Individuare scritti gramsciani anonimi in un corpus giornalistico. Il ruolo dei metodi quantitativi

This contribution shows the results obtained by a heterogeneous group of scientists (one linguist and four physician/mathematicians) who were asked to try to discover the Gramscian writings hidden in the huge corpus of unsigned articles published by the newspapers where Gramsci usually wrote. The study adopts quantitative authorship attribution methods, which were extensively tested for this specific study on these texts, yielding very interesting results. A very brief history of quantitative methods for authorship attribution is drawn, with a focus on those used for this study: both based on a mathematical model of texts and the author/text relationship, and both using similarity distances (the first compares the statistics for sequences of n characters – n-grams – in the texts, while the second is based on the concept of entropy of a symbolic sequence). These methods have until now usually been in the toolbox of textual scholars and historians who could fear being dispossessed of their role and interpretative authority, but things are far more complex.