Ever since Hume passed his striking criticism on metaphysics, there seems to be an unbridgeable gap between metaphysical concepts and empirical testability. From that time on more and more philosophers began to eye metaphysics with suspicion, what finally culminated in what can only be described as a call for banishment of metaphysics from the field of scientific research for the sake of progress in empirical science and philosophy. As a consequence, philosophers of science avoided reference to metaphysical concepts and tried to get as far as possible without employing them. But in recent years, however, the received opinion has become much more metaphysics-friendly and as the extensive debate on Hempel's deductive-nomological and inductive-statistical models has shown, there are phenomena in our world that cannot be explained without knowledge of their underlying causal structure. Thus Hume's initial question arises again: Can we discover causal relations and, if the answer is an affirmative one, how can we detect them? Fortunately, there have been recent approaches (causal graph theory) to connect empirically accessible concepts with concepts that allow an intuitive representation of causal structures, which are intended to bridge the gap between metaphysics and empirical science. In my talk I will do three things: (1) I will present a frequently discussed philosophical problem - the flagpole problem. By dint of this problem I will demonstrate why reference to causation is needed in scientific explanations of a certain kind. (2) I will explain the fundamental terms and principles of causal graph theory that are needed for the last part of my talk. (3) I will solve the flagpole problem using an algorithm developed by Spirtes et al. In doing so, I will give an exemplary case of how empirical data can be used to determine causal structures in the world - I will show how it is possible to cross the border between empirical science and metaphysics.
Diskussion zum Vortrag.