I had the chance to take part in the conference of the Danish Philosophical Association in Aalborg, early March 2014. For the occasion, I decided to try out again the idea of ‘variational reasoning’ in causal modelling. If you haven’t came across it already, it is actually quite simple. One question concerning causal reasoning is what notion(s) guide model building and model testing. This is an epistemological question. My answer is that we reason around variations. In other words, without variations we cannot detect causes at all. Of course, detecting variations is not enough to establish causal relations, and that’s why we need to impose further constraints, for instance regularity or invariance. If it is so simple, why is it so important? My view is that it sheds light on important aspects of causal epistemology, and it helps putting other causal notions (e.g., manipulation, regularity, …) in the right place of the #causalmosaic.