Models and the Semantic View
Philosophy of Science
Note to the reader I wrote this paper in . It has circulated quite widely since then, and I have presented various parts of it in talks at various times. One part has been published: "Models and the Semantic View," Philosophy of Science, December 2006, 73: 524-535, is a descendant of section 2 of this paper, along with improved versions of some supporting material from section 1 and the introduction. I have plans to publish other parts of the paper, but I am archiving it here in the
... ving it here in the meantime in response to requests from colleagues who wish to be able to cite some stable public version of it. This version di!ers from the 1997 version in only a small number of superficial ways: I have corrected some typos, filled in one or two references, updated my surname (which changed from Jones to Thomson-Jones in 2004), and updated the bibliographical entries for works which were forthcoming at the time and have since appeared in print. As will be obvious, however, I have made no attempt made to update the bibliography (or the content of the paper) to reflect the numerous additions to the literature which have appeared in the interim. There are of course many changes I would now make, but there is one significant confusion I should address here. Section 1 lays out, at some length, a taxonomy. I say in more than one place (including the title of the section) that the taxonomy is to be taken as a taxonomy of notions of model. Understood that way, however, it is woefully inadequate. There are (and were) plenty of notions of model in the literature which are neither used nor mentioned in the laying out of the taxonomy; indeed, many such notions are discussed in the paper. The right way to understand the taxonomy is as a taxonomy of models. The aim is, in part, to have all the objects picked out (or purportedly picked out) by every widespread and coherent use of the term 'model' in the philosophy of science, and in the sciences themselves, fall into one of the categories included in the taxonomy. I think the taxonomy I present comes close to succeeding in this aim, and I hope to improve it so that it comes even closer. (One outstanding is-sue is to think about how computer models fit in, or how the taxonomy should be modified so that they do.) Of course, the aim in question could be achieved all too easily in various trivial ways, but the rest of the paper makes it clear, I hope, that the distinctions employed in the taxonomy are philosophically useful ones.