Coevolution in trophic interactions is often considered as a major factor underlying diversification in interacting species. Most focus hitherto has however been on bottom–up processes where host-associated differentiation drives diversification, and less on top–down processes through enemy-associated differentiation. In this paper, we contrast bottom–up and top–down processes as drivers for diversification based both on theoretical reasoning and empirical studies. As a framework, we use the oscillation hypothesis which posits that local differentiation and network rewiring following geographic range shifts are major drivers for diversification. Our conclusion is that speciation through bottom–up and top–down processes likely act on different spatial scales but also that conditions for the interaction are different for hosts and enemies. For example, enemies can select their resource, leading to a specialized attack, but hosts cannot select their attacker, leading to a generalized defense. These differences have the consequence that local adaptation likely proceeds at a faster pace in enemies, and particularly in parasitoids, compared to hosts. Our analysis results in several testable hypotheses but we also argue that further theoretical evolutionary models are needed that combine ecological processes and genetic mechanisms.