Xavier Aparicio (University of Picardie Jules Verne – CRP-CPO)
Multilingualism and cognitive control: the language switching hypothesis
Most of the studies focusing on the access of multilingual lexicon are in line with the hypothesis of an integrated lexicon, containing all the information on words from different languages, as well as their characteristics. It seems widely accepted that during word recognition, lexical candidates from all the known languages are activated (Aparicio & Lavaur, 2013). A central question of research is to understand how multilinguals manage to control the activation of several languages, and keep the interferences between them very low. Therefore, language control implies to manage the interplay between activation and inhibition of lexical candidates (Abutalebi & Green, 2007). In addition, several studies have brought to light that bilinguals are more efficient in comparison with monolinguals in tasks involving cognitive control (e.g. Heidlmayr et al., 2013; Kroll & Bialystok, 2013).
A possible explanation accounting for better cognitive control in bilinguals in comparison with monolinguals is their ability to switch, with a relative ease, from one language to another (Aparicio & Lavaur, 2013; Green & Abutalebi, 2013). Indeed, switching from one language to another necessitates an inhibition of the irrelevant language, in order to limit interferences, and this increase in language control abilities are somehow shared with domain-general processes (Green & Abutalebi, 2013). In addition, an asymmetry in language switching was largely demonstrated (larger switch cost when switching from L2 to L1 than from L1 to L2), because of the greater automaticity of L1. However, inhibition efficiency could be modulated by proficiency, as well as the age of acquisition of the additional languages. Consequently, the impact of inhibition during language switching remains controversial (Bobb & Wodniecka, 2013) and requires studies taking into account several components of inhibition, as well as different populations of multilinguals. With this aim in view, we developed two sets of studies.
In the first one, we focused on language switching mechanisms in successive French-English-Spanish trilingual participants (native speakers of French, with a late acquisition of English (10 years old) and Spanish (12 years old). In Experiment 1, participants had to perform several lexical decisions (decide if the presented letterstring was a word or not) involving words from one, two or three languages. The results showed that reaction times and error rates increase as a function of the number of languages involved in the experimental list, due to an increase of language switches in different directions. In Experiment 2, we used masked translation priming, with targets in English or Spanish. Targets were primed by a translation or a non-related word in one of the two other languages. Interestingly, switch costs are larger when the target is preceded by an unrelated prime, in comparison with translation priming. Switching also requires more cognitive resources when switching from the two non-native languages. These results showed that language switching relies on inhibition, and is widely influenced by language proficiency.
In a second set of studies, we investigate the role of inhibition in tasks involving language control, cognitive control and motor control. In Experiment 1, we compared performances of bilinguals and simultaneous interpreters, differing in terms of expertise in task switching, but with similar performances in terms of proficiency in L2. They had to perform two tasks (language decision and Stroop), each one involving a component of active inhibition, but language decision also involves overcoming of inhibition, which requires more cognitive resources. Results showed similar performances for both groups in the Stroop task (similar performance in active inhibition), but interpreters performed better in language decision task (overcoming of inhibition). In Experiment 2, bilinguals and monolinguals had to perform a motor task of antisaccades. The hypothesis was that bilinguals, used to switch from one language to another (language control), would be better than monolinguals to exert a motor control in a non-linguistic task. Results revealed better performances for bilinguals in trials involving overcoming of inhibition, corroborating the hypothesis of a link between linguistic and motor components of inhibition.