On reciprocal causation in the evolutionary process

Abstract

Recent calls for a revision of standard evolutionary theory (SET) are based in part on arguments about the reciprocal causation. Reciprocal causation means that cause-effect relationships are obscured, as a cause could later become an effect and vice versa. Such dynamic cause-effect relationships raise questions about the distinction between proximate and ultimate causes, as originally formulated by Ernst Mayr. They have also motivated some biologists and philosophers to argue for an Extended Evolutionary Synthesis (EES). The EES will supposedly expand the scope of the Modern Synthesis (MS) and Standard Evolutionary Theory (SET), which has been characterized as gene-centred, relying primarily on natural selection and largely neglecting reciprocal causation. I critically examine these claims, with a special focus on the last conjecture and conclude - on the contrary- that reciprocal causation has long been recognized as important both in SET and in the MS tradition, although it remains underexplored. Numerous empirical examples of reciprocal causation in the form of positive and negative feedbacks are now well known from both natural and laboratory systems. Reciprocal causation have also been explicitly incorporated in mathematical models of coevolutionary arms races, frequency-dependent selection, eco-evolutionary dynamics and sexual selection. Such dynamic feedbacks were already recognized by Richard Levins and Richard Lewontin, well before the recent call for an EES. Reciprocal causation and dynamic feedbacks is one of the few contributions of dialectical thinking and Marxist philosophy in evolutionary theory, and should be recognized as such. I discuss some promising empirical and analytical tools to study reciprocal causation and the implications for the EES. While reciprocal causation have helped us to understand many evolutionary processes, I caution against uncritical extension of dialectics towards heredity and constructive development, particularly if such extensions involves attempts to restore Lamarckian or \"soft inheritance\".