Fitness and evolvability are central yet often weakly connected concepts in evolutionary theory, in part because they have been deployed across different disciplines. This project advances a “grain-sensitive” strategy for clarifying the fitness–evolvability nexus by explicitly tracking how traits, types, environments, and evolutionary success are characterized in various contexts using different grains of description.

Building on earlier work by Bourrat, Deaven, and Villegas (2024), I will work with Cristina Villegas to refine the quadripartite framework we proposed, distinguishing (i) environments of reference vs. not-of-reference and (ii) types of reference vs. not-the-type of reference, thereby separating classical fitness (adaptedness), robustness of adaptededness, and evolvability. A central aim is to re-evaluate our framework without assuming weak ergodicity. While weak ergodicity can make long-run success comparatively tractable, it sits uneasily with evolvability: by rendering long-run behavior largely insensitive to path-dependent features, it effectively washes out historical structure that many accounts treat as constitutive of evolvability. We will therefore propose a updated framework that retain evolvability’s distinctive connections to novelty, variance generation, and the historically conditioned accessibility of phenotypic space.