Cognitive flexibility and stability at the task-set level: A dual-dimension framework

Metacontrol coordinates goal-directed information processing, giving rise to cognitive flexibility and stability. However, the structure of flexibility and stability in metacontrol has long been subject to an overlooked assumption that these states vary on a single spectrum. This unidimensional structure gives rise to an obligatory flexibility-stability tradeoff: Becoming more flexible must come at the cost of lower stability. Although a “unidimensional” framework such as this has intuitive appeal, a great deal of recent work reveals that flexibility and stability can vary independently. Here, we review evidence that is challenging for the unidimensional framework to account for. As an alternative, we propose a dual-dimension framework (DDF) whereby flexibility and stability are assigned to separate dimensions, each ranging from low to high and capable of varying independently. In addition, we describe processes by which people shift along both dimensions. Theoretical benefits of adopting the DDF include a more fine-grained explanation of observed variability in behavior. Possible applications include strategies for better aligning metacontrol states with situational demands. In light of these implications, combined with the available data, we propose that the DDF might offer a better way to describe the structure of flexibility-stability metacontrol.