Elucidating the relationship between linguistic from and literary function in a corpus of French literary texts from the second half of the nineteenth century using text analysis software

Abstract

The intention was to compile a collection of texts, which was representative of both 'standard French' and argot from 1865 to 1895. It was decided at the outset that the corpus would contain only written texts. The literary component of the corpus was compiled using random selection procedures from a bibliographical list. Since there was no readily available list of electronic literary texts published between 1865 and 1895, it was necessary to compile one using a number of library and academic text archive sites. The use of random sampling ensured that all texts in the population had a chance of being selected, and thus removed the possibility of their being any subjectivity or personal bias involved in the selected process. In addition to random sampling, another method of sampling was used in the construction of the corpus: stratified sampling.