INetwork's integration Year: 2020
NFields of research: discourse analysis, natural language processing, medical humanities, medicine, digital humanities
This project focuses on the automatic analysis of discourse, seeking to move methods developed so far for the analysis of discourse and testimony in literary texts on texts deemed a priori not literary.It proposes that in the overall study of a corpus of texts from medical practitioners (doctors, caregivers, students), which can be the subject of statistical analysis, account be taken of the narrative, its enunciation, emotion and time.It is therefore well at the intersection of medical and digital humanities.It is therefore part of the broad context of Life course analysis: the life course analysis that has penetrated the social sciences since the early 2000s. It is divided into two parts:
The study of a first contemporary corpus, composed of nearly 3,500 reports written by medical students after an experiment with the « SAMU Social », will be an opportunity to develop a method of semantic excavation adapted to this type of discourse.A system of annotations based on rules can be developed in order to identify and categorize within the corpus the different types of experiences (descriptive, emotional, sensory, positive/negative, etc.). In the longer term, such work will lead to the development of an interface, offering an exploration of the texts according to the nature of the experience and the affects engaged.Our method will have the advantage of being reusable on contemporary corpora of wider scope, such as medical social networks.
This first part of the project is envisaged in order to serve at the same time the treatment of an unprecedented patrimony constituted from the historical fund Charcot, held by the Jubilothèque of the Faculty of Sciences of Sorbonne University.In view of their exploration, these archival documents will first be the subject of a complete digitization, then of a statistical and semantic analysis based on our method of exploration.