At what level is racialism accepted in the scientific community?

Great Debates
221

I don’t like wasting time on semantics, so I will introduce a neat trick. I will define Parker race (P-race) as divisions that are “substantially greater in magnitude than any other divides in the continuum of genetic similarity” and few in number and I will define Darwin races (D-races) simply as divisions in which individuals are arranged according to pedigree. (Note P-races are a subtype of D-races)

I will then make the following claims:

  1. The Darwin race concept corresponds better with the dominant historic and modern biological race concept than does the P-race concept (i.e., the P-race concept is a straw man or is a more narrow understanding).
  2. The following groups constitute continental level Darwin races: Negroids, Caucasoids, Mongoloids, Amerindians, and Oceanians.

There are many lines of evidence for (2). One is that these groupings emerge from cluster analysis at some level of analysis. This indicates higher within than between group genetic similarity. Which indicates higher within than between group inbreeding. Which indicates closer within than between propinquity of descent. To be clear when I say that these are Darwin races, I am not saying that these are THE races. I am saying that were one to carve up the human species, this would represent a valid biologically natural division (i.e., one based on genealogy) – ergo biological races.

As for (1), one can simply look at historic and contemporaneous usages in biology. If you go back to the very early theories (e.g., Buffon) you will notice that racial divisions of all degrees were acknowledged:

Kant (1775): Those features belong to varieties – which are, therefore, by themselves hereditary (even if not always) – can through marriages that always take place within the same families, even produce, in time, something that I call family stock. These features ultimately become rooted in the reproductive power so characteristically that they come near to forming a variation in the way that they perpetuate themselves…If nature, when undisturbed (without the effect of migration or foreign interbreeding) can effect procreation everywhere, she can eventually produce an enduring stock at any time. The people of this stock would always be recognizable and might even be called a race, if their characteristic features does seem too insignificant and so difficult to describe that we are unable to use it to establish a special divisio.

Dohzhansky (1946): One may perhaps question the desirability of applying the term ‘racial differences’ to distinctions as small as those that can be found between populations of neighboring villages and as large as those between populations of different continents. Might one modify the definition of race by specifying that the differences in gene frequencies be above a certain minimum magnitude? Such a modification is undesirable for two reasons. First, since all magnitudes of difference are found among populations, any specified minimum can be only arbitrary. Second, it is most important to realize that the differences between the ‘major’ human races are fundamentally of the same nature as the relatively minute differences between the inhabitants of adjacent towns or villages.

Leroi (2005): Yet there is nothing very fundamental about the concept of the major continental races; they’re just the easiest way to divide things up. Study enough genes in enough people and one could sort the world’s population into 10, 100, perhaps 1,000 groups, each located somewhere on the map. This has not yet been done with any precision, but it will be. Soon it may be possible to identify your ancestors not merely as African or European, but Ibo or Yoruba, perhaps even Celt or Castilian, or all of the above.

See also: Garn, S. M., & Coon, C. S. (1955). On the Number of Races of Mankind1.American Anthropologist, 57(5), 996-1001.

Generally, it’s hard to reconcile the claim that race really means/meant deep genetic divisions when biological race was/is typically conceptualized – by racial theorists, not opponents of the concept – in a much broader sense. Perhaps you are (1) confusing the race concept with the zoological subspecies one and (2) assuming that there is some commonly accepted genetic difference criteria for zoological subspecies? Whatever the case, D-races are what Wade and I mean when we talk about the e.g., mongoloid race.

222

Your original statement was: “Denial of the history of **white supremacy in the United States **as way to minimize the likelihood that so-called “nurture” factors explain outcome disparities here.”

Since we are discussing the historic effects of U.S. White Supremacy on African American cognitive ability, the performance of Black migrants, who constitute at least 10% of this population, needs considering. If first and second generation Black immigrants from Africa and the West Indies don’t perform much worse than 10th generation Black Americans then this surely suggests that there is more to the deficiency than historic US white supremacy and years of being under White rule. For data we have for example. Now as for selection, Black immigrants are positively selected in human capital, so we would expect the scores of those left behind to be lower. This doesn’t damage my point. As for identification, this is based on self identification with U.S. government classifications:
http://www.whitehouse.gov/omb/fedreg_race-ethnicity Alternatively, one can use national origin for those surveys that provide such data. It’s possible the only low aptitude Black immigrants ID as Black, but if so, one would see a very skewed distribution. I guarantee that this doesn’t exist – though you could check.

In the article:

“The analysis approach we have adapted is a logical extension of estimation methods based on pedigrees. It allows estimation of additive genetic variation that is captured by SNP arrays and is therefore informative with respect to the genetic architecture of complex traits. The estimate of variance captured by all of the SNPs obtained in GCTA is directly comparable to the heritability estimated from pedigree analysis in family and twin studies, as well as the variance explained by GWAS hits, so that missing and hiding heritability can be quantified.”

This has been [discussed](file:///C:/Users/John/Desktop/Testing%20the%20key%20assumption%20of%20heritability%20estimates%20based%20on%20genome-wide%20genetic%20relatedness.%20Journal%20of%20human.pdf) by some, but my original point was that GCTA estimates are not confounded by epigenetics. I think it’s unlikely that they are non-trivially confounded by cryptic shared environment (in the causal sense), not to be confused with external outcomes.

223

I could easily publish a study along these lines in e.g., Intelligence. In the NLSY 97 parents were asked about their parents’ ethnicity. And the skin color of the student respondents was measured. As expected, Black students (raced defined by parents) whose parents reported having White/European + Black parents where lighter in color than those with parents who reported having only Black parents. They also had higher IQs. This is utterly expected since individuals with one Black and one White parent have intermediate scores relative to those with two parents of the same self-identified race. All of this merely points again to the fact that race differences are passed on inter-generationally. (On average, a person with 2 B parents inherits ~100% of the disadvantage, 1 B, 1 W parent ~50%, 1 B, 3 W grandparent ~25%).

The problem is that these types of analyses don’t disentangle shared environmental and genetic effect. And we know the the B/W gap is some combination of the two.

224

By this logic, pedigree analysis – or Mendelian genetics – is not genetic. This confuses “genetic” with “DNA”. What you mean is that there is no DNA based evidence. But even this is no longer true because a few well replicated cognitive SNPS are known and the frequencies of these vary regionally in a correlated fashion.

What is this supposed to mean?

This was a survey of intelligence researchers on matters related to intelligence. The relations to race (in the sociological sense) is that there are race (in the sociological sense) differences in intelligence,

I fail to understand why you think that biologists or anthropologists would know better. Do they need to be consulted regarding sex differences in the same trait?

225

Because eliminating outcome differences–and not just in the US, with its particular history–among races are so stubbornly resistant to manipulation of nurturing variables, the tack for egalitarians in opposition to “scientific racialism” has been shifting from “everybody has about the same genes driving about the same functional outcomes,” to “race is a very poor taxonomy from a genetic perspective.”

This is why these arguments become so convoluted. The pro-racialist group wants to hold the question to this:
Are the average observed outcomes between self-identified races a result of average gene differences in addition to nurturing variables?

The anti-racialist group wants to side-step that question and turn the debate into one of biological taxonomy:
*Are the average observed differences between self-identified races of any significance at all if “race” is a poor biologic taxonomy?
*

Science is not helping the anti-racialists. Gene frequencies can be quantified, and it is becoming readily apparent that there is a reasonable dividing point at out-of-africa for some genes. Self-identified races, which by definition are a purely social construct, can be quantified for gene frequency and shown to have average continent of recent origin. Gene functions are slowly being teased out for variants which in turn can be shown to have differing frequencies even at the self-identified race grouping level (see data for ACTN3 frequency differential by race, e.g.).

A poor understanding of what average group outcome means is not helping the pro-racialists. Not only are we so obviously admixed and diverse that a group average does not define an individual; between groups the range of variability is broad enough and the overlap of outcome extensive enough that at an individual level “blacks are genetically superior at power sprinting” becomes inane when two random individuals from each group are juxtaposed. Worse, lumping all self-identified blacks together becomes absurd when individual lines of diversity are teased out. The Mbuti and Kalenjin (using self-idenitification again as a proxy for some biological average differences) are not going to be taking over the NBA even given perfect nurturing. It is a quirk of historical migration anchor points that “blacks” are lumped into a single clumping.

The problem is that the present clumps of “races” are deeply ingrained, and because average outcomes differ within those clumps despite efforts to erase the gaps, the very social groupings of race which are eschewed by anti-racialists are promoted into persistence out of an effort to track and enforce race-based equality of representation across all societal parameters.

It’s not so much that race is a great biological taxonomy. It’s that, once that self-identification exists, and once we decide the fairest society needs to have all self-identified groups fairly represented, then we drive that race paradigm.

To the extent that average gene frequencies which happen to exist within those self-identified races drive outcomes, we are trapped within a never-ending cycle:
Mandate self-identification so we can make sure your group has its share of seats at the table (at which point we can stop self-identifying), but those seats will never be spontaneously equally distributed among races because average differences of gene pools drive average differences in outcomes.

226

You’ve still failed to demonstrate that this is anything beyond just a hypothesis, especially considering the published and peer reviewed data that shows African admixture among self-identified black people has no correlation to test scores.

We could accept that every self-identified race has different genes for every single human characteristic (we don’t, of course, by the way), and this still wouldn’t tell us anything about which group has inferior genes for intelligence. Test scores don’t and can’t tell us anything about genes, full stop. So there’s no point in going on and on about the possibilities of different lineages and different gene pools.

227

This is absolutely false, and further, contrary to the ‘out-of-Africa’ theory of human origins. Imagine many thousands of years ago, group A in East Africa splits into two groups: A1 and A2. A1 stays in East Africa, and A2 moves a bit north. A1 further splits into numerous other groups: A1a, A1b, A1c, etc., which settle all around central and southern Africa. A2 also splits into multiple groups. A2a moves to north-eastern Africa, and a later split, A2aX, leaves Africa, while A2aY stays in Africa. A2aX, now in the Middle East, also splits at some point, and various descendant groups (A2aX1, A2aX2, etc.) settle in the Middle East and western Asia. An African group like A2aY is more closely related, genealogically speaking, to non-African groups like A2aX1 than it is to other far flung African groups like A1b. And this is exactly what the genetic data shows – some African populations are, genetically speaking, more closely related to various non-African populations then they are to various far-flung African populations.

So there’s no way that a group like “Negroids” could include all sub-Saharan African populations without including many (and possibly all) non-SSA-populations.

This is just common sense, and it’s been tested. And it shouldn’t be surprising at all – Somalians are genetically more closely related to Middle-Easterners than they are to Namibians. Malian black people are more closely related to North African Tuaregs then they are to South Africans of Zulu descent. So there’s no biological grouping that could include both Somalians and Namibians that would not also include many non-African groups, and perhaps all of humanity.

228

Not only is genetic variation based on geographic distance (not racial groupings), but since the variation is more dense within Africa: Two non-Africans are much more likely to be genetically similar then two Africans (within the same km radius). But good luck teaching population genetics to someone like this:

229

And as Greg Laden could tell you, you only have a correlation that it is more likely related to physical looks than intelligence.

Exactly that, climate change deniers, creationists and yes, scientific racists like to cherry pick and do “science by survey”. To be useful you have to survey the ones that are involved with the mechanism and the genes, not the ones that just look at correlations and then remain stuck there.

The whole of the science is actually missed when the best you can do is to pump a survey of researchers that is not directly involved with the relevant science.

And the misleading ideas continue, you are confusing the issue with the cause of it. The point most geneticists, biologists and anthropologists do is that the cause of that difference is not likely to be genetic or very subtle in the end.

230

Yes. But, of course, that’s not the bifurcation I was discussing. Race would be a much more plausible genetic category if the races corresponded to genetically distinct populations, but they don’t.

You try to get there by reasoning from the stubbornness of the gap that the problem must be genetic and therefore these categories must make genetic sense, but this is doubly bad reasoning. First, we have actual genetic evidence that contradicts this claim, which is much better than your circumstantial inference from the gap. And second, lots of social problems are extremely stubborn and quite obviously not genetic.

But the point is setting the cluster analysis to 3 or 4 or 5 is arbitrary unless the improvement in the fit to the data is substantially greater at one increment than another. If 2 is a terrible fit, and 3 is marginally better, and 4 is marginally better, ad nauseum, then you’re simply wrong to assert that “groupings emerge from cluster analysis.” They don’t emerge from the analysis. They are an independent variable established by the scientist, not a dependent one.

If your assertion is simply that some populations have different likelihoods of having different genes, that’s trivially true. It doesn’t get you anywhere close to being able to make these claims about racial intelligence. Indeed, these grouping do not even correspond to the historical races.

Obviously my comment was made prior to the introduction of your idea of testing migrants. In that context, you have to consider the effects of white supremacy elsewhere, as well as the affects of American White Supremacy upon arrival, among many many other factors.

As for the rest, I am–to put it mildly–unimpressed with the scientific rigor of the cited paper. Among the myriad factors you need to control for, surely the most obvious when using English-language tests is English fluency. Yet the only discussion in that paper of that critical topic is the observation that third generation Hispanic immigrants don’t speak English much better than their first and second-generation counterparts. I’m not sure what that is supposed to demonstrate, but it sure as shit does not show that immigrants who self-identify as Black are as effective at taking English tests as native English speakers.

Do you not see the difference between that paragraph and your claim?

And you’ve not done anything to prove that point beyond citing the fact that this test is useful for estimating genetic contribution. I’d be much more interested in either (1) an actual explanation for how it controls for epigenetics or co-correlated environmental influences or (2) a published paper actually saying this.

231

This is like saying speed test scores for greyhounds don’t and can’t tell us anything about genes. full stop.

Only a confused individual would hold such a position.

If average speed test scores are different between all “greyhounds” tested and all “all other dogs” tested, then the analysis turns to nurture, number tested, reasonable representation for group, and so on. And from that analysis can come a very good conclusion about whether or not genes are at play.

For humans, it’s well accepted that genes drive physiologic, neurologic and superficial outcomes, however one wants to define those. It is not beyond the reach of science to examine variables and reach conclusions, nor is it beyond the reach of science to uncover mechanisms (such as human migration patterns and evolution) which might–in fact, more likely than not, would–drive descendant lines apart from one another. It is not beyond the reach of science to determine that gene frequencies and gene-driven outcomes do differ among populations.

When one looks at the overall picture, it is far more likely than not that separated population lines have disparate gene pools, and far more likely than not that those disparate gene pools have driven disparate outcomes.

The notion that evolution and migratory patterns would somehow drive exactly egalitarian results among populations is quite naive, and what test scores tell us when we normalize for things like nurturing opportunity (say, SES and educational opportunity, e.g.) is that the predicted divergence of descendant lines is, in fact, consistent with what a genetic underpinning would create: disparate outcomes.

232

This does not have a thing to do with whether or not average gene frequency differences in the total gene pool of a given race drives disparate outcomes among those races.

See my earlier explanation if you need more help understanding this.

While quantity of variation may be an argument to develop a different taxonomy, IF we use the present taxonomy of (roughly) continental origin, THEN there is average gene frequency difference which drives disparate average outcomes. This is why genes for physical appearance are a major driver of self-identification: IF we self-identify as “black” or “white” THEN we belong to a pool where the genes driving appearance differ in average frequency.

233

Humans are not greyhounds, and intelligence is not speed. It’s not even well accepted that there are any tests that accurately measure intelligence. Humans really are different, and intelligence really is a different kind of characteristic.

Fine (notwithstanding the previous), though in this case “nurture” (as we’ve discussed over and over again) has not come close to being corrected for.

None of this has anything to do with intelligence.

It’s not “far more likely than not that those disparate gene pools have driven disparate outcomes”, especially when the populations lines are not really separated.

I’ve never stated this. Quit repeating this fucking straw man.

No one has yet normalized for “things like nurturing opportunity”. Not even close.

234

I’m sure this has all been done a dozen times, but I am fascinated by how it is that you cannot see how bad this reasoning is, Chief Pedant.

Your premises seem to be:

  1. Different human populations have different likelihoods of having different genes.

  2. There are outcome gaps between self-identified races that involve traits that may have some genetic component.

  3. We have tried and failed to close these gaps.

Those are all true. But that is an extremely weak circumstantial argument for the proposition that the outcome gaps between self-identified races are a result of different genetics. So you add some more premises:

  1. If the gaps were the result of non-genetics, we would have closed them by now.

  2. Self-identified race correlates to clusters of genetic similarity/diversity such that we should expect there to be differences in gene frequencies for genes other than those tied up in the self-identification.

  3. Intelligence is a trait determined in large part by genetics.

Those are all much more dubious. And even if you take them as true, you have at most a medium-strength inference to make from circumstantial evidence. You don’t have an actual argument from the genetics. You haven’t identified intelligence genes or shown that they have multiple alleles that produce different traits. Instead it’s a series of inductive inferences. And your eagerness to make those inferences seems suspect, given their history.

235

You seem to be under the impression that I’m even addressing your bloated nonsense. Interesting.

236

It has been done many times, but thank you for a great post – phrased quite differently than I and others have in the past.

237

I am not aware of any evidence refuting a claim that genes are at play for the disparate outcomes observed among races.

The evidence shows:

  1. Migration histories show a bifurcation point that largely isolated sub-saharan africans from eurasians for tens of thousands of years.
  2. Genetic studies show introgression of archaic lines into eurasian, but not african, populations (on average the modern persistence of those archaic genes is something like 1-4%, and some researchers think as much as 20% of the Neandertal genome persists in eurasian–but not african–populations. (Not 20% per given individual, but 20% of the original genome.)
  3. Frequency studies for any number of genes (I have noted MCPH1 haplogroup variant D, ACTN3 and DARC as examples) also vary by african/non-african, and one can further break down non-african depending on how many races you want to define.

In summary, the average gene pools are different, and there is a plausible explanation why the gene frequencies differ.

You seem to be stuck on the idea that races should reflect “plausible genetic categories,” and I think this is where you are confused about the basic argument.

The pro-racialist contention is (or should be):
IF we self-identify as (current) race groups, THEN are there gene frequency differences that emerge as a result of that self-identified grouping.

That is unequivocally true.

So the further contention that those average gene frequency differences drive average outcome differences is not rendered moot by an argument that the groups themselves are not plausible genetic categories. How plausibly appropriate the categories are genetically is a matter of definition; an discussion for splitters and lumpers.

The current race categories are self-defined. What’s plausible–indeed, easily demonstrable–is that people self-identify into categories which have a migration and evolutionary history that have created average gene frequency differences.

238

Yes you are. You just continually ignore it.

239

It’s generally acknowledged the two studies done employed inadequate methodologies. This is basic math. If a method is incapable of disconfirming a hypothesis a null finding can’t be counted as evidence against it. For example, in the Scarr et al study, a negative but non-significant correlation between African genes and IQ was found 0.05. The authors considered this evidence against a strong genetic hypothesis, but in fact it wasn’t because such a hypothesis predicts a correlation no higher than this. Note, the authors said: “An extrapolation from the contrast between extremes within the hybrid group to the average differences between the races predicts that not more than one third of the observed difference between the races could be due to genetic differences. In view of the negligible correlations between estimated ancestry and intellectual skills even this seems unlikely.” But it’s trivially easy to show that the correlation found was consistent with a strong genetic hypothesis.

Here is the paper. Here is why the results are consistent with a genetic hypothesis:

  1. According to recent analyses, the mean African admixture is 20% and the standard deviation of admixture is 15%. According to Zakharia, et al. (2009). With this we can calculate an expected IQ-ancestry correlation.

  2. In this instance the genetically conditioned difference between Blacks and White would be 0.75 SD, since we are proposing that 75% of the gap is genetic; the ancestry difference would be 5.3 SD, which is the number of SDs separating Blacks who are 20% White and Whites, given that 1 SD of admixture equals 15% Whiteness ((100-20)/15=5.3). The correlation between test scores and genotypic ancestry, would then be 0.75/5.3 or 0.14.

  3. This would be the correlation for an index that had perfect reliability. Correcting for the unreliability (see 1b), the the correlation between IQ and the index would be 0.085

  4. This would be the correlation between IQ and Scarr’s index, assuming that the within population heritability was 1, as a lower within population heritability will attenuate the correlation. According to Scarr et al., the within population heritability was 0.48. Correcting for the lowered correlation between IQ and genes, we get 0.06 (0.085*SQRT(.48)), which is approximately the correlation found. (We should also correct for the test reliability which is typically 0.9 –correcting for this, the expected correlation would be 0.05.)

  5. The difference between the upper and lower thirds would then be .13, which was approximately what was found.

Alas that the sample sizes was too small! Otherwise, we would have indisputable evidence of an IQ-ancestry correlation! Nonetheless, the results are consistent with a HH and thus suggestive.

Now, I’m guessing that you will say: “But Chuck the authors didn’t say that!” Ok, but this is basic math. This is what a HH would predict (Jensen, 1981) in this situation. Whatever, as said, given the well establish negative correlation between educational attainment and African genomic ancestry in the AA population and the well established correlation between IQ and educational attainment in the same population it’s difficult to see how IQ and African Ancestry would not negatively correlate (granted correlations are not transitive).

You end up having to dismiss the correlation between color and IQ, between self reported EU ancestry and IQ, between parental self reported race and IQ, between genomic ancestry and the IQ correlates of education, income, and occupation, and harping on a study which failed to confirm (but didn’t disconfirm) a genetic hypothesis to make your point.

I guess we should just agree to disagree on this point then.

I think other readers will recognize that I am being more rational and sensible on this matter: (1) I am acknowledging the totality of the evidence and (2) acknowledging that the fining of a ancestry IQ correlation can be rendered compatible with an environmental hypothesis.

My point here is that (1) supports a HH because some environmental models predict otherwise. It just doesn’t prove it because other environmental models make the same prediction.

240

I already commented on this. There are well replicated alleles that vary in a consistent manner between populations, thus supporting the position of differential selection. See: Estimating strength of polygenic selection with principal components analysis of spatial genetic variation | bioRxiv

The logic is:

(1) Selection on IQ/education would exert an impact across the genome (this is polygenetic selection).

(2) Thus if there was differential polygenetic selection, IQ/education-associated alleles would vary across populations in a correlated fashion. (If there was no such differential selection they would vary willy-nilly.)

(3) Well replicated IQ/education associated alleles do vary in a consistent manner across populations.

(4) Thus this is evidence of differential polygenetic selection for IQ/education associated traits.

The problem is, as I said, the # of well replicated alleles is few. Thus this evidence can still be disputed, rendering it evidence, not proof.

Whatever the case, the “no IQ genes” argument needs to be dropped.