Conférence donnée par Alex Cristia dans le cadre du "Colloquium de l'Institut d'étude de la cognition".
Hearing children make great strides in attuning their perception to fit the ambient spoken language within the first year of life. In this process, the spoken input plays a key role, as infants growing up in Japanese-speaking communities attune to Japanese and not English. It is therefore of crucial importance to gain a good understanding of the input to language acquisition. In this talk, I summarize some of my work, approaching this problem from multiple perspectives, albeit concentrating in phonology. Corpora analyses reveal that infant-directed speech is neither uniformly hyperarticulated nor a linguistically messy, purely emotional signal. A combination of corpora analyses and perceptual experiments sheds light on the next step, how the acoustic signal is perceived by the human infant. The acoustic characteristics of individual caregivers' speech are found to predict their infants' sound discrimination, but the strength of this association is dependent on the auditory model used to describe the signal. While this bivariate association is ambiguous by definition, artificial grammars and laboratory-based exposures reveal the extent to which infants' perception is shaped by the distributions in the speech they are exposed to, but also that infants' induction can be under- and over-specific. A final research strand assesses the brain networks that are involved in infants' speech sound processing, primarily using fNIRS.
Un colloquium du DEC.
Alex Cristia est chercheuse au Laboratoire de Sciences Cognitives et Psycholinguistique (LSCP) où elle travaille notamment sur les mécanismes psychologiques d'acquisition des fonctions cognitives typiquement humaines comme le langage et la conscience.Cliquer ICI pour fermer
Dernière mise à jour : 02/09/2016