Published online by Cambridge University Press: 01 January 2022
Van Fraassen argues that data provide the target-end structures required by structuralist accounts of scientific representation. But models represent phenomena not data. Van Fraassen agrees but argues that there is no pragmatic difference between taking a scientific model to accurately represent a physical system and accurately represent data extracted from it. In this article I reconstruct his argument and show that it turns on the false premise that the pragmatic content of acts of representation include doxastic commitments.
I am particularly grateful to Roman Frigg, Alexandru Marcoci, F. A. Muller, Bryan Roberts, and two anonymous referees for this journal for extensive feedback on earlier drafts of this article. Thanks also to audiences at Mathematizing Science II at the University of East Anglia, the BSPS Annual Meeting 2014, and the PSA Biennial Meeting 2014 for constructive comments.