Linguistic Approaches to Bilingualism - Volume 8, Issue 4, 2018

The container-content relation represents a set of nominal configurations unexplored in the acquisition literature. Whereas in English the switch from a noun-noun compound (water bottle) to a noun-prepositional phrase (bottle of water) is associated with a semantic shift from container to content, Spanish and Arabic adopt single canonical configurations for both conditions, noun-prepositional phrase and noun phrase, respectively. Importantly, Spanish, Arabic, and English display structural overlap in the content condition maintained by head-first isomorphic strings. In the container condition, they show structural dissimilarity; whereas English uses a head-final construction, Arabic and Spanish consistently use head-first constructions. Results from an elicited sentence-reordering task demonstrate that advanced late learners pattern native speakers when tested in Spanish but not when tested in English. Additionally, when tested in English, Arabic-speaking and Spanish-speaking learners overextend their L1 canonical configurations to both conditions. Furthermore, bilingual native speakers do not perform at ceiling, suggesting bidirectional cross-linguistic influence.

For native speakers, congruent gender marking on determiners and adjectives facilitates recognition of subsequent nouns, while incongruent marking inhibits recognition (e.g., Bates et al., 1996 ). However, there is conflicting evidence regarding whether second language learners demonstrate this effect. We investigated this issue in Swahili. Native speakers and English-speaking L2 learners of Swahili in their 3rd-5th semester completed two word repetition tasks, one examining gender and one number. Participants heard verb-noun phrases in Swahili with verbal marking that was congruent, incongruent or neutral with respect to gender or number. The time to repeat each noun was recorded. Both language groups appeared sensitive to number marking; however, only native speakers appeared sensitive to gender marking. The findings suggest the lack of a feature in the L1 may impede online processing in the L2, while the presence of a feature may mean that native-like processing is possible, even at early levels of proficiency.

Intra-sentential code-mixing presents a number of puzzles for theories of bilingualism. In this paper, we examine the code-mixed English-German utterances of a young English-German-Spanish trilingual child between 1;10 – 3;1, using both an extensive diary kept by the mother and audio recordings. We address the interplay between lexical and syntactic aspects of language use outlined in the usage-based approach (e.g. Tomasello, 2003 ). The data suggest that partially schematic constructions play an important role in the code-mixing of this child. In addition, we find, first, that the code-mixing was not mainly the result of lexical gaps. Second, there was more mixing of German function words than content words. Third, code-mixed utterances often consisted of the use of a partially schematic construction with the open slot filled by material from the other language. These results raise a number of important issues for all theoretical approaches to code mixing, which we discuss.

The central research question is how intensive exposure to the majority language at school affects the development of vocabulary. In a cross-sectional study with a longitudinal subsample follow-up, we investigated the development of vocabulary during the first four years at school, as vocabulary development is considered an important factor regarding bilingual children’s success at school. The sample consisted of 126 bilingual Russian/German-speaking children aged 6;0 to 10;11 years in Germany, who were tested for expressive and receptive vocabulary using a picture naming task.

Our results show that while the majority language is acquired at an expected rate, the heritage language’s extensive vocabulary does not develop further over the course of primary school attendance. The overlap of the vocabularies increases. Additionally, the number of items that are named exclusively in the majority language increases, whereas the number of items that are named exclusively in the heritage language decreases.