A Letter to Dr. Michael White

There are debates in linguistics about how to categorize languages and dialects; Nicholas Wade has reignited the debate over how to categorize human populations.

Michael White’s recent article is titled  “Why Your Race Isn’t Genetic,” although at the end of his essay, he writes, “Without natural genetic boundaries to guide us, human racial categories remain a product of our choices. Those choices are not totally arbitrary, biologically meaningless, or without utility.” So, perhaps a better title would have been “Why Your Race Isn’t Only Genetic.”

I recommend you read the article before reading the following letter.

– – – – – –

Dr. White,

In your article, you cite Templeton’s “Biological races in humans,” where Templeton argues that all humans share a common lineage and that races are not sub-species because the five major ‘races’ of humans account for only 4.3% of cross-population human genetic variation—well below the 25% threshold set for sub-species categorization. But, later, Templeton himself writes that “this finding does not mean that all human populations are genetically identical. Past founder events, isolation-by-distance, and other restrictions on gene flow ensure that human populations are genetically differentiated from one another, and local adaptation ensures that some of these differences reflect adaptive evolution to the environmental heterogeneity that our globally distributed species experiences” (9).

So, there are genetic differences between human populations, but the argument offered by you and Templeton is that those differences don’t meet the standard for classifying different human populations as sub-species. I accept this argument completely insofar as “sub-species” is given an objective cut-off point, but it still doesn’t tell us how to classify (or whether we should classify) the differences that do exist between human populations. With that point in mind, here’s my first extended question:

I mentioned dog breeds on Twitter, and it seems that variation across breeds is near 27%, while human genetic variation has been found to be somewhere on the range of 5-10% (Parker et al. 2004), though, as just noted, Templeton puts it at 4.3%.

However, all of these numbers take large swaths of humanity (or dog breed-dom) into consideration. Ostrander and Wayne 2005 note (Figure 2) that within certain clusters of breeds, there is considerably less variation between one breed and the next. Two other papers (Erdogan et al. 2013 and Ye et al. 2009) have shown that cross-breed genetic variation drops well below 25% in certain contexts. Genetic variation between labs and springer spaniels, for example, is set at 0.09.

That dog breeds are the results of artificial breeding is inconsequential for this discussion about categorization. We know a priori that the notion of “breed” in dogs is a valuable classification system. So, if it’s true that among certain breeds, cross-breed variation drops well below 25%, then why isn’t it possible to have such a classification system to describe variation among human populations, which likewise drops below the 25% threshold for sub-species categorization?

My next extended question is related to this idea of variable genetic distance:

You make much of the fact that human populations are fuzzy and not distinct; the reticulating nature of our human family tree makes any kind of intra-human categorization moot:

. . . . But as it turns out, our species’ family history is not so arboreal. Geneticists have methods for measuring the “treeness” of genetic relationships between populations. Templeton found that the genetic relationships between human populations don’t have a very tree-like structure, while chimpanzee populations do. Rather than a family tree with distinct racial branches, humans have a family trellis that lacks clear genetic boundaries between different groups.

But doesn’t the truth of this statement wax and wane depending on which parts of the human family tree you’re talking about? I could be wrong here, so it’s an honest question. There can be fuzzy boundaries in Northern Europe and fuzzy boundaries in Southwestern Africa, but does that mean that the boundary between populations in Northern Europe and populations in Southwestern Africa is equally fuzzy as when comparing within those geographic boundaries? Isn’t this the point of Figure 2 in Templeton’s paper?

There’s a lot of fuzziness between dialectical boundaries in English and dialectical boundaries in Ojibwe, but not nearly as much fuzziness between English and Ojibwe. If boundaries are as universally unclear as you imply here, what’s the use of FST scores, and how is it that scientists manage to know a person’s ancestry down to a small geographic area?

Templeton argues for an isolation-by-distance model of human genetic variation, and I don’t disagree at all. But isolation-by-distance plus small but not negligible amounts of allele frequency variation between populations . . . . sounds a bit like allopatric speciation to me, at least when you’re comparing the far ends of that isolation cline?  But then, I’m a linguist, not a biologist, so I’m willing to be corrected.

Final question:

We can all agree, I think, that there is variation in allele frequencies between human populations, and that geography is a decent proxy for the occurrence of those frequencies. (This is all that I, and most people, mean by “race.”)  Where we disagree is on whether or not that variation is worth codifying with a classification system. Some people think it is; you think otherwise.

Not all human populations are genetically identical, and insofar as some of us think the study of genetic differences in human populations is interesting, we need a word for those differences. If you want to abandon “race,” fine, but what word would you use? Or would you not use any word because you don’t think these differences are meaningful or worth studying?

9 responses

  1. We can always use whatever word the census and government uses. If they use race, we use race. If they use ethnicity, we use ethnicity. Whatever terminology of human population subgroupings they utilize to administer affirmative action quotas, heritage plus and minus factors for admissions, diversity requirements or disparate impact litigation, and so on, well, that’s the word you use, with the cover that it’s the ‘officially correct’ word to use to describe the thing we’re trying to describe.

    But anyway, who cares about degrees of genetic variation? Just like with dog breeds, sometimes a lot of variation produces dogs that seem very close, and sometimes a few genes, or even a single mutation in a single gene, can produce dramatic differences. Maybe we need ‘effective race’ to go along with ‘genetic race’. For example, Jews could be considered an ‘effective race’ by some measures.

    What we’re all really wondering about is whether HBD matters. Yes, human population groups emerged by being more or less endogamous and more or less genetically isolated from each other for a long time, and for a variety of practical and historical reasons. They diverged genetically to some extent, especially in regards to the frequencies certain versions of certain alleles tend to exist within their population groups.

    Ok. But the question is whether those differential allele frequencies, no matter how many or few, or what you call them, have a definite impact on the differential means and variances of phenotypic and behavioral expressions between those population groups.

    In other words, is there a true stereotype? A strong ad meaningfully predictive statistical relationship that allows us to distinguish between two subgroups based on a heavily heritable and genetically determined quality. A stereotype can be as simple as “East Asians almost all have epicanthic folds, whereas very few South Asians do.” How many bell curves can you draw in which the groups are distinguishable by at least half a standard deviation?

    If you have a lot of separation on lot of important bell curves (however ‘important’ is defined, maybe in terms of life outcomes metrics), and all of those are strongly correlated to the heritage of genetically isolated groups, then you’ve got a good candidate for a ‘effective race’, no matter how many genes actually differ.

    June 27, 2014 at 3:51 pm

    • Sorry, forgot the last paragraph.

      It’s when there is good overlap between ‘genetic race’ and the above definition of ‘effective race’ that we enter into the realm of HBD and DE insights and red pills that contradict progressive orthodoxy on human neurological uniformity.

      June 27, 2014 at 4:14 pm

  2. The isolation by distance and the few detectable cases of selection driven fixation or change in allelle frequency are not wholly connected. The Wade case for the genetic component of race tries to bridge the gap between lets say five rather distinct human clusters and the stories he’d like to tell about the achievements of somewhat segregated populations in the past 300 years, but its hooey.

    in the past 300 years clear artificial selection did create a bunch of domesticated breeds, so yes there are many profound characteristics that separate the french poodle from the british bull dog; but the british empire or industrial revolution is not the outcome of some distinct set of selection pressures that gave the british a bunch of “empire building genes” (now sitting languid) and then finally resulted in the building of an empire bigger than the frenches. Thats not how natural selection works, but it is how artificial selection works. British people can drink milk and look generally more british than someone from north east china, but that really doesn’t tell you a whole lot beyond that.

    and what have we seen in the last 300 years? a stunning amount of admixture. Linguistically, economically, politically the results have been clear, but genetically just a lot of mashing about. Good luck telling me where in the united states a child is from.

    June 27, 2014 at 4:12 pm

  3. Pingback: A Letter to Dr. Michael White | Reaction Times

  4. fnn

    On the other hand:


    …Not only are the genetic differences within any one dog breed greater than the differences between any of them, the differences within a single dog breed, such as Chihuahuas, is actually greater than the genetic difference between two different recognized SPECIES of animals – wolves (Canis Lupus) and coyotes (Canis latrans). To quote James Serpell’s The Domestic Dog:
    ”Recently using genetic and biochemical methods researchers have shown domestic dogs to be virtually identical . . . to other members of the genus . . . Results using mtDNA (mitochondrial DNA) data . . . reveal startling similarities among canids . . . Greater mtDNA differences appeared within the single breeds of Doberman Pinscher or poodle than between dogs and wolves . . . to keep things in perspective, it should be pointed out that there is less mtDNA difference between dogs, wolves, and coyotes, than there is between ethnic groups of human beings.” (pp. 32-33)
    So contrary to Wise there are larger genetic differences between a West African, a Northwest European, and an Northeast Asian, then there is between two separate species, a wolf and a coyote, or between a Shih Tzu and a German Shepherd.

    …Wise also completely invented the fact that human races are less genetically differentiated than animal races. Using Sewall Wright’s genetic distance statistic, FST, the value of differentiation for human races is typically 12-15%, Wright suggested this figure is on the high to moderate end of genetic differentiation among other animal races. This figure, by the way, clumps together a number of low differentiation populations and drags the number down; if we compare just the two largest sub-divisions of the human species – Africans and non-Africans – we get a number of 25-30%, which is huge. So at the very least we can say the human species has two enormously genetically differentiated races compared with the rest of the animal world.

    June 27, 2014 at 7:21 pm

  5. fnn

    Wolves, Coyotes and Dogs don’t exist:


    …Now three researchers—Javier Monzón of SUNY Stony Brook, Roland Kays of the North Carolina Museum of Natural Science, and Daniel Dykhuizen also at SUNY Stony Brook have investigated the genetic ancestry of 427 canids from Ohio and the eastern states above the Mason and Dixon Line and found that all are admixtures of wolves, dogs, and coyotes, with the wolves representing eastern and western subspecies of gray wolf, and with the Great Lakes serving as a mixing/contact zone. Their article, “Assessment of wolf-coyote-dog admixture using ancestry-informative diagnostic SNPs,” appears in the January 2014 issue of the journal Molecular Ecology (subscription required).

    …All the animals that the researchers sampled showed evidence of wolf-dog-coyote admixture although the percentages of each varied by area. The mixing was not uniform; Ohio’s canids, widely believed to be pure coyotes, were 24 percent wolf and 10 percent dog.

    Those percentages are fairly constant across the study area, although they do show some variation based on geography and food supply, the researchers said. Thus, where forests and deer populations are dense, the canids are more wolflike, probably due to a steady diet of deer, not a regular part of the western coyote’s diet but standard fare for wolves. In areas more disturbed by human activities, coyote genes seem to increase.

    June 27, 2014 at 7:30 pm

  6. Pingback: This Week in Reaction | The Reactivity Place

  7. Rollory

    “Wolves, Coyotes and Dogs don’t exist”

    A statement such as this is known as reductio ad absurdum.

    There’s absolutely no need to read any of the evidence presented, because the conclusion is so visibly false.

    July 15, 2014 at 2:59 pm

  8. Pingback: This Week in Reaction – The Reactivity Place

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s