An Instrumental Bayesianism Meets the History of Science
As this conference brought together historians and philosophers to examine a topic of common interest, in this response I critically examine Hartmann’s Bayesian approach to analyzing science by studying how far it may be used to model real scientific fields and by locating areas requiring further philosophical work or generalization. Most broadly, Dr. Hartmann’s paper seeks to find a consistent way to explain and characterize the continuity and coherence which the author finds to empirically characterize science in the past and present. In this response I first outline the program which Hartmann pursues and analyze the strategies used to pursue it. Next, while largely sympathetic to the project and underlining that it points towards ways of representing complex and partial forms of correspondence, I argue that the Bayesian formalism deployed is incapable of capturing the essential diversity of approaches and fluidity characteristic of theoretical work in real, living fields of scientific inquiry. It functions as a limited, post hoc formalism, and thus should be considered as one tool applicable for describing “normal science”. It should be noted that, in this discussion, while I enter into the language of Bayesian discourse at various points, I also intentionally make statements which do not readily fit within the assumptions underlying Bayesian analysis, especially concerning the temporal characteristics of the proposed program.
Unable to display preview. Download preview PDF.
- Bovens, L. and Hartmann, S. (2003) Bayesian Epistemology. Oxford: Oxford University Press.Google Scholar