Published online by Cambridge University Press: 01 April 2022
Duhem's and Quine's holistic theses, when properly understood, allow methodologically responsible ways of resolving a conflict between a theoretical system and experience; they only deny the possibility of doing it in an epistemically persuasive way. By developing a “string” model of scientific tests I argue that the pattern of interaction between the elements of a theoretical system arising in response to multiple adverse data can be helpful in locating a “weak spot” in it. Combining this model with antiholistic arguments of Popper, Greenwood, and Lakatos significantly reinforces their joint power.
I am indebted to J. T. Cushing, G. Gale, J. Leslie, E. McMullin, and P. Quinn for many helpful comments, and to an anonymous referee for insightful criticism and suggestions. Special thanks are due to S. V. Illarionov for stimulating discussions of the D-thesis.