Understanding the evolution of early nervous systems is hazardous because we lack good criteria for determining homology between the systems of distant taxa; the timing of the evolutionary events is contested, and thus the relevant ecological and geological settings for them are also unclear. Here I argue that no simple approach will resolve the first issue, but that it remains likely that animals evolved relatively late, and that their nervous systems thus arose during the late Ediacaran, in a context provided by the changing planktonic and benthic environments of the time. The early trace fossil provides the most concrete evidence for early behavioural diversification, but it cannot simply be translated into increasing nervous system complexity: behavioural complexity does not map on a one-to-one basis onto nervous system complexity, both because of possible limitations to behaviour caused by the environment and because we know that even organisms without nervous systems are capable of relatively complex behaviour.