In this paper we want to compare the results from monolingual children with object omissions in bilingual children who have acquired two languages simultaneously. Our longitudinal studies of bilingual Dutch–French, German–French, and German–Italian children show that the bilingual children behave like monolingual children regarding the type of object omissions in the Romance languages. They differ from monolingual children with respect to the extent to which object drop is used. At the same time, the children differentiate the two systems they are using. We want to claim that the difference between monolingual and bilingual children concerning object omissions in the Romance languages is due to crosslinguistic influence in bilingual children: the Germanic language influences the Romance language. Crosslinguistic influence occurs once a syntactic construction in language A allows for more than one grammatical analysis from the perspective of child grammar and language B contains positive evidence for one of these possible analyses. The bilingual child is not able to map the universal strategies onto language-specific rules as quickly as the monolinguals, since s/he is confronted with a much wider range of language-specific syntactic possibilities. One of the possibilities seems to be compatible with a universal strategy. We would like to argue for the existence of crosslinguistic influence, induced by the mapping of universal principles onto language-specific principles – in particular, pragmatic onto syntactic principles. This influence will be defined as mapping induced influence. We will account for the object omissions by postulating an empty discourse-connected PRO in pre-S position (Müller, Crysmann, and Kaiser, 1996; Hulk, 1997). Like monolingual children, bilingual children use this possibility until they show evidence of the C-system (the full clause) in its target form.

In this study the visual recognition of inflected, derived and monomorphemic Swedish nouns in monolingual Swedish and bilingual Finnish–Swedish speakers was investigated. While bilinguals were slower overall, the inflected items yielded disproportionately longer reaction times in the bilingual group. The derived items, on the other hand, elicited fastest reaction times in both groups. The observed processing cost associated with inflectional morphology indicates that bilingual language background can affect the recognition process for inflected words, possibly by leading to morpheme-based recognition which is slower than full-form recognition. Further studies are needed to examine whether this effect is specific to the language background of our bilinguals (including Finnish which is a morphologically very rich language) or whether it could be a more general processing feature in bilingual speakers faced with regular inflected forms.

Tip-of-the-tongue states (TOTs) in proficient Hebrew–English bilinguals were compared to those of age-matched monolinguals. Monolinguals retrieved words in English, and bilinguals retrieved words from both languages. Results showed an increased TOT rate in bilinguals. However, bilinguals demonstrated comparable rates of spontaneous resolution, and similar ability to access partial information about target words. Interestingly, bilinguals named the same number of targets as monolinguals when naming an item in either language was counted as a correct response. Besides bilingualism, other factors that predicted TOT rate included word frequency (only for bilinguals), and age (younger participants had more TOTs). Unexpectedly, TOTs for Hebrew targets were not characterized by increased access to grammatical gender and number of syllables relative to control states, thus contrasting notably with TOTs for Italian and English targets respectively. We discuss these results in terms of their relevance for constraining models of bilingual lexical access and models of TOT.

This paper compares the development of tense morphology and verb-second in different learner populations. Three groups of Swedish pre-school children are investigated longitudinally; ten L1 children, ten L2 children and ten children diagnosed with Specific Language Impairment (SLI). Data was collected twice, with an interval of six months. The results at Time I reveal a significant difference between normally developing L1 children on the one hand and L2 children and children with SLI on the other. The L1 children use verb-second correctly in topicalized declaratives, whereas both L2 children and children with SLI use structures with the verb in third position (XSV structures) as an intermediate step towards verb-second. There is a clear development between the two data collection sessions for the L2 children and the children with SLI, diminishing the difference between them and the unimpaired L1 children. The similarity that is found between L2 children and children with SLI in this study bears important implications for the discussion of the role of transfer in L2 research and for the question of a defective linguistic representation in SLI research.