Theories as Complexes of Representational Media

In this paper, we review two standard analyses of scientific theories in terms of linguistic and non-linguistic structures respectively. We show that by focusing exclusively on either linguistic or extralinguistic representational media, both of the standard views fail to capture correctly the complex nature of scientific theories. We argue primarily against strong versions of the two approaches and suggest that the virtues of their weaker versions can be brought together under our own interactionist approach. As historical individuals, theories are complex consortia of different representational media: words, equations, diagrams, analogies and models of different kinds. To replace this complexity with a monistic account is to ignore the representational diversity that characterises even the most abstract physical theory.