Retour vers liste

Détail de la contribution

Auteur: Joao COSTA

Titre:
The robustness of language development: the role of crosspoulation and crosslinguistic studies


Abstract/Résumé: The literature on language acquisition from the last two decades reveals that most syntactic operations and configurations are acquired very early on, and that parameter values are set in the first year of life. These findings provide a good piece of evidence in favor of the innate nature of language. It is however known that not all syntactic knowledge is acquired so early, and interesting asymmetries can be observed between very similar constructions. For instance, in A-bar dependencies, object dependencies are acquired later than subject dependencies, and in the domain of referential dependencies, pronouns are harder to master than anaphors (Chien and Wexler 1990). A detailed analysis of the types of dependencies that develop slower reveals the relevance of fine-grained linguistic investigation (e.g. Friedmann, Belletti and Rizzi 2009 show that object dependencies are only problematic in intervention configurations). Furthermore, crosslinguistic variation is found: for example, pronouns do not pose problems for comprehension when they are clitic (McKee 1992, Padilla 1994). Variation in the pace of acquisition and crosslinguistic variation in acquisition provides a very interesting window into the universality of language development, but, as I will argue, do not challenge its robustness. In this talk, I aim at discussing the importance of crosslinguistic and crosspopulation studies in order to determine what can be explained in language development in terms of fundamental language knowledge, language-specific characteristics or general processing capacities non specific to language. Crosspopulation studies: I will show that certain difficulties in language development – particularly in the parsing of movement and referential dependencies – do not follow from a lack of language knowledge, but from different working memory capacities (for this, I will report on studies comparing typically developing children, SLI children, deaf children, and patients with Alzheimer’s disease). The point to be made is that certain tendencies found in children are also found in populations without any language disorder. This argues for the robustness of linguistic knowledge, and the slower development can be independently motivated. Crosslinguistic studies: I will show that difficulties found in certain languages in the development of specific constructions are not universal, and follow from language-specific properties, and that the crosslinguistic asymmetries can be explained once the role of interfaces is taken into consideration.