Abstract

It is argued that the concept of observational equivalence can be made coherent, and that observationally equivalent theories can be theoretically incompatible. However, it is also argued that observationally equivalent, theoretically incompatible theories are difficult to generate in a way that leaves even a prima facie epistemological problem, for they will usually involve superfluous expansions. An account, involving a discussion of scientific realism, is given as to why superfluous expansions are not epistemolgically troublesome. Furthermore, a method is given that allows one to solve the epistemological problems of those theories that are observationally equivalent, theoretically incompatible and not superfluously expanded