In this chapter we review three research programs that address the view that emotions, like perceptual similarity, functional equivalence, and theories, serve as a kind of conceptual coherence, or an organizing theme for objects and events in the world. We begin by suggesting that emotions organize perception and can prime emotion-congruent information in the visual field. Results of four lexical decision experiments and one word naming experiment demonstrate that emotions affect low-level perceptual processes, and suggest that emotional response categories correspond to specific, or basic, emotions, not merely to positive and negative responses. The second line of research demonstrates that when people are emotional they tend to put things together into categories that share a common emotional association. For instance, when experiencing strong emotion, people are more likely to say that a "kiss" is more similar to a "fortune" (which is similarly related to happiness) than to "handshake" (which shares a taxonomic relation). Moreover, it appears that people do not do this in a messy and random way (e.g., form a category of all things emotional), but rather group together things that are associated with happiness and group things together that are associated with sadness. Finally, the last research program was designed to evaluate the possibility that the mediating mechanism in emotional response categorization is selective attention. Multidimensional scaling analyses reveal that in making judgments of similarity, individuals who are experiencing distinct emotional states give greater weight to emotional, as compared to non-emotional, dimensions of multidimensional stimuli in their judgments of similarity.