Abstract
In a provocative and important article, Robert Solomon argues that emotion researchers should abandon the notion of valence: it is used many different ways, and no single construct captures the pretheoretical distinction between positive and negative emotions. I echo Solomon in arguing that some of the most popular theories of valence are unlikely to succeed, though my case against these constructs comes from a noncognitive, as opposed to cognitive, perspective. I then argue that there is one notion of valence, related to reward and punishment, that may apply to all emotions, and align with pretheoretical taxonomies