This chapter argues that popular narrativist and constructivist positions in theory of history – represented by Arthur Danto, Hayden White, Frank Ankersmit, Leon Goldstein and Paul A. Roth – do not offer satisfactory answers to the central question of the present book: how should one understand the relation between past and present in historical research? The chapter argues that this shortcoming has two different roots. First, narrativists tend to construe retrospective description, which is indeed one central aspect of the historian’s work, as if it was the only way historical research relates to the past. Second, constructivists tend tacitly to rely on positivist conceptions about ‘the real past’ and thereby fail to understand the proper function of that concept in historical research, and consequently the reality of the historical past. In conclusion, the chapter argues that relations to the past in history, and the role of the concept of the real past, can only be properly understood by attending to the presuppositions and a priori concepts internal to the historical method.