PhD funding for thesis on language development in humans and artificial intelligences (M/F)

Description of the thesis topic: 

Most research on early language acquisition has documented entry and learning cues in only a handful of cultures, the assumption being that the mechanisms postulated to explain acquisition in these cultures are universal. However, some previous work suggests that the conditions under which young children acquire their mother tongue vary considerably from culture to culture (e.g., Cristia, 2022). 

Nowadays, it has become relatively easy to capture both spoken input and children's vocal productions via small recorders worn by the child for one or more representative days. These large-scale recordings are analyzed using a combination of professional linguistic annotation and automated big data analysis. This technique allows for more accurate documentation of the “real” input of infants' experience of language in a variety of cultures and social settings.

Candidates are invited to develop a research project building on the thousands of hours of audio, collected from hundreds of children who are growing up in diverse environments, using methods from developmental psychology, linguistics, data science, and/or machine learning. The precise research question will be defined together with the candidates, but some of the potential research questions are:

• Are there differences in the developmental profiles exhibited by children as a function of the typological properties of the language(s) they are learning? 
• What types of experiences are necessary and/or beneficial for vocal development? 
• How does early exposure to other children's or adults' voices correlate with morphosyntactic development, and what may this suggest about learning mechanisms involved in language acquisition? 
• What cognitive mechanisms are necessary to process the kind of speech input available to children across cultures? 
• What is the shape of the statistical relationship between experiences and outcomes, and what does that indicate with respect to the robustness of learning mechanisms involved in early language learning? 

Candidates may also ask to be involved in additional data collection, in which case additional research questions are possible, including: how do these novel measures of development relate to more standard measures of language development?

There is funding for 2 PhD students from the ERC project ExELang (exelang.fr). Self-funding (in which case a broader set of topics can be envisaged) may be obtained via ED3C.