Code-switching in bilingual children (review)

Both empirically and theoretically, Katja F. Cantone's ambitious study on language mixing in young German-Italian bilinguals is a promising, though occasionally exasperating, addition to the field. Empirically, C's treatment of experimental design and data collection is meticulous. As regards theory, she advances an elegant story for children's language mixing. The book's introduction and eight substantive chapters develop the intriguing premise that young bilinguals' mixed utterances are not due to developmental factors and therefore should be analyzed no differently from adult code-switching. Insofar as C's target hypothesis is supported, this work obtains an even wider scope than its title denotes, providing insights not available from monolingual accounts for the role of functional categories in sentence building.

Ch. 1 articulates bilingual first language acquisition (BFLA)—children exposed to two languages simultaneously from birth who consequently acquire two first languages—and delineates historical baselines of BFLA research. C lays the foundation for her nondevelopmental code-switching account, arguing against the unitary-language-system hypothesis (i.e. bilingual children possess a single linguistic system in the first stages of language acquisition). Since the pivotal term 'system' is not discussed, readers should assume a generative-grammar interpretation whereby a language system consists of modules (e.g. syntax, phonology, etc.) with complex interactions functioning together in both the representation and the production of language. This said, most BFLA arguments supporting either one or two systems are inevitably circular because the evidence advanced is never related to a totality of intersecting modules. The present study, like so many others, is unable to escape this shortcoming. But more crucially, C's case against the single-system hypothesis is not compelling because her syntactic arguments at the sub-sub-modular level are simply inconclusive. In any event, the two-system hypothesis (i.e. bilingual children start the acquisition task with separate language systems) is hailed as the cornerstone of C's analysis at chapter's end. This outcome may be perfectly valid, yet given the absence of critical scrutiny of the two-systems account, a biased tone is set in the chapter that seems at odds with the author's general intent of scientific rigor.

Ch. 2 surveys early language mixing (ELM), touching on dual lexical acquisition, translation equivalents, and language dominance. Standard developmental accounts hold that mixed utterances prior to age three indicate impoverished grammatical, lexical, and/or pragmatic competence. C attributes ELM to language-separation difficulties (i.e. sociolinguistic-based focus on the use of two languages), claiming that ELM amounts to flawed methodologies and poorly defined language contexts during data collection. As with most complex questions, the accurate picture probably involves some combination of developmental and sociolinguistic factors. Consequently, although C attempts to demonstrate the superiority of a language-differentiation perspective (i.e. grammatical focus on two discrete language systems in the mind/brain), her questionable supporting evidence forces continued discussion of sociolinguistic aspects as well as the unitary-system hypothesis.

Ch. 3 outlines the minimalist program (MP) (Chomsky 1995, 2000, 2001). Tenets central to C's analysis include MP's lexicalist assumption that all items come fully inflected from the lexicon, and the functional parameterization hypothesis, which holds that parameters of syntactic variation are associated with F(functional)-categories. C also claims that 'Select', the operation that takes items from the lexicon and places them into the lexical array, plays a crucial role in child mixed utterances. C supplies straightforward German and Italian examples integrating these core MP notions. She motivates theory-internal differences in the languages, such that Italian main clauses do not project complementizer phrases (CP), whereas German always projects CP. These conditions have consequences for the role of F-categories in building distinct word orders, and will prove equally relevant for C's bilingual analysis. Ch. 3 ends with an overview of L1 acquisition theories based largely on maturation and continuity frameworks. C selects the weak continuity hypothesis as most compatible with MP assumptions and therefore most suited to her Ch. 6 bilingual data...