Lexically linked domains in language allow a speaker to formulate incompatible rules. How should they be represented theoretically? We argue that a speaker has a set of mini-grammars for different domains so that, in effect, every speaker is bilingual. It is argued that Tense or Agreement Checking, V-2 for quotation, and resumptive pronouns, all lead to bilingual representations. In addition, this perspective on Theoretical Bilingualism suggests that optionality and stages in the acquisition of an initial grammar should also be characterized as a form of bilingualism.

In this paper, I argue that current approaches to modeling of concepts in bilingual memory privilege word representation at the expense of concept representation. I identify four problems with the study of concepts in bilingual memory: conflation of semantic and conceptual levels of representation; scarcity of methods targeting conceptual representation; assumption of the static nature of the conceptual store; and insufficient acknowledgment of linguistic and cultural specificity of concepts. Basing my arguments on recent developments in the fields of neurolinguistics, linguistics, psychology, linguistic anthropology, and second language acquisition, I suggest new approaches to the study of concepts in bilingualism, based on notions of concept comparability and concept encoding. Subsequently, I discuss various ways in which concepts could develop and interact with each other in bilingual memory and address possible individual, psycholinguistic, and sociolinguistic constraints on conceptual representation and interaction in bilingual memory.

In this study we address the question of how lexical selection is achieved by bilingual speakers during speech production. Specifically, we test whether there is competition between the two lexicons of a bilingual during lexical access. In two picture–word interference experiments we explore the performance of two groups of bilinguals, English–Spanish and Spanish–English proficient bilinguals while naming pictures either in their L1 (Spanish) or in their L2 (Spanish). Picture naming was facilitated when the name of the picture and the distracter word were the “same”, regardless of the language in which the distracter was printed: same-language (e.g., mesa–mesa [table in Spanish]) or different-language pairs (e.g., mesa–table). The magnitude of this facilitatory effect was similar when naming in L1 (Experiment 1) and in L2 (Experiment 2). We also found that naming latencies were slower when the distracter word was semantically related to the picture's name (e.g., mesa–chair), regardless of the language in which the distracter was printed. The results suggest that there is no competition between the two lexicons of a bilingual during lexical access for production. This interpretation favors a model of lexical access in which lexical selection is language-specific both when speaking in L1 and in L2.

This study had two purposes: (i) to see if bilingual children can differentiate their languages with respect to the ability to form compound nouns and (ii) to test the validity of previous explanations of the acquisition of compounds. Compound nouns are right-headed in English and left-headed in French. If the French–English bilingual child in this study could differentiate between the two compounding rules, his compounds should show differential order based on the language of the semantic head. The analysis was based on the child's spontaneous compound productions from 2;9 to 3;3. The results showed that the language choice of the semantic head noun predicted the order of his compounds, suggesting that he had two distinct compounding rules. The pattern of errors made by the child cannot be accounted for by any previous explanation alone. It is suggested that children use various cues to learn compound structure.