I’m actually surprised that the article below was published in The Proceedings of the National Academies of Science (PNAS), one of the more high-quality science journals, just a tad below Science and Nature in prestige. It has had a reputation for being “progressive” (e.g., woke), one that I discussed last year when Steve Pinker had an email exchange with National Academy of Sciences (NAS) President Marcia McNutt.
After McNutt, along with the Presidents of the National Academy of Medicine and of the National Academy of Engineering, issued a pro-affirmative-action and pro-DEI statement on June 30, 2023, Pinker wrote McNutt pointing out that such statements are incompatible with the NAS’s mission. His email (reproduced at the link above) contained this bit:
I would like to express my disquiet at the recent NAS Statement on Affirmative Action. The desirability of racial preferences in university admissions is not a scientific issue but a political and moral one. It involves tradeoffs such as maintaining the proportion of African Americans in elite universities at the expense of fairness to qualified applicants who are rejected because of their race, including other racial minorities such as Asian Americans. Moreover it is a highly politicized policy, almost exclusively associated with the left, and one that majorities of Americans of all races oppose.
It’s not clear to me how endorsing one side of a politically polarizing, nonscientific issue is compatible with the Academy’s stated mission “providing independent, objective advice to the nation on matters related to science and technology”.
The problem is worse than being incompatible with the Academy’s mission; it could substantially harm the Academy’s goal of promoting politicians’ and the public’s acceptance of science. Extensive research has shown that rejection of the scientific consensus on evolution, anthropogenic climate change, and other scientific topics is uncorrelated with scientific literacy but predictable from political orientation: the farther to the right, the greater the rejection of evolution and climate change.
McNutt wrote back, but declined to have her answer reproduced on this site. Nevertheless, from Pinker’s response to her response, you can gather that she defended the stand of the original three-President statement, apparently written to criticize the Supreme Court’s decision that college admissions could not be based on race.
Steve said this, among other things (again, see the whole of his email at the site):
Even more concerning, the statement could have been lifted out of the pages of any recent left-wing opinion magazine, since it reiterates the current conviction that racial inequities are primarily due to “past and current racial discrimination and structural, systemic, and institutional racism in education” and to “individual bias and discrimination.” Entirely unmentioned are other potential causes of racial discrepancies, including poverty, school quality, family structure, and cultural norms. It is surprising to see a scientific organization attribute a complex sociological outcome to a single cause.
Finally, the statement, and your letter, equate diversity of ideas with diversity of race. The advantages of intellectual diversity are obvious (though I have not seen any statements from the Academy addressing the shrinking political diversity among science faculty, nor the increasing campaigns that punish or cancel scientists who express politically unpopular views). The assumption that racial diversity is the same as intellectual diversity was exactly what the Supreme Court decision singled out and struck down, since it carries with it the racist assumptions that black students think alike, and that their role in universities is to present their race-specific views to their classmates.
Dr. McNutt replied, but again did not give permission for her letter to be reproduced.
I have to give McNutt credit, then, for allowing the two-page piece letter to be published, as it contains a pretty explicit criticism of McNutt, especially of a later piece by McNutt and Crow, “Enhancing trust in science and democracy in an age of information,” published in Issues in Science and Technology. McNutt and Crow bemoan the detachment of science from society and society’s ethical values and make this statement, which is debatable:
Therefore, we believe the scientific community must more fully embrace its vital role in producing and disseminating knowledge in democratic societies. In Science in a Democratic Society, philosopher Philip Kitcher reminds us that “science should be shaped to promote democratic ideals.” To produce outcomes that advance the public good, scientists must also assess the moral bases of their pursuits. Although the United States has implemented the democratically driven, publicly engaged, scientific culture that Vannevar Bush outlined in Science, the Endless Frontier in 1945, Kitcher’s moral message remains relevant to both conducting science and communicating the results to the public, which pays for much of the enterprise of scientific discovery and technological innovation. It’s on scientists to articulate the moral and public values of the knowledge that they produce in ways that can be understood by citizens and decisionmakers.
While the good part of McNutt and Crow’s message is their call for scientists to explain the scientific results of their work to the public, it’s a different matter to ask scientists to “produce outcomes that advance the public good.” That can be an explicit aim of science, as in producing golden rice or Covid vaccines, but many scientists doing “pure” science are motivated by simple curiosity. That curiosity, too, can have salubrious social outcomes, but most of the time it just enriches our knowledge of the universe.
Further, it seems excessive to asks scientists to also “articulate the moral and public values of the knowledge that they produce.” Are scientists experts in morality? And what are “public values”—the latest ideology of the times? One might think from this piece, and the correspondence above, that McNutt does favor the politicization of science, but along the lines of “progressive” politics.
Thus I was pleased to see this letter, by evolutionary molecular biologist Ford Doolittle, appear as an opinion piece in the latest PNAS. Here he takes issue not only with the politicization of science, but explicitly with McNutt and Crow’s article. You can read the letter by clicking on the screenshot below, or read the pdf here:
But Doolittle begins with a thesis that I find dubious: that “group selection”—the differential reproduction of genetically different human groups—has led to our drive to understand nature—indeed, to selection on many species to “understand” their environment. But, says Doolittle, group selection has not led to the drive to integrate science and social values. (Other species don’t really have “social values” anyway). Bolding is mine:
Most humanists and scientists now agree that science is special in its relationship to the real world, more special than are other human activities—religion and politics, for instance. But philosophers of science keep arguing about why that should be. There is, I believe, a good evolutionary explanation of why—one that incorporates what is often called group selection (1). But group selection will only move humans closer to the truth if researchers and others take care to ensure that social values don’t distract or mislead.
So, my plea is that scientists and others ensure that science remains independent from social values. Social values are constraints—limitations on the evolutionary process. I worry that mixing science and social values hampers scientific progress.
and this from Doolittle’s piece:
My evolutionary argument starts with the contention that there is a selective advantage at all levels to having a better map of reality. Having a better understanding of the world promotes fitness. Living things at all levels (genes, cells, multicellular organisms, species, multispecies communities, tribes, nations of humans, and even broader cultural frameworks) that have such a better map of the world leave more progeny or last longer than living things that don’t, all else being equal. This has been true from the beginning of life.
. . . And, of course, human groups—tribes, nations, and broader cultural collectives—that have better knowledge of the natural and cultural world have a better chance, all else being equal, than those that have less adequate knowledge.
This is a bit mixed up, for evolutionary group selection is a genetic phenomenon, not a cultural one, and in this case would argue that some groups of humans genetically endowed with better knowledge of the environment would survive and reproduce better than less-informed groups. And, over time, this would spread the genes for acquiring more and more accurate knowledge about the universe.
The problem, as always with group selection, is that, because it depends on the differential survival and reproduction of groups, it is much slower than selection acting on individuals harboring genes producing an ambition to know. Those genes would spread within groups and there is no bar to having individuals with such genes. (I think Doolittle’s misconception here is that only groups can differ in their urge to understand.) Group selection is usually invoked to explain the evolution of traits that are advantageous to groups but not individuals, like pure altruism towards nonrelatives. But over time, group selection has fallen out of favor; see this eloquent critique by Pinker on Edge: “The false allure of group selection“).
Doolittle notes that occasionally Darwin was a group selectionist, but in fact A. R. Wallace, in his first exposition of natural selection, published simultaneously with Darwin’s, was even more of one!
But I digress; natural selection acting on genes (Dawkins’s “replicators”) and the bodies bearing them (the “vehicles”) is sufficient to produce the drive to know. Still, in the end it hardly matters. Humans are curious creatures, and there’s doubtlessly a big effect of evolution on that trait.
And it doesn’t even matter whether our drive to know is evolutionary rather than purely social if one argues, as Doolittle does, is that mixing science and politics is bad for science. Here’s Doolittle’s peroration about why mixing science and ideology is bad:
But outside certain limits, society is not ethically uniform, and important values are not shared. We are so politically polarized now that there is an ever-present danger of “weaponizing” the pursuit of knowledge, and thus of the results of earnest inquiry being dismissed by those whose social values disagree with those of scientists. We embrace political polarization to the detriment of both scientists and the scientific enterprise.
Science is based on the assumption that our collective understanding of the world, though always imperfect, generally improves over time and that there is no trade-off between what we think we should do and the scientific truth. As the 18th-century philosopher David Hume noted, you can’t derive “ought” from “is.” The consilience of scientists’ personal social values (which surely have changed over time) and modern, fundable science is precisely why I see current trends in politicization as dangerous to the scientific enterprise—a worry underscored when these trends are viewed through an evolutionary perspective going from genes to individual cells to tribes to broader cultural frameworks.
We scientists should be even more careful not to allow what we think is “right” (what we ought to do) to influence what we think is “true” of the world. What we think is right changes with time and context, but what we think is true should be our eternal goal.
Doolittle notes that “it is inevitable that science which does not agree with some aspect of society’s current value system has little chance of getting funded,” but that isn’t 100% true. Sure, if you want to show that there is “structural racism” in an academic field, then your grant may well get funded, but it could also get funded if you’re studying the systematics of ants, or string theory, or the migration distance of Drosophila. Those kinds of studies get funded based on merit, not on “society’s current value system”—unless, that is, you define “value system” tautologically as “what people want to fund”.
In the middle of the article, though, he’s careful not to go too hard after McNutt. But, again to her credit, she let this be published:
As an ethical constraint, the sentiments of Marcia McNutt, the president of the National Academy of Sciences, and her coauthor Michael Crow, president of Arizona State University, might serve as a contemporary example (10). They write that science must “produce outcomes that advance the public good,” citing the Columbia University philosopher Philip Kitcher to remind us that “science should be shaped to promote democratic ideals.” Science, in other words, should be constrained by human social values. Perhaps they meant by this that science functions best (that is, provides better understandings of the world) in democratic societies, rather than arguing that democracy is best for our species. The former is an epistemic value, but the latter is a social value and thus an unnecessary constraint.
McNutt and Crow’s social values are mine, too, and those of many scientists, I hasten to add. . . .
As I said, if you want to stretch “ethical values” to become “the idea of what sorts of questions need answering,” then of course the science that people do, and especially the science that gets funded, will generally comport with social values. But McNutt, Crow and Doolittle are talking, I think, about prioritizing science that matches our current ideology (i.e., justifying DEI initiatives, documenting inequities, or trying to show that indigenous “ways of knowing” are coequal to modern science). Alternatively, McNutt and Crow might urge us not to do forms of science carrying any possibility that they could have bad social consequences (the classic example is studying group differences in IQ).
But it would have behooved Doolittle to give more examples of the kind of science that people are objecting to now. I’ve written a lot about the ways that ideology is intruding in science in detrimental ways: two examples are my paper with Luana Maroja on ‘The ideological subversion of biology” and also the Abbott et al. paper “In defense of merit in science.”
I see this has been a rather rambling post, involving group selection, the debasing of science by politics, and debates in the scientific literature. So be it, and again I’m pleased that NAS President McNutt has allowed an op-ed to be posted in “her” journal that explicitly takes her to task. That is in the finest tradition of allowing open discourse in the literature.
h/t: Anna, Luana
I agree this is a good sign for PNAS. I too was surprised to see this published, given the treatment we received with our merit paper — Jerry wrote about it here:
https://whyevolutionistrue.com/2023/06/02/our-wall-street-journal-op-ed-free-at-last/
The idea that science should be driven by morality is wrong and dangerous. Morality keeps changing, but scientific truth endures. There are many moral systems around, as exemplified by different religions and belief systems, but the laws of nature transcend belief systems. Previous efforts to subjugate science to ideology — as in the USSR — led to disastrous results for science and society.
Say what you will about this country, there are many salutary conversations going.
Why can’t group selection be considered just one more evolutionary mechanism? Selfish genes may run the show, but when both individuals and groups die their genes die with them. Consider a stochastic catastrophe like meteorite which blasts an island and its entire endemic population to oblivion, or a newly emerged pathogen that wipes out all within a related group. What about a culture whose practices clash with those of their neighbors such that a genocidal war erupts? All those generations of selected genes have come to naught!
Stochastic events are the reason why group selection is so much slower than natural selection among individuals. By definition there are fewer groups than there are individuals. Selection works more effectively in large than in small populations because the effect of selection on allele frequencies is less easily opposed by the stochastic loss of advantageous alleles due to nonselective deaths. The population of groups among which group selection can act is always much smaller than the population of individuals among which natural selection can act.
In a large population and a small population with the same allele frequencies, stochastic effects (genetic drift) are stronger in the small population because the stochastic loss of a single advantageous allele is relatively large. Selective effects (adaptation) are stronger in the large population because the stochastic loss of a single advantageous allele is relatively small.
So yes one can imagine a scenario in which “generations of selected genes have come to naught” but on average natural selection will have a larger effect on the net rate of adaptation.
Group catastrophes may be uncommon, and thus slow in the grand scheme of things, but they sure get the job done. I think that anything that takes a gene out of the gene pool should qualify as natural selection.
Thanks. Was helpful for this non-scientist’s understanding.
I suggest that you read Pinker on group selection at the link above. And can you name any trait that evolved by group selection that could NOT have evolved by individual selection? I cannot.
What about the differences between say Europeans and Australian aborigines? I was assuming the lighter skin colour was a benefit at higher latitudes for absorbing sunlight for vitamin D.
Also lactose tolerance appears in certain groups (some African tribes, most Europeans. These were groups who had domesticated cattle).
Those examples are explainable by selection acting on individuals within a group – e.g., an individual who is lactose-tolerant in a society that raises cattle has access to more sources of nutrition, and presumably better chances of having children. Of course, when that selection is active on all or most individuals, it shifts the value of the trait averaged over that group.
Group selection is different: it refers to traits that allow a group to thrive, even if they may be neutral or disadvantageous to individuals, e.g., fighting bravely to defend your village from attackers, rather than running away. A village where everyone displays that behavior may have better chances of repelling an attack instead of being wiped out, but there’s usually a free-rider problem: one individual who finds a way to avoid fighting will not usually make a difference in the outcome for the group, but will have higher chances of survival than his peers in the first line of defense, and of course the first individual to evolve the “fight bravely and sacrifice yourself for the others” gene in a group of cowards will tend to have a very short life, making the evolution of such traits tricky to explain.
OK. Now I understand. Thanks.
Physicist here (now working in IT); I concur with @bruce morgan.
I have just read Pinker’s piece and all the comments including JC’s and Dawkins’ ones.
To my mind, it’s actually counter-intuitive NOT to “conceive of groups of organisms as sitting at the top of a fractal hierarchy with genes at the bottom, with natural selection applying to each level in parallel ways”.
@Mike here says “group selection is so much slower than natural selection among individuals”. But so what? The velocity of a molecule of air at room temperature, due to thermal agitation, is of the order of 500 m/s. The velocity of most tornados is of the order of 20-80 m/s. The phenomena more striking to us is actually the “weaker” one and is “emergent” upon a finer structure on a “lower” level.
This does not amount to negate the validity of a “reductionist” approach; but we don’t even explain all chemistry by speaking in terms of quantum mechanical interactions, let alone biology! So while in principle the gene, as the replicator, is the “fundamental” unit, I don’t see why when speaking of vehicles we should limit ourselves to some aggregates (individuals) but not others (groups).
Dawkins wrote in reply to Pinker: “is it helpful to call it group selection? Not for my money it isn’t. It ‘s no more helpful than to say that a lion is a product of ‘group selection’ because a lion is a ‘group’ of cells. (…)
why is it helpful to talk of a group phenotype?”
He is not claiming that it would be WRONG, but that he doesn’t see the USEFULNESS of doing so. Well, why couldn’t it be useful? Why should the extended phenotype include a gene’s effect on things that an individual builds (a dam), effects on OTHER individuals (a gene expressed in male that influences a behaviour in females) but not when the other individuals “influenced” simply belong to the same group (which implies, if not kinship, at least a certain amount of genetic similarity)? What about characteristics that can be defined only in the context of an interaction between individuals, e.g. a tendency to cooperate for example?
In the same comments Joseph Henrich wrote: “Rejecting group selection models is like banning spherical coordinates because you prefer to do your verbal reasoning in Cartesian coordinates” and Michael E. Price “multilevel selection theory seems vital for explaining the major transitions in the evolution of life”
It seems to me there is an historical (i.e. biased, prejudicial) rejection of the WORDS “group selection”. But this is just linguistic.
David C. Queller: “I think of these less as alternative theories that make different predictions than as two different languages describing the same world”.
John Tooby: “Readers might detect a note of testiness here and there in the responses to these proposals. The reason why is that those who have worked long and hard to clear out the jungle, and to lay a rigorous foundation for modern evolutionary biology see this progress jeopardized by the rhetorical conflation of faulty or implausible (and dormant) ideas with these new models (…)
traits can evolve in which a functional phenotype is expressed across individuals or by a group (…)
Collaborative hunters must have adaptations in individual brains that cause them to hunt in groups, and must also hunt in groups (a group-level functional phenotype) to reap the advantage. Individual selectionists consider this a case of individual selection (…) Group selectionists consider it a case of group selection (…) THE APPROACHES ARE DIFFERENT IN NAME ONLY”
Theoretically, group selection could work, but if we want real EVIDENCE for it, as opposed to individual selection, then you should point to a trait that could only have evolved via group selection and not by individual selection: i.e. a genetic trait that is adaptive for the reproduction or survival of groups but not of individuals. If g.s. is common, then such traits could be common.
They are not. In fact, I can’t think of one morphological, physiological, or behavioral trait that is good for the group but deleterious for the individual possessing it. (Speciation rates may be a group trait, as I discuss in Chapter 12 of my book.
If you think you have a defense of group selection, I urge you to write a paper on it and publish it in a good evolution journal, giving possible examples.
I’m obviously not qualified enough 😛 I haven’t even finished reading all books I intend to about evolution
But if I may, I think your answer is misleading. The point is precisely that group selection, properly intended, i.e. when it is NOT a resuscitation of obsolete ideas, is NOT a different thing from the “gene’s eye view”. Of course it is “ultimately” advantageous to the gene, but this is not the same as it being advantageous to the individual, as the concept of kinship already considers. “Inclusive fitness” already means averaging on all the individuals where the gene is expressed; since the “selfish” gene is actually pretty cooperative with the other genes in the pool (because it has to build a functional body), and those genes are distributed in the bodies of the individuals in the group, I contend that the mathematic is already considering the group.
You may find traits overly important, but the core of evolution, ultimately, is almost tautological (which is not to deny the idea was pure genious): replicators that have left more copies of themselves are more present in the environment. The question to me is then how you define the aggregates of genes (why individuals) and which is the “proper” time scale to consider?
Referring to Dawkins idea of the lion as a group of cells: it may seem provocative to him, but it’s not clear to me why we shouldn’t regard it precisely in that way. The immune system capacity of attacking cancerous cells can be seen precisely as group selection, why not?
If one restricts himself to traits I suspect it would be difficult to explain the major transitions; I agree that speciation rates could in fact be one of the applications.
Nobody said that group selection could not work; the question is how often does it work. And there’s no evidence that it is at all common. You clearly didn’t understand my answer, so let’s end this exchange. I’d urge you to read more, as you admit that you’re not qualified to discuss this question.
Brownian motion vs. tornado is not a good analogy. Also what you’re calling velocity is speed. Air molecules under Brownian motion have high speed but their velocities are vectors with ~infinite different directions, so the net movement is ~zero. Air molecules in a tornado have lower speeds but their velocities are vectors with ~same directions so the net movement is large and damaging. This speed vs. vector difference is not like the difference between natural selection vs. group selection.
Glad to see a rejoinder to McNutt and Crow published in PNAS. I don’t see how the group selection argument fits in and I don’t see how it helps. Too much baggage if you asked me. And I don’t agree with Philip Kitcher’s point that science should be “shaped to promote democratic ideals,” but I don’t know the context in which he develops that thesis, so my disagreement is provisional. (I know him a bit from my graduate student days.)
To my limited mind, the principle to be applied is that science institutions and science journals should stay out of politics and should focus on the dissemination of science. This is not to say that science is immune from political and social influences. It is not immune. But let’s not institutionalize such biases as a matter of policy. That would seem to be the wrong turn to take.
Maybe Doolittle should talk to Berlinski about group selection to see who is most wrong? hehehe. No disrespect intended – fine people can be wrong.
Steven P’s letter is, as always, top notch quality. He is one of the people I admire most.
D.A.
NYC
I found it rather difficult to parse some of what Doolittle was saying. But that bit about how “of course” groups with better knowledge of the natural and cultural world have a better chance …. That seemed very likely wrong to me. What he seemed to say was that such groups would have better fitness under natural selection. But populations with high education tend to have fewer children, don’t they?
It’s come to something when a scientist attracts praise for not suppressing or censoring an article that mildly criticises her own hypothesis.
+1
Yes, why praise her?
About subscribing to The Free Press: I found you, Frau. Thank you!! Where there’s a will there’s a way… That’s really helpful information. Appreciate it
Did you get my message? I couldn’t reply directly because your comment was too deeply nested.
TFP is $8 /month or $80 /year. I go monthly because annual subscriptions don’t usually refund if you cancel.
Yes. I got it. Thank you so much!
I believe research scientists should do their work based on interest and curiosity.
But in a representative democracy, it is the citizens who should determine what is valuable for tax dollars to invest in and promote. The citizens, when presented with scientific research and discoveries, should debate the ethical and cultural issues.
What actually happens is that capitalism makes these decisions. Who has the most money to invest and where will they profit the most often “wins” any debate. So I can understand the motivation of a scientist wanting ethical scientists to take the reins.
However, this is just another form of control: their values determine the direction of research vs those who make a profit. I still believe we the people should direct important ethical issues that impact our society, not scientists or billionaires.
Capitalism can make such decisions well, and works based on what a broad market of individuals want. Capitalism is a “we the people” mechanism, and profit is the guidance. Ideas for things that people don’t want to use (and buy) don’t get funded. It is voting by we the people. I may disagree with some of the outcomes but customer voice wins.
I’m no Marxist, but I don’t believe the market = democracy in action. There is more to citizenship and ethical debate than selecting my consumer commodities thoughtfully.
I totally agree. I was only pointing out that capitalism does allocate resources to R&D in an efficient manner in an economy. I’m also of the opinion that both capitalism and citizen-led decision making via our elected representatives have error-correcting mechanisms.
I know nothing about George Mason University, but its response to the Supreme Court’s decision on affirmative action, which was posted in the comments below the WEIT article containing Pinker’s emails, is in interesting contrast to the NAS presidents’ statement: https://archive.is/Hb2De
Very good economics program at GMU.
Plus they have Tyler Cowen.
Ramble On
-Led Zeppelin
Slightly different, but hey – why not!