Science, realism and ontology
The debate over scientific realism exists in an atmosphere of confusion. Within the literature, arguments proceed as though the nature of realism--what it is and what it argues for--is so clear that there is no need for any precise statement of the position. Disparate positions are identified as "scientific realism," and a variety of arbitrary theses are habitually tied to it. Resolving this predicament is a two-fold task. It calls for (1) bringing the debate more in line with the historic metaphysical discussion of realism in general and freeing it from the miscellaneous epistemological and semantic doctrines to which it is often erroneously tied; and (2) putting philosophers more in touch with actual scientific practice. Proposal (1) can be approached by emphasizing the ontological distinctions among empiricism, instrumentalism, and realism. The differences among the first three lie in their semantics, not their ontology: essentially, all deny that the theoretical entities or unobservables of science exist or that we are justified in believing that they exist; however, they have differing theories of meanings for theoretical terms. What distinguishes realism from these three is its ontology: it is committed to the existence of most of the firm posits of well-accepted theories. The widely-accepted notion of scientific realism according to which theories are statements and the aim of science is to produce theories ever closer to the truth, is a neo-positivist conception and fails to confront the fundamental issue of realism: ontology. Moreover, as an essentially ontological doctrine, realism is not dependent on any particular theories of truth or meaning. Proposal (2) requires acknowledging that, for the most part, philosophers of science are conspicuously preoccupied with theory, to the virtual exclusion of any substantive consideration of experimentation. This "armchair" approach has prompted both a spurious picture of science and a rift in the dialogue between philosopher and scientist. Any attempt at a substantive analysis of realism (or of science in general) must involve adequate consideration of the broader spectrum of actual scientific practice. The link between proposals (1) and (2) hinges, on one hand, on the means by which realism, properly interpreted, accounts for the dual aspects of science and, on other hand, the means by which experimentation can support ontological commitment by means of demonstration and indication.
Klein, Ronnie Boriskin, "Science, realism and ontology" (1994). ETD Collection for Fordham University. AAI9416671.