Causal inference in the abstract or seven myths about thought experiments

I analyse and criticise the following seven commonly held, but to my mind, mistaken beliefs about thought experiments: (1) The history of science is full of significant thought experiments; (2) A good thought experiment provides evidence in its own right; (3) We learn from thought experiments in essentially the same way as we learn from concrete experiments; (4) It is puzzling that thought experiments allow us to learn about the world without providing new empirical data; (5) Thought experiments make acceptance of their result(s) compelling; (6) Mental experiencing is essential to thought experimentation; (7) Thought experimentation involves intervention. After clearing the ground in this way, I sketch a positive theory of the thought experiment. The basic idea of the new theory is to integrate thought experiments into a broader inductive scientific methodology. Within such a broader methodology, thought experiments can assume a number of functions, four of which I briefly discuss: (a) concept formation, (b) establishing a causal hypothesis, (c) nomological refutation and (d) suggestion of “new works”.