The study of the variation of pronunciation contributes to the scientific comprehension of language functioning and linguistic change. In this context, the socio-linguistic approach enables to understand the connections between linguistic and social factors in the effective use of speech for the purpose of expressiveness, differentiation or categorization. Previous research in the field of peripheral varieties of French brought progress in the comprehension of change forces in action and highlighted the relation of French speakers towards standard French. The project presented is in line with this framework and pursues a twofold objective. First, the project consists in a study of regional prosodic variation of French in Belgium on the basis of four varieties representative of distinct dialectal areas (Brussels, Gembloux, Liège and Tournai). Considering that in a formal situation some speakers make use of “standard Belgian French†the comparison between the varieties will be completed by a comparison with standard practice in order to give a precise picture of the present situation of French in Belgium. Second, the project aims at explaining which prosodic factors (duration, fundamental frequency, intensity, vocal quality) are responsible for the perception of accent. This dimension brings a fundamental character to the project as at present no model could predict which acoustic measures are relevant for explaining the differences perceived between varieties of French. In view of the above, the project aims at improving the comprehension of the link between acoustic phenomena and auditory perception of accents along with taking up the challenge of categorization of data for comparability purposes. As a result, it is expected to make progress in the modelling of French prosodic system and in the development of prosody analysis tools. ; L’étude de la variation de la prononciation contribue à la compréhension scientifique du fonctionnement du langage et du changement linguistique. Plus précisément, ...