

Historians of philosophy, by contrast, may suspect that miscellany gives us reason to try to understand causation not by way of analysis, but by way of genealogy. There is something that unifies what David Lewis calls the "gruesomely disjunctive miscellany" of cases of causation, and conceptual analysis is required to identify that unifying factor.

Many metaphysicians see in this a reason to seek a conceptual analysis of causation. A pool ball strikes another, causing it to move my failing to water the plant causes it to wilt your willing to imagine an elephant causes your imagining of the elephant and on and on. An account of causation in general must, on the face of it, be able to handle all of the many kinds of causation. How much better, then, to understand causation itself! The problem - well, one problem - is that there are so very many wildly disparate cases that are thought to be instances of causation. The reason seems to be that causation is one of the precious few relations fit to serve as explanation in many domains, to understand a phenomenon just is to know its cause. Philosophers are drawn to causation like moths.
