But perhaps the most important reason for building bridges is that human supremacism may not be an effective strategy for fighting dehumanisation after all. There is growing evidence that shows that a belief in human supremacy and species hierarchy aggravates, rather than alleviates, the problem of dehumanisation. The more people believe that humans are superior to animals, the more likely they are to dehumanise immigrants, women, and racial minorities.
Interestingly, the link between species hierarchy and dehumanisation is causal, not just correlational. For instance, when participants in studies are given a newspaper story reporting on evidence for human superiority over animals, the outcome is the expression of greater prejudice against human outgroups. By contrast, those who are given a newspaper story reporting on evidence that animals are continuous with humans in the possession of valued traits and emotions become more likely to accord equality to human outgroups. In short, reducing the status divide between humans and animals helps to reduce prejudice and helps to strengthen belief in equality amongst human groups.
On this basis he concludes that, "human supremacism is not only unnecessary to counteract dehumanisation, but is in fact counterproductive."
Unfortunately, as a basis for this specific conclusion, this is much more speculative than Kymlicka really recognizes. (That he is being more careful than most can be seen by his references to empirical research like that of Costello and Hodson, however; most discussions of dehumanization are based on models and schemes that are themselves highly speculative.) The research on which Kymlicka is basing his conclusions is mostly WEIRD; that is, it draws heavily on Canadian and American undergraduates, mostly white. There seems to have been very little research done on cultural variation on this point, and there's practically nothing on whether there are any differences if the notion of human superiority over animals is combined with other ideas. As a little thought on Kymlicka's example of the newspaper story shows, the kind of causal link the evidence he refers to shows is not the kind that we can assume has large-scale, stable effects; it is small-scale, possibly transient effects, that register in voluntary studies of this sort. Indeed, one of the problems with this kind of study is that it's not clear that the result has much to do with human-animal comparison itself, as opposed to just talking from the beginning about superiority and inferiority -- to consider just one possibility, it may well be that people are more likely to classify other human beings as inferior if they happen to have inferiority in general on the mind at a given moment; but for all that on its own tells us, this could be a momentary effect that can also be had if we talk about how one brand of toothpaste is superior to another. I have no idea if that's actually the case; but it's an obvious kind of thing that would need to be studied before we could get from the evidence that serves as the basis to the conclusions Kymlicka wants to draw. Maybe Kymlicka is right; but it's a lot of conclusion to base on small patches of evidence.
What is more, Kymlicka's argument fails completely to consider the complexity of human judgments on these matters. If you ask whether a bloodhound is superior to human beings in smelling, almost everyone will agree to this; indeed, if anything, they may well exaggerate the superiority of bloodhounds. (Human beings are unusually good at distinguishing things by smell; indeed, it is something of a mystery how we can be so good at it given that our physiological apparatus for it appears to have no obvious superiorities, and it is usually thought to be related to our linguistic ability to classify things. In scent discrimination tests, human beings can outperform almost all animals for almost all scents that we can actually smell. But there is a straightforward way in which bloodhounds are superior to us in smelling: they are very good at a range of scents we cannot easily smell.) Likewise, if you ask if a human being is superior to a cheetah in speed, people will say no. On the other side, almost everyone will say that human beings are superior to dogs in studying philosophy, writing poetry, doing mathematics, and thinking of the big picture. So is it specifically moral superiority that matters? Is it specifically being superior in the sense of being more morally salient, involving more duties and obligations? (Related to this, Kymlicka also doesn't consider issues that have arisen in terms of dehumanization, namely, that there is research that at least suggests that there are very different kinds of dehumanization that appear differently in different situations -- e.g., there are studies that suggest that people are much more likely to dehumanize for instrumental benefit than when specifically moral questions about desert are considered.) Are there combinations that matter, or is it just one or two particular types, or is it just the superiority and inferiority we happen to be thinking about at the moment?
There's another interesting feature of the research in question that I don't think Kymlicka has actually considered at all, even though he would need to consider it. In much of the work that has been done, there is an important asymmetry that is directly relevant to all of the conclusions -- thinking of animals as like humans seems to get you less dehumanizing results, thinking of humans as like animals gets you more dehumanizing results. This raises the worry that the asymmetry also affects the broader conclusion here: Kymlicka's own argument makes clear humans-as-like-animals moves. For instance, he says:
One reason is that supremacism is difficult to defend in our post-Darwin world. Darwin’s theory of evolution showed that humans and other animals are continuous with respect to our interests and capacities. If humans share ancestry with other primates, and if we share 98.8 per cent of our DNA with chimpanzees, then whatever makes human lives valuable or worthy of respect almost certainly has analogues in the lives of other animals.
This is framed directly in terms of human beings being like animals, which is the direction that tends, in the very studies Kymlicka is considering, to yield the worrying result. And it raises a question that Kymlicka seems to assume does not arise: How do you oppose human supremacy in the right way? He has not established, contrary to what he seems to have assumed, that his own framing of the issue is the correct way to do it. Kymlicka's humans-are-like-animals approach is itself at least suspect, on the very grounds to which he appeals; and it may well be that the solution is not anything like Kymlicka's argument but simply more empathetic learning about animals and the impressive things they can do.
The bigger issue, though, is it's unclear what normative conclusion can be drawn from any of it. Let us suppose that the phenomenon is completely established and widespread across different cultures, that it is not affected by other beliefs (e.g., by the belief that animals should be treated well). It is still a leap from this to the strong conclusions that Kymlicka wants to draw. For instance, it's probably true that the belief that human beings have significant differences from each other is part of the causal chain leading to treating some human beings as subhuman; at least, it's difficult to see how you could do it without assuming that human beings have significant differences. But it's a stretch to say that this gives us any reason to avoid thinking of human beings as having significant differences from each other; it's just obviously true that human beings have significant differences from each other, and, what is more, living a moral life actually sometimes requires taking such significant differences into account -- in accommodating disabilities, for instance. The fact, if it is one, may give you reason to be more cautious about it, and it may give you good reason to look into the possibility of compensating for this tendency, but on its own it doesn't actually give us any insight into what to do about it. So here. Besides eliminating a belief, there are many other things that you can do that can change behavior: attenuate it, add beliefs that compensate for it in some way, simply be more cautious about acting on it. In various circumstances, any of these may in fact be the better way to go. Nothing in Kymlicka's argument actually rules out any of these other options.
The Civil Rights Movement, as Kymlicka notes, deliberately went the route of making the species line significant; it did so not arbitrarily or by happenstance, but because the species line was already significant -- our very notion of 'subhuman' comes in part from eugenic discussions of races as intermediate between animals and humans proper. Kymlicka does nothing to consider this problem -- a problem that still exists -- or how his proposal would affect it. He merely appeals to vague considerations and speculations -- admittedly based on evidence, but it's not evidence that clearly removes the conclusions from the realm of speculation. He does nothing to consider the question of whether the Civil Rights Movement's experience that this was an effective line of appeal should also be counted as evidence on the opposing side.
We are not, after all, talking about general strategies for opposing dehumanization, which may go in any direction; Kymlicka's argument comes to the specific conclusion that "human supremacism is not only unnecessary to counteract dehumanisation, but is in fact counterproductive" despite his recognizing that our understanding of rights of human beings heavily depends on it. This is the conclusion that Kymlicka does not establish at all. He has not shown that it is all-things-considered counterproductive; he has not shown that it is genuinely unnecessary in practice; the evidence he appeals to shows, at best, that we should add a note of caution. No doubt there is more to be said; but we are a long way from any substantive and adequate argument.