Note: this is blogpost number 100! A relatively meaningless milestone, as it includes 25+ articles for The Hindu BLink, some re-blogs, and some housekeeping posts, but a milestone nonetheless. To mark it, here is the second half of an essay that I wrote recently about the role of competition and cooperation in biology and among biologists. The first half of the essay was adapted from this post. The second half of the essay is about the recent controversy surrounding the evolution of extreme cooperation and what, if anything, we learn from bitter scientific disputes. If you already understand the textbook versions of kin selection and eusociality, skip ahead to the section titled “Is Cooperation on Shaky Ground?“. Hope you like it!
If helping someone else helps you too, then cooperation can be pragmatic. Many examples of cooperation in both the natural and the human worlds take this sensible form.
In contrast, the evolution of altruism seems paradoxical—an altruistic individual helping someone else doesn’t, by definition, gain anything in return. Under the principles of natural selection, an organism should do everything it can to maximize its own survival and reproduction (or “fitness,” for short). Being altruistic, however, benefits someone else’s fitness, while harming one’s own. How then could altruism evolve?
But altruism does evolve. Some of the most abundant animals on the planet exhibit an especially extreme form of altruism called eusociality. In a eusocial colony, some individuals forego all their fitness by not reproducing; instead, they work to help a single colony member produce an enormous number of offspring. All sorts of insects (ants, wasps, bees, termites, and even some beetles), a few shrimp, and even two species of mole rat, live in colonies of this sort.
From the point of view of a worker in a eusocial colony, who will never bear offspring of her own, eusociality must seem like the ultimate exploitation. But if you think, instead, about the genes that encode the information from which organisms are built, altruism and eusociality can begin to look consistent with natural selection. An allele, defined as a particular variant of a gene, can be said to succeed if there are more copies of it in future generations than there are at present. For the most part, an allele’s fitness increases if it confers some advantage to the individuals that carry it. This leads the individual it lives in to have lots of offspring, some of which will carry the allele as well.
Let’s assume that the allele we’re talking about causes its bearers to cooperate, and consider two siblings who both have the allele. The frequency of the allele will increase equally if, say, one of its bearers has four offspring and another has none, or if both bearers have two offspring each. Cooperation among these two siblings to ensure that one of them has more offspring is, from the perspective of the allele’s success, not inconceivable. And if the allele confers an extra advantage to cooperating siblings—suppose cooperating individuals have a total of six offspring instead of four—then the evolution of cooperation may even be likely.
This logic is captured in a simple equation1 proposed by W.D. Hamilton in 1964, according to which cooperation between two individuals can be favoured by natural selection if the following relationship holds between the costs (c) and benefits (b) of cooperation, and how closely the two individuals are related to one another (R):
Hamilton’s Rule: R > c/b
This equation suggests that, when thinking about the evolution of cooperation from the perspective of an individual, we cannot restrict our measurement of fitness to just the individual’s own offspring. Fitness must stretch to include the offspring of one’s relations as well, weighted by the appropriate measure of relatedness.This expanded version of fitness is described as “inclusive”. So an individual with no offspring of its own, and thus with no direct fitness, can still have inclusive fitness via its relatives. And a behaviour that spreads due to its inclusive fitness benefits is said to evolve by the process of kin selection. So, according to this equation, we expect that for a worker in a eusocial colony, the benefits of bearing her own children are far outweighed by the costs. She instead works to maximize her inclusive fitness by helping raise sisters, directly benefitting her mother’s fitness. Through the lens of kin selection, eusociality is the ultimate collaboration.
For the longest while, I accepted all of the above as fact, unaware of the uncertainty among scientists about how exactly kin selection and inclusive fitness work. But this uncertainty rumbled below science’s surface, out of most biologists’ sight, until the publication in 2010 of an explosive paper by Martin Nowak, Corina Tarnita, and E.O. Wilson2. From the fallout from this explosion, it has become clear that we don’t understand cooperation in the natural world all that well
Is Cooperation on Shaky Ground?
In their paper, Nowak et al.2 claim that the contributions of kin selection and inclusive fitness to biology “must be considered meagre.” Provocatively, they suggest that the inclusive fitness theory hasn’t even been tested properly, and that it probably won’t be, because actually calculating inclusive fitness is not straightforward. This is especially true in populations where interactions among individuals are so complicated that their effects on fitness cannot simply be added together. Imagine you and your friend, separately, do something good for me—I can add your contributions together to measure the total benefit I’ve received. But now suppose you and your brother work together with me to help my cousin, but that you refuse to similarly help my uncle—adding together the benefits and costs that I’ve received from you and your brother becomes much trickier.
Additive interactions, Nowak et al.2 argue, are rare in nature, so “inclusive fitness theory…only works for very restrictive scenarios.” In other words, Nowak and colleagues think that studying inclusive fitness isn’t worth the trouble.
But measurement problems are just one entry in the list of Nowak et al.’s2 complaints. A crucial question that Nowak et al.2 pose is about causality. Did the close relatedness among members of a group of ancestral bees, for example, cause them to evolve eusocial cooperation? Or was the close relatedness among members of a hive simply “the consequence rather than the cause of eusociality”? Nowak et al.2 claim the latter is true, and this claim formed the centre of several critiques of their paper, which set into motion a seemingly endless exchange of responses.
The Frustration of Learning from an Argument
Reading the protracted back and forth between the challengers and defenders of kin selection is like watching a tennis match in which the ball abruptly changes shape, size, colour, and direction every time it crosses over the net. Five comments3-7 written by over a hundred biologists were published in response to Nowak et al.’s2 paper, accompanied by a response8 from the original authors. Then, in 2015, Xiaoyun Liao, Stephen Rong, and David Queller published a paper9 that served as an extended comment on Nowak et al.’s2 original; a response from Nowak and Ben Allen10 followed, as well as a response to the response11.
You’d hope that reading this lengthy conversation would help to clarify the points of dispute among these scientists, but it doesn’t. As insect biologist and photographer Alex Wild put it in an essay12 responding to the original paper, “You might think that scientists who study cooperation ought [to] show signs of being good at it themselves. But you’d be mistaken.” Consider, for example, the discussion surrounding how to test the hypothesis that high relatedness causes the evolution of eusociality.
In an idealized version of the scientific method, one advances science by testing hypotheses about how the world works. And, implicitly or explicitly, each hypothesis is tested by comparing it to an alternative—we discover what is true by gradually figuring out if one explanation is more true than another, and rejecting the explanation that is less true. Nowak et al.2 contend that biologists studying of inclusive fitness theory have failed to “consider multiple competing hypotheses.” Instead, “when the data do not fit, elaborations of inclusive-fitness theory can be constructed that make them fit.” But Nowak et al.’s2 scathing critique of the absence of alternative hypotheses does not seem to extend to themselves. In constructing mathematical models to understand the evolution of eusociality, Nowak et al.2 don’t factor in variation in relatedness at all. They therefore don’t actually test the hypothesis that high relatedness can cause eusociality to evolve. As Carl Zimmer analogizes13, “it would be as if a team of researchers carried out a study on the effects of diet and exercise on health. Their subjects get different amounts of exercise but stay on the same diet. In the end, the experiment might show that exercise makes people more healthy. But it would not make any sense to also conclude that diet plays no role.” In fact, Nowak et al.2 did the equivalent of feeding their participants the healthiest possible diet—they assumed that relatedness within their hypothetical groups was as high as it could possibly be. As a result, critics9 conclude that Nowak et al.’s2 models “have nothing to say about the importance of relatedness in the evolution of eusociality,” while Nowak et al.2 themselves argue emphatically that “relatedness does not drive the evolution of eusociality.”
Liao et al.9 respond to this problem directly by modifying Nowak et al.’s2 models to incorporate variation in relatedness, and find that eusociality evolves more readily in groups with higher relatedness among its members than in groups with low relatedness. Liao et al.9 see this as a win for the majority opinion, saying that Nowak et al.’s2 “modelling strategy, properly applied, actually confirms major insights of inclusive fitness studies of kin selection.” But Nowak and Allen10 respond by saying that Liao et al.9 “fail to analyse their own models with inclusive fitness theory” and that Liao et al.’s9 work has “no relevance for evolution of eusociality,” because only the unlikeliest of biological scenarios could lead to relatedness varying in the way that Liao et al.9 describe.
Juxtaposed thus, the extent of mutual misunderstanding is remarkable. How can the two groups of scientists, who in principle are united in their goal of understanding eusociality, disagree on something as fundamental as whether their own models are at all relevant to eusociality? Equally baffling disagreements of similar scope pervade the rest of this discussion as well.
As a relatively un-invested reader, I found it nearly impossible to understand if either side was worth listening to. It seemed that both sides were making problematic assumptions, but the consequences of these assumptions came to light only when pointed out by their rivals. And equally frustratingly, neither side engaged with their rivals’ most damning criticisms. Remember the objection, raised by Nowak et al.2, about being unable to measure inclusive fitness in situations complex enough that we can’t simply add bits of fitness up? Not one of the kin selection defenders addresses this criticism properly. More generally, any sort of objectivity, any generous engagement with the opposite side’s views, was hard to come by. I found it difficult to escape the conclusion that this bitter dispute was fuelled mostly by the desire of both sides to win the competition. The unwritten rules of this competition seemed to indicate that winning depends on not conceding any of one’s own weaknesses or mistakes. Winning the competition did not seem to involve much learning, and our understanding of cooperation in the natural world had stagnated as a result.
And I’m not alone in finding this sort of competition futile. “The partisans have become more interested in discrediting the other side than in advancing mutual understanding,” wrote Alex Wild in his exasperated essay12 about this dispute, “small statements are taken out of context and destroyed in straw-man arguments, studies are cherry-picked for rhetorical effect, quotes are mined, and the result is a downwardly tribalistic spiral as frustration grows and everyone starts to hate everyone else. A fine pickle for cooperation research if you ask me.”
Recasting the Impasse
I was all but ready to give up on competition among scientists as a source of good in scientific advancement, when a paper written by two philosophers, called “Kin Selection and Its Critics,” came to my attention14. I was encouraged by both the title, which indicated a consideration of both sides, and the authors’ distance from the dispute, which hinted at the possibility of some objectivity.
And the paper did not disappoint. Beyond simply summarizing the last five years of argument, Jonathan Birch and Samir Okasha14 diagnose why the two sides seem incapable of fruitful communication—they are, in fact, talking about very different things.
In the half century since its formulation1, Hamilton’s Rule has taken on different meanings to different people. The differences between these versions have to do with what exactly the parameters c and b mean in this equation:
R > c/b
In the special version of Hamilton’s Rule, the parameters c and b represent the exact costs and benefits to individuals in a population. The special version applies to a population only when these costs and benefits can be added up easily (when you and your friend, separately, do something good for me, for example). It’s this special version of Hamilton’s Rule, which rarely holds in natural populations, that Nowak and colleagues are arguing against.
But the version that almost every other biologist is defending is a far more general one. The general version can apply to all sorts of situations (like when you and your brother work together with me to help my cousin, but refuse to similarly help my uncle), because all the complexities of different types of interactions get subsumed into the values of b and c. This makes b and c impossibly difficult to actually calculate. However, the general form of Hamilton’s Rule has the mathematical advantage of always being true.
It’s no wonder then, that the two sides of this debate are incapable of finding any common ground. One side is attacking a version of kin selection that is rarely true. The other side is defending a version of kin selection that is always true, but that may not be useful.
But why does the difference between these two versions of Hamilton’s Rule matter to the rest of us? What does it say about how we understand the evolution of cooperation? The answer lies in the gap between our intuition and reality. For most of us, our intuitive understanding of how kin selection works depends on being able to talk about individual costs and benefits (flip back to my initial description of kin selection and inclusive fitness, if you don’t believe me). It’s how we are taught about it, it’s how we read about it in textbooks, and it’s probably how we will teach our students about it. But this dependence on talking about individual costs and benefits means that our intuition describes the special version of Hamilton’s Rule. Our intuition is therefore usually wrong. Birch and Okasha’s14 clarification of the kin selection dispute tells us that we can retain our intuition about kin selection or believe that kin selection applies broadly to the natural world—we cannot have both.
One would hope that any future argument about kin selection between biologists takes heed of their conceptual misunderstandings as well as their biological disagreements. As Birch and Okasha14 say in their paper’s conclusion, “progress is achievable if rival camps of researchers are able to communicate and cooperate” But any communication or cooperation would be impossible without conceptual clarity, and these philosophers offer us that. They also offer us a model for engagement, if we can imitate their “even-handed approach that identifies what both critics and defenders of kin selection have got right.”
So I suppose I’m beginning to see that there may be purpose to a bitter scientific competition after all, if it brings attention to a subject that would otherwise languish, unclarified. I asked both Birch and Okasha about the extent to which the protracted, high-profile dispute influenced their decisions to study kin selection and inclusive fitness. For Birch, it was a direct influence. “I already had an interest in kin selection for other reasons,” he said in an email to me, “but I think it was the controversy that persuaded me that kin selection merited philosophical attention in its own right.”
For Okasha, pondering the controversy has led him to reconsider his views. “I was originally of the view that Nowak et al. were ‘surely wrong’, but more recently have come to the conclusion that they actually made some good points, albeit rather overstated,” he wrote. And though Okasha signed one of several critical responses3 to Nowak et al.2, along with more than a hundred other biologists, he said that “in retrospect, I would not have signed the letter.” If a dispute can spur anyone to change their mind after careful consideration, then the dispute certainly has some value.
Birch also agrees that this sort of competition can be productive, “in so far as it causes the underlying philosophical assumptions of different research programmes to be brought to the surface, potentially giving philosophers of science and other impartial observers a clearer picture of how the two programmes differ and how they might be reconciled. This can then feed back into the science in a productive way, provided philosophers of science make an effort to make their work visible to scientists (and provided scientists pay attention to it!).” Nevertheless, Birch seems doubtful of his impact so far. “It’s hard for me to tell whether I’ve succeeded in influencing either side.”
- Hamilton, W.D. 1964. The genetical evolution of social behaviour. Journal of Theoretical Biology 7: 1–52.
- Nowak, M.A., C.E. Tarnita, and E.O. Wilson. 2010. The evolution of eusociality. Nature 466: 1057–1062.
- Abbot, P, et al. 2011. Inclusive fitness theory and eusociality. Nature 471: E1 – E4.
- Boomsma, J.J., M. Beekman, C.K Cornwallis, A.S. Griffin, L. Holman, W.OH. Hughes, L. Keller, B.P. Oldroyd, and F.L.W. Ratnieks. 2011. Only full-sibling families evolved eusociality. Nature 471: E4–E5.
- Strassmann, J.E., R.E. Page Jr., G.E. Robinson, and T.D. Seeley. 2011. Kin selection and eusociality. Nature 471: E5–E6.
- Ferriere, R., and R.E. Michod. 2011. Inclusive fitness in evolution. Nature 471: E6–E7.
- Herre, E.A., and W.T. Wcislo. 2011. In defence of inclusive fitness theory. Nature 471: E8–E9.
- Nowak, M.A., C.E. Tarnita, and E.O. Wilson. 2011. Nowak et al. reply. Nature 471: E9–E10.
- Liao, X., S. Rong, D.C. Queller. 2015. Relatedness, conflict, and the evolution of eusociality. PLoS Biology 13: e1002098.
- Nowak, M.A., and B. Allen. 2015. Inclusive fitness theorizing invokes phenomena that are not relevant for the evolution of eusociality. PLoS Biology 13: e1002134.
- Queller, D.C., S. Rong, and X. Liao. 2015. Some agreement on kin selection and eusociality? PLoS Biology 13: e1002133.
- Wild, A. September 14, 2010. What’s the big deal with Nowak, Tarnita, and Wilson? Myrmecos
- Zimmer, C. August 30, 2010. Science square off on evolutionary value of helping relative. New York Times
- Birch, J., and S. Okasha. 2014. Kin selection and its critics. Bioscience 65: 22–32.
Thanks to my colleagues Pavitra Muralidhar, Yoel Stuart, Dan Rice, Mara Laslo, and Jonathan Losos for their thoughts/feedback on the topics discussed here and on the piece itself!