Craig's theorem, Ramsey-sentences, and scientific instrumentalism (Q2560802)

This paper is concerned with a problem in philosophy of science. The theoretical terms used in scientific explanation and prediction can be taken as referring to entities which actually exist (scientific realism) or as `non-referring symbolic devices that function solely as integral parts of inference rules to warrant inferences among observational statements' (scientific instrumentalism). The author analyses in considerable detail two arguments in support of instrumentalism which are based on the possibility of eliminating theoretical terms from theories. One appeals to Craig's theorem, in accordance with which only empirical terms are needed, and the other invokes Ramsey sentences, with existentially-quantified variables taking over the functions of theoretical terms. The conclusion is that the counter-arguments that have been advanced against these arguments are not conclusive, and scientific realism remains unjustified. The debate between realists and instrumentalists is still open.