This is the second of three guest posts by Professor Jonathan Kaplan of Oregon State University (see Part One here). Professor Kaplan is a noted philosopher of biology who has published extensively on biological race and IQ among other topics.
Part II. A paper about policy that doesn’t engage with policy
Cofnas’s recent piece,1 published in journal Philosophical Psychology, is problematic for a different reason than Winegard, Winegard, and Anomaly – it simply fails to do what it claims. The very title of the paper “Research on group differences in intelligence: A defense of free inquiry,” points towards its absurdity. As many people immediately pointed out, research on group differences in intelligence has been pursued and published regularly, and there are no limits on “free inquiry” around it.2 At least, no more so than there are limits on “inquiry” surrounding climate change skepticism or claims about the dangers of routine childhood vaccinations. In each case, what there is, instead, is a broad consensus among experts who actually engage with research of the relevant kind that the claims being made are too often wildly ill-supported (often already having been shown to be wrong), and the implications drawn by supporters on the basis of these ill-supported claims are often far too strong. When researchers argue that reasonable people of goodwill should not pursue bad research that purports to support certain kinds of conclusions about group differences in intelligence, or certain kinds of conclusions about anthropogenic climate change, or certain kinds of conclusions about the safety of vaccine, they are not arguing against “free inquiry,” but rather in favor of not pursuing grossly irresponsible research.
Here, again, I want to focus not on the sloppy and irresponsible research that Cofnas makes use of in his piece per se, but rather on the claimed goals of the paper. The abstract ends with what a reader might reasonably take to be the paper’s main point:
The present paper argues that the widespread practice of ignoring or rejecting research on intelligence differences can have unintended negative consequences. Social policies predicated on environmentalist theories of group differences may fail to achieve their aims. Large swaths of academic work in both the humanities and social sciences assume the truth of environmentalism and are vulnerable to being undermined. We have failed to work through the moral implications of group differences to prepare for the possibility that they will be shown to exist.
In a response to Charles Murray‘s book, Human Diversity, Turkheimer notes on his blog that one of the standard tropes of hereditarians can be summed up as: “Soon we are going to have personal flying cars, and society had better start building appropriate traffic control structures, or everything is going to be a mess when it happens!” The stated aim of Cofnas’ article is very much in line with this ‘project.’
So what are these “social policies predicated on environmentalist theories of group differences”? What “large swaths of academic work” are going to be undermined? (Well, obviously, I suppose that all the work arguing that there is good evidence for environmental differences, and no good evidence for genetically mediated differences of a certain sort, would be made moot, but that hardly seems relevant here, as that is just what we would expect if the question was definitively settled in that direction!) What kinds of “moral implications” need we work through? These are the questions that the paper’s abstract tells us the author will confront, and answer.
The paper starts with a summary of “the scientific controversy.” Again, my purpose here is not to critique the science, but the summary is remarkably shoddy, and ignores much of the relevant literature. And of course, one might ask, why is this summary necessary? If the paper is merely suggesting that work on this topic should be done, because future discoveries might be relevant to extant policy, the current lay of the land would seem somewhat tangential.
The next sections of Cofnas’s paper lay out his claim that at least some researchers are against research in this area, at least quasi-independently of the possibility of the research being reasonably high-quality. (Many researchers are, undeniably, against the usual poor quality research, but that’s a different matter.) These are, roughly, the same few researchers who are cited by hereditarians on this topic every time, and in fact, the exact same few passages (some of which only seem to be suggesting that even high-quality research should not be pursued if they are interpreted rather naively or out of context).
But let’s move on to what is supposed to be the heart of the paper – the programs that, if the hereditarian hypothesis is true, will be undermined, and the moral dangers lurking in our not having seriously considered the possibility that it might be true.
Cofnas’s first example is the Head Start Program:
If we had followed Jensen’s recommendation in 1969 to devote money to programs that were tailored to the strengths of different groups rather than to Head Start, around two hundred billion dollars would have gone to improving lives instead of accomplishing nothing that can be detected. This is one of many examples of how basing social policies on questionable scientific premises can have enormous opportunity costs.
There are a number of very serious problems here. First, the idea that we now have, still less had in 1969, any good ideas about what it would mean to “tailor” programs to “the strengths of different groups” is patently absurd. How would the “strengths” of different groups be discovered? How would education be tailored to those different strengths? What “groups” is Cofnas even referring to here? None of these questions can be sensibly answered at all in this case, still less answered in a straightforward way, still less were answerable in 1969. Second, the idea that Head Start was premised on the denial of group-level average differences is extremely odd (in fact, it is just false) – the premise, such as it was, was that students from disadvantaged backgrounds would benefit from early intensive intervention. This does presume that environmental interventions can make a difference, but makes no assumptions at all about whether there are group-level genetic differences that matter, or for that matter about whether individual genetic differences matter. All that was necessary for the program to ‘make sense’ was the claim that early environmental interventions might matter, whatever else was true about the world.
Cofnas cites work that he takes to show that Head Start was entirely useless, and takes that work to be definitive. It is not. There is indeed much controversy around how to measure the outcomes and the overall effectiveness of Head Start (as well as around what versions of Head Start should be measured), with many researchers finding extensive and long-term gains (see here, here, here, or here), Others, the only ones that Cofnas cites, finding little of value (here or here). In what would be an ironic twist if it weren’t so frankly horrible, the very source that Cofnas cites as showing that all measurable gains fade by 3rd grade noted a few marked exceptions, one of which was some key gains made by disadvantaged Black children (see here, p.xxxiv). (Cofnas might respond that these gains were not gains in IQ test-taking ability, which is his only interest, but if so, that should be made explicit; it was not for example the only interest of proponents of Head Start, nor is it an obvious reading of his claim about “nothing that can be detected”!)
There are of course legitimate questions about whether Head Start is “worth it” – it might be that, for example, the money spent on it would be better spent if simply given to poor families to do with as they please. Or it might be that to maintain some of the important gains made by children in Head Start, we would need to ensure that the schools to which the children eventually went were of much higher quality than they in fact are, and that their families were kept out of poverty. But to claim that the research unequivocally shows it to have been a failure is simply inaccurate.
But even if that particular program was a failure, the claim that this shows that environmental interventions are in general useless is, again, simply wildly out of touch with reality. If we know anything at all, we know that environmental changes can result in massive changes in IQ test-taking performance, Educational Attainment, and other similar measures (if this weren’t true, the “Flynn Effect,”which found that IQ have increased throughout the 20th century, couldn’t exist). We may not know what the most effective ways are of making those changes, or precisely what environmental changes matter, but we do know that such changes can have a profound effect.
If one wished to criticize the Head Start program, there are a number of reasonable avenues, and reasonable people can disagree on how big and how important (if at all) are the impacts that the program has had on outcomes. But there is absolutely nothing in the program that presupposes the falsity of the hereditarian hypothesis, and if somehow the hereditarian hypothesis were known to be true it would have no impact whatsoever on the justification for the program’s continuing, and if the hereditarian hypothesis were known to be false, this would have no impact whatsoever on the arguments in favor of ending it and trying something new. Indeed, to claim that the truth of the hereditarian hypothesis would undermine the potential usefulness of a program like Head Start is to make precisely the mistake that Lewontin warned against in his “Analysis of Variance and Analysis of Causes.” Just because a difference in outcomes is associated with genetic differences does not mean that the trait in question will not be sensitive to environmental intervention. And quite obviously, a lack of group-differences, or even a lack of heritability, could not imply that early intervention of the sort Head Start used would be effective; the malleability of IQ test-taking ability based on developmental environment might be very high, but the particular kind of intervention pursued by Head Start still might not do much work at all. It might for example be the case that the program is entirely useless for fixing problems associated with race and racism, as too much of the cognitive damage has already been done by environmental harms associated with racism before children start it. Or the interventions might just be of the wrong sort entirely. In general, the only way to tell if a program that makes particular interventions would be effective, is, as Lewontin suggested, to try them and see.
Cofnas then goes on to list some of the potential dangers of discovering that racial differences in average IQ scores are caused by differences in the frequencies of genes that act straightforwardly on neurological development across a range of developmental environments (here, I interpret Cofnas’ position rather charitably, as those are distinctions he fails to make in this paper and perhaps fails to truly appreciate). But these worries have nothing to do with policy, in the end. Cofnas might well be right that the discovery that differences between populations in frequencies of some alleles that are causally responsible for differences in cognitive ability will not inevitably result in another Holocaust, and that such discoveries would not inevitably result in increased prejudice and stereotyping. But given that Cofnas himself has just argued for cutting funding for a program aimed at helping disadvantaged youth, on the basis of either a wild misunderstanding of the basic science or just flat-out dishonesty, I think it is fair to say that there are indeed potential dangers here. Even if a policy recommendation (“stop spending money trying to help poor black children do better in school”) does not in fact follow from accepting the hereditarian hypothesis, that’s exactly the conclusion Jensen wanted us to draw, and a conclusion that Cofnas apparently continues to support. So the claim that we should view the aims of the proponents of the hereditarian hypothesis with less suspicion and fear of bad consequences is rather belied by the very paper making that claim.
As far as I can tell, Cofnas in the end makes no suggestions whatsoever for policies or practices that we might put in place now such that if there turned out to be the kind of differences hereditarians posit, we’d be better off for having implemented those policies. If his goal really was to “work through” the moral implications of such potential findings, he has utterly failed to accomplish it. In some ways, this failure is entirely expected. The “moral implications” of group differences, if any, are determined entirely by the details of how those differences are realized; in the absence of knowing those details (or even having assumed some details as a premise), we can ultimately say nothing of value. In a talk, Eric Turkheimer once had the audience reflect on what would happen if we found two populations that differed in their average IQ test-taking ability, and discovered that the source of the difference was that in one population there was a very high rate of untreated phenylketonuria (PKU). PKU is the result of a rare genetic disease that can cause a metabolic disorder, one result of which can be impaired cognitive development; detected early, it can be very effectively managed with a special diet. Arguments about subtle statistical effects would of course fall away entirely and the heritability of IQ test-taking ability would cease to matter at all to the analysis. More importantly, the policy implications would be startlingly clear – testing infants for PKU, and managing those who tested positive via the usual dietary interventions, would be called for (and these implications would be the same for both populations!).
But perhaps it isn’t entirely true, after all, to say that Cofnas makes no policy suggestions. The few policy suggestions that he does make can best be summarized as a negative – please ignore racism, even when it is really obvious, and let’s also pretend that racism doesn’t actually harm people. He writes:
Linda Gottfredson (2005b) observes that “currently, racial parity in outcomes is often treated as the ultimate standard for fairness and lack of parity as a measure of White racism” (p. 317). Those who deny that there is evidence in favor of hereditarianism are forced to conclude that phenotypic differences between groups “must be artificial, manmade, manufactured. Someone must be at fault” (p. 318). However, if hereditarianism is true, then it may be that no one should necessarily be blamed for different average outcomes among groups. There is no theory of justice that says it is right to falsely blame a group of people for wrongs they did not commit because confronting the exculpatory evidence causes us discomfort.
I do not think that I am being unfair or exaggerating when I say that only someone utterly blinded by their own racism could write the above. There is, to the best of my knowledge, no case, no case whatsoever, where a failure for there to be “racial parity in outcomes” is treated as the “ultimate” measure of White racism. The “ultimate” standard for discovering White racism is and always has been finding places that Whites treat otherwise similar Blacks and Whites differently, in ways that disadvantage Blacks. We should be ashamed that such cases are nearly ubiquitous, and so easy to find. To pretend that it might be the case that differences in e.g. average Black and White household wealth in the U.S. are caused not by the legacies of racist practices (including the state-supported destruction and theft of Black wealth) but rather by the ‘natural’ outcome of cognitive differences, is simply obscene. Claiming that this might be the case is roughly the equivalent of, if after you’d caught me on video breaking into your home, stealing all your valuables, beating you half-to-death, and then stealing all the money in your bank account, I claimed that your only evidence of wrong-doing on my part was the current difference in our households’ wealth, which could after all be due to almost anything. This, it seems to me, is more or less what Gottfredson is pretending, and where Cofnas is following her. I feel strongly that morally decent people who deliberately ignore real racial injustices, both historical and contemporary, or pretend that they don’t matter, should be ashamed, and that the proper attitude to hold towards such people is one of profound disgust.
So far from providing a list of current policies that are premised on the assumption that whatever genetic basis there is that impacts directly on IQ test-taking performance across a wide range of developmental environments, that basis is distributed equally in all socially ascribed racial categories, Cofnas has in the end provided nothing. And the claims that he does make about the value of pursuing such research go beyond being merely “speculative.” He writes that “We cannot understand the nature and evolution of intelligence unless we can explain why selection favored different levels of general intelligence, or specific intelligences, under different conditions” (140). This claim is simply baffling. If Cofnas is merely suggesting that some diversity in levels of “intelligence” (on some operationalization of that term) among living things is necessary to understanding the evolution of intelligence, this might be true, but it is completely irrelevant to the topic of the paper. If the claim is that absent interesting group differences in evolved intelligence on some operationalization, we cannot understand the evolution of intelligence, this simply seems false. Our difficulties with understanding the precise evolutionary history of human intelligence has little do with the lack of diversity between extant groups, and there is no particular reason to believe that the discovery of such groups would itself yield interesting information about the evolution of intelligence. (Again, everyone writing about the evolution of human intellectual ability should keep Lewontin’s “Questions We Will Never Answer” in mind, and be prepared to explain why the difficulties he presents are surmountable. They might be surmountable, but to simply ignore them goes beyond wishful thinking.) One might reason that if a sufficiently large variety of populations were found with interesting differences of the sort Cofnas imagines, and if we knew enough about the different environments that they had encountered during the relevant parts of their evolutionary history, then we might be able to use comparative methods (making use of appropriate phylogenetic corrections, etc.) to generate partial understandings of the adaptive significance of the traits (and differences in the traits) in question. But of course, we are not in anything like that situation, and there seems no hope whatsoever that we will discover information that makes that picture of extant human diversity seem even remotely plausible.
In the end, Cofnas provides just another weak rehashing of the same tired evidence usually cited to support the hereditarian position, the same promissory note that despite all evidence to the contrary and history of abject failure researchers will soon learn enough about the genetics and developmental biology of the traits underwriting IQ test-taking performance that we’ll be able to clearly identify the relevant pathways with confidence, and the same claims that despite all the papers (like his own) being published defending the existence of group differences of the sort hereditarians are convinced exist, there is somehow an unfair barrier to the publication and dissemination of this kind of ‘research.’ The stated goals of the paper – to show the ways in which current policy is shaped by the assumption that the hereditarian hypothesis is false and the to show ways in which if the hereditarian hypothesis were shown to be true it would require a kind of moral reflection we have not prepared for – go not only unmet, but almost entirely unengaged.
1 Some readers will no doubt be familiar with the controversy surrounding the piece discussed here, Nathan Cofnas’ “Research on group differences in intelligence: A defense of free inquiry.” After it was published in the journal Philosophical Psychology, a group of scholars (myself included) wrote a reply, criticizing both the paper itself, and the review process that led to what was in our view a hopelessly flawed piece being published. We describe here our reasons for being dissatisfied with how the journal handled our reply, and after some social media attention, one of the editors (Cees van Leeuwen) announced his resignation, and remaining editor and editorial board proposed publishing the commentary as a “letter to the editor.” See here for additional details about this story).
2 Impertinent editorial footnote by jpj: I was one of those people!
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.