Special Bisanar Seminar Series

Abstract

Universal Darwinism is the observation that any system that undergoes variation, selection and heredity—whether it is living or not—is the subject of Darwinian evolution. The purpose of this talk is to explore the quantitative implications of this bare-bone idea. Taking my cues from seminal contributions of R. Fisher in genetics and statistics, I will show  that fitness distributions are subject to statistical universality: in the space of all possible distributions, fitness distributions are attracted to a low-dimensional manifold of limiting (and generally non-Gaussian) shapes.

This result, I will argue, is the central mathematical prediction of Universal Darwinism, and as such is directly testable from evolutionary data of any kind. I will present an illustration from molecular evolution and contrast my findings with other quantitative approaches such as "fitness wave theory".