How much would Bill Murray have liked to be able to speak Japanese! Bill Murray's character in the movie Lost in Translation exemplifies the way we feel when trying to communicate with someone that does not speak the same language. Often, in such cases, the exchange of information is disrupted and even translation does not seem to capture the communicative intention of the interlocutors. Thus, to be able to speak two languages at will is obviously a worthy skill to have. However, there is also a potential drawback, namely, bilingual speakers need to control their production in such a way that the two languages do not end up mixed in an inappropriate manner during the discourse. For example, if Bill Murray would have been an English–Japanese bilingual, he would have had to be careful not to use English words when speaking to the director of the commercial. This poses interesting problems to researchers in cognitive psychology: How does a bilingual speaker control her two languages during speech production? How do bilingual speakers manage to avoid massive interference from the language they are not using? What is the role of the language-not-in-use during lexical retrieval and phonological encoding? The articles included in this issue aim at discussing the answers that have been put forward to some of these questions.

Bilingual speech requires that the language of utterances be selected prior to articulation. Past research has debated whether the language of speaking can be determined in advance of speech planning and, if not, the level at which it is eventually selected. We argue that the reason that it has been difficult to come to an agreement about language selection is that there is not a single locus of selection. Rather, language selection depends on a set of factors that vary according to the experience of the bilinguals, the demands of the production task, and the degree of activity of the nontarget language. We demonstrate that it is possible to identify some conditions that restrict speech planning to one language alone and others that open the process to cross-language influences. We conclude that the presence of language nonselectivity at all levels of planning spoken utterances renders the system itself fundamentally nonselective.

In this article we discuss different views about how information flows through the lexical system in bilingual speech production. In the first part, we focus on some of the experimental evidence often quoted in favor of the parallel activation of the bilinguals' two languages from the semantic system in the course of language production. We argue that such evidence does not require us to embrace the existence of parallel activation of the two languages of a bilingual. In the second part of the article, we discuss the possibility that the language-not-in-use (or the non-response language) is activated via feedback from the sublexical representations and we devise some experimental procedures to assess the validity of such an assumption.

Models of bilingual speech production generally assume that translation equivalent lexical nodes share a common semantic representation. Though this type of architecture is highly desirable on both theoretical and empirical grounds, it could create difficulty at the point of lexical selection. If two translation equivalent lexical nodes are activated to roughly equal levels every time that their shared semantic representation becomes activated, the lexical selection mechanism should find it difficult to “decide” between the two (the “hard problem”) – yet in some cases bilinguals benefit from the presence of a translation equivalent “competitor”. In this article, we review three models that have been proposed as solutions to the hard problem. Each of these models has difficulty accounting for the full range of findings in the literature but we suggest that these shortcomings stem from their acceptance of the assumption that lexical selection is competitive. We argue that without this assumption each proposal is able to provide a full account of the empirical findings. We conclude by suggesting that the simplest of these proposals should be rejected before more complicated models are considered.

Phonological encoding is the process by which speakers retrieve phonemic segments for morphemes from memory and use the segments to assemble phonological representations of words to be spoken. When conversing in one language, bilingual speakers have to resist the temptation of encoding word forms using the phonological rules and representations of the other language. We argue that the activation of phonological representations is not restricted to the target language and that the phonological representations of languages are not separate. We advance a view of bilingual control in which condition-action rules determine what is done with the activated phonological information depending on the target language. This view is computationally implemented in the WEAVER++ model. We present WEAVER++ simulations of the cognate facilitation effect (Costa, Caramazza and Sebastián-Gallés, 2000) and the between-language phonological facilitation effect of spoken distractor words in object naming (Hermans, Bongaerts, de Bot and Schreuder, 1998).

Recent work in the bilingual literature suggests that naming pictures in a second language (L2) differs from naming pictures in the first language (L1) because of effortful lexical retrieval. This finding has received some support in the neuroimaging literature (De Blesser et al., 2003). In the current study, twelve Spanish-English bilinguals, who had learned English later in early adulthood, were asked to name pictures covertly in either Spanish or English while being scanned with functional Magnetic Resonance Imaging (fMRI). Picture naming in the second language (L2) relative to the native language (L1) revealed increased activity in the right insula, anterior cingulate gyrus and dorsolateral prefrontal cortex (DLPFC) and the left fusiform gyrus. These results are consistent with the view that picture naming in a less proficient L2 requires increased effort to establish links between motor codes and visual forms.

The typical speech of (fluent) bilinguals in monolingual settings contains few switches into the non-target language. Apparently, bilinguals can control what language they output. This article discusses views on how bilinguals exert control over their two languages in monolingual tasks, where participants only have to implicate one of their languages in performing the task, and in translation and, especially, simultaneous interpreting, tasks that can only be performed if both languages are addressed. A distinction is made between “global” control, where control involves the activation and/or inhibition of complete language systems, and “local” control, where control impacts on a restricted set of memory representations. A number of studies suggest that bilingual control is a special case of the control of action in general. This insight suggests an opportunity to incorporate relevant work in the field of translation studies in the study of bilingual control, embedding it in the same theoretical framework.

This contribution discusses findings and hypotheses from empirical data of naturally-occurring codeswitching. The discussion is framed by some comparisons of the approaches of contact linguists and psycholinguists to bilingual production data. However, it emphasizes the relevance of naturally-occurring codeswitching to the theoretical questions asked by psycholinguists. To accomplish this, relevant grammatical structures in codeswitching are exemplified and analyzed. Analysis largely follows the Matrix Language Frame (MLF) model, but differing approaches are mentioned.