Code mixing in early bilingual acquisition: dominance, language modes, and discourse strategies. A case study of bilingual acquisition of Norwegian and English

One of the main questions that arises in cases where children code switch or code mix often in their early production is whether mixing can be considered a sign of confusion, of competency, or of neither. Earlier research on language mixing in bilingual first language acquisition often pointed to the idea of a “unitary language system,” with code mixing being considered a sign of confusion. More recent and systematic research contradicts the claim of the hypothesis, however. More recently, researchers have proposed that bilingual children use code mixing as a way to fill lexical gaps where translation equivalents are not yet present. This study examines the code mixing of a bilingual girl, Hedda, growing up in northern Norway and acquiring Norwegian and English from birth. Hedda’s father is American and consistently speaks English with her, while Hedda’s Norwegian mother consistently addresses her in Norwegian. The parents speak Norwegian with each other. Thus, although her father is a consistent and ample source of English input, Norwegian is clearly the dominant language in Hedda’s surroundings. The thesis uses corpus data collected between the ages of 2;3-3;3 to examine Hedda’s mixing habits when conversing with native speakers of both Norwegian and English in order to compare her code mixing habits in each language. The data show that Hedda code mixes extensively when communicating with English speakers, while code mixing is almost absent in Norwegian contexts. Hedda’s mastery of English lags behind her Norwegian competence (the latter is comparable to that of her monolingual Norwegian peers). Norwegian is clearly her dominant language. Moreover, her day-to-day life rarely brings her into contact with monolingual English contexts, making a monolingual English mode a rare occurrence for her. Based on this, the thesis examines possible contributing factors to account for Hedda’s code mixing patterns