Where is the limit?

 A common description of the reviewers was that Heirmet's work comprised too many years and too many arguments in this chapter of American history. Even though the reviewers treated the book as an intellectual work, some of them had some harsh comments on his thesis and narrative so that  they gave the impression that the line between history and fiction was very fine in Heirmert's account.
E.S. Morgan made a sharp argument against the intention of Heirmert to read, not only the declaration in the sentences, but navigate the semantics of the arguments  in order to discover not what the author said but what he/she meant, thus Morgan claimed that Heirbert failed to interpret "what was really said" at the light of context and empirical evidences.
How much is a historian willing to interpret and shape historical evidence in order to support his/her arguments?