The demand that historical explanation conform to the model used in the sciences — subsuming an event under a general law — enjoyed a period of prestige and has few defenders now, for reasons worth separating. The objection is not that history lacks regularities; plenty of regularities can be stated, and some are informative. It is that the laws general enough to be true are too weak to explain any particular episode, while the generalisations specific enough to explain an episode are supported by that episode and perhaps two others. A historian explaining why a particular regime fell is offering a narrative in which each step is intelligible given the preceding one, which is a different intellectual achievement and not an inferior one. The pressure to reformulate it as law-like inference produced very little.