Some have argued recently that there's no such thing as race. Anthony Appiah's In My Father's House and Color Conscious are probably the two most notable discussions. Scientists often take this view without realizing all the philosophical leaps in reasoning they've made to get to the view (see, for example, this article, registration required). Others think it's merely a social category (most philosophers who write on the topic, usually on the same basis as the scientists above but with more sensitivity to issues about human language and social catgegorizations. There are also two possible positions according to which it's a genuine biological category, one of which I think is easily refuted by the data. The other seems to me to be a legitimate view that hasn't been discussed by philosophers or scientists to my knowledge (though I think economist Thomas Sowell may have suggested such a view -- I'm not quite sure yet). Below are the arguments for developing and sorting out these various views.
Before you read this it might be helpful to look at my first two posts in this (sort of) series. First is a set of cases to test your intuitions on racial classification, with the followup giving the data from my students' answers to those same questions.
Appiah presents three pieces of information worth considering just to get the scientific data. (The above-linked science discussion also has some helpful stuff about species variation when compared with variation between sub-species in other animals. The conclusion is that racial groupings in humans have much less variation than sub-species in other animals, genetically speaking. This, however, is a poor argument for the conclusion that there are no groups such as races. All it should show is that there aren't nearly as many or as significant variations between human races than sub-species in other animals. It doesn't follow that groups such as races don't exist.)
Appiah (In My Father's House, pp.35-36) presents one important fact. There's a measure called J -- homozygosity. It measures the likelihood of a randomly selected gene being held in common by any two randomly selected people of a given population. It's .852 for the human population as a whole and .857 for the population of Caucasions. So the chances of differing on any randomly chosen allele on a randomly chosen gene between two randomly chosen Caucasians is 14.3%, while it's only 14.8% for any two randomly selected human beings, not a great difference.
In Color Conscious, Appiah gives further data. On p.69, in a footnote, he says that .2 percent of human genetic material will differ with any two randomly selected people (and yes, I know this differs from the above, so perhaps it's been revised with further research, or maybe it's measuring something different -- I know little enough genetics to say). Of that .2 percent difference, 85% of it will be variations within local populations, 9% variations within ethnic groups that aren't assigned different racial status (e.g. Italians vs. French), and only 6% due to differences between racial groups (e.g. Europeans vs. Asians). So that's 6% of .2%, which is .012% of all human variation. These facts make it difficult to think of race as purely genetically defined, since racial differences seem like such a small component of human diversity. Yet again it doesn't necessarily follow that there are no races, as we'll see.
Another fact on the next page (of the second book) complicates this for the racial distinctions between African Americans and white Americans. 40% of African Americans have Native American ancestry. 5% of white Americans have African ancestry. 20-30% of the genes of African Americans come from European and Native American ancestors. So even if we could identify groups based on ancestry, it would be difficult to assign a number of people full status in those groups if that's all it's based on. The mixed race case is a special case, and some say such people have no race but others do, but it turns out to be a special case that includes most black people in the U.S., a very large special case.
Now a few issues will be relevant to how you're going to respond to the data. This doesn't flat-out show that race is non-existent, because that depends on what race might be. There are at least four views I can identify, one of which I haven't found anyone holding.
1. Race is purely biological, and you can read it off our DNA. This would be the racial essence view. There's some element of racial groups that gets passed on genetically, and that thing is present in all descendants of people of their race. The racist version based on narratives of racial purity has black essences trumping white essences so that black ancestry makes someone black and therefore not white, even if white ancestry is present. A non-racist version might allow someone to be both black and white. I think both versions are clearly refuted by the above data, though I'm not here going to trace out carefully how that goes. This is the view many people on a popular level seem to have, though, but I've never seen a philosopher trying to argue for it recently. Immanuel Kant is a good example of someone who seemed to say this sort of thing, though without talking about DNA. In fact, many philosophers give him the dubious honor of first expressing this sort of view and thus list him as the author of the modern concept of race.
2. Race is based on biological characterists that are almost arbitrarily singled out over other characteristics that could just have easily been singled out. Social practices and historical factors do the work in singling these out. So it's a biological category, but it's picked out in a way that other categories that are equally significant genetically aren't picked out (e.g. the category that includes just those who have the gene for blue eyes, the gene for tongue curling, the gene for six-fingeredness, and the genes necessary for having a height of over 6 feet, most of which are not always actualized with the traits in question). We could have arbitrarily picked any such criteria and grouped people according to those categories rather than bone structure, hair type, and skin color. But since we didn't, the reality of the groups we have picked out does remain and is based in real biological characteristics. To differ from #3 below, this view would have to argue that these biological characteristics are the main ways of distinguishing between races and not some of the other factors in Mills's 10 cases (from Blackness Visible, ch.3; see the links to my earlier posts above for the cases). I'm not quite sure how to sort through those issues yet, but I've never seen a philosopher arguing for this view, and it seems to me to be worth having on the table at least.
3. Race is purely a social grouping on the order of liberals, basketball teams, philosophy departments, first-year college students, and fomer presidents of the United States. It's based not so much on people's views about the nature of race as on the social practices of assigning racial categorizations. This view would have to distinguish itself from the second view by looking at cases like those of Charles Mills and showing that those traits named in view 2 aren't the only or even the main issues involved in racial classification. This is Naomi Zack's view as far as I can tell (as judged by Thinking About Race), though she thinks with borderline and mixed cases the person has no race (see her Race and Mixed Race for those arguments). I haven't seen her Philosophy of Science and Race yet, but that has more detailed discussion against a biological view. Most philosophers who write about this subject take this view or one similar enough to it (e.g. Charles Mills and Linda Alcoff more straightforwardly place themselves in this group, though Mills has a more complicated taxonomy than mine). The main idea seems to be that our terms get their meaning from how we use them, not what we believe about them (which would lead to view 1 for most people). If we do indeed use them in ways that classify people into genuine groups, even if there are borderline cases, then it seems to me that this view is at least on the table.
4. Race is purely fictional. There's nothing that corresponds to our notion of race. Our notion of race is #1 above, and nothing fits that, so the terms must not refer to anything. This is Appiah's view. To argue for this view, you'd need to look at Mills's cases and show that our notion of race doesn't allow us to hold either 2 or 3.
The main obstacle to this view has to do with the nature of language and how our terms refer to things in the world. We have many terms whose concepts associated with them don't match up to reality. This is particularly true with science. When people thought of water as an element, does that mean their terms for water didn't refer to water? When people thought of heat as a substance, does that mean their terms for heat didn't refer to average kinetic energy? Most philosophers think there's an external aspect of how our words get their meaning, having to do with what the best candidate is really out there to refer to based on our usage of the term. The terms 'evening star' and 'morning star' refer to the same thing, the planet Venus, since that's the thing that people point at when they use those terms. At one point people who looked at the sky didn't know they were the same thing. Do their concepts make it that they're not the same thing? Most philosophers say no.
If you then apply this to race, then you have to say we should find the best candidate that does exist that our terms refer to. To take view 1, there needs to be such a thing as a racial essence, so that doesn't seem an option. View 2 is possible, since we can classify people according to biological traits, even if we think they're arbitrarily selected. Still, those groups do exist, even if there are borderline cases (which there are of most groups, and it doesn't stop there being such a group; it just leads us to wonder whether certain people or things are genuine parts of that group). The question is whether that group is the most likely referent of our terms for racial classification, and that might require looking at how we actually use the term. It seems as if the social groups that allow more factors to be considered than just biology would seem to be a possible candidate also. You have to sort out which more closely fits our usage. To argue for view 4, you'll need to show that both groups are very poor candidates for reference of these terms or that somehow the history or concepts involved with view 1 infect the terms still in such a way that this argument will get resisted. That's what Appiah tries to do.
One other argument in Appiah is, I think, misguided. He says all sorts of morally illegitimate attitudes are attached to our racial terms. Therefore, we should stop using them. (This is a different point from saying that no group corresponds to those concepts.) The problem with this is that we can't address racial injustices that still go on, against any group, if we can't admit that the group exists. We have to refer to the injustice by naming it and describing it. Patricia Williams frequently makes this point, mostly when talking about normative whiteness and how the very idea of stopping talk of race is used to hide unconscious racist practices, including racial narratives, that will no longer be mentionable or observable to those who remove race terms but will still go on without the terms. In some ways that seems even worse. She envisions a time, possibly, when that will happen, but we're not there. Also, it doesn't allow us to affirm people for who they are if the groupings do exist, and that's as crucial for solving racial problems as it is to point out faults. [My own take on such issues is that there might well be something evil built into some notions while still it being necessary to use those notions looking forward to a time when those notions are no longer necessary. That may well be what's going on here.]
Also, if the group exists then we're not speaking falsely when using these terms, and that's the primary issue I'm looking to answer here. Even if it's morally illegitimate to use the terms, the groups might still have reality, and if so then the terms still refer to those groups. Therefore the moral argument doesn't give us any way to say there's no such thing as race, just to say that we hope there will one day no longer be social practices that allow us to use such terms. The other arguments would need to do the job to show that there aren't such groups.
In future posts, possibly, I'll look at the various suggestions I've made here to sort out more specific questions. This is the general state of the topic as I've been able to observe in current discussion, and as I develop my own work on the issue perhaps I'll have more to say.