Danny Yee >> Anthropology >> anthro-l
Essential reading: Cosma Shalizi on g "general factor" intelligence and the heritability and malleability of IQ.

The race/intelligence/IQ debate

December 1997

Claims that "Blacks" are dumber than "Whites", a la The Bell Curve, crop up every so often on anthro-l. In a way I'm responsible for the latest outbreak, since it began when I posted a review of Jared Diamond's Guns, Germs and Steel.

In order to save time and energy next time this comes around, and to provide a resource for people interested in the issue, I am collecting what I consider to be the most interesting contributions. In particular, I have left out messages which are overly abusive, or purely ad hominem. Some messages have been edited slightly, to remove personal comments and redundant, irrelevant, or less interesting sections. The full text of the originals, and all the other contributions to the debate, can be found in the archive.

My own biases should be obvious. (See also my review of Meritocracy and Economic Inequality.)


J. Philippe Rushton

(an extract from a much longer review)

Although independent researchers have repeatedly confirmed: (1) The geographical distribution of intelligence, (2), the relationship between intelligence and brain size, (3), the geographical distribution of brain size, and (4), the heritability of intelligence, Diamond, the author of The Rise and Fall of the Third Chimpanzee, is like a composite of the three wise monkeys and does not want to see, hear, or say anything about these topics. Therefore, I will briefly summarize them. Readers seeking a more extensive summary can consult The Bell Curve, and for a complete discussion of how brain size and IQ explain much of human behavior and are in turn explained by human evolution, see my Race, Evolution, and Behavior.

1. The geographical distribution of intelligence. One hundred years of research has established that East Asians and Europeans average higher IQs than do Africans. East Asians, measured in North America and in Pacific Rim countries, typically average IQs in the range of 101 to 111. Caucasoid populations in North America, Europe, and Australasia typically average IQs from 85 to 115 with an overall mean of 100. African populations living south of the Sahara, in North America, in the Caribbean, and in Britain typically have mean IQs from 70 to 90 (see Lynn, 1997, for a comprehensive review).

Parallel differences are found on relatively culture-free tests such as speed of decision making. All children can perform the task in less than one second, but children with higher IQ scores perform faster than do those with lower scores. Asian children in Hong Kong and Japan average faster than do European children from Britain and Ireland, who in turn average faster than do African children from South Africa. This same pattern of racial differences is also found in California.

2. The relationship between intelligence and brain size. Diamond neglects to mention any of the remarkable discoveries made during the 1990's "decade of the brain" using magnetic resonance imaging (MRI). Such MRI studies, which construct three-dimensional models of the brain in vivo, show a correlation of about 0.40 between brain size and IQ, as replicable a set of results as can be found in the social and behavioral sciences. The first MRI/IQ studies were published in the late 1980's and early 1990's in leading, refereed, mainstream journals like Intelligence (Willerman et al., 1991) and the American Journal of Psychiatry (Andreasen et al., 1993).

3. The parallel geographical distribution of brain size. Racial differences in brain size have been established recently using wet brain weight at autopsy, volume of empty skulls using filler, and volume estimated from head sizes. Using endocranial volume, for example, Beals et al. (1984, p. 307, Table 5) analyzed about 20,000 skulls from around the world. East Asians averaged 1,415 cm3 (SD = 51), Europeans averaged 1,362 cm3 (SD = 35), and Africans averaged 1,268 cm3 (SD = 85). Using external head measures to calculate cranial capacities, Rushton (1992) analyzed a stratified random sample of 6,325 U.S. Army personnel measured in 1988 for fitting helmets and found that Asian Americans averaged 1,416 cm3 (SD = 104 cm3), European Americans 1,380 cm3 (SD = 92), and African Americans 1,359 cm3 (SD = 95). Moreover, a recent MRI study found that people of African and Caribbean background averaged a smaller brain volume than did those of European background (Harvey, Persaud, Ron, Baker & Murray, 1994).

Contrary to purely environmental theories, these racial differences in brain size show up early in life. Data from the U.S. National Collaborative Perinatal Project on 35,000 children found that Asian children average a larger head perimeter at birth than do White children who average a larger head perimeter than do Black children, even though, at age seven, Asian children average smaller body size (and Africans larger body size) than do Europeans. Further, head perimeter at seven years correlates with IQ at age seven in all three racial groups (see Rushton & Ankney, 1996, for review).

4. The heritability of intelligence. As discussed in The Bell Curve and Race, Evolution, and Behavior, the heritability of intelligence is now well established from numerous adoption, twin, and family studies. Particularly noteworthy are the heritabilities of around 80% found in adult twins reared apart (Bouchard, Lykken, McGue, Segal & Tellegen, 1990). Moderate to substantial genetic influence on IQ has also been found in studies of non-Whites, including African Americans and Japanese. Even the most critical of meta-analyses find IQ about 50% heritable (Devlin, Daniels & Roeder, 1997).

Transracial adoption studies suggest a genetic contribution to the between-group differences. Studies of Korean and Vietnamese children adopted into White American and White Belgian homes show that, although as babies many had been hospitalized for malnutrition, they grew to excel in academic ability with IQs 10 points or more higher than their adoptive national norms (Frydman & Lynn, 1989). By contrast, Weinberg, Scarr and Waldman (1992) found that at age 17, Black and Mixed-Race children adopted into White middle-class families performed at a lower level than the White siblings with whom they had been raised.

Liz Snyder

Because I am not entirely familiar with the literature being discussed (i.e. I have only read secondary sources on this subject), I will not attempt to make any valid assessment of the works themselves. However, I would just like to note that the populations being discussed in these works are not genetically determined categories.

For instance, the largest argument involving "race" and IQ in the United States involves intelligence testing of African Americans. The category "African Americans" has absolutely no basis in genetic categorization; that is, African Americans are *not* an example of intra-specific variation within Homo sapiens sapiens. In fact, the African American "genotype", if there is such a thing, is a conglomerate of several genetically distinct groups, *including* European "whites" (which, by the way, is another wastebasket category). To say that any generalized characteristics of African Americans can be assessed is a biased statement in and of itself, but to say that such characteristics are genetically based is pure hogwash, and every physical anthropologist with his/her head screwed on straight knows it.

The same is true of any population that is not derived from a single evolutionary source. Any conclusions found using geographically mixed populations are dubious at best. In addition, the emic category of "race" will never serve to accurately reflect the actual biological populations present. When you look at this data, just remember the dictum, "Garbage in, garbage out."

Liz Snyder
SJSU Dept. of Anthropology

Mike Salovesh

As far as I know, there just aren't ANY human populations separated by as much as 40,000 years.

Let's see now: European and African populations have been interbreeding directly for a VERY long time. Hannibal's army, of course, came from Africa, and those Africans left more descendants in Europe than Hannibal's animals left elephant flops. Yes, I know, Hannibal's Carthage was in NORTH Africa; so? There's a long, almost continuous record of North African populations interbreeding with those of central Africa. (There are also reasonable records of people in classical Egypt interbreeding with dark-skinned peoples whose origins were much farther south than the farthest reaches of pharaonic rule.)

Whatever the populations of Europe may be, they surely haven't been genetically isolated from people of central and even southern Africa for millenia. Hannibal's armies are just a passing example of the long history of breeding contacts. You don't even have to buy the "out of Africa" hypothesis of the origins of anatomically modern humans to conclude that Europe and Africa have NEVER been the homelands of totally separate breeding populations.

All right, let's look to the east. I usually mention Ghengis Khan in this connection: thousands of very fierce warriors emerging from Central Asia and invading to the heart of Europe. They left direct linguistic evidence of their presence in the languages of Finland, Hungary, and Turkey. How likely is it that they did not impregnate one helluva lot of local women as one of the rewards they gave themselves for their conquest? That, of course, was only one of many waves of European invasion from the east over several centuries.

My examples deliberately take off from that comment about "predating the European expansion": they are examples of non-Europeans expanding into Europe. I may have rushed to some kind of prejudgment of what you're saying, however -- I just sort of assumed that the European expansion you were talking about began, more or less, in the fifteenth century with, say, the Portugese under Phillip the Navigator.

Perhaps you meant the expansion of Rome. In that case, try looking at an old, old popular book about classical archaeology: "Rome beyond the imperial frontiers". (I don't feel like looking up author and publisher; I think it was by Mallory and the edition I read many years ago was a Penguin paperback.) It certainly is clear that Rome was engaged in direct trade with China. Well, you may have heard the old saw, "the flag follows trade"; in this case, it also is clear that interbreeding did, too. All right, drop back a little farther in history and you'll come across Alexander the Great and his "Drang nach Osten". His armies actually made it to the Indian subcontinent. Of course, his father's old enemies, the Persians, did a lot of interbreeding with Greeks at one end of their empire and with peoples of Afghanistan and Inner Asia at the other.

Once more, whatever the populations of Europe may be, they surely haven't been genetically isolated from people of central, southern, or even southeast Asia for a very long time. Asia and Europe probably never have been the homelands of human breeding isolates of any more than local and short-term significance; they certainly haven't been for the last three millenia or so.

The two populations we think of as most separated from the rest of Homo sapiens once were the sole occupants of Australia and the Americas. But the inhabitants of Australia never were completely cut off from breeding with the rest of humanity. Recent archaeology in Australia turns up more and more evidence every year that there was lots more contact with the rest of the world than we used to think. The evidence now goes back 50,000 or more years. That's close enough to the date we used to accept for the origins of modern H. sapiens sapiens for me to say that, effectively, the Australians have ALWAYS had significant reproductive interaction with the rest of humanity.

Gosh, that only leaves one group that might fit your comment about a human population that has been separate for 40,000 years. Well, I don't want to get into a side argument about when humans first entered what Eurocentric geography calls "the New World". I don't think that kind of date matters much in this discussion. What counts is the other end of the entry of humans to the Americas. As far as I can see, whenever that movement started, serious scholars all agree that there was substantial migration from Asia into the Americas via some kind of Bering Straits land bridge something on the order of ten to fifteen thousand years ago. (I'm deliberately sloppy about my dates to avoid getting sidetracked into discussing exactly when that major population movement ended or exactly how long it lasted. That's irrelevant here.) Thus the native peoples of the Western Hemisphere were doing close interbreeding with Asian populations at least as recently as 15,000 B.P.: their ancestors were parts of those Asian populations. There's no room for 40,000 years of separation there.

Personally, I would argue that the Americas and Asia have never been reproductively isolated from each other. The breeding links are most clear with the Inuit and various Alaskan groups; it is at least arguable that the Athabaskans provided their own links back to Asia. Then there's the peculiar anomalies of such plants as sweet potatoes moving west from South America to the southwest Pacific area. If cultigens were being moved by ocean-crossing humans, I'd bet the farm that when those travelers got to their destinations they had some pretty good celebrations that MUST have left some biological descendants.

There's no sense in looking for the biological or genetic results of 40,000 years of separate development of isolated human groups before European expansion. That separation simply did not happen.

We are a single interbreeding species, and we have been at least as long as our ancestors have been fully anatomically modern humans.

That doesn't deny that specific hereditary human traits show distributions along geographic clines. Of course they do. Those traits, however, still are subject to independent assortment. The clines exhibited by two independent traits cannot be expected to be exactly the same.

The facts of variation in human herditary traits are interesting, and they can have important consequences. I'm all for studying them. But I don't expect that we'll get very far if we assume a priori that the distribution of those traits must coincide with the breeding boundaries of isolated populations.

Once more, the clines overlap, but they do not duplicate each other. No human population -- NONE !!! -- known to the archaeological and historical record has ever maintained itself in reproductive isolation from its neighbors for more than a few centuries.

In biological terms, it's a fair conclusion that there are no "pure races" in Homo sapiens AND THERE NEVER HAVE BEEN.

--  mike salovesh                               <salovesh@niu.edu>
    anthropology department
    northern illinois university                     PEACE !!!

Race, Evolution, and Behaviour (J. Philippe Rushton)

Phil Rushton here continuing the race thread --
  1. Although African Americans, who average about 20% White genes, have IQs averaging about 85, Africans from Africa average IQs of about 70. See reviews of this literature in The Bell Curve or my book (http://www.bookworld.com/rushton)
  2. East Asians average IQs on the same Euro-American designed tests of about 107 regardless of whether tested in the U.S. or in Asia.
  3. Brain size correlates about 0.44 with IQ tests established using Magnetic Resonance Imaging (about 8 modern studies).
  4. The racial populations differ in average brain size with East Asians averaging about 17cm3 (1 cubic inch) larger than Whites who average about 80 cm3 (5 cubic inches) larger than Blacks (established by autopsies, endocranial volume, external head size measures, and one MRI study).
  5. In the USA, brain size differences are there at birth, 4 months, 1 year, and 7 years. The differences are not due to body size because by age 7 Blacks are taller and heavier and East Asians shorter and lighter, than Whites (on average).
  6. Testosterone differences differentiate the races (Blacks most; Asians least; Whites intermediate).
  7. Violent crime rates within the U.S. are the same as those found internationally as assessed from INTERPOL Yearbooks (Blacks most violent; Asians least violent; Whites intermediate).
  8. Sexual activity (age at first intercourse, intercourse frequency, sexually transmitted diseases, including AIDS) shows the same pattern internationally as it does within the USA (Blacks most active, Asians least active, Whites intermediate).
  9. Speed of physical maturation (dentition, age to walk, age of puberty) shows the same pattern globally as it does within the USA (Blacks fastest to mature; East Asians slowest to mature; Whites intermediate).
  10. Rate of dizygotic twinning (woman produces two eggs in monthly cycle) is 4 per 1,000 births in Asians, 8 per 1,000 in Whites, and 16 per 1,000 in Africans.

Can purely cultural factors account for the world wide pattern? Why do east Asians avergae the largest brains and the lowest two-egg twinning rate, Afriocans the smallest brains and the highest two-egg twinning rate, and Whites intermediate?

My book uses r-K life history theory from evolutionary biology to explain the race differences. Alternative theories welcome. Name calling won't change the facts.

Dwight Read

D. Read

Liz Snyder

Dr. Rushton,

The only compelling evidence you present is in regards to the rate of dizygotic births among different "races"... However, you do not seem to understand that the *correlation* between dizygotic twinning and all the other *social* and *developmental* factors which you so offhandedly name HAS NOT BEEN ESTABLISHED.

Furthermore, you cite the rate of AIDS infection among "racial populations" as evidence for increased sex drive among "Blacks". How do you explain that the largest population infected with the AIDS virus is in Southeast Asia, and not in Africa [MMWR 42 (53): 17]. I thought that "Asians" (or should I say "mongoloids", Mr. Rushton?) were supposed to have the lowest sex drive, and therefore the lowest rate of infection...

In addition, I question your racial categorization. Your generalized view of racial categories is inherently biased by your social views of human populations. No zoologist who wanted to keep his name would ever name three subspecies in a population with as little intra-specific variation as Homo sapiens sapiens. In comparison to most groups, we are an *extremely* homogeneous population. You are assuming that certain characters found among populations (be they culturally or genetically determined) are evolutionarily significant when they are clearly not. If brain size were a factor in human development, then I would be more intelligent than Albert Einstein. Don't I wish... From this case in point, it is clear that, whereas overall encephalization in relation to body size is an important aspect of inter-specific comparison, it is not significant to the intra-specific variation found among Homo sapiens.

I would also like to point out that your term "blacks" is not a valid term to use in an academic discussion. Pray tell, just WHO are you referring to? Do you include indigenous populations of New Guinea? If you do, then your theories involving dizygotic twinning will fall apart-- the Dani have just about the lowest incidence around. What exactly are you talking about when you say Blacks, Whites and Asians... I personally can't seem to find one iota of continuity between skin color and genetic makeup.

One of the reasons that you will find a greater incidence of dizygotic twinning among populations indigenous to Africa (to put a finer point on it than you managed to, Mr. Rushton) is because of the diversity of evolutionary development found among those populations. Assuming that Homo sapiens first moved along the arc out of Africa at about 100,000 years ago (which is my view of things, though there is some controversy on this point), the populations of Homo sapiens moving out of Africa would become somewhat geographically isolated as they moved father and father away from home populations in Africa. However, while populations in Europe and Asia underwent isolation, effectively homogenizing those populations, the populations in Africa continued to intermix, and diversity increased. Thus, populations in Africa can be found that are more different from ONE ANOTHER than they are from European or Asian populations. In example, the Bantu and the !Kung are genetically more different from one another than they are from European whites. To say that there is one "black" genotype is like saying that you are a psudo-scientist afraid of losing your place in the Western hegemony. Please don't waste our time.


Liz Snyder
SJSU Dept. of Anthropology

Danny Yee

Prof. J. Philippe Rushton wrote:
> 1. Although African Americans, who average about 20% White genes,

Please give us a genetic definition of 'African American'. Please define 'White gene'.

And please explain why you are completely inverting the procedure used by biologists studying genetic differences. Which is to start with clearly defined characteristics fairly directly linked to genetic differences (blood groups, nucleotide sequences, proteins, etc.), to study biogeographical distribution of variation in these, and thence to come to conclusions about demic structure, hybridization, divergence times, the extent of gene flow and genetic isolation, etc. On the basis of these, subspecies status, or 'population' status, might then be assigned.

J. Philippe Rushton

Phil Rushton replying to the queries about definitions.

Most of the empirical research itself, especially from the U.S. requires social definitions as in self- or other-identification. Thus my study of cranial capacity in a random stratified sample of 6,325 U.S. Army personnel published in the 1992 issue of the journal INTELLIGENCE, used the military personnel's own classification system which I believe was self definition. I found self-defined Asian Americans averaged larger craniums than self-defined Whites who averaged laeger craniums than self-defined Blacks. In a follow up study in the 1994 issue of INTELLIGENCE, I examined International Labour Office data from African countries, European countries, and North East Asian countries, and found the same average differences in brain size estimates as within the U.S. My data only replicate what others have found before me using various procedures like autopsies, endocranial volume, etc. For those who don't want to read the thorough review in my book Race, Evolution, and Behavior (http://www.bookworld.com/rushton) there is a good summary of my position in the February 1996 issue of CURRENT ANTHROPOLOGY devoted to a review of The Bell Curve.

As for White and Black Genes, well there is the entire Out-of-Africa model based largely on molecular DNA data (corroborating fossil data). Cavalli-Sforza and many others (the late Alan Wilson from Berkeley) who looked at mutation rates in DNA and similarities and differences in blood proteins, etc to find that modern humans evolved in Africa about 200,000 years ago with an exodus out of Africa about 100,000 years ago, and a Mongoloid-Caucasoid split about 40,000 years ago. Selection pressures were different in hot sub-Sharan Africa than in cold, cloudy Europe, than in Siberian cold east Asia. Thus the different populations were selected for various traits like brain size and intelligence and reductions in sex hormones (and twinning rates) as they diverged north and east.

Interestingly, the high-tech DNA data corroborate social definitions and others based on morphology going back to Linnaeus, Darwin, and others. Today forensic anthropologists routinely identify race of victims and perpetrators, even from decomposed bodies or drops of semen or hair. These are increasingly sophisticated eg identifying Black-White versus Black-Asian hybrids. Check out a paper on Population Specific Alleles in the 1997 issue of American Journal of Human Genetics.

Liz Snyder

Self-identification is clearly a social factor. Why do you insist that this correlates with genetic data? It clearly does not.

Alan Wilson's hypothesis concerning recent African origins actually supports the homogeneity of Homo sapiens. Many other hypotheses place the splitting date much earlier (i.e. at the time of Homo erectus). Furthermore, just to let you know, that data was derived from *mitochondrial* DNA. In addition, Alan Wilson and Rebecca Cann, who first wrote on the Recent African (RA) origins hypothesis, would *never* contend that human populations have been separated by 40,000 years. I don't know where folks come up with this figure, but it is DEAD WRONG. Please go back and read Mike Salovesh's post titled "That 'race' thread". I don't feel the need to repeat what he's already said. Ok, well, maybe I do. In brief: The only populations that *could* have been totally isolated for that long were in Australia and New Guinea. This corroborates the genetic evidence, including a study done on blood hemoglobin. However, all other populations were totally isolated for a period of no more than 15,000 years. The genetic evidence also corroborates the relative homogeneity of all human groups.

I think you are confusing "race" differences with genetic causality. Because a self-identified group shares characters such as intelligence (which is culturally determined depending upon what is asked of the subject), sexual drive (which is culturally determined based upon what the subject thinks the researcher *wants* to hear), and brain size (which is largely meaningless, and is dependent on neonatal maternal care), it does not mean that this self-identified group is genetically linked. They can be, but they don't have to be.

Case in point: All self-identified Buddhists are vegetarians. Therefore, Buddhists are a genetically determined group that have evolved away from the omnivorous population of Homo sapiens sapiens. They are Homo sapiens sapiens vegetarianus. I am a self-identified vegetarian. Therefore, I must have some genetic link to Buddhists, who evolved largely in East and Southeast Asia. Ergo, I must be part Asian.

You can see that this argument is ridiculous when put in these terms. You are assuming that every *self-identified* characteristic is linked to genetic characteristics without doing the appropriate research.

Racial Definitions

Phil Rushton replies to an obviously peeved Liz Snyder who claims:
> Self-identification is clearly a social factor. Why do you insist that
> this correlates with genetic data? It clearly does not.

But, Liz, if it did not, how could taking a hair from your head tell us that you are a non-white, non-asian, and probably much more too.

If you were reduced to a jaw bone how could forensic anthropolgists determine that you were a white female?

And, true, Alan Wilson's original work (oh, yes, and Rebecca Cann's too) used mtDNA not nuclear DNA, but dozens and dozens of other studies have confirmed the approximate splitting times of the populations using all sorts of techniques../...NOT to say that this doesn't mean there has been ISOLATION. Of course genes have been traded back and forth as in migrations, conquests. But Nelson Mandela's DNA will never be confused with Mahatma Gandhi's.

How do you explain that forensic anthropologists can differentiate Black-Asian hybrids from Black-White hybrids (there are many of the former now in the U.S. as a result of recent migrations from Vietnam -- the offspring of offspring from the Vietnam War Era. A very different social definition, but a very predicatble DNA fingerprint!

Danny Yee

> Selection pressures were different in hot sub-Sharan Africa than in
> cold, cloudy Europe, than in Siberian cold east Asia.

Care to explain how warm climates select against intelligence? Care to explain why they do that in Africa but not in equatorial Asia? Care to explain how this is consistent with the origins of H. sapiens (including massive increase in brain size) in equatorial Africa?

Does the term "Just So story" ring any bells?

> Nelson Mandela's DNA will never be confused with Mahatma Gandhi's.

But neither will Nelson Mandela's DNA ever be confused with Winnie's! More than 90% of human genetic variation is intra-regional.


Forensic scientists don't distinguish "Black-Asian hybrids from Black-White hybrids" - they distinguish individuals with different blood groups, immunoglobin proteins, etc. These characteristics have correlations with physical surface features such as melanin concentration and jaw shape, which have some bundling into physical types and some correlation with geographic location and ascriptions of race. But to start with the races is to put the cart before the horse.

It would make much more sense if you tried to make "intelligence", along with blood groups, etc., diagnostic of race (though you can't make it diagnostic of physical type). Races are socio-culturally defined, using bundles of surface physical features. These features are what forensic anthropologists work with, along with blood groups and other genetic markers. By all means throw "intelligence" in with those - it's not much use to forensic science, obviously, and it's obviously a lot harder to work with (and more controversial) than the physical and molecular markers, but at least you'll be doing something which isn't epistemologically bogus right from the beginning.

I also sense a kind of typological reasoning which suggests metaphysical confusion as well as epistemological. Just like species (see Ghiselin, Metaphysics and the Origin of Species), sub-species or races (if they exist) are ontological individuals, NOT classes. Thinking of them as classes will produce all kinds of confusions.

anthro-l << Anthropology << Danny Yee