What Is a Task and How Do You Know If You Have One or More?

Abstract

The burgeoning interest in how voluntary actions are organized into goal-based behaviors has produced a rich body of data showing that task representations are complex, integrating not only stimulus and response information but also internal (e.g., goals and relationships among actions) and external (e.g., task-irrelevant stimuli) context. The dominant description of task representations is the task set, a collection of SR associations, whose presence is inferred primarily through task-switch costs. However, we argue that this approach has serious limitations that are often ignored. First, task switch costs likely reflect numerous processes, including those relating to attention and inhibition, that vary across experimental procedures and complicate their interpretation. Second, the switch cost measure is coarse in that it groups transitions into a small number (usually two) of categories. As we demonstrate empirically, this procedure can produce misleading results. Observing a performance cost when a putative task boundary is crossed may be too coarse a measure to adequately describe how a task is organized. So, what exactly is a task? How do researchers know when two responses belong to the same or different tasks, and how do we determine when an effect relates to a change of task instead of some other central operation, such as attentional shifts or rebinding? In this chapter, we will examine the construct of a task and whether it is useful for understanding voluntary behavior.