Response to Reviewers of Williams Rule

David Sloan Wilson

February 26 2024

I thank the reviewers for their overall approval of the manuscript and their constructive comments, to which I reply in detail below.

Reviewer 1. I thank R1 for his/her overall positive comments and many specific points. R1 states that reference to Williams’ rule isn’t necessary to make the point that MLS theory is widely applicable across topic domains. I respectfully disagree. Adaptation and Natural Selection is too important, both historically and conceptually, to ignore. One goal of my article is to distinguish Williams’ Rule from his second strong claim that higher-level selection is invariably weak compared to lower-level selection. Once the second claim is rejected, then Williams’ Rule can be properly seen as a way of evaluating levels of selection on a case-by-case basis. It is very important for Williams’ Rule to be seen in this light--both within evolutionary biology, where he is widely misinterpreted, and for readers who are learning about him for the first time.

R1 offers the evolution of adaptive genetic mechanisms (such as recombination and mutation rates) as a topic that might challenge Williams’ Rule. I believe that Williams’ Rule will be validated but the topic area is beyond the scope of my article, which is intended largely as a basic tutorial. I look forward to pursuing the matter further with R1, perhaps resulting in a second article. 

R1 points out that higher levels of selection can disrupt lower levels of selection (using sexual recombination as an example) in addition to the reverse. I agree, but this reinforces my point that levels of selection must be determined on a case-by-case basis, as opposed to the categorical rejection of between-group selection as invariably weak compared to within-group selection (Williams’ empirical claim). That said, the evolution of sexual reproduction is one of the few cases where some evolutionary biologists were prepared to acknowledge a role for higher-level selection. 

R1 takes issue with my statement that “Individual-level adaptations are frequently not good for the group, species, or ecosystem.” I am referring to the classic comparison between altruistic and selfish behaviors and have made this more clear in the text. 

At R1’s suggestion, I have acknowledged in a footnote that Williams’ was not the only person critiquing “for the good of the group” thinking, citing David Lack, John Maynard Smith, and Borello’s history of the subject. 

R1 writes at length about avoiding the use of the word “law”, but I have already avoided the word in favor of the word “rule”, which R1 finds more acceptable! 

I am happy to add sexual reproduction to the list of group-level adaptations accepted by Williams and thank R1 for the references. 

On p5, I qualify my statement “evolution is all about relative fitness” to read “evolution within groups is all about relative fitness”. This makes my point about the result of Muir’s experiment without triggering the other issues raised by R1 for this passage. 

With respect, I do not find R1’s comment on my water strider example helpful. I am providing a clear case of a behavioral polymorphism in which aggressiveness is being maintained by within-group selection and passivity by between-group selection. Discussions of evolvability and DNA/RNA polymerases are not germane to this example. Please remember that I am providing a tutorial and the examples must be kept simple while remaining accurate! 

I have revised “when the conditions specified by Williams’ Rule are met” to “when selection operates at the appropriate level.” 

Even though I appreciate R1’s comments, I find that they often take me “into the weeds” in ways that are inappropriate for the accessible tutorial that I am trying to write. On page 9, I am making the simple point that even multicellular organisms are vulnerable to disruptive within-organism selection (e.g., cancer). It would be distracting to bring in topics such as the evolution of senescence. 

Similarly, on p 10, I’m trying to make the simple point that the study of evolution became gene-centric for most of the 20th century, and R1 is pointing out that technological limitations favored within-population analyses over metapopulation analyses. Perhaps this is true, but it is a different subject! 

R1 and I are in agreement that within-group cooperation and between-group hostilities characterized indigenous societies in addition to modern Western societies. 

On p12, R1 and I are in agreement that the invisible hand metaphor does not do justice to the views of Adam Smith. 

On p12,  I am pointing out that the selective disadvantage of prosocial behaviors at the most local scale must be counterbalanced by a selective advantage at a larger scale, which requires a population of groups. For me, this is a more specific requirement than the term “environmental context”. 

I appreciate R1’s comments about the simplifying assumptions of models of sexual reproduction, without needing to incorporate them into the article. 

Reviewer 2. I thank reviewer 2 for his/her constructive comments. 

On p 4-5, I have briefly elaborated on nested series of fitness comparisons.

On p5, I have removed the level of tournaments of tournaments for the Monopoly example. 

P7, thanks to R2 for pointing out my use of jargon in the beaver example, which I have attempted to remedy. 

P 7, I have revised the wording of the political and biological regimes, as suggested by R2. 

P 8, I have reminded the reader about the 2nd chicken experiment. 

p 10, I have removed the section on intragenerational selection as superfluous. 

P 11, I have improved the paragraph on Ostrom in relation to Williams’ rule. 

P 12, I have rewritten the section on CAS1 and CAS2 systems and regard it as much improved.  

P 12-13, I can see that R2 is knowledgeable about the group selection literature and hope that he/she can appreciate my need to touch upon disagreements among the experts in this section without getting too much into the weeds. In my paragraph on the underlying unity of all theories of social evolution, my point about calculating the fitness of a focal individual applies to all models of social evolution (for example, multigenerational groups), not just ephemeral group models associated with the term trait-group. I have therefore removed reference to my 1975 article. I do point out definitions of altruism based on absolute fitness in the following section and reference my own 1990 article distinguishing between weak and strong altruism. 

P 15, R2 makes a good point that highly cooperative groups are not necessarily examples of between-group selection. I have revised the paragraph accordingly. 

P 16, I have elaborated on how cultural evolution can take place at the planetary scale as a decision-making process with planetary welfare as the criterion of selection, as opposed to a raw process of evolution among multiple groups. 

Reviewer 3: R3 is very familiar with the literature and I thank him/her for the very constructive and substantive comments. 

P2. Thanks for pointing out that so-called naïve group selection. One is to invoke higher-level adaptations without invoking higher-level selection. The other—which includes Wynne Edwards-- is to invoke higher-level selection but to regard it as far more pervasive than warranted by the evidence. I now acknowledge and discuss this in a footnote. 

P2. On cancer, I have added the Howe et al 2022 reference but I don’t think that R3’s point about cancers evolving complex adaptations bears upon the basic classification of cancer as within-group selection. Also, from my reading of Aktipis, is it not the case that the high mutation rates of cancers do result in the rapid evolution of multiple traits? 

P3. I have added the Gardner 2013, 2014 references in a footnote acknowledging Maynard Smith’s contributions along with Williams. 

P4. I have added the Birch 2019 reference for equivalence and acknowledge that it is more complex than I am representing in this article, without wanting to get into the weeds. For the record, saying the IFT is the one and only approach to individuals as designing agents requires an a priori commitment to regarding them as designing agents, which is problematic and can’t deal with cases of within-individual selection. 

P4 bottom: For the record, I do not regard the MLS1 vs MLS2 distinction as primary and worth introducing in a tutorial of this sort. All of the core early models of group selection were MLS1 models, centered on the evolution of altruism, which is an individual level trait. Some traits can only be measured in groups, such as group size, but their mechanisms are measurable in individuals, such as cannibalism in the experiments of Mike Wade. This topic is worth an article in its own right but is beyond the scope of the current article. 

P10: I have removed the section on intragenerational evolution, as also suggested by R2.

P10: On Major Evolutionary Transitions, This quote from Strassman and Queller (2010 p 605) also represents my own view: “What makes an organism is high and near-unanimous cooperation among its parts, with actual conflicts among those parts largely absent or controlled.” Szathmary’s 2015 article titled  “Toward Major Evolutionary Transitions Theory 2.0” includes the following section on humans titled “So is it a Major Transition?”: “We see key elements that are highlighted in other transitions; cooperation (including reproductive levelling and food sharing), a form of eusociality, a powerful novel inheritance systems, and living in groups”. I therefore feel on safe ground calling human evolution a MET. I have added these quotes to the text and acknowledged in a footnote that the human MET also differs from other METs in some respects, citing the references provided by R3. As an aside, even if the human case fails to qualify as a MET by some criteria, humans can still be regarded as a highly group-selected species.  

P 12. I have revised the section on CAS1 and CAS2, as also suggested by R2. 

P12, bottom. Dawkins had plenty of time to change his mind prior to 1994 and did not change his mind afterward! One reason I don’t cite something more recent is because Dawkins stopped contributing to the academic literature. 

P14 bottom. I have added that heritability at the microbiome level is an empirical matter.  

P15-16. I have added text that establishes expectations for this section provides references where the readers can learn more. I have elaborated on the possibility of global governance (as also suggested by R2. Finally, I have elaborated on the importance of small groups as the “cells” of multicellular society.