The extended life cycle: a multilevel–multiscale modelling framework for extended evolutionary dynamics

Abstract A complete explanation of evolutionary change requires reconciling processes that operate across multiple levels of spatial/temporal organization. Development, the processes by which traits are generated, unfolds over an individual’s lifetime; heredity, encompassing the diverse forms of information transmission, occurs across generations; and population and ecological change often take place across longer horizons. Capturing these layered dynamics in a formal mathematical framework is essential for advancing evolutionary theory in line with recent conceptual developments, such as extended evolutionary theory (EET) or the extended evolutionary synthesis (EES). In this work, we introduce the extended life cycle (ELC), a multilevel–multiscale mathematical modelling framework that centres the life cycle as the fundamental unit of evolutionary analysis. In contrast to traditional gene-centric approaches, the ELC formalism captures the multilevel causal architecture of biological systems by modelling development, heredity, population and ecological dynamics across distinct, but interacting ontological levels. Each level is expressed as a stochastic state-space model, with multiple scales within each level. We demonstrate the utility of the ELC through a Bayesian learning application to multiscale habitat selection and show how our framework can predict latent trajectories generated from complex multilayered dynamics.