Bilingual inhibitory control: N-2 language repetition costs unchained

Project Information

BIC

Grant agreement ID: 706128

DOI

10.3030/706128

Project closed

EC signature date 14 March 2016

Start date 1 October 2016

End date 30 September 2018

Funded under

• EXCELLENT SCIENCE - Marie Skłodowska-Curie Actions

Total cost

€ 173 076,00

EU contribution

€ 173 076,00

Coordinated by

CENTRE NATIONAL DE LA RECHERCHE SCIENTIFIQUE CNRS
France

Periodic Reporting for period 1 - BIC (Bilingual inhibitory control: N-2 language repetition costs unchained)

Reporting period: 2016-10-01 to 2018-09-30

Summary of the context and overall objectives of the project

According to bilingual research, language selection is the principal challenge for the bilingual mind, as both languages are typically activated in parallel, and thus compete with one another. In the present project, we investigated whether inhibition is involved as a crucial underlying mechanism for the resolution of this cross-language competition in production and comprehension, as suggested by several models (e.g. Green, 1998).
Prior research has provided evidence that bilingual inhibitory control is implemented, which is in line with several bilingual models, but these measures can typically also be explained without inhibition (for a review, see Declerck & Philipp, 2015). In the current proposal, we mainly relied on a marker of bilingual inhibitory control that can only been explained with inhibition, but that has not yet been investigated extensively, namely n-2 language repetition costs. More specifically, we set out to investigate whether there is bilingual inhibitory control during production and comprehension using this measure, and whether the bilingual process is persisting (i.e. continues into the next words) and/or reactive (i.e. more inhibition is required for more active languages).
In this project, we found evidence for persisting inhibitory control during production, but the evidence for comprehension shows that this is not always the case. We also found evidence for reactive inhibitory control, but the evidence showed an opposite pattern of what most models assume: lesser activation of the non-target language leads to more inhibition of this language.
This line of research is important on a societal level for several reasons. First off, because more than half world’s population is proficient in two or more languages (Grosjean, 2010), this project allowed us to gather knowledge on how language processing occurs for most people. Another reason is that a better understanding of the language control process will also lead to a better understanding of bilingual aphasic patients, who pathologically switch languages.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

We first set out to investigate whether n-2 language repetition costs, which is a marker of persisting inhibitory control, could be observed during production and comprehension (Declerck & Philipp, 2017). The results indicate that persisting inhibition is implemented during bilingual production. However, this study also indicate that inhibitory control is possible, but not necessary, during bilingual comprehension. This was also further discussed in Declerck, Meade, and Grainger (2018) and will be further explored in Declerck and Philipp (in preparation).
Because of the sparse evidence for inhibitory control during comprehension in Declerck and Philipp (2017), we came up with a paradigm to further investigate this issue. The new bilingual flanker paradigm allows bilinguals to perform a classification task on a central word that was flanked by words from the same language or another language (for a closer look at a trial in this paradigm, see Figure 1). Two studies have been published showing that inhibitory control during comprehension can be observed with this paradigm (Declerck, Snell, & Grainger, 2018; Eben & Declerck, 2018). This paradigm was also presented at the 3th International Meeting of the Psychonomic Society (Amsterdam, Netherlands).
We further investigated how language control in more general terms would occur during comprehension. In two studies, covering eight experiments, we found no comprehension-based language-switch costs. Comprehension-based language-mixing costs were observed in one experiment, but none of the others, which also puts some doubt on whether language control is necessarily implemented during comprehension. One of these studies has been submitted for peer review (Declerck, Koch, Duñabeitia, Grainger, & Stephan, submitted) and was presented at the 20th ESCOP conference (Potsdam, Germany), while the other study is being written up (Mirault, Grainger, & Declerck, in preparation).
Next to persisting inhibition, we also set out to investigate whether inhibitory control is reactive. To this end, 24 French-Italian-English trilinguals practiced naming pictures in French or English prior to the n-2 language repetition block. The results indicated that more activated (i.e. practiced) languages are inhibited less.
We also performed several related studies, such as a meta-analysis on asymmetrical switch costs, which is the measure most often cited for bilingual inhibitory control. While it is very often used as a measure of inhibitory control, it is sometimes not observed. A preliminary analysis, on the other hand, shows that this is a very reliable effect. More specific analyses are run to confirm this finding and we project that this study will be written up by the end of 2018 (Gade, Declerck, Philipp, & Koch, in preparation) and will in turn be submitted for publication.
A second related study examined a controversial issue, namely whether bilingual inhibitory control experience enhances non-linguistic inhibitory control. To investigate this issue, we related bilingualism (language proficiency, age-of acquisition, and usage of both languages) to non-linguistic inhibition by presenting a battery of ten non-linguistic inhibition tasks (e.g. Simon task and Stroop task) to a group of 160 bilinguals. The results are still being analyzed (Gade & Declerck, in preparation).
We also examined whether language control would be similar to control processes implemented in a different linguistic context. To this end, we let French-English bilinguals perform a language switching task and contrasted this against a task in which they would switch between naming pictures with a formal name (e.g. boy) or an informal name (e.g. kid). The results indicated an overlap in control processes, but also differences. This study was presented at Psycholinguistics in Flanders (Ghent, Belgium), and will be submitted soon (Declerck, Ivanova, Grainger, & Duñabeitia, in preparation).
Finally, a workshop about bilingual inhibitory control was set up during this project at the Aix-Marseille University (18-09-2018). For this workshop, Andrea M. Philipp (RWTH Aachen, Germany) was invited to give a talk.

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

This project has led to some interesting findings that were not entirely established yet in the production literature. First, we generalized the claim that bilingual inhibitory control persists during production. While evidence was observed for persisting inhibitory control, our results indicate that a larger activation leads to less inhibition, which is the opposite of what has been assumed.
We also put quite some effort in establishing more knowledge about bilingual inhibitory control during comprehension. The results of these studies indicate that bilingual inhibitory control is not necessary, but can be implemented. In this line of research, we have also developed a new paradigm, namely the bilingual flanker paradigm, which could lead to many novel breakthroughs in the field.
The main impact of this project is that researchers of bilingualism have a much better insight into how inhibitory control impacts bilingual language processing. This could help, for example, with providing better help for bilingual aphasic patients, who pathologically switch languages.

Example of a trial with the new bilingual flanker paradigm