The paper takes its departure from the following question: How is it possible that, by a reconstructive history of science as is practised for example by Thomas Kuhn, we can learn something from history, i.e. not only about the facts that may be described and causally-externally-explained by value-free science but also about the reasons,or norms,of practicing good science? Must not this enterprise lead to a dilemma: i.e. either to deriving norms from empirical facts and thus committing a naturalistic fallacy,or to presupposing already norms of scientific rationality in the evaluation of the reasons and norms of the scientists to be understood, and thus to committing a logical circle in the attempt of learning something new about norms of good science.The response my paper provides to this challenge is presented by introducing the — non-vicious — hermeneutic circle, as a central methodical device of normative hermeneutics. This approach is then illustrated through an argument with Karl Popper's "Third World Hermeneutics' (and, in passing, with Imre Lakatos'). In this context, the method of the hermeneutic circle as I understand it, amounts to a synthesis between Third World Hermeneutics, taken as a heuristics, and hermeneutic experience, taken as a semiotically specific kind of empirical evidence which is different from empirical evidence in the natural sciences.