I suppose one could argue that in an absolute philosophical sense there is no causation that we can have knowledge of, only differing amounts of correlation. It would be an extreme position to take but - like solipsism - a difficult one to argue against.

(Anyone with an actual background in epistemology care to help me out here?)

I always considered causality less as a binary thing that was always 0 until $somethreshold was crossed, and more of a bayesian (or other) probability model.

Since in many cases we don't have the evidence for a deductive conclusion, we must often rely on inductive logic, even while still seeking the evidence until we can switch to deductive.