Coevolution, reciprocal adaptation between two or more taxa, is commonly invoked as a primary mechanism responsible for generating much of Earth's biodiversity. This conceptually appealing hypothesis is incredibly broad in evolutionary scope, encompassing diverse patterns and processes operating over timescales ranging from microbial generations to geological eras. However, we have surprisingly little evidence that large-scale associations between coevolution and diversity reflect a causal relationship at smaller timescales, in which coevolutionary selection is directly responsible for the formation of new species. In this synthesis, we critically evaluate evidence for the often-invoked hypothesis that coevolution is an important process promoting biological diversification. We conclude that the lack of widespread evidence for coevolutionary diversification may be best explained by the fact that coevolution's importance in diversification varies depending on the type of interaction and the scale of the diversification under consideration.