New paper: "Nonlinear social evolution and the emergence of collective action"
Ben Allen
Hello LEG,
I am writing to share my paper “Nonlinear social evolution and the emergence of collective action”, by myself and Emmanuel College colleagues and students, which has just been published in PNAS Nexus: https://doi.org/10.1093/pnasnexus/pgae131. This was also the subject of my talk at the 2023 LEG Forum.
This paper represents my effort to bring together the approaches of population genetics, evolutionary game theory, multilevel selection, and inclusive fitness into a cohesive mathematical framework for modeling the evolution of social behavior. The framework is flexible in the variety of social interactions and population structures it can represent—from altruism between diploid relatives to evolutionary games on heterogeneous networks—with a particular focus on nonlinear, multilateral interactions.
We derive a simple, general condition for when selection favors one allele over another. This condition has the form of an inclusive fitness effect, with fitness effects due to actors weighted by relatedness to each possible target. However, in contrast to standard inclusive fitness theory, the “actors” in this case are collectives rather than individuals.
It’s no secret that my previous work has been critical of inclusive fitness theory. I acknowledge that the actor-centric perspective of inclusive fitness theory is a useful thinking tool for many. My criticism focused on the limitations of this approach under nonlinear and multilateral interactions, which do not cleanly separate into effects caused by individual actors. I do not believe linear regression solves this issue, since regression coefficients do not necessarily correspond to causal effects (correlation is not causation).
How, then, to extend the actor-centric perspective of inclusive fitness theory to nonlinear interactions? Our work suggests a novel resolution: Extend the notion of “actor” to collectives. In this perspective, selection acts simultaneously on the behavior of individuals and collectives at multiple scales. This does not lead any particular actor to act as if maximizing inclusive fitness, but rather to compromise and conflict between competing individual and collective interests.
I do not expect that everyone will agree with this proposed resolution, but I do hope it lays the groundwork for more productive conversations going forward. At the very least, I hope that a general mathematical modeling framework, which encompasses the models favored by inclusive fitness theorists and population geneticists with those studied by evolutionary game theorists, can provide a common language for these discussions.
I welcome any feedback and am happy to discuss further with anyone who is interested.
All the best,
Ben
--
Ben Allen
Associate Professor of Mathematics
Emmanuel College
http://plektix.blogspot.com