We used a visual lexical decision task to explore control processes in proficient German–English bilinguals. Participants pressed a “yes” button if the letter string was a word in English and a “no” button if it was not. Our critical stimuli were interlingual homographs such as the low-frequency English word TAG. In German, TAG means “day” and is a relatively high frequency word. Overall, our participants responded more slowly to an interlingual homograph than to a control word matched to its English frequency. As we expected, the size of this interference effect depended on various factors. First, including “pure” German words in the stimulus list increased interference. However, participants were able to reduce the degree of interference over time even in the presence of such words. Second, in the absence of pure German words, informing participants about the presence of interlingual homographs from the start of the experimental trials allowed them to reduce interference. We examined the locus of these control effects by analysing carry-over, i.e., reaction times on word trials immediately following an interlingual homograph or its matched control. We inferred from the patterns of interference and carry-over that the primary locus for reducing interference is external to the bilingual lexico-semantic system. We consider the implications of these data for theories of control.

The aim of the present study was to examine potential effects of early exposure to sign language on the use of communicative gestures by a bilingual hearing child of deaf parents. Data collected monthly during the first two years were analyzed in order to identify types and tokens of communicative gestures, words, and signs and the ways in which they were combined. These data are compared with those obtained from 12 monolingual hearing children observed at 16 and 20 months of age who were exposed only to spoken language. Findings suggest that while exposure to sign language does not seem to provide the bilingual child with an advantage in the rate of early linguistic development, it does appear to influence the extent to which he communicated in the manual modality and made use of its representational and combinatorial potential.

This study investigates whether age of onset of bilingualism has an effect on ultimate attainment, and documents incomplete acquisition and attrition of tense/aspect morphology in Spanish. A group of 20 monolingual Spanish speakers, 31 US-born bilinguals, and 8 Latin America-born bilinguals participated in the study. Sixteen of the US-born participants were simultaneous bilinguals who had been exposed to Spanish and English from ages 0 to 3, the remaining 15 learned English between ages 4 to 7 (early child L2 learners), and the Latin-America-born bilinguals learned English between 8 and 12 years of age (late child L2 learners). Results of an oral production, a written completion, and two meaning-interpretation tasks showed that the simultaneous bilingual and the early child L2 groups differed significantly from monolingual Spanish speakers, while only a few individuals in the late child L2 group showed erosion with some aspects. Results suggest that morpho-phonological spell-outs and semantic features of functional categories are affected by incomplete acquisition and language loss, and that many of the patterns of divergence found resemble stages of second language acquisition.

In this paper, we show how some proposals within the Minimalist Program are compatible with a model of codeswitching that recognizes an asymmetry between the participating languages, the Matrix Language Frame model. Through our discussion of an analysis of NPs in a Spanish–English corpus, we illustrate this compatibility and show how recent minimalist proposals can explain the distribution of nouns and determiners in this data set if they adopt the notion of Matrix Language as the bilingual instantiation of structural uniformity in a CP. We outline the central premises of the Matrix Language Frame model, and introduce the Uniform Structure Principle which requires that the structure of constituents be uniform at an abstract level. We then review previous applications of the Minimalist Program to codeswitching.

Much recent research in minimalism has focused on issues related to feature checking. Earlier approaches to feature checking required matching of features in grammatical structures, although more recent proposals consider distinctions in the values of features and in types of matching. Because phi-features for grammatical gender in Spanish and English differ, an analysis of NPs in this corpus of naturally occurring Spanish–English conversations provides a test for minimalist applications to codeswitching. We present our general findings of the distribution of types of NP constituents and then consider explanations of these distributions in light of minimalist proposals. It is possible to explain these distributions in a recasting of the Matrix Language Frame model in minimalist terms, if the construct of the Matrix Language is maintained. The requirement that one language, the Matrix Language, provide an abstract grammatical frame in bilingual constituents corresponds to the type of uniformity that Chomsky (2001) suggests is necessary for the explanatory study of language and variation in language.