Outils pour utilisateurs

Outils du site


From Sound to Discourse: Computer-based modelling of large-scale data to study linguistic change over time

IdEX Emergence en Recherche, January 2021 - December 2022
PI - Ioana Chitoran

The Son – Discours project proposes a multi-level analysis of spoken language at the interface of linguistics and computer science. The goal is to study linguistic variation: (a) at the fine-scale level of sound structure (phonetic and phonological variation); (b) at the larger scale of discourse (variation in the organization of discourse). The working hypothesis is that language is variable, but variation is not random. Variable patterns develop within the context of a linguistic system which is both flexible and constraining. Some patterns may remain circumscribed to the system and disappear, others may evolve by moving further away from the system, leading to a linguistic change. Understanding how variation emerges, and what factors may favor or constrain it, is the general goal of our project. We propose a hybrid methodology, where the hypothesis is tested by both behavioral experimental methods and automatic classification techniques borrowed from automatic speech recognition, in joint, incremental approaches combining linguistic analysis and automatic processing. We use machine learning methods derived from data mining to test the likelihood that a given variation pattern may be related to linguistic change.

Project team:

  • Ioana Chitoran, Giuseppina Turco, Hiyon Yoo (Université de Paris)
  • Ioana Vasilescu, Lori Lamel (LISN - Paris Saclay)
  • Elisabeth Degand, Anne-Catherine Simon (Université Catholique de Louvain)

Phonetics and Phonology in Europe - PaPE 2021
Workshop:From speech technology to big data phonetics and phonology: A win-win paradigm

user/ioana_chitoran/from_sound_to_discourse.txt · Dernière modification: 2021/11/30 02:57 (modification externe)

Outils de la page