Based on the seeming evidence of their eyes, most non-biologists believe there are real human races. Relying on this, as we saw above, Sarich and Entine ignore the proper definition of ‘race’ and tell you, in the name of science, that you can believe what you see. When you have a distorted perception, but you nevertheless believe that what you supposedly see is real, you are caught in a Flat Earth Fallacy.

And yet it is important to understand that Sarich and Entine, in a remarkable two-for-one, are also committing the Unicorn Fallacy. (As explained earlier, you are caught in a Unicorn Fallacy if you believe that something exists in the world merely because you can imagine it.)

Let us consider more closely Vincent Sarich’s position, which Entine relies on, as stated on page 110 of Taboo:

“That there may be no universal agreement on the name, number, or precise qualities that define each race does not undermine the reality of race. To escape this trap, Vincent Sarich has suggested that it might be useful to think of races as akin to what mathematicians call  ‘fuzzy sets’. . . .”

Useful to whom? Who needs to “escape this trap”?

A ‘fuzzy category’ (or, as mathematicians put it, a ‘fuzzy set’) is simply a category that doesn’t have crisp boundaries. For example, take the categories YOUNG and OLD. Is a 37-year-old man in the category YOUNG or in the category OLD? There is no clear answer because, although people gradually lose membership in the first category and gain it in the second as they age, there is no point at which they make a clean jump between YOUNG and OLD. Cognitive scientists call such categories ‘fuzzy’ because their boundaries are just that—fuzzy.

Entine concedes that human races blend smoothly into each other, and also that this creates a trap for him that he must escape. Why? Because Entine still wants to convince you that human races do exist, despite the facts. So although “there may be no universal agreement on the name, number, or precise qualities that define each [alleged] race,” Entine asserts that this supposedly “does not undermine the reality of race” because some categories, after all, can have fuzzy edges.

You’ve just been conned.

Yes, some categories can have fuzzy edges, but so what? The issue here is not what kinds of categories we can construct in our minds, but what sorts of things really exist in the world. In the real world biologists have a real vocabulary according to which a real race exists only if you can identify a real discontinuity. Since the meaning of ‘discontinuity’ is ‘the opposite of smooth change,’ it follows that the category RACE, as defined by biologists, is not fuzzy. And since it isn't, if you want to answer the question, “Are there biological human races in the real world?” it doesn’t matter one whit whether Vincent Sarich can construct fuzzy categories in the privacy of his own mind, as undoubtedly he can.

What Sarich and Entine are trying to sell you is a clever version of the Unicorn Fallacy. By shifting the focus of attention from the measurement of human variation in the real world to the making of categories in the human mind, Sarich is trying to convince you that a real human race exists if you can imagine a fuzzy one. If that argument were valid, then unicorns would exist, because I have no trouble imagining a unicorn.

Do Sarich and Entine have an argument of last resort here? It would have to be something like this:

“Okay, Francisco, but you are missing the point. There are, after all, genetic differences among humans. For example, no man is genetically identical to his father. People in one area will have some average genetic differences from people far away, and so forth. The point we are making is that differences in people’s physical appearance (skin color, hair texture, etc.) correlate to differences in their overall genetic makeup, including things we cannot easily observe. Okay, so our lay categories of ‘black,’ ‘white,’ and ‘yellow’ are not exactly ‘races’ (if you insist) as biologists use the term. But understood as fuzzy categories with a prototype, they do help us identify and keep track of changes in overall human genetic variation by looking at changes in surface appearance.”

Entine does make this argument. To understand it properly, consider briefly the category MIDDLE AGE, which is a ‘fuzzy category with a prototype.’ It is fuzzy because it is impossible for me to say where exactly MIDDLE-AGE begins. And yet I have no trouble saying what I think is the perfect example of a middle-aged person: ‘fifty years old.’ So for me ‘fifty’ is ‘typical’ of MIDDLE AGE—or, to use a more imposing term, prototypical. Categorization theorists refer to the ‘best representative’ of a fuzzy category as its ‘prototype.’

Entine wants us to think of human variation as being organized into fuzzy categories with prototypes (he calls these prototypes “core races”):

“Since the gene flow as a result of intermingling on the fringes of population pockets was only a trickle [at the beginning of alleged racial differentiation], relatively distinct core races would likely have been preserved even where interbreeding was common.”[1]

Entine’s point here is geographic. He is saying that black people are in Africa and white people are in Europe. Where Africa and Europe come close together we have “the fringes of population pockets.” And away from these fringes we have the so-called “core races.”  In other words, just as a 40-year-old man is not entirely young and not entirely middle-aged, the people in North Africa are neither completely black nor completely white, because, says Entine, they are “intermingling on the fringes of population pockets.”

Unfortunately for this argument, a scientist wishing to construct fuzzy categories with prototypes that are rigorous must answer the question: What are the prototypes at the center? What, for instance, is the ‘best example’ of ‘black’? That is, who is a ‘pure black’ and therefore a member of the ‘core race’? Kenyans? Why not Pygmies? Hmm…

Neither Entine nor Sarich ever answers this question. In fact, they never even ask it.

How come? Surely the issue is crucial. If features that we can see—skin color, hair texture, etc.—are indicators for overall biological differences, then the largest ‘racial’ biological differences (in athletic ability or whatever) should surface when comparing the prototypes or “core races.” The way to test Entine’s argument would be, first, to find these supposed “core races” and then show that the differences in athletic performance (or whatever) between core races are larger than the differences at the fringes. But neither Entine nor Sarich appear the slightest bit serious about even attempting to construct ‘fuzzy races’ with prototypes and then testing their scientific usefulness. Merely two pages after introducing the idea, Entine wanders off to another issue.

Why? Did he merely intend to plant the notion that human races do exist in the minds of lay readers, his target audience? Is he counting on the high probability that these lay readers won’t have expertise regarding categorization, on the one hand, or the biological concept of ‘race,’ on the other, much less both simultaneously?

Since Entine has dropped the ball, let us pick it up for him. What would Entine and Sarich have to do if they were really serious about using this concept they put forward—this ‘fuzzy race with a prototype’? Well, Entine’s first task would be to explain away the embarrassing fact that Moroccans and Algerians are the kings of speed in intermediate distances, and Ethiopians are among the best distance runners. Why is this embarrassing? Because they are at the fringes of what Entine calls the ‘black’ and ‘white’ races. So this is a contradiction. If ‘blacks’ are better at ‘sports,’ as Entine claims, then those who are ‘less black’—Moroccans and Algerians—should not be running faster in mid-distances than both ‘whiter’ and ‘blacker’ athletes. And neither would we expect those who result from a melting pot of African and Middle-Eastern mixing—Ethiopians—to win more Olympic gold medals in track events than any other African country except for Kenya and Morocco (and that’s despite missing the 1976, 1984, and 1988 games!).

Even if there was such a thing as a black “core race,” Moroccans, Algerians, and Ethiopians would be the least likely candidates. So much, then, for the argument that “black athletes dominate sports.”

Not for the first time, Entine’s arguments have failed after we generously granted all of his assumptions and did his work for him. And, not for the first time, the problems are even deeper. To see why, let us move beyond the issue of sports performance to ask the question: Is it even possible for Entine to say what the “core race” might be, in each case?

Ask yourself, what is the ‘best example’ for ‘white’? Scandinavians? Spaniards? Turks? Of these three, which one did you choose? I bet you picked Scandinavians because they appear most different from so-called ‘black’ people (and if so then the best example of ‘white’ would be in fact Icelanders, who flirt with being albino).

This, however, will not do. A scientist needs to choose a prototype on the basis of the relative representation of traits, not the gut feelings of laypeople. A rigorous and systematic examination of the morphological variation (differences in external traits) among so-called ‘whites’ will reveal that the appearance of Scandinavians is not typical (especially in the case of Icelanders), so they cannot be the prototype on the basis of typicality.

But neither are Scandinavians central (another possible criterion of prototypicality). ‘Fifty’ is central to my own category MIDDLE AGED because numbers become less good examples of MIDDLE AGED going in both the younger and older directions. Since Scandinavians are at one of the poles of variation of what Entine would like to call the ‘white race,’ rather than at the center, they cannot be the prototype on this basis either.

Who then might be representative of ‘white’? As it turns out, choosing a central type for what people call ‘white’ or ‘black’ in a rigorous, scientific fashion is a fool’s errand. This is because the different traits (e.g. nose width, eye-type, lip thickness, skin color, etc.) do not covary systematically in bundles, which is what you would need to find the center and endpoints of such ‘fuzzy races.’

What does this mean? Two variables ‘A’ and ‘B’ are said to covary if high numerical values of ‘A’ tend to correspond to high numerical values of ‘B.’ So, for example, if thicker lips were almost always associated with flatter noses, then those two traits would be said to ‘covary highly.’ But instead of finding this sort of thing, what we find is that the physical traits vary rather independently of each other in different directions. Our brains try to fit people into neat categories and therefore find it hard to notice this. But here is one obvious example of how traits can be combined in ways that violate our racial prejudices: the skin of people in southern India is just as dark as that of Africans, and yet they have straight hair. Another example: entire populations in southern Africa have epicantic folds over their eyes, just like people in Asia, and yet they have dark-skin and wiry hair.

An argument of last resort here could be that we don’t need to find the typical whites but those who are most ‘white.’ If Scandinavians are at one pole of variation among ‘whites’ then perhaps this is because they are less mixed, and therefore truly, purely ‘white.’ Ah, but how can we know for sure? We cannot just assume it, can we? We will have to look at the genes and see whether Scandinavians are both ‘whiter’ than other Europeans and also less related to non-‘white’ populations.  

Well are they? No. The Finns, for example, are certainly light-skinned Scandinavians, and to all appearances European. But among that small part of human genetic stock that does vary, “one quarter of the Finns’ genetic stock is Siberian” (and of course, Siberians look very ‘Asian’).[2]

One must resist the temptation to be shocked by this factoid. The genes you can see on a Finn are the tiniest fraction of the total number of genes inside him, because only a very few genes (those responsible for skin color, lip thickness, etc.) have effects of the sort that a human can easily perceive from the outside. Therefore, it is the easiest thing in the world for the Finns to be closely related to the Siberians and for this not to be in any way apparent.

Well, if we can’t trust our eyes. . .

If we can’t trust our eyes but we wish to know if there are human ‘races’ as modern biologists use this term, all we can do is look at the genes of humans around the world and see if they parse into any races. Morphology (outward appearance) was only supposed to be a guide to finding these supposed races anyway, so let us look at the genes directly.

We may do this. But even if such a project were to uncover genuinely biological human races, it is important to point out what it could not do. As we have seen, the lay concepts of ‘black,’ ‘white,’ and ‘yellow’ cannot be turned into rigorous morphological categories. Therefore, the boundaries of any races, if found, would surprise ordinary Americans, because such boundaries would contradict what Americans think they see.

This is the umpteenth death-knell to Entine’s argument, for he is not merely trying to convince you that there are some human races, whatever they may be. His project, rather, is to convince you that the racial categories which you already think you can see correspond to genuinely biological (i.e. genetic) racial cuts of the human species. That argument fails, but so does the broader claim that any races exist. For, as we shall see below, there is simply no biological cut of the human species that allows us to speak of any human races.

Biologists have concluded that “there are no human races.” Entine, the journalist, disagrees...
____________________________________

Luigi Luca Cavalli-Sforza and colleagues conducted a celebrated world-wide survey that is the best assay we have so far of the distribution of human genetic variation.[3] Their analysis includes many different genetic loci (this word refers to the location on a chromosome where a particular gene will be found) and is therefore not biased (as our eyes are) by the very few loci that underlie the human traits that we can see, such as skin color. In Taboo, Entine refers to Cavalli-Sforza’s data and claims that it proves human biological races are real.

This requires some nerve.

There is very little genetic variation in the human species. The argument that humanity can be separated into races that developed on the continents of Africa, Europe, and Asia requires that a significant portion of the little genetic stuff that does vary show relatively sharp changes along the continental boundaries. This, however, is not the case. Here is what Cavalli-Sforza and his colleagues reported in 1997:

“It is often taken for granted that the human species is divided in rather homogeneous groups or races, among which biological differences are large… [but] Differences among continents represent roughly [just] 1/10 of human molecular diversity, which does not suggest that the [alleged] racial subdivision of our species reflects any major discontinuity in our genome.”[4]

Since a “major discontinuity” is precisely what is required for there to be a human ‘race,’ as biologists use this word, it follows that, based on the findings of Cavalli-Sforza and his associates, there are no human races.

Now, Cavalli-Sforza is recognized as the worldwide expert on this topic, so Jon Entine is forced to describe him, on page 108 of Taboo, as:

“. . .[the] renowned. . .Stanford University geneticist L. Luca Cavalli Sforza…[whose 1994] thousand-page classic of population analysis, The History and Geography of Human Genes, [was] coauthored with Paolo Menozzi, and Alberto Piazza. . .”

However, having admitted that Cavalli-Sforza is the authority on the geographical distribution of human genes, Entine writes:

“Although Cavalli-Sforza and his colleagues go out of their way to say that ‘the concept of race has failed to obtain any consensus,’ their research belied such an assertion.”

So the journalist says that the renowned population biologists are wrong about their own data. Not only that. According to Entine, “Cavalli-Sforza and his colleagues go out of their way” to misread their own data, so Entine is apparently suggesting that these scientists are driven by a political agenda concerning the question of ‘race’ (savor the irony). 

Earlier we saw that, in Taboo, Jon Entine tries to prove the supposed sports-superiority of blacks by using maps that specifically refute his arguments. Now he does the same with Cavalli-Sforza’s data, and once again cheerfully walks right off a precipice.

On page 109 Entine produces a “Genetic Tree of the World’s Populations,” adapted from Cavalli-Sforza et al. (1994), and claims that it shows the human species divided into biological races. But it doesn’t.

Cavalli-Sforza’s trees show, for any geographically defined human population (say, North Asians) whether it is genetically closer to a second population (e.g. South Asians) than it is to a third (e.g. Europeans). But what these trees lack entirely is any information concerning the magnitude and sharpness of the differences between any two populations, and it is precisely this information that is needed to decide if a population is a biological race.

To see this a little better, consider the following. It stands to reason that my brother and I are more genetically similar than either of us is to our third cousin, but that hardly means my brother and I are in one race, and our third cousin in another. The same is true with populations. Cavalli-Sforza’s trees are a bit like the genealogical tree that would show my brother and I as more closely related to each other than to our third cousin: they show that two local populations are more genetically similar than either is to a third population which is farther away. However, these trees include no information about the magnitude of genetic differences between populations, which is why they can neither support nor undermine the claim that biological human races exist.

Entine, therefore, is presenting non-evidence as his ‘evidence’ that races exist…

Nevertheless, Cavalli-Sforza’s trees do have some important information relevant to Entine’s specific claims. As you will recall, Entine is not only saying that there are human races, he is saying these supposed races correspond to the ordinary American categories of ‘black,’ ‘white,’ and ‘yellow.’ For Entine’s argument to be correct, Cavalli-Sforza’s trees of relative genetic similarity between populations should look very similar to trees based on the ‘black,’ ‘white,’ and ‘yellow’ categories. But in fact they look quite different. Indeed, there is no arbitrary set of morphological features (features that we can see with our eyes) that corresponds to the overall patterns of biological variation documented by Cavalli-Sforza. For example, although our eyes intuitively group North Asians together with South Asians, Cavalli-Sforza’s data actually place North Asians as genetically closer to Europeans than to South or Southeast Asians. Similarly, though Australians and New Guinea Islanders look quite African indeed, they are more closely related to the ‘yellow’ populations of Southeast Asia.

And I note that the tree from Cavalli-Sforza et al. (1994), which Entine proudly reproduces on page 109 of Taboo, in fact shows very clearly the two points I make above concerning North Asians and Australians. So we might say, expanding on a previous metaphor, that Entine has walked his readers to Cavalli-Sforza’s tree, plucked an orange, held it high for all to see, bitten fiercely and then, with his mouth full of orange peel, exclaimed joyously: “Apples: De-li-cious!”

The final nail in Entine’s coffin is that, although the sweep of the world does show that there are some genetic differences between regions (how could there not be?), there are no sharp boundaries, only smooth clines. But as you will recall, by definition races do not exist unless we can find relatively sharp boundaries, and as Cavalli-Sforza notes in the excerpt quoted earlier, our best data “does not suggest that the [alleged] racial subdivision of our species reflects any major discontinuity in our genome.” The upshot is that human races do not exist—neither the ones Entine argues for, nor any others. (A clear and summarized exposition of these points, using the same data, can be found in Boyd & Silk’s book How Humans Evolved, chapter 16).

Now, if our best recent data supports the claim that human genetic variation is very smooth, and that our eyes mislead us as to the relative genetic proximity of various populations, why does Vincent Sarich—Jon Entine’s guru—insist that we can divide humans into races with an eyeball test, the same as we do with dog breeds?

I take up this question next.

»» Continue to Chapter 4:
http://www.hirhome.com/rr/rrchap4.htm
________________________________________________________

Footnotes
________________________________________________________

[1] Taboo (p.111)

[2] “As early as the 1960s and ’70s, Finnish researchers made the significant discovery that one quarter of the Finns’ genetic stock is Siberian, and three quarters is European in origin.” -- from “Where do the Finns come from?”, Written for Virtual Finland by Christian Carpelan, Licentiate in Archaeology and a researcher at the University of Helsinki;
http://virtual.finland.fi/finfo/english/where_do.html

[3] Cavalli-Sforza, L. L., P. Menozzi, et al. (1994). The history and geography of human genes. Princeton, NJ, Princeton University Press.

[4] Guido Barbujani, Arianna Magagni, Eric Minch, and L. Luca Cavalli-Sforza (1997) “An apportionment of human DNA diversity.” Proc. Natl Acad Sci U S A. 1997 April 29; 94(9): 4516–4519.