In this article, the author addresses the question of how the mind represents two languages in simultaneous bilingualism. Some linguistic theories of intrasentential code switching are reviewed, with a focus on the Minimalist approach of MacSwan (1999b); the author concludes that evidence from code switching suggests that bilinguals have discrete and separate Lexicons for the languages they speak, each with its own internal principles of word formation, as well as separate phonological systems. However, the author argues that computational resources common to the two languages generate monolingual and bilingual syntactic derivations alike. Advantages of the Minimalist Program for the analysis of code switching data are discussed at some length.

No abstract

Although some educational researchers have appealed to “semilingualism” or “limited bilingualism” to explain differences in student achievement among language minority students, in this article the author argues that the construct contributes much more to the malady than the remedy in the education of linguistic minorities. The author reviews four kinds of reputed evidence for semilingualism (from studies of language variation, linguistic structure, school performances, and language loss) and concludes that all of it is either spurious or irrelevant. The author argues that semilingualism is essentially indistinguishable from classical prescriptivism and that Cummins’s Threshold Hypothesis, which incorporates the semilingualism thesis, should be abandoned on empirical, theoretical, and moral grounds. An alternative account of the descriptive facts Cummins sought to explain is presented, and implications for education are discussed.

Translanguaging is a new term in bilingual education; it supports a heteroglossic language ideology, which views bilingualism as valuable in its own right. Some translanguaging scholars have questioned the existence of discrete languages, further concluding that multilingualism does not exist. I argue that the political use of language names can and should be distinguished from the social and structural idealizations used to study linguistic diversity, favoring what I call an integrated multilingual model of individual bilingualism, contrasted with the unitary model and dual competence model. I further distinguish grammars from linguistic repertoires, arguing that bilinguals, like monolinguals, have a single linguistic repertoire but a richly diverse mental grammar. I call the viewpoint developed here a multilingual perspective on translanguaging.

This article presents an empirical and theoretical critique of the Matrix Language Frame (MLF) model (Myers-Scotton, 1993; Myers-Scotton and Jake, 2001), and includes a response to Jake, Myers-Scotton and Gross's (2002) (JMSG) critique of MacSwan (1999, 2000) and reactions to their revision of the MLF model as a “modified minimalist approach”. The author argues that although a new structural definition of the Matrix Language (ML) makes the MLF model falsifiable, its empirical predictions are inconsistent with the facts of codeswitching (CS). The author also identifies significant theoretical problems associated with the MLF model, and suggests that it be rejected on empirical and theoretical grounds, and on grounds of scientific parsimony. In addition, the author contends that JMSG's critique of the Minimalist approach to CS rests on conceptually significant misreadings of published research in CS and in the theory of syntax, and that JMSG fail to motivate their central claim, namely, that a Minimalist approach to CS cannot succeed without incorporating the ML construct. It is further shown that an analysis of the CS facts which JMSG consider as their test case may be straightforwardly pursued within the Minimalist Program without appealing to the ML construct or any other aspect of the MLF model. The author concludes with the recommendation that research on CS depart from the formulation of general CS-specific constraints like the MLF model and engage in the task of analyzing language contact phenomena in terms of independently motivated constructs of linguistic theory.