This is the third of three guest posts by Professor Jonathan Kaplan of Oregon State University (see Part One here and Part Two here). Professor Kaplan is a noted philosopher of biology who has published extensively on biological race and IQ among other topics.
Part Three: When a Paper Doesn’t Take Its Own Thesis Seriously
I’ll end this series with the paper with which I have the longest history – Jonathan (“Jonny”) Anomaly’s 2017 “Race Research and the Ethics of Belief.” I was one of the Very Mean ReviewersTM that Anomaly wrote a blog post about, and yes, in my review I suggested that the paper was unpublishable. Clearly I was wrong, as it made its way into print in more or less the same form that I originally read. But pace Anomaly’s interpretation, my main complaint about the paper, as a reviewer, had little to do with the paper’s (weak) defense of hereditarianism (though my scorn for Nicholas Wade’s shoddy defense of scientific racism was openly and forcefully expressed in my review). Rather, my main complaint was focused on what I viewed as the paper’s more or less entirely incoherent argument structure.
I’ll note at the outset that some readers may find this account to be, as John Jackson put it, a little bit “inside baseball.” The gap between what the paper purports to do, and what it actually goes, is stark, but making sense of what the paper purports to do does get us a bit into actual philosophy (that isn’t even about race).
The paper starts off suggesting that its primary goal will be to develop and defend a version of John Stuart Mill’s harm principle, but as applied to beliefs; the paper, Anomaly states, will defend the view that “it is morally defensible to hold socially consequential beliefs for which there is imperfect evidence only if doing so is unlikely to impose significant, uncompensated harm on other people” (289). This is to be seen as part of a more general approach based on the claim that “our epistemic and practical goals can come apart,” such that while a particular claim might be rational thing to believe (given all the evidence), one might be better off overall not believing that claim.
This is obviously an interesting claim, and equally obviously a quite controversial claim; it is also one with a rich history and associated literature. Perhaps the “classic” text in this case is David Pears 1984 monograph Motivated Irrationality, but a survey of the field would also prominently include work by Alfred Mele (here or here), some key elements of classic work by Kavka and Gauthier (here), and the game theoretic approaches of for example Skyrms (here) and Gintis (here), to name a few. It is rather a surprise, then, to find none of these authors cited (especially after, as part of my Very Mean ReviewTM I noted this issue, and provided citations to some of the key pieces). Indeed, while it wasn’t his main interest, Kavka’s work provides numerous examples (many admittedly rather far-fetched) of places where it would be in an agent’s rational self-interest to adopt a belief that they in fact have good reasons to believe is false; this would be one way of solving the problem in “The Toxin Puzzle,” for example (if one could come to believe, falsely, that the only way to get the money was to drink the toxin, one would ‘win’), but as Kavka notes, insofar as “our beliefs are constrained by our evidence,” one, alas, cannot just “believe whatever one wants to believe” (36). (The literature on the “epistemology of ignorance,” especially the ways in which ignorance of key facts about the world can be deliberately cultivated, would also be relevant here, and again, Anomaly’s paper makes no effort whatsoever to connect his stated position on belief acquisition and revision to these broader philosophical conversations.)
Aside from its failure to engage meaningfully with the literature surrounding this problem, Anomaly’s paper also fails to appreciate key distinctions, or even to recognize them. Consider the following passage:
Suppose ‘P’ is the proposition that ‘my client is guilty,’ and it is held by a defence lawyer. The belief may be justifiable in light of the evidence and ultimately true. But if a defence lawyer allows herself to form this belief, especially if her client is accused of an especially gruesome crime, she might exert less effort defending her client in court. Similarly, if ‘p’ is ‘my basketball team will lose,’ and it is believed by most members of the team, they may not try as hard to win the game. Whether I ought to believe that my client is guilty or my basketball team will lose depends in part on what the point of this belief is—to form a justified belief about the world, given my current evidence, or to accomplish the goal of defending my client or winning a game. (288)
It should not need pointing out that these are really two quite radically different kinds of cases. In one case, believing that one’s client is factually guilty might change the likelihood that they will be found to be legally guilty, but it can have no bearing whatsoever on whether they in fact committed the crime in question! In the other, believing that the team on which I am competing will lose in fact changes the likelihood that my team will lose. While in both cases, I might have non-epistemic reasons for adopting a belief that is not the best fit to the current evidence, only in one case does my so-adopting that belief change the likelihood that the belief itself is true. Anomaly not only fails to be clear about why these two cases are considered together, or why treating these two cases as similar makes sense in this context, but the paper fails to even recognize that the critical difference between these cases.
But in any event, none of this matters, since on any remotely reasonable interpretation the paper ends up having no real interest in the harm principle as applied to beliefs after all. The structure of the paper’s argument, in a nutshell, is as follows:
- Thesis: [Where there is not dispositive evidence in favor of a belief] We are sometimes morally required to refrain from believing a thesis for which there is good evidence, when so-believing that evidence would cause [significant, etc.] harm to others [and do no similar weighty good].
- Beliefs about differences in e.g. average cognitive abilities between different racial groups being genetic in origin might fit the definition in (1) of a belief that does harm, where there is non-dispositive evidence in its favor.
- But in fact, properly understood, no such harms would accrue to believing in the existence of differences in average cognitive ability (mediated by straightforward genetic causes) between groups, if that is where the evidence in fact points.
- So the harm principle does not apply, and if the evidence were to point in the direction of there in fact being average cognitive differences between racial groups that were genetic in origin, we ought to believe that.
- [BTW: the evidence does so point in at least some cases, though this has nothing much to do with the argument about the harm principle.]
This structure is, on the face of it, truly bizarre. There is no reason to introduce a controversial proposal such as the harm principle as applied to belief acquisition and revision, only to show that the principle does not apply in the only case that you are considering in detail.
Or perhaps there is a reason after all, albeit one that is (deliberately?) elided. One claim often made by hereditarians is that those people arguing against them are doing so either in bad faith, or are self-deluded; hereditarians claim, in other words, that those of us opposed to their views know, or ought to know, that they are right, but so dislike [what we see as?] the potential social consequences of their being right that we pretend not to believe them (see e.g. Sesardic 2010 – “Nature Nurture Politics”). But of course, the argument that those of us opposed to hereditarianism make is not in fact “we don’t like the consequences of this, so we choose to think it is false,” but rather “we reject the hereditarian position, because there is no remotely strong or compelling evidence in favor of that view, and some good reasons to think that the view as usually expressed is deeply empirically problematic, and in addition, good reasons to think that environmental differences are likely driving at least most of the observed developmental differences of interest” or something like that. Not being able to find someone who rejected hereditarianism on the basis of a harm principle for belief, it might appear to a reader of his paper that Anomaly simply made one up, in his own person, to refute. This would also serve as an explanation of why the view developed regarding belief acquisition and revision was so weak, and why it failed to engage meaningfully with the relevant literature – defending that view was never in fact what the paper was actually about!
The paper has one more oddity that works to undermine any claim to be seriously engaged with its own thesis. While it is true that opponents of the hereditarian hypothesis base their opposition in large part on the very poor quality of the evidence that hereditarians are able to produce, as I hope the last blog posts made clear, much of the anger at hereditarians is generated precisely by the policy positions that many hereditarians in fact claim follow from their view. So for example, Jensen explicitly linked his hereditarian views to opposition to better school funding for minority students, and to opposition to programs designed to help disadvantaged students do better; Cofnas follows him there, despite the truth or falsity of the hereditarian view being entirely independent of and irrelevant to judgments regarding the success or failure of those programs. Part of the point of The Bell Curve was to convince the reader that in fact our society was (mostly) fair, and that the radically different outcomes experienced by different racialized groups (as well as differences within racial groups) were the result of differences in average cognitive abilities between those groups playing out in a system of mostly fair equality of opportunity within a constrained market, and not to e.g. discrimination and the legacies of explicitly racist policies and practices. Gottfredson (2005), as noted in the previous post, suggests that accepting the hereditarian hypothesis would end, or at least result in less, “recrimination” and fewer accusations of racism, and that this would be a good thing, apparently because she herself doesn’t believe racism is all that serious a problem. In other words, modern “Jensenism” has always been about justifying the (terrible) status quo for Black Americans, by claiming that it is not the result of the gross racism and discrimination faced by Black Americans, but just the ‘natural’ result of genetically mediated cognitive sorting in a (mostly) fair environment (or something like that). (As in the previous post, note well that reasonable people do not think that the best evidence we have for the continued existence of racism, and profound effects of past racism, is differences in outcomes, but rather direct evidence of continued racism itself, and clear historical evidence of past racism and its effects. One reason, as I suggested in my paper “Ignorance, Lies, and Ways of Being a Racist,” to believe that many hereditarians are in fact racist is their apparent inability or unwillingness to perceive racism, and its importance, even when the evidence is staring them in face, and their willingness to outright lie about the existence of equal opportunities in cases where it is blindingly obvious that none in fact exist or have existed.)
So if one were going to attempt to apply the harm principle with respect to belief in racial differences, that would be the place to do it – might it be the case that even if there were evidence that favored a straightforward genetic hypothesis about the average difference in Black/White IQ test-taking ability, we would be better off ignoring it? Could one make an argument that we should ignore evidence that favors hereditarianism, given the easily observable racism that actually exists in the U.S. today and that clearly impacts people’s life-chances, and the obvious existence of legacies of past obviously racist policies, and the ways in which evidence in support of that position is supposed to undermine the need for policies to redress those obvious wrongs? First, of course, the question itself would seem to suggest that the answer is “no” – redressing those obvious wrongs is clearly the right thing to do, whatever the situation would be in the absence of those wrongs, so the point is moot. But oddly, a reader of Anomaly’s paper will not get any answer to those questions, as Anomaly chooses instead to focus on the difference in average I.Q. test-taking ability between Ashkenazi Jews and other populations.
This is, on the fact of it, an extremely odd decision – no one, to the best of my knowledge, has ever thought that much hinges on this difference. There are no arguments that we should cut school funding for people who aren’t Ashkenazi Jews, for example. Nor was Harvard’s (and other ‘elite’ universities’) past attempts to limit the number of Jewish students admitted a consequence of a belief in the intellectual superiority of Ashkenazi Jews – antisemitism was not based primarily on the belief that Ashkenazi Jews were more intelligent, still less on their supposed superior IQ test-taking ability. And indeed, Anomaly’s discussion of antisemitism is oddly ill-informed and ahistorical. Anomaly claims, without referencing any sources, that:
For example, many racists and conspiracy theorists continue to explain Jewish success by appealing to their moral depravity—the idea that theft, conspiracy, or even sorcery are the sources of Jewish success. When beliefs like these are widely shared, they are more likely to lead to anti-Semitic actions (like the pogroms in Russia) and policies (like the Nuremberg laws in Germany).
On the basis of this, Anomaly goes onto claim that far from leading to social ills, accepting the (provisional) truth of the hereditarian hypothesis with respect to Ashkenazi intelligence would be socially beneficial, “if it gives a better explanation of the relative success of different groups in different domains” (294).
These claims are, frankly, simply bizarre. The range of antisemitic stereotypes was quite broad, and varied enormously in these different locations & times. The idea that the Russian pogroms were motivated by unexplained “Jewish success” is not supported by most historians’ reading of the record (rather, political shocks & economic instability, scapegoating, and various antisemitic religious beliefs seemed the main drivers). Nor were the Nuremberg laws a response to a misunderstanding about Jewish intelligence. To suggest, as Anomaly seems to, that if only the truth of Jewish superior intelligence had been better understood, Jews would have been subject historically to less violence, seems fundamentally out of touch with the history of antisemitism.
But again, whatever the arguments about the potential harmlessness of accepting the (wildly unsupported) claim that Ashkenazi Jews perform better, on average, on IQ tests, for straightforward genetic reasons, these are really not, and have never been, the focus of hereditarian debates. It seems misleading in the extreme to write a paper that 1) purports to defend the thesis that a harm-principle might require us to refuse to accept a claim that we have evidence in favor of, and 2) takes as its example a version of a claim that is, in the form expressed in the paper, not usually associated with particularly morally problematic policy decisions, and 3) then goes on to say that the claim in question does not support the harm principle. While it is perhaps impossible to know what an author was thinking, taking this paper at face-value makes it out to be absurd. Taking it to be defending a hereditarian view, and opposing a view that, while not actually defended by opponents of the hereditarian view, is attributed to those opponents by many hereditarians, at least makes the paper make sense, albeit at the cost of making it seem rather grossly intellectually dishonest.
In the end, these are all disappointing articles. Not because they defend an unpopular claim, but because they do so in ways that fail to meaningfully engage with the very debates in which they purport to be interested. Research into the so-called “hereditarian hypothesis” has not progressed much, despite the many researchers who have tried to find ways to support it. Compared to the progress made in e.g. supporting hypotheses regarding the evolution of lactase persistence, hypotheses regarding purported psychological differences remain stuck at the very first step – establishing that there is a difference mediated by straightforward genetic differences, or even properly operationalizing the phenotype in question in a meaningful way. Until researchers can suggest ways of testing their hypotheses using the same kinds of tools used by competent evolutionary biologists, there is no good reason to continue to pretend that the project in which they are engaged is a reasonable one, driven by something other than a straightforwardly racist agenda.
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.