Language impairment in Swedish bilingual children - epidemiological and linguistic studies

The aim of this thesis was to provide both epidemiological and linguistic data on bilingual children with language impairment (LI), since few data on this group exist. Data from 438 bilingual and monolingual children, referred over a period of 12 months, were compared in the first epidemiological study. The results showed that the bilingual children with suspected LI were referred significantly later than monolingual, and by more diverse referral sources. The risk of being discharged owing to non-attendance was significantly higher in the bilingual group, and the risk increased with increasing severity of LI. In the second epidemiological study, data from 698 assessed bilingual and monolingual children were compared. They were referred over a period of 24 months. Most risk factors for LI relative to degree of severity seemed to apply for both bilingual and monolingual children. Parental stress and short attention span were risk factors in both groups. Gender was a risk factor in the monolingual group, but did not appear as a risk factor in the bilingual group until triggered by environmental factors. Some risk factors were specific for the bilingual group, as maternal arrival within + 1 year in relation to child birth and parental need of interpreter after >5 years of residence in Sweden. In the linguistic studies, 10 Swedish-Arabic children with severe LI and 10 matched controls were tested in both languages with grammatical and phonological tests, designed to allow for comparisons between the languages. The results showed that all bilingual children followed the same developmental sequence on their way towards the target language, both cross-sectionally and longitudinally. The children with severe LI were able to develop both their languages, although at a much slower pace. They also seemed more vulnerable to a limited exposure of both Swedish and Arabic. The phonological study showed that the children with and without severe LI, tended to develop Swedish and Arabic phonology in the same way as monolingual children, with some exceptions. After approximately 1;6 years of exposure to Swedish in preschool (span 1;2-2;0), the children without LI scored at high grammatical levels in Swedish, and the number of phonological processes decreased substantially. This was not the case with most of the children with severe LI. This time limit has the potential of becoming clinically significant when assessing bilingual children with suspected LI.