Fine-tuning Stylometric Tools: Investigating Authorship and Genre in French Classical Theater

This paper is concerned with stylometric classification applied to French seventeenth-century plays. It reports on ongoing investigations into parameter setting and its impact on classification of such texts by author, genre, or form, using Eder & Rybicky's stylometric scripts for R (Eder & Rybicky 2011). Based on an investigation into the Corneille-Molière controversy, several methodological issues standing in the way of reliable results have been identified. One issue concerns the degree to which such authorship classification tasks are influenced by genre (here, comedy or tragedy) and form (here, verse or prose). Investigation of this issue shows that input parameters have indeed effects on the relative influence of authorship, genre and form in the classification of plays.