This study argues that much research of the emotions has been misguided. It attempts to show that "emotion" encompasses psychological states of very different, and thus not comparable, kinds. Some emotions, such as a brief flaring up of anger in response to some experience, are evolutionary ancient, reflex-like responses which appear insensitive to culture. Others, like moral guilt, differ importantly across cultures, despite their long history in humans, and affinity to behaviour seen in other species. Yet other emotions appear to be the acting-out of today's psychological myths, as ghost possession acted out the metaphysical myths of past centuries. These three kinds of responses have different evolutionary origins, different adaptive functions, different biological bases, and different roles in human psychology. The concept that binds them together, emotion, plays no useful role, since there is no object of scientific knowledge that corresponds to it. A detailed overview of the relevant theoretical approaches is provided in this text, assessing the relative merits of three main theoretical approaches: affect programme theory, evolutionary psychology, and social constructionism.