In the last decade there has been a significant surge of interest in the psycholinguistics of bilingual memory which has resulted in a number of models of the relationship between words and concepts in the bilingual lexicon. Recently, authors such as Paradis (1997) and Grosjean (1998) have pointed out that many of these models do not distinguish between the semantic and the conceptual level, and that they suffer from confusion between processing and representation.

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.

This study examined the recognition of English words by groups of native speakers of Italian who differed in age of arrival in Canada and amount of continued native language use. The dependent variable was the number of words correctly repeated in English sentences presented in noise. Significantly higher word recognition scores were obtained for early than late bilinguals, and for early bilinguals who used Italian seldom than for early bilinguals who used Italian relatively often. A hierarchical regression analysis showed that the native Italian participants' ability to perceive English vowels and consonants accounted for a significant amount of variance in the word-recognition scores independently of age of arrival, amount of L1 use, and length of residence in Canada. The native language use effect was interpreted to have arisen from differences in the extent to which the early bilinguals' Italian phonetic system influenced the representations they developed for English vowels and consonants.

We contrasted the effect of instruction-induced expectancies and language intermixing in an English lexical decision task performed by Dutch–English bilinguals. At the start of the experiment, participants were instructed to respond to interlingual homographs and exclusively English words by giving a “yes” response, and to English non-words and exclusively Dutch words by giving a “no” response. In the first part of the experiment the stimulus list did not contain any Dutch words. In the second part of the experiment, Dutch items were introduced. No significant differences were found between interlingual homographs and controls in the first part of the experiment, while strong inhibition effects were obtained for interlingual homographs in the second part. These results indicate that language intermixing rather than instruction-based expectancies drives the bilingual partipants' performance. Consequences for current views on bilingual word recognition are discussed.