Poorer performance in conditions involving task repetition within blocks of mixed tasks relative to task repetition within blocks of single task is called mixing cost (MC). In 2 experiments exploring 2 hypotheses regarding the origins of MC, participants either switched between cued shape and color tasks, or they performed them as single tasks. Experiment 1 supported the hypothesis that mixed-tasks trials require the resolution of task ambiguity by showing that MC existed only with ambiguous stimuli that afforded both tasks and not with unambiguous stimuli affording only 1 task. Experiment 2 failed to support the hypothesis that holding multiple task sets in working memory (WM) generates MC by showing that systematic manipulation of the number of stimulus-response rules in WM did not affect MC. The results emphasize the role of competition management between task sets during task control.Keywords: task switching, switching cost, mixing cost, task ambiguity, working memory One of the prominent questions in psychological research concerns behavior control. Whereas the behaviorists' tradition has ascribed most behavioral control to the environment, recent cognitive theorists have emphasized more internally driven, top-down forms of control. This idea of cognitive control involves concepts such as goal-directed behavior, initiation, executive control processes, and so forth. A key concept in the context of top-down control is the task-set. According to Rogers and Monsell (1995), the control of task-sets is manifested in the ability to configure processing resources to perform one rather than another of the many cognitive tasks that a stimulus affords.A common paradigm for studying task-set control is the taskswitching paradigm. In the original version of this paradigm, performance in blocks of trials in which a task is repeated is compared with performance in blocks in which the participants switch between two different tasks (Allport, Styles, & Hsieh, 1994;Jersild, 1927;Spector & Biederman, 1976). More recent studies have used a modified paradigm that makes it possible to contrast task-switch trials and task-repetition trials within blocks of mixed tasks (e.g., De Jong, 1995bGoschke, 2000;Mayr & Keele, 2000;Meiran, 1996;Rogers & Monsell, 1995). In most cases, switching tasks is accompanied by a robust performance cost, seen both in reaction time (RT) and error rates, indicating switch cost.A recent conceptualization that incorporates knowledge from the variety of task-switching paradigms elaborates and sharpens the understanding of the switch mechanisms by differentiating between several cost components. The difference in performance between switch and repetition trials (in mixed-tasks blocks) is termed switching cost, and the difference between repetition trials (in mixed-tasks blocks) and single-task trials (in pure blocks) is termed mixing cost (