The interaction between lifetime learning and evolution

Abstract

The impact of learning on evolution has been subject to significant debate and analysis for over a century. Much progress has been made in explaining how learning, as a non-inherited, highly adaptive form of plasticity, affects evolutionary trajectories with the potential to become genetically assimilated via the Baldwin Effect. To date, most computational and mathematical models devoted to understanding the effect of learning on evolution only focus on how environmental input impacts the expression of a small number of independent traits. However, West-Eberhard suggests learning mediates selection pressures over many phenotypic traits thereby driving genetic correlations; these correlations, in turn, may increase the effectiveness of the learning, which with circular reinforcement, increases the effectiveness of genetic evolution. Here this concept is extended so that learning has the capacity to guide the evolution of linkages between innate behaviours leading to genetic assimilation of that learning through canalisation. This feedback is bi-directional: whilst learning improves the fitness signal to evolution, correlations between innate behaviours channel learning to enable the discovery of optimal phenotypes that would not normally be found by evolution alone. This dynamic is explored through novel computational models of two different causal paths between learning and evolution; one where learning has an indirect effect on the innate behavioural traits by changing the experienced environment and one where learning acts directly on the expression of innate behaviours. Both these models show that guided by learning, the evolution of correlations between innate behaviours produce high-fitness behavioural phenotypes faster and more consistently than without learning. Further, it is shown that the introduction of linkage between innate behaviours can relax the conditions under which learning can become genetically assimilated. All this work supports the emerging view of the phenotype as a vital actor in the process of evolution.

More information

Identifiers

ORCID for Richard Watson: orcid.org/0000-0002-2521-8255

Catalogue record

Export record