Non-fiction overwhelmingly adopts the form of the long narrative - the form favored by the genres of testimony, biography and autobiography. It's hardly surprising, then, that the study of factual literature is dominated, at least in France and in the tradition of Gérard Genette's work, by questions of narratology. There is, however, another kind of factual literature which, in order to write the real, invents formal alternatives to narrative, its continuity and its power of integration. Factographies present and juxtapose materials taken from extra-literary reality: archival documents, notations, interviews, conversations seized on the fly. In this way, they question the notion of representation, and invite a shift in the way theory looks at factual literatures. These forms escape the two main questions that narratology poses to factual literatures: that of the criteria of their literarity and that of the criteria that distinguish factual from fictional statements. They thus suggest that a theory claiming to account for the entire field of factual literature would have to rely on other approaches. Following on from Leona Toker's work on Gulag literature, in which she proposes to define factual works as "multifunctional" objects, this article argues that an operative definition of the notion of factual literature is essentially a pragmatics of reception.