"Science shouldn't offend". So says Nature Magazine, which used to focus on science publishing

As this guest article explains, the new editorial policy of Nature Human Behavior basically trashes free and open inquiry and scientific merit.

Critical Mass was founded on the idea that free speech and open inquiry are among the greatest gifts of modern civilization, allowing creativity in both science and culture to flourish. They also allow all of us the opportunity to have our own biases confronted, with the possibility of actually changing our minds.

Quillette Magazine, edited by the redoubtable Claire Lehmann, is a magazine that provides a great example of this because it frequently confronts political correctness and conventional wisdom, producing strong reactions online by both the right and left. All the better, I say. Some on the left seem to think that Quillette is a right wing shill, but that is probably because they don’t read it. For example, last year, Quillette published several articles criticizing ‘anti-vaxxers” and got deluged with complaints from some right-leaning readers.

For full disclosure, I have written numerous articles for Quillette, and I continue to be impressed by the journal’s willingness to uphold its standards of free speech and open inquiry, regardless of which online mob might be provoked. It is a source of useful information and analysis, and that is what counts.

Yesterday, Bo Winegard published a compelling piece in Quillette entitled The Fall of Nature, which exposed a new editorial policy of Nature Human Behavior that is suggests that scientific articles that pass the filter of quality and accuracy may nevertheless be withheld from publication if the research results they describe might inadvertently cause some individuals or groups to ‘feel’ stigmatized, especially groups that are most vulnerable to “racism, sexism, ableism, or homophobia.”.

Weingard carefully parses the new policy and exposes the logical and scientific problems that can result, leading inevitably to the sense that, as he puts it, “academia becomes more like a partisan media organization than an impartial institution”.

The article is worth reading, and nicely complements a number of pieces I have written in Critical Mass. As a result, I am particularly grateful to Claire Lehmann for giving me permission to reprint it here in its entirety.

The Fall of ‘Nature’

A once-respected journal has announced that it will be subordinating science to ideology.

“And science, we should insist, better than other discipline, can hold up to its students and followers an ideal of patient devotion to the search to objective truth, with vision unclouded by personal or political motive.”
~Sir Henry Hallett Dale

Although the modern prestige bestowed upon science is laudable, it is not without peril. For as the ideological value of science increases, so too does the threat to its objectivity. Slogans and hashtags can quickly politicize science, and scientists can be tempted to subordinate the pursuit of the truth to moral or political ends as they become aware of their own prodigious social importance. Inconvenient data can be suppressed or hidden and inconvenient research can be quashed. This is especially true when one political tribe or faction enjoys disproportionate influence in academia—its members can disfigure science (often unconsciously) to support their own ideological preferences. This is how science becomes more like propaganda than empiricism, and academia becomes more like a partisan media organization than an impartial institution.

An editorial in Nature Human Behavior provides the most recent indication of just how bad things are becoming. It begins, like so many essays of its kind, by announcing that, “Although academic freedom is fundamental, it is not unbounded.” When the invocation of a fundamental freedom in one clause is immediately undermined in the next, we should be skeptical of whatever follows. But in this case, the authors are taking issue with a view very few people actually hold. At minimum, most academics will readily accept that scientific curiosity should be constrained by ethical concerns about research participants.

Unfortunately, the authors then announce that they also wish to apply these “well-established ethics frameworks” to “humans who do not participate directly in the research.” They are especially concerned that “people can be harmed indirectly” by research that “inadvertently … stigmatizes individuals or human groups.” Such research “may be discriminatory, racist, sexist, ableist, or homophobic” and “may provide justification for undermining the rights of specific groups, simply because of their social characteristics.” Because of these concerns, the Springer Nature community has worked up a new set of research guidelines intended to “address these potential harms,” explicitly applying ethics frameworks for research with human participations to “any academic publication.”

In plain language, this means that from now on, the journal will reject articles that might potentially harm (even “inadvertently”) those individuals or groups most vulnerable to “racism, sexism, ableism, or homophobia.” Since it is already standard practice to reject false or poorly argued work, it is safe to assume that these new guidelines have been designed to reject any article deemed to pose a threat to disadvantaged groups, irrespective of whether or not its central claims are true, or at least well-supported. Within a few sentences, we have moved from a banal statement of the obvious to draconian and censorious editorial discretion. Editors will now enjoy unprecedented power to reject articles on the basis of nebulous moral concerns and anticipated harms.

Imagine for a moment that this editorial were written, not by political progressives, but by conservative Catholics, who announced that any research promoting (even “inadvertently”) promiscuous sex, the breakdown of the nuclear family, agnosticism and atheism, or the decline of the nation state would be suppressed or rejected lest it inflict unspecified “harm” on vaguely defined groups or individuals. Many of those presently nodding along with Nature’s editors would have no difficulty identifying the subordination of science to a political agenda. One need not argue that opposing racism or promoting the nuclear family are dubious goals in order to also worry about elevating them over free inquiry and the dispassionate pursuit of understanding.

Suppose someone discovers that men are more likely than women to be represented at the tail end of the mathematical ability distribution and therefore more likely to be engineers or physics professors. Does such a finding constitute sexism, if only by implication? Does it stigmatize or help to negatively stereotype women? Are the authors of the editorial contending that journals should not publish an article that contains these data or makes such an argument? The very vagueness of these new guidelines allows—or rather requires—the political biases of editors and reviewers to intrude into the publishing process.

As the editorial proceeds, it becomes steadily more alarming and more explicitly political. “Advancing knowledge and understanding,” the authors declare, is also “a fundamental public good. In some cases, however, potential harms to the populations studied may outweigh the benefit of publication.” Such as? Any material that “undermines” the “dignity or rights of specific groups” or “assumes that a human group is superior or inferior over another simply because of a social characteristic” will be sufficient to “raise ethics concerns that may require revisions or supersede the value of publication.”

But no serious scientist or scholar contends that some groups are superior or inferior to others. Those who write candidly about sex and population differences, such as David Geary or Charles Murray, routinely preface discussion of their findings with the unambiguous declaration that empirical differences do not justify claims of superiority or inferiority. Nevertheless, the editorial is a warrant to attack, silence, and suppress research that finds differences of any social significance between sexes or populations, regardless of whether or not such differences do in fact exist. The empirical claim that “men are overrepresented vis-à-vis women at the extreme right tail of the distribution of mathematical ability” can therefore be rejected on the basis that it may be understood to imply a claim of male superiority even if no such claim is made, and even if it is explicitly disavowed.

Sensing the dangerous and censorious path they are walking, the authors pause to offer a sop to those of us who still believe in the importance of academic freedom:  

There is a fine balance between academic freedom and the protection of the dignity and rights of individuals and human groups. We commit to using this guidance cautiously and judiciously, consulting with ethics experts and advocacy groups where needed. Ensuring that ethically conducted research on individual differences and differences among human groups flourishes, and no research is discouraged simply because it may be socially or academically controversial, is as important as preventing harm.

This is not at all reassuring. Asking ethicists to assess the wisdom of publishing a journal article is as antithetical to the spirit of science as soliciting publication advice from a religious scholar. Who are these “ethics experts” and “advocacy groups” anyway? I am skeptical of ethical expertise. I am especially skeptical of ethical expertise from an academy more inclined to reward conclusions that support progressive preferences than those that emerge from empirical study and rational thought. I am more skeptical still of advocacy groups, which exist to pursue a political agenda, and are therefore, by their very nature, a good deal more interested in what is useful than what is true.

Imagine the outcry on the Left if a journal announced it would be consulting pro-life advocates before publishing an article about the effects of abortion on wellbeing. Or if it decided to consult conservative evangelicals when evaluating an article about the effects of adoption by homosexual couples. The journal is effectively announcing the employment of sensitivity readers, who it can safely be assumed, will invariably recommend the risk-averse option of suppression whenever the possibility of controversy arises.

Before they set out their new guidelines, the authors take a moment to self-flagellate, with a cookie cutter denunciation of science for its dismal history of inequality and discrimination. Still, “with this guidance, we take a step toward countering this,” they say as if it were an act of atonement. I find that I am more positive about the science of the past than the editorial’s authors, and more gloomy about the social-justice-oriented science of the future they are proposing. Yes, humans are flawed and fallible and always will be, so we must accept that science will forever be an imperfect endeavor. But the best way to correct its imperfections is not to demand the capitulation of science to ideology, but to remain alive to our biases and devise mechanisms that can compensate for them. Trying to counter past bias by replacing it with a new kind of bias is self-evidently nonsensical—like trying to conquer alcohol consumption by replacing beer with hard liquor.

Predictably, the proposed editorial guidelines focus on the needs and sensitivities of groups perceived to be marginalized and identified by race, ethnicity, class, sex, and sexual orientation, religious and political beliefs, age and disability. And naturally, the guidelines themselves are as vague and troubling as the rest of the editorial. The authors reiterate that they want to extend protections for research participants across the entire publishing process. “Harms,” they note, “can also arise indirectly, as a result of the publication of a research project or a piece of scholarly communication—for instance, stigmatization of a vulnerable human group or potential use of the results of research for unintended purposes (e.g., public policies that undermine human rights or misuse of information to threaten public health).”

Like almost everything else in the editorial, this claim is unhelpfully ambiguous and politically contentious. Furthermore, possible real-world harms (or benefits) that result from the publication of academic papers are incredibly, perhaps prohibitively, difficult to anticipate and measure. Would a paper that finds homosexual men to be more promiscuous on average than heterosexual men result in the “stigmatization of” or “harm to” a “vulnerable human group”? The answer would depend in no small part upon the respondent’s view of homosexuality and how capacious or otherwise their definitions of “stigmatization” and “harm” are.  

The notion that homosexual men are more promiscuous than straight men might produce some negative stereotypes about the former. But it could also raise awareness of the disproportionate dangers posed to homosexual men’s sexual health by unprotected promiscuity, which might in turn lead to a reduction in the rate of sexually transmitted infections. We simply do not know. This is precisely why peer review should only consider the plausibility and theoretical importance of articles, not their unknowable political and moral effects.

The new guidelines state that even if a project were to be reviewed and approved by appropriate committees, editors “reserve the right to request modifications” or even “refuse publication … or retract post-publication” if it contains content that:

Is premised upon the assumption of inherent biological, social, or cultural superiority or inferiority of one human group over another based on race, ethnicity, national or social origin, sex, gender identity, sexual orientation, religion, political or other beliefs, age, disease, (dis)ability, or other socially constructed or socially relevant groupings (hereafter referred to as socially constructed or socially relevant human groupings).

Or:

Undermines—or could reasonably be perceived to undermine—the rights and dignities of an individual or human group on the basis of socially constructed or socially relevant human groupings.

Or:

Embod[ies] singular, privileged perspectives, which are exclusionary of a diversity of voices in relation to socially constructed or socially relevant human groupings, and which purport such perspectives to be generalisable and/or assumed.

No examples are adduced, of course, so it is difficult to know what kind of content would commit these retractable iniquities. Could a discussion of group differences in cognitive ability “reasonably” be perceived to undermine the “rights and dignities of an individual or human group”? Would an exploration of sex differences in homicide rates? Would an analysis of political differences in cognitive rigidity? Would a test of the association between religiosity and pro-sociality? And who is to be the judge of what is and is not “reasonable”? And what does or does not constitute “undermining”?

Ambiguity is piled upon ambiguity to expand the capricious purview of the censor. It does not require clairvoyance to predict that these criteria will not be consistently applied. It may be considered racist to point out that a disproportionate number of crimes are committed by black Americans, but it will surely not be considered misandrist to point out that a disproportionate number of crimes are committed by American males. Even those who work ardently for the triumph of progressive ideas and values should shudder. Not only will these guidelines further degrade the already embattled prestige of science, but they offer remarkable deference to the idiosyncratic moral concerns of editors and reviewers which are subject to change at short notice. As radical feminists have recently discovered, those who sit within the progressive Overton window today may find themselves thrust outside of it tomorrow—victims of a censorious system they thought they were erecting in their own interests.

The guidelines intended to combat racism begin by announcing that race and ethnicity are sociopolitical constructs. This is a contentious claim (even if we could agree on what is meant by “sociopolitical construct”), and it is one that I happen to think is unsupported by either the data or by sound philosophical argument. Even so, the section goes on to assert that:

Biomedical studies should not conflate genetic ancestry (a biological construct) and race/ethnicity (sociopolitical constructs): although race/ethnicity are important constructs for the study of disparities in health outcomes and health care, empirically established genetic ancestry is the appropriate construct for the study of the biological aetiology of diseases or differences in treatment response.

This convoluted reasoning will surely only aggravate existing double standards in discussions of race and ethnicity—those who contend that society is teeming with racism can point to disadvantages experienced by racial groups, but those who contend that disparities are caused by behavioral differences are flatly told that race does not exist. Would these standards be consistently applied to a paper that examined racial disparities in police shootings and a paper that examined racial differences in crime rates?

“Racism,” we are told, “is scientifically unfounded and ethically untenable. Editors reserve the right to request modifications to (or correct or otherwise amend post-publication), and in severe cases refuse publication of (or retract post-publication), racist content.” But since “scientifically unfounded” material can be rejected on that basis alone, there is no need to invoke potential harms to vulnerable groups as an additional justification. The authors’ implication seems to be that “racism” should be understood (unlike the “reverse” variety) to apply to some groups and not others, and that what the authors wish to oppose is research that might discredit the efficacy or justness of, say, affirmative action. But since the editorial and its guidelines provide no examples of supposedly racist content, it is difficult to know.

The section on sex, gender, and sexual orientation is similarly vague and tendentious. The authors claim, for example, that, “there is a spectrum of gender identities and expression defining how individuals identify themselves and express their gender.” Well, maybe. But this is an ideologically provocative claim—and certainly one with which many people across the political spectrum will strongly disagree. Brazenly avoiding any pretense of objectivity, the authors then itemize the usual laundry list of putative gender identities, “including, but not limited to, transgender, gender-queer, gender-fluid, non-binary, gender-variant, genderless, agender, nongender, bi-gender, trans man, trans woman, trans masculine, trans feminine and cisgender.” Gender norms, we are told, are “not fixed but evolve across time and space. As such, definitions will require frequent revisiting…” It is hard to imagine that more than five percent of conservatives would agree to this, but that is evidently of no concern to the authors. The chief purpose of this section seems to be to signal to other progressives, “We are on your side,” and to send a corresponding signal to conservatives: “You are not our people.”

The editorial closes by declaring that, “Researchers are encouraged to promote equality in their academic research,” and that editors reserve the right to retract articles that are “sexist, misogynistic, and/or anti-LGBTQ+.” Again, no examples of these retraction-worthy crimes are offered, and so familiar objections resurface. Is a paper that contends that men are physically stronger than women “misogynistic”? Is a paper that examines the correlation between trans-identity and other mental illnesses “anti-LGBTQ+”?

Science is a human activity, and like all human activities, it is influenced by human values, human biases, and human imperfections. Those will never be eliminated. The banner of science has undoubtedly been waved to justify, excuse, or otherwise rationalize appalling crimes and atrocities, from the racial pseudoscience of the Nazis to the blank slatism (and Lysenkoism) of the communists. But the correct response to these distortions is not to endorse a highly partisan vision of science that promotes a progressive worldview, alienating all those who disagree and further encouraging doubt about the objectivity of scientific endeavor. The correct response is to preserve an adversarial vision of science that promotes debate, disagreement, and free inquiry as the best way to reach the truth.

-Bo Winegard