We know that infants can extract regularities from the speech signal through statistical learning (SL) and that this is a fundamental mechanism in language learning. Approaches based on SL assume that statistics are automatically computed over all available input information. However, the language learning situation is so rich and multidimensional that infants need to pinpoint the appropriate subset of dimensions relevant for computing statistics. How does the baby navigate through the immense search space, being such space dependent on each language?In this project, I will explore the relationship between mechanisms of attention and cognitive control, and language acquisition. I will approach the issue through two different strategies. First, I will analyse how possessing better mechanisms of attention and control contributes to tuning the language processing system to the environmental languages specificities. To this end, I will start by measuring language and attention development in 11- and 30-month-old children. Then, I will evaluate language and attention in a group of adults, capitalising on the important individual differences existing in non-native speech perception. Second, I will take the opposite perspective and explore how specific language exposure (namely, bilingualism) sculpts mechanisms of cognitive control. In this research line, I will focus on how bilingual exposure alters mechanisms of attention and control in preverbal infants (of which meagre evidence exists). The project will focus on (but will not be constrained to) the way the phoneme inventory of the native language is established. Phonemes are one of the pillars of the language system that are tuned early on to the language of the environment and upon which fundamental computations are performed, yielding the discovery of words and morphosyntactic properties.
ERC ADVANCED GRANT. 01/09/2013 - 31/08/2019 PI: Nuria Sebastian Galles