Bas C. Van Fraassen

The opposition between empiricism and realism with respect to science is old: it appeared clearly in the seventeenth century sense of superiority of the ‘mechanical philosophy’ to Scholastic metaphysics, and continued for the next three centuries’ debates over the philosophical foundations of physics. Empiricist views developed by the logical positivists of Vienna and Berlin were defeated by the emergence of scientific realism in the mid‐twentieth century. This defeat was largely due to the inadequacy of the positivist theories of meaning and language. To present a viable empiricist account of science, we need conceptions of theory structure, truth, and empirical adequacy that are essentially independent of a theory's linguistic formulation. Correlative to this discussion of the relationship between theory and world is the question what it is to accept a scientific theory––how much belief is involved? And what is involved other than belief? The position to be developed here will view scientific activity as one of construction rather than discovery: construction of models that must be adequate to the observable phenomena, and not discovery of the truth about postulated unobservable parts of nature. Acceptance of a scientific theory so viewed will have an irreducible pragmatic dimension in addition to involving only a modest amount of belief.

The attempt to formulate a viable empiricist and non-foundationalist epistemology of science faces four problems here confronted. The first is an apparent loss of objectivity in science, in the conditions of use of models in applied science. The second derives from the theory-infection of scientific language, with an apparent loss of objective conditions of truth and reference. The third, often cited as objection to The Scientific Image, is the apparent theory-dependence of the distinction between what is and is not observable. The fourth and last is the loss of the possibility of objective evaluation of rationality in scientific methodology. It is argued that each of these problems is illusory

Scientific theories do much more than answer empirical questions. This can be understood along empiricist lines only if those other aspects are instrumental for the pursuit of empirical strength and adequacy, or serving other aims subordinate to these. This chapter accordingly addresses four main questions: Does the rejection of realism lead to a self‐defeating scepticism? Are scientific methodology and experimental design intelligible on any but a realist interpretation of science? Is the ideal of the unity of science, or even the practice of using distinct theories in conjunction, intelligent on an empiricist view? What sense can an empiricist position accord to those theoretical virtues––such as simplicity, coherence, explanatory power––that are not reducible to empirical strength or adequacy? The answers to these questions rely strongly on the pragmatics of scientific inquiry, and advocate a ‘Clausewitz doctrine’ of experimentation as a continuation of theorizing by other means.

In "Unnatural Science"(1) Catherine Elgin examines the dilemma which David Lewis sees posed by Putnam's model-theoretic argument against realism. One horn of the dilemma commits us to seeing truth as something all too easily come by, a virtue to be attributed to any theory meeting relatively minimal conditions of adequacy. The other horn commits us to "anti-nominalism", some version of the ancient doctrine that language must "carve nature at the joints": that there are natural kinds or classes which alone qualify as referents (extensions) for our predicates. Elgin offers a searching critique of Lewis' response (accepting the second horn) and an illuminating defence of its contrary: "we cannot construe (mere) truth as the end of scientific inquiry. Not ... because truth is too hard to come by, but because it is too easy" (p. 301)

Why-questions and how-possibly-questions are two common forms of explanation request. Answers to the former ones require factual assertions, but the latter ones can be answered by displaying a representation of the targeted phenomenon. However, in an extreme case, a representation could come accompanied by the assertion that it displays the only possible way a phenomenon could develop. Using several historical controversies concerning statistical modeling, it is argued that such cases must inevitably involve tacit or explicit empirical assumptions.