The first part of the article deals with the theoretical problem of historical fiction as a means of exploring the historical past. According to the consensual view in the theory of fiction, fictional worlds are inhabited only by fictional entities, so that the Tallinn of Jaan Kross’ historical novels is a fictional Tallinn, although one with a factual counterpart. In this view historical novels cannot be used for exploring the past. In the best case they can be used as sources about the times of their composition. On the other hand there is a tradition represented by certain novelists as well as philosophers who prefer blurring the line between fiction and history, claiming that poetic imagination and inventiveness is a necessary skill as much for the historian as for the novelist. Moreover, in their opinion the novelist can provide a better, more intimate insight into the experience and motivation of historical actors. This divide between isolationists, who separate fiction and history into water-tight compartments, and the writers who prefer to see a continuum between the two genres is parallel to the distinction between modernism and postmodernism, as the former is often interpreted as a separation of mental spheres and the latter as a blending of different genres and blurring of boundaries. This distinction also has a geographical counterpart, as according to Andrew Wachtel there has been a continuous intergeneric dialogue between historiography and fiction in the Russian culture, whereas in the West the two have grown apart since the mid-19th century. The second part of the article revisits the debates of the 1970s and 1980s between the authors Lennart Meri and Jaan Kross on the one hand and their critics among historians on the other. Both Lennart Meri and Jaan Kross, in their different ways, claimed for their imaginative conjectures the status not only of historical verisimilitude but even veracity. Although both were rebuked by academic scholars, their speculations have entered the popular view of history or collective memory.