The object of study in this research is the topics of representation and narrativity in the human sciences, starting from a comparison with the field of the theory of art. The hypothesis proposed here is that the classic notion of mimesis, which supported the theory of representation for a long time, went through a crisis with regards to its theoretical bases during the 19th and 20th centuries. After a series of theoretical interventions – in which I highlight the importance of romanticism, the perspective of social anthropology and linguistics – the mimetic paradigm loses its strength in favor of a theory of representation focusing on the dimension of language. My objective is to analyse the consequences this epistemological change has for the discursive regime that separated literature and the human sciences into opposite fields. As such, I intend to demonstrate that the scientific principle of rejecting poiesis obeys an obsolete conception of language, in which empiria and poetics are considered mutual threats.