Ressources linguistiques pour l'analyse de corpus : le projet ATISHS

International audienceNowadays, most corpus processor software applications analyze texts as if they were constituted by sequences of graphical wordforms. However, in users of these software who are working in the social sciences and the humanities are looking for information in the corpora they study expressed by linguistic units of meaning. It is therefore necessary to establish a link between the graphical wordforms that constitute the text files and the units of meaning that are of interest to the users of these software. We show that creating this link requires precise linguistic methods and resources. We present the linguistic resources developed with the NooJ linguistic platform and demonstrate how they are used by the new ATISHS corpus processor to provide more reliable statistical analyses for the digital humanities