Over the past three decades research has overwhelmingly supported the notion that positive affect promotes global, abstract, heuristic information processing whereas negative affect promotes local, detailed, and systematic processing. Yet despite the weight of the evidence, recent work suggests that such a direct relationship may be highly tenuous. In line with the affect-as-information account, we maintain that affective cues are adaptive and serve to provide individuals with information about their current psychological environment. We argue that these cues do not directly produce specific processing styles, but instead confer value on whatever processing inclination is dominant at the time. Positive affect (e.g., happiness) tends to promote reliance on currently dominant processing inclinations, whereas negative affect (e.g., sadness) tends to inhibit such reliance. Thus the impact of affect on processing is highly malleable and depends on both the type of processing that is currently active and the information provided by affective cues.