Abstract

Is methodological theory a priori or a posteriori knowledge? It is perhaps a posteriori improvable, somehow. For example, Duhem discovered that since scientists disagree on methods, they do not always know what they are doing. How is methodological innovation possible? If it is inapplicable in retrospect, then it is not universal and so seems defective; if it is, then there is a miracle here. Even so, the new explicit awareness of rules previously implicitly known is in itself beneficial. And so, improved methodology may make for improved methods. Hence, methodology is in part descriptive, in part prescriptive. Knowing this, a methodologist might improve his own studies. For example, Popper would then not hasten to conclude from the fact that past scientists depended on positive evidence that they had better do so in future as well; perhaps a lesser concern with confirmation may increase the productivity of scientific inquiry