It ought to be conceded, as an empirical fact, that there are seldom, if ever, perfect regularities in nature. Generalizations, instead, have to be made ceteris paribus. If there is no perfect regularity, however, this still does not mean that there is no causation. Causal claims can instead rest on recognizable tendencies. Tendencies can come in various degrees of strength, some very strong and some very weak. Ceteris paribus laws could be understood in terms of tendencies, which involve less than necessity but more than pure contingency. A tendency cannot be identified with a statistical incidence, however. Instead, we can think of any such incidences as being produced by the underlying tendencies.