what is most simply called the literariness of non- literary phenomena. Qualities often thought to be literary turn out to be crucial to non-literary discourses and practices as well. For instance discussions of the nature of historical understanding have taken as a model what is involved in understanding a story. Characteristically historians do not produce explanations that are like the predictive explanations of science: they cannot show that when X and Y occur Z will necessarily happen. what they do. rather, is to show how one thing led to another, how the First World War came to break out, not why it had to happen. The model for historical explanation is thus the logic of stories: the way a story shows how something came to happen. connecting the initial situation, the development, and the outcome in a way that makes sense The model for historical intelligibility, in short, is literary narrative. We who hear and read stories are good at telling whether a plot makes sense. hangs together, or whether the story remains unfinished. If the same models of what makes sense and w counts as a story characterize both literary and historical narratives. then distinguishing between them need not seem an urgent theoretical matter. Similarly. theorists have come to insist on the importance in non-literary texts- whether Freud's accounts of his psychoanalytic cases or works of philosophical argument-of rhetorical devices such as metaphor, which have been thought crucial to literature but have often been considered purely ornamental in other sorts of discourses. In showing how rhetorical figures shape thought in other discourses as well, theorists demonstrate a powerful literariness at work in supposedly non-literary texts, thus complicating the distinction between the literary and the non-literary. But the fact that l describe this situation by speaking of the discovery of the literariness of non-literary phenomena indicates that the notion of