The semantic or model-theoretic view of theories and scientific realism

The semantic view of theories is one according to which theories are construed as models of their linguistic formulations. The implications of this view for scientific realism have been little discussed. Contrary to the suggestion of various champions of the semantic view, it is argued that this approach does not make support for a plausible scientific realism any less problematic than it might otherwise be. Though a degree of independence of theory from language may ensure safety from pitfalls associated with logical empiricism, realism cannot be entertained unless models or (abstracted and/or idealized) aspects thereof are spelled out in terms of linguistic formulations (such as mathematical equations), which can be interpreted in terms of correspondence with the world. The putative advantage of the semantic approach - its linguistic independence - is thus of no help to the realist. I consider recent treatments of the model-theoretic view (Suppe, Giere, Smith), and find that although some of these accounts harbour the promise of realism, this promise is deceptive.