Sentences Have Been Models All Along

Ron Giere and others have recently written a number of papers suggesting that science proceeds best by “modelling” phenomena, not by describing phenomena in sentences. The main idea (I think; at any rate, one of the main ideas) is that whereas sentences have to be true (or approximately true) in order to provide good representations of the world, models represent the world not through truth but through some other property — something like goodness of fit.

But if certain forms of pragmatism are right, sentences represent the world in the same way as models — not because models can be true or false, but because “true” and “false” as applied to sentences are best seen as indications of goodness of fit and nothing else.

Maybe this paper’s already been written, but I haven’t seen it or heard of it. Maybe somewhere in the Wittgenstein literature? (Of course Wittgenstein himself already said something like this a long time ago, but I’m looking for something that’s said it since the recent interest in models by Giere and others.)