Participatory and Interdisciplinary Analysis of a Suicide Prevention Chat – APIPreSui

The APIPreSui project (Participatory and Interdisciplinary Analysis of a Suicide Prevention Chat) is an interdisciplinary participatory research of a digital suicide prevention chat system. The analysis is co-constructed with the prevention association, its trainers, psychologists and volunteers on the one hand, and researchers in language sciences (sociolinguistics - discourse analysis - lexicometry - NLP), computer science and social psychology on the other.
The first part of the project consists of a discourse analysis of the 17 years of the chat corpus (2005-2022), combining qualitative and quantitative approaches. The analysis of the entire corpus in its sociolinguistic and discursive characteristics aims to better understand the functioning of these sensitive conversations. The lexical, enunciative and interactional exploration will make it possible to bring to light the conditions of felicity and dysfelicity (good/dysfunction) of the listeners' discourse. This will enable us to propose ways of improving the discursive functioning of the system and the training given to volunteers. Regular work seminars bringing together the trainers and volunteers of the association and the researchers will allow the different phases of analysis of the corpus to be worked on systematically, by mutually combining the field and research experiences of the whole team. The second part consists of comprehensive interviews and an ethno-photographic survey of the trainers, listeners and callers of the prevention association, and an online survey of the callers about the chat system. The third part concerns the dissemination and publication of the knowledge co-constructed by the project in the form of enhancement and training works, with the active participation of the third-party monitor, in addition to the planned scientific publications.
The central problematic concerns the understanding of the functioning of this aid device, its conditions of good functioning and sometimes of dysfunction, and to what extent these conditions are due to the interaction through this digital discourse in particular, whose particularities will be analysed: comparisons with corpora of e-mails from the same aid association are planned, as well as with the telephone call device (the latter through the metadiscourses of the actors and the sociolinguistic practices in the chat corpus, the telephone corpora not being available). Indeed, digital discourse occurs without the voice, whose prosody may be lacking, in these sensitive contexts of great fragility. At the same time, this absence of emotionality through the voice is often what is sought by callers, and even helps them to formulate their suffering in writing, before being able to do so orally. This is even more true sincethe chat presents a very formal register and of a very distant conceptional relief. As this aid system has only recently become digital (the association started in 2005), its evolution will be studied and modelled with a view to participating in the training of the listeners, as interaction in a prevention chat, in a sensitive context, needs to be finely understood and transmitted, as a discursive genre in its own right.
The meetings, facilitated by a third-party monitor, with the trainers and callers will be organised over a period of two years by analysis and hetero-confrontation workshops to ensure the intense mutual learning that has already taken place over the last four years between practitioners and researchers.
These analyses of the conditions for handling calls from people in distress have very concrete implications for the daily lives of these people and their families, in France and elsewhere, and could constitute a laboratory for suicide prevention by chat.