There is a philosophical point I enjoyed, which doesn't generalize at all but is still interesting, about the idea that the past is a good guide to the future. (For example, "How do you know the sun will come up tomorrow? Well, it always has in the past.")

Someone pointed out that the relevance of the past to the future is a pure assumption. Even the usual argument that an otherwise unmotivated heuristic is valid -- "it's always worked in the past" -- can't legitimately be used for this point, since it assumes the truth of the question under consideration.