That which must not be named

Constance Holden was at an October NIH workshop, titled, “Workshop on Ethical, Legal, and Social Issues in Natural Selection Research.”

I study recent selection in human populations. As I see it, my ethical obligation is to formulate and test clear hypotheses and to communicate clearly with the public about my research.

So I was laughing out loud reading Holden’s report. Apparently some of the participants don’t agree that “clear communication” is ethically or socially desirable:

Nothing makes scientists more nervous than the topic of "race," so much so that they'd like to find a way not to talk about it at all.
Everyone at the meeting agreed on the need for non-"fraught" terminology--"geographic ancestry," for example, instead of "race." But specifying such ancestries is also a minefield. "Amerindian," for example, is offensive to Native Americans, according to one speaker. "Caucasian" is also unacceptable because it implies racial rather than geographic ancestry. Some speakers even advised that it is inappropriate to refer to a "European allele" for lactose tolerance, because it also occurs in other groups.

So let me get this straight: The terms that scientists have made up in the past are now unacceptable because either (a) they’re strange and arbitrary names that only a scientist could think up (“Caucasian”, “Amerindian”), or (b) they actually have meanings that people think they understand (“race”, “European allele”). The solution to both problems? Make up new arbitrary terms that are so strange only a scientist could think them up (“geographic ancestry,” for instance, which like race substitutes a hypothesis – “ancestry” – for an observation – present geographic location).

The article illustrates the absurdities of this attitude:

Carlos Bustamante, a population geneticist at Cornell University, said that when his group published a study in Nature this year indicating that European-Americans had more deleterious gene mutations than African-Americans, some publications touted the report as suggesting that blacks are fitter than whites.

That’s a contradiction of a near-Orwellian level. The plain meaning of “deleterious” is “fitness-reducing.” They concluded that “European-Americans had more deleterious gene mutations than African-Americans” and we’re obviously supposed to know that this means something other than “blacks are fitter than whites”?

I had a major problem with that study when it came out, and I’ve submitted a letter on it – I think there are major errors in the analysis. But what exactly is the complaint here? It can’t be that the public has misunderstood the concept of race, because the authors explicitly divided their sample into race categories. That part is perfectly clear. You may disagree with their categories, but then its the authors who are confused, not journalists!

Biological variation and moral status

I would say that the misunderstanding here is a straightforward biological one: Many people – including far too many scientists – equate “fitness” with “value”. That is, “fitter” equals “better.” But in biology, fitness is a measure of relative reproduction. “Fitter” means “characterized by a higher intrinsic rate of population growth.”

The problem here is not that people don’t understand race, it’s that some people assume that there’s a value relation between fitness and moral status. And to that point, I cannot emphasize enough my fundamental agreement with Dobzhansky on the moral import of the science of human diversity:

Man is a highly variable polytypic and polymorphic species. The genetic variability affects behavioral traits no less than physiological and structural ones, and it is false to imagine that these three categories are clearly separable. The chief reasons why so many people are loath to admit the genetic variability of social and culturally significant traits are two. First, human equality is stubbornly confused with identity, and diversity with inequality, as though to be entitled to an equality of opportunity, people would have to be identical twins. Human diversity is not incompatible with equality. Secondly, it is futile to look for one-to-one correspondence between cultural forms and genetic traits. Cultural forms are not determined by genes, but their emergence and maintenance are made possible by the genetically conditioned human diversity.

I’ve written about these issues before. I think biologists (and others) make a serious mistake when they accept the argument that biological variation should inform our moral judgments. In particular, we often hear the claim that human races are 94% identical, or that people are 99.9% identical, or that we are all 98% identical to chimpanzees, and that therefore races, or people, or chimpanzees, should be treated as our moral equals.

This argument is fallacious, and moreover is dangerous. Moral judgments are normative, rule-giving, not empirical. I will not concede that some arbitrary level of biological similarity should dictate my moral obligations.

Why race is not irrelevant

There has been much discussion of race this week, after the U.S. Presidential election. The shorthand version of much of this coverage has been “race has become irrelevant.”

But of course that is emphatically not what has happened. Race in the United States remains a very important determinant of sociological, economic, and medical outcomes. If we want to explain these outcomes, or ameliorate them, we cannot assert dogmatically that either social or biological influences are irrelevant to them.

What has happened is that the voting public’s moral judgment changed. This did not happen instantaneously, and it did not happen this year. Nor has race has not become irrelevant – if the voting public truly had made race irrelevant, then the networks would hardly have needed to show images of tearful Jesse Jackson and Oprah Winfrey. But there’s no question that different in this case is better: during the last 30 years, people have acted more and more as if moral equality was not conditional upon race.

Some would say that this is a triumph of anthropology – that the post-WWII UNESCO statement on race, the

I disagree. Biological anthropologists did not prove that races don’t exist. They washed their hands of the race concept. For some, this may have been an attempt to purify the science from its racist history. For others, it was doubtless a way of making their studies of human variation seem more objective, less determined by social whims – despite the fact that many continued to describe and define supposedly biological populations based upon social characteristics like language, residence patterns or subsistence practices.

Perhaps some health disparities will be explained by recent natural selection. For example, maybe Type 2 diabetes risk may be partly attributable to recent selection associated with cereal agriculture. But any test of this hypothesis will have to deal with the social impacts of race: dietary, cultural and socioeconomic differences among living people.

Like I said, I don’t believe the study’s conclusion. If I did, the natural prediction is that Europeans and Africans therefore should differ in reproductive potential, or – depending on recent environmental changes – that they may have done so in the past. If we are skeptical of that prediction (as I am), then that should suggest that the study is incorrect in its conclusion.

Here’s the thing: There’s no question that “race” is a concept that inevitably includes both social and biological elements. And many people emphasize the biological elements while conveniently forgetting the social ones.

But these scientists are not giving people nearly enough credit. When I talk to the public about race, I find that almost everyone has experienced the contradictions that emerge, and readily share stories about them. It is easy to explain the difference between descriptive characteristics and predictive ones, and to illustrate why differences between groups do not always predict things about individuals.

More than anything else, people are suspicious when scientists refuse to use the word race. Because that’s usually a sign that the speaker is not being entirely truthful.

The most important finding about recent natural selection is that so many of the traits that distinguish human populations strongly from each other are very new. So it is especially important that we formulate and test clear evolutionary hypotheses that involve the elements

Instead, some geneticists would like to substitute a new version of the myth of pure races. They approach human variation with the viewpoint that

What I would say at such a meeting

What are the practical effects of a meeting like this? Clearly, NIH is not about to abandon race-based sampling schemes in its large medical studies. Billions of dollars (yes, billions) have been invested in such studies, which are ongoing. Some human geneticists have been trying to increase their influence on such investments, by promoting various kinds of ancestry testing.

The idea behind ancestry testing is that genetic markers can test the proportion of admixture of different populations in any one person’s genealogy. But then, that method itself assumes that there existed once pure races that have mixed in recent times. (I wrote about that problem here.) You can see this from the strong emphasis that many researchers place on population bottlenecks. Their logic is that ancient bottlenecks caused populations to become genetically divergent, and that those divergent populations later mixed with each other. In that theory, those ancient populations must have been isolated to have become so different. Therefore, they must have existed in different geographic locations. In other words, the present genetic variation in the world results from the mixture of ancient geographic races.

I would propose that we replace ancestry testing with a fuller understanding of population history. Ancestry testing is a statistical method that is trivial to run and overly simplistic in its assumptions. Its simplistic application encourages simplistic assumptions about population history. In fact, there is no anthropological evidence for continent-scale bottlenecks. Meanwhile, there is abundant evidence for recent natural selection. Selection may reflect geographic boundaries (like seas) but often it will not. It may

My fear is that meetings like this one, discussing the “ethics” of natural selection research, will serve only to reinforce the false idea that population history is irrelevant to biomedical outcomes. Research into the roots of natural selection on recent human populations can tell us where and when alleles of biomedical importance arose, and can give us some idea of why they were valuable to ancient people. Recent selection has mainly affected dominant genes of strong fitness effects. This argues that their phenotypic effects on some health outcomes may also be large – although in many cases, a health outcome may be a side effect of their once-selected function. And if an allele wasn’t selected, well, it takes research into population history to determine that as well!

Clearly, we should be careful how we communicate our results. But I think that obfuscation is itself unethical. As the article discusses, the public already has well-defined reasons why they are willing to fund biomedical research, and they are going to

Participants acknowledged that however they characterize their findings, they can't control what the public makes of them. "When translated into popular culture, society reads whatever term we pick as 'race,' " said Timothy Caulfield, a health law professor at the University of Alberta in Edmonton, Canada.

Ancestry is not the same as geography. The context of recent natural selection is not purely or even primarily geographic. Lactase persistence – an example in the article – is not a result of geography; it is a result of culture history through subsistence practices. Malaria, perhaps the strongest source of selection in recent history, is geographic only in the sense that the mosquito species that serve as vectors don’t winter well. But its severity and incidence in different human populations have been consequences of land use patterns (e.g., swidden agriculture, swamp draining), technology (e.g., pottery containers, stilt construction), and human movements (e.g., circum-Mediterranean trade networks). Even pigmentation – the quintessential “geographic” trait in humans today – is not only about geography. If it were, then Siberians would have blond hair and blue eyes.

If we pretend that ancestry is mainly about geography, then we ignore the biocultural interactions that have driven human history. I happen to think that those interactions should receive much more attention, because they have generated biomedically important variation in living populations. I think that understanding past environments – including cultural environments – may be very important as we try to understand and cure human disease.


Holden C. 2008. The touchy subject of "race". Science 322:839. doi:10.1126/science.322.5903.839a

Dobzhansky T. 1963. Anthropology and the natural sciences -- the problem of human evolution. Curr Anthropol 4:138+146-148.