The constitution of a field of autonomous research around the concept of discourse fits in a general way within the framework of the evolution of the sciences of the language, and more broadly in that of the human and social sciences, from the years 1960. Discourse analysis has complex relations with linguistics, which are in constant redefinition. For it is more of a scientific movement at the crossroads than a well-circumscribed discipline, with its object, its methodological framework and its concepts. Despite their diversity, all approaches in discourse analysis assume that utterances do not present themselves as sentences or sequences of sentences but as fragments of discourse and thus converge to a single definition of discourse. A speech is then a specific form of organization that must be understood as such by relating it to its context and its conditions of production.