation of attitudinal racism is made. The perpetrator pleads innocence, and offers some evidence of neutrality. The allegation is then switched to one of systemic racism, against which that defence is invalid. But this switch also neutralizes the moral force of the allegation. If the perpetrator points this out, then the switch gets made back to an insinuation of attitudinal racism, as the most likely explanation for the systemic racism. And so on.
The reason for all this shiftiness, I think, has to do with a particular frustration experienced by those who are convinced that racism is ubiquitous, but have difficulty producing any evidence of it in day-to-day life. This difficulty is no doubt due, in part, to the fact that many racists have figured out that their attitudes are not, broadly speaking, acceptable, and so have learned to keep their opinions to themselves and avoid any overt displays. This leads to the suspicion that there is a significant undercurrent of racism out there, but that is difficult to pin down. And so critics switch their allegation to one of “systemic racism,” which as I have mentioned, is much easier to prove, you just look at statistics on outcomes. Unfortunately, it is also not necessarily wrong, and so in order to preserve the sense of wrongness, they insinuate that the differences in outcome are evidence of attitudinal racism (mediated, for instance, through “implicit bias”).
This conceptual muddle is not a huge problem, so long as one’s only desire is stand on a soap box and condemn racism in the abstract. Where it proves inadequate, however, is when one wants to take concrete steps to actually eliminate racial injustice. At this point, the question of whether the “racism” in question involves disparate treatment or disparate impact becomes extremely important, and so you need to draw the relevant distinctions. Under these circumstances, the sort of bobbing and weaving that Paradkar is doing becomes supremely unhelpful.
P.S. I am aware that there are other, more sophisticated conceptions of “systemic racism” in the academic literature. I am not criticizing those, but am commenting rather on a specific confusion that occurs in public debate.