Exploiter un corpus de données textuelles sans post-traitement : l’écriture burlesque de la Fronde

Abstract

The texts printed during the Fronde (“mazarinades”) and written in “burlesque” style are a set of documents well known to literary and historical scholars but little studied in a systematical way, mainly because they are difficult to access and turn into a corpus. Our first task here is to delimit this corpus and to define contrastive corpora in order to reveal the specificities of “burlesque writing” in this political context. Then, the problem of acquiring textual data arises: the process of digitizing old printed matter and its evaluation are thus exposed. We choose to exploit these raw data, without post-processing or text correction, which still seems to allow us to obtain results (regarding lexicon, metre or enunciation), with various methods (textometry and observation of character n-grams). We thus show the specificity, in the burlesque writing of the Fronde, of some literary motifs or rhymes. The aim is therefore to present the approach of a corpus in all the steps of the process: from the selection of data, their acquisition, and their transformation into a corpus, to their statistical and algorithmic exploitation, and their interpretation through text mining.

Keywords

• literature
• digitisation
• corpus building
• data acquisition
• textual analysis
• stylistics