Verbal short-term memory and vocabulary development in monolingual Dutch and bilingual Turkish-Dutch preschoolers

Abstract

With increasing immigration, bilingualism has become part and parcel of the everyday lives of many children. Although research indicates that under favourable circumstances bilingual children can become balanced bilinguals, especially immigrant children seem to have difficulty coping with the language of schooling. In the current thesis, 67 Turkish-Dutch immigrant children and 72 Dutch native speakers were followed during their kindergarten years and tested at 4, 5 and 6 years of age. For most Turkish immigrant children in the Netherlands, Turkish is their home language and school enrolment at the age of 4 marks their first submersion in a primarily Dutch language environment. The Turkish-Dutch children in our study did not catch-up with their Dutch peers in Dutch vocabulary level; their rate of development was even somewhat slower. As vocabulary at the start of kindergarten is an important predictor for later literacy development and overall school success, it is of urgent practical interest to disentangle different factors influencing bilingual vocabulary development from an early age. In the current thesis, it was shown that the cognitive factor referred to as verbal short-term memory is a crucial interface between language input and vocabulary acquisition not only in monolingual, but also in bilingual children. Although this factor has been suggested to be highly heritable, our results showed that it is influenced considerably by the language input children experience in their homes, and that all developmental improvements in the capacity of verbal short-term memory in the preschool period can be attributed to growing long-term knowledge support. Because Turkish-Dutch children receive less input in Dutch in their home environments due to the bilingual situation, they have less support from long-term knowledge of Dutch, and consequently more difficulty in remembering new words in that language, which constitutes a double risk. Especially knowledge of the statistical distribution of phoneme clusters in Dutch, so called phonotactic knowledge, seems to be less available in their second language. The children had some support from phonotactic knowledge in Dutch, but performance was far below that of the Dutch group. As the Turkish-Dutch and Dutch children did not differ in nonverbal intelligence, visuospatial short-term recall, and verbal short-term recall for highly automated digit knowledge, it was not likely that these group differences could be attributed to differences in general cognitive abilities. These results suggested that because of a disadvantage at the very basic level of phonotactic knowledge acquisition, the Turkish-dominant bilingual children had more difficulties with sustaining new aspects of the Dutch language in their memory for short-periods of time and as a result were hampered in their construction of long-term representations. Based on these results, it is advisable to develop better tailored special educational programs by including verbal short-term memory training to improve the capacity to learn from input, artificial speech stream training to improve Dutch phonotactic knowledge development, and encouraging teachers to use shorter instructions and teach compensatory memory strategies to reduce verbal short-term memory demands.