Abstract

It has been argued, for instance based on a Quinean holism, that the same term occurring in competing scientific theories cannot refer to the same (theoretical) object or property. A popular response is that the causal theory of reference avoids this threat. I suggest that the causal theory of reference is inadequate as a rational reconstruction of scientific language and argue that Psillos's causal-descriptive theory of meaning is a special case of the descriptive theory of meaning. Assuming the descriptive theory, I show that intertheoretically shared terms can stand for the same concepts when the analytic and synthetic components of a theory fulfill Carnap's conditions of adequacy. To do so, I provide a compact reformulation of Carnap's conditions and show that, contrary to arguments by Winnie and Demopoulos, the conditions need not be supplemented. Specifically, the axiomatization and the observational vocabulary of a theory determine its synthetic component uniquely but leave some freedom in the choice of its analytic component. Besides throwing new light on the process of the rational reconstruction of theories, this result allows for analytic components of theories that provide fine control over which theories share which concepts.