Since Ehrlich & Raven's seminal paper 50 years ago, coevolution has been seen as a major driver of species diversification. Here, we review classical and more recent case studies on the coevolution of plants and associated insects, to examine whether the coevolutionary component holds as an explanation of their current diversity. We discuss the main dogmas in coevolution and argue that coevolutionary processes should not be considered as major drivers of diversification in plants and insects. Instead, we suggest that coevolution essentially occurs through relatively short 'interludes', making the pattern difficult to detect. We also criticize the use of comparative phylogenetics to investigate coevolutionary processes, as coevolution may not necessarily produce congruent phylogenies among interacting lineages and, in turn, other processes may produce patterns of codivergence. Finally, we propose new lines of investigation for future research.