Abstract
Categorization is a crucial process in language acquisition. In order to be able to construct and analyse words and sentences, the child has to build up a system of phonemes, i.e., abstract representations of speech sounds, as well as grammatical categories such as ?noun? and ?verb?. Research on early language acquisition in children suggests that statistical-distributional analysis of speech is crucially involved in inducing the phonemes of the native language. There are also strong indications that statistical-distributional analysis plays a role in the formation of grammatical categories. The current proposal starts from the hypothesis that there is a single neurocognitive mechanism underlying the statistical-distributional analysis of spoken utterances, supporting categorization in both phonology and grammar. The two postdoc projects in the programme comprise a series of experiments testing statistical and distributional analysis at the phonological and grammatical levels in children between 5 and 30 months of age. The following questions are addressed: (1) Do corresponding manipulations of distributional-sequential properties of input at the levels of phonology and grammar have similar effects on categorization performance? (2) Is children?s performance on phonological categorization predictive of their performance at the level of grammar (and vice versa)? (3) Do children with a developmental language delay perform more poorly than typically developing children on grammatical and phonological categorization tasks? Two groups of children will be involved: typically developing children, and children with a genetic risk of dyslexia. On average, language development in at-risk children is delayed, and the current proposal hypothesizes that this is due to a dysfunctioning distributional-sequential analysis mechanism underlying linguistic categorization. The resulting deficiencies in, specifically, phonological categories have a detrimental effect on the acquistion of an alphabetical writing system.