Emotions have a bad reputation among scholars of cognitive control. Clearly, emotions can be powerful distractors that challenge controlled behavior. Not surprisingly, theoretical accounts often describe emotions as antagonists to controlled behavior. In this talk, I will focus on a different perspective, suggesting that emotions have a signaling function, because they are both output and input for controlled processes. I will present meta-analytic data addressing this controversy by means of experimental manipulations of affect on cognitive control in conflict tasks (Stroop, Simon, flanker). Results of two primary meta-analyses found no clear evidence that emotional stimuli modulate cognitive control in general. Yet, moderator analysis suggested that specific aspects of the task, stimuli, and testing conditions show enhanced control with emotional stimuli. Based on this moderation, I will discuss how different theoretical positions on the emotion-control interface might be reconciled.