The central executive component of working memory is a poorly specified and very powerful system that could be criticized as little more than a homunculus. A research strategy is outlined that attempts to specify and analyse its component functions and is illustrated with four lines of research. The first concerns the study of the capacity to coordinate performance on two separate tasks. A second involves the capacity to switch retrieval strategies as reflected in random generation. The capacity to attend selectively to one stimulus and inhibit the disrupting effect of others comprises the third line of research, and the fourth involves the capacity to hold and manipulate information in long-term memory, as reflected in measures of working memory span. It is suggested that this multifaceted approach is a fruitful one that leaves open the question of whether it will ultimately prove more appropriate to regard the executive as a unified system with multiple functions, or simply as an agglomeration of independent though interacting control processes. In the meantime, it seems useful to continue to use the concept of a central executive as a reminder of the crucially important control functions of working memory.