Bayesian Provisionality Applied to an Epistemology of Science

Abstract

Kuhn's work on the socio-epistemology of science gained far wider acceptance among cultural commentators than it did within science, where a strong adherence to Popper's principle of 'falsifiable predictions' remains strong. Many see this as a failure by science to relinquish its modernist traditions of truth and objectivity in the face of postmodern developments in philosophy, despite the highly speculative nature of science's theorising at the frontiers of knowledge. Our paper interrogates this fractured state of epistemology within science and proposes a novel approach based on provisional Bayesian constructs adapted from the field of sensory perception. Written from the perspective of a scientist with an interest in philosophy and a philosopher with an interest in science, our new approach is designed to overcome many of the concerns and limitations of both modern and postmodern philosophies. It could also be regarded as an important step towards reconciling the apparently incommensurable Kuhnian and Popperian models of science to encourage a more inclusive transdisciplinary approach to research across the arts, sciences and humanities.