Empirical Tests Are Only Auxiliary Devices

Abstract

We test a scientific theory empirically by deducing from it a 'prediction' and comparing it with observation. In Popper's view of science, theories can be refuted this way, that is through tests, but can never be rendered more probable. In his view, even when a theory has passed all the tests to which we have subjected it, it still remains as a conjecture with zero probability (Popper [1962], p192). Popper's view of science includes a form of skepticism. According to Popper, we can propose theories (in answer to problems) and try to refute them; in doing so we hope we come closer to the truth, but we will never know whether we have actually done so (Popper [1959], p415; Hattiangadi [1983]). Suppose we accept Popper's argument that theories cannot be rendered more probable by empirical tests. Do we have to accept his form of skepticism? I think not. Popper's argument would lead to skepticism only if empirical testing constitutes the central mechanism in the evaluation of theories.' If it is not the central mechanism; if there is some other way of evaluating theories more important than tests; skepticism may not follow. In this paper, I argue that tests are but auxiliary devices. Though definitely useful, they are dispensable. There is a more important method of evaluating hypotheses. This method of evaluation I call the method of progressive evaluation, or MPE. MPE is the central method of evaluating hypotheses in science-and in other kinds of empirical investigations. When we apply this method in evaluating an hypothesis, if the hypothesis is true (without ourselves knowing), we will become progressively more certain that it is. If the hypothesis is a good approximation, the investigative strategy, IVS, associated with this method will enable us to improve upon the approximation. If the hypothesis is false, and seriously defective to boot, MPE will allow us to eliminate it. In a word, MPE together with IVS, allows us to uncover knowledge gradually. Whether this will satisfy all skeptics is open to question; however, if we do have this method, we will have some indication as to whether we are uncovering knowledge. I will explain MPE and IVS later on in this paper. 1 Theories can be evaluated in other ways. For example, logically, to see whether they are internally consistent. * I wish to thank an anonymous referee for helpful comments on an earlier version of this paper.