Realism, Structuralism Realism and Instrumentalism

The question this paper addresses is one in what might be called the theory of theories. What is theoretical knowledge in the sciences, knowledge of? There are two profoundly opposed answers to this question. One is the thesis of scientific realism which asserts that, if we believe that our scientific theories are at least approximately true and our evidence for this belief is precisely the correctness of their experimentally checkable predictions, then we must believe to that same extent in the existence of those hypothetical entities, those theoretical entities, which those theories assert are responsible for the experimental evidence. But the Ramseyfied version a scientific theory is observationally equivalent to the original theory and makes no reference to theoretical entities, so what could be the justification for the claim that the evidence supports the existence of theoretical entities, when it equally supports a theory which makes no mention of them? The claim that there is no justification is the thesis of scientific instrumentalism. A new view has emerged in this old debate which seems attractive in that it captures what seems correct or insightful in both views but without abandoning the cognitive significance of theory. This view is called Structural Scientific Realism. It is a realist view because it accepts that scientific theories describe in part an unobservable reality but it also accepts from the instrumentalist position that the cognitive content of theories is not what the realist traditionally has thought that to be. I argue that these claimed advantages are illusory and that the new position is untenable.