Abstract

This thesis examines the theoretical and philosophical underpinnings of the concept of natural selection which is pervasively invoked in biology and other ‘evolutionary’ domains. Although what constitutes the process of natural selection appears to be very intuitive (natural selection results from entities exhibiting differences in fitness in a population), this conceals a number of theoretical ambiguities and difficulties. Some of these have been pointed out numerous times; others have hardly been noticed. One aim of this work is to unpack these difficulties and ambiguities; another is to provide new solutions and clarifications to them using a range of philosophical and conceptual tools. The result is a concept of natural selection stripped down from its biological specificities. I start by revisiting the entangled debates over whether natural selection is a cause of evolutionary change as opposed to a mere statistical effect of other causes, at what level this putative cause operates and whether it can be distinguished from drift. Borrowing tools from the causal modelling literature, I argue that natural selection is best conceived as a causal process resulting from individual level differences in a population. I then move to the question of whether the process of natural selection requires perfect transmission of types. I show that this question is ambiguous and can find different answers. From there, I distinguish the process of natural selection from some of its possible products, namely, evolution by natural selection and complex adaptation. I argue that reproduction and inheritance are conceptually distinct from natural selection, and using individual-based models, I demonstrate that they can be conceived as evolutionary products of it. This ultimately leads me to generalise the concepts of heritability and fitness used in the formal equations of evolutionary change. Finally, I argue that concepts of fitness and natural selection crucially depend on the grains of description at and temporal scales over which evolutionary explanations are given. These considerations reveal that the metaphysical status of the process of natural selection is problematic and why neglecting them can lead to flawed arguments in the levels of selection debate.