As promised last week, I've looked over this report on Canada and the experiences of immigrants, in some detail. As I probably should have guessed, part of the problem I had with it was the idiot headline-writers who decided that this report showed there was something wrong with Canada's multiculturalism policy. The real objections, raised near the end of the report, are that Canada's multiculturalism policy doesn't go far enough, not that it's no good.
I'll start with a quick summary. The report opens with a brief discussion of immigrants and economic success, measured basically by income. The authors allow that there are a lot of potential explanations for the data that suggest immigrants fare worse than native-born Canadians, and so move on to a more central point they want to make, namely that immigrants, even into the second generation, experience discrimination that negatively impacts on Canadian society. This is really a three-stage argument. First, the authors try to establish that immigrants perceive that they are discriminated against. Second, they suggest that these perceptions may be supported by more direct measures of racism. Third, they argue that this negatively impacts on something important: namely, social cohesion. Finally, the authors conclude with some discussion of potential policies that could be enacted or improved upon in order to better the situation.
It's pretty evident to me that the second stage of the main argument outlined above is weak. There's a lot of ways to explain away perceptions of racism -- or, indeed, perceptions of anything -- and, to their credit, the authors gesture towards at least one of them at a couple of points, i.e., that there may be different expectations on the part of immigrants with regard to appropriate treatment. In other words, what's good for the goose may not be good for the gander: treatment that a native-born Canadian might shrug off, an immigrant might take to heart. We could also explain away perceptions on the basis of genuine misunderstanding, and this is particularly prevalent when it comes to racial issues. One has to choose one's words (and deeds) with extreme care, else will cause inadvertent offense.
Of course, this doesn't mean that the conclusions drawn by the authors are false, only that they aren't well-supported. Again, to their credit, they do recognize that the second claim is the weak one, and that work needs to be done in this area. I would suggest, though, that the problem is worse than it appears. I see at least two serious problems in trying to nail down features that track genuine racial prejudice and discrimination. The first is a conceptual problem regarding race itself. The second is a problem of model-building.
The conceptual problem is this. It's pretty much agreed at this point that race is a social construct. There are certain broad, coarse-grained physical differences between people, but they don't fall into neat categories. Where the lines are drawn between the categories is pretty arbitrary, and which of a given person's features are prominent enough to count them as one race rather than another is also somewhat arbitrary. Compare this to, say, classifying minerals.
The fact that race is a social construct doesn't mean that it's impossible to track, of course. The greatest (for a given value of "great") social construct is money, and we track that all the time. But money is an odd case: it obeys rules that are basically dictated from "on high". There is no parallel relation between race and something as there is between money and economic policy/institutions. So, the construction of race is a matter of social practice, not social rules, and therein lies the problem. Social rules fix conceptual boundaries: the reason that shells don't count as money (they very well could, after all) is that we have a clearly-defined rule that says what counts as money and what does not. But what, exactly, is the rule for counting someone as, say, black? We certainly have a grab-bag of judgemental dispositions and rules of thumb that we can use to categorize people, but they're relatively easy to flummox. Tiger Woods, for example, has many physical features that I associate with the category "black"; however, I also know that his father has, in addition to black ancestry, also Chinese and Native American, and his mother has a mix of Thai, Chinese, and Dutch (so, white) ancestry. So, which is Tiger Woods: black, white, Asian, Native American, some combination? My rules don't really help me here; the initial judgement based on his appearance conflicts with a series of judgements based on his genetic heritage, and I find myself with no clear answer.
Thus, if we're facing the problem of trying to track race, we are, in a very real sense, trying to hit a target obscured with fog -- a target which, to make matters worse, doesn't actually have any distinct and clear boundaries. I suppose that makes it rather like trying to hit the fog itself, rather than something in fog. So, because of the nature of the concept of race, we're going to have to allow for a certain amount of fuzziness to the variable we're tracking. Moreover, it seems almost inevitable that this fuzziness will be inherited by any results we draw from our tracking, should it be successful. That is, if we can successful model race somehow, we're going to have to allow that this model will support only loose generalizations, and not any particularly strong claims. If the model of race is supposed to support policy decisions, then these decisions will have to be themselves fairly loose and open to interpretation.
Moving now to the second problem. My understanding of the social sciences suggests that anything that can't be measured directly should be measured indirectly, by means of a model. That is, take, say, the effects of globalization. There is no such thing as a globalization thermometer which can be gently applied to people in order to determine exactly globalization's effects upon them. The very idea is nonsense. So, we have to look for something that we can measure which would be affected by globalization: economic prosperity, for example, or level of international trade.
So, suspending the conceptual worries about race itself, let us suppose we can come up with some rough-and-ready working definition of race and thus racism. What could we use to track racism? On the face of it, we have a huge methodology problem: unlike economic features, which are relatively objective and measureable, anything we could use as a proxy for racism is itself fraught with conceptual or logical difficulties. We could ask people if they had experienced racism, but, as said earlier, the perceptions of racism are highly sensitive to factors other than racism itself; so, we'd run a serious risk of measuring something other than what we're trying to. It's possible to control for this with carefully constructed questioning on parallel to the malingering measures that exist in psychological test batteries. That is, anyone who is faking psychological symptoms will answer questions in certain abnormal patterns, e.g., describing themselves as more disabled than all but 1% of the population, giving inconsistent answers to highly similar questions, etc.
We could maybe define a series of acts as inherently racist, but this has even more problems. First, we could inject our own perceptual biases into the equation, just as would be present if we simply quizzed people. Second, we have a grain problem: that is, we have to be able to specify acts (and observe them in order to check if they meet the specification) with sufficient detail that we ommitted innocent cases -- for example, a white man is hired rather than a black man because the black man failed to pass a necessary security check -- but caught non-innocent ones -- a white man is hired rather than a black man because the black man failed to pass a necessary security check, and the white man was never tested. Third, if we're looking for general trends, the more specific we get -- in order to solve the grain problem -- the less able we are to observe any overall pattern to how particular groups of people are treated.
In principle, it seems, any potential model for racism is rife with difficulties. So, in addition to the fuzziness that attaches to the concept itself, any conclusions drawn from this model will also probably inherit uncertainies due to the uncertainties and hedging present within the model.
This seems to spell disaster for policy recommendations. Either we end up where we currently are, with a vaguely-worded Act that declares multiculturalism policy without any specific mechanisms or goals, or we end up with only cosmetic changes. Nothing significant can be done without a stable concept of race and a good means of tracking racial discrimination.
Unless, of course, we're willing to just chuck the whole enterprise and try to solve problems for everyone. Let me explain. The conclusion this study reaches is that something special needs to be done for immigrants. I've been trying to make trouble for the justification of the claim that immigrants are obviously less well-off than native-born Canadians. However, if policy were shaped to provide blanket benefits for everyone, then my objections would be defeated. That is, there are basically three things we could do, given the report and my critique.
(1) Do nothing, keep the status quo. This runs the risk of being branded the racist solution. There is very probably some racial problem in Canada, even though, for technical reasons, we're going to have a very hard time telling what it is and how bad it is.
(2) Give immigrant groups special resources, the favoured solution of the report. The problem with this, however, is that it assumes the report's results are well-founded and sound, and this is far from obviously true. If we can't justify the claim that immigrants are particularly badly-off for reasons not of their own devising, then we can't justify diverting some portion of society's resources for their particular benefit.
(3) Give everyone more. This avoids the problem of being a racist solution, as the aim is to give every person, regardless of race, an equal set of resources or opportunities to gain resources. This also avoids the problem of the questionable nature of the particular results, for everyone is benefitting. No one is being singled out for special treatment, so there is no special consideration that stands in need of justification. Furthermore, if there really is a significant racial inequality, then persons who are at a significant disadvantage would be more likely to use more of these extra resources than those who are already advantaged, and thus don't need them. Thus, even if we can't measure the disadvantage, we can still correct it.
Finally, (3) also helps solve the cohesion problem, although not in a way the authors of this report likely saw. Their claim is that society needs to be cohesive -- that is, needs to have some sense of "common bonds" -- in order to enact important collective projects and, indeed, to continue to hang together as a society. This is probably true (but it could use some rigorizing). Their solution is targetted welfare or redistribution of resources, a solution which, as I have said several times now, is not sufficiently justified. However, untargetted welfare or redistribution would also solve the cohesion problem. In order for any system of this kind to work, there have to be common bonds among persons. For if there are not, then there will not be sufficient resources available to redistribute widely, and there will not be sufficient will on the part of individuals to take advantage of collective resources when they are required. In other words, everyone has to put something into the pot when they can, in order for everyone to be able to take something out when they need to. This is a collective goal, and a goal which could, as far as I can tell, very easily pull a society together and develop the cohesion that the report's authors consider important.