Phasic pupillary responses reveal differential engagement of attentional control in bilingual spoken language processing

Abstract

Language processing is cognitively demanding, requiring attentional resources to efficiently select and extract linguistic information as utterances unfold. Previous research has associated changes in pupil size with increased attentional effort. However, it is unknown whether the behavioral ecology of speakers may differentially affect engagement of attentional resources involved in conversation. For bilinguals, such an act potentially involves competing signals in more than one language and how this competition arises may differ across communicative contexts. We examined changes in pupil size during the comprehension of unilingual and codeswitched speech in a richly-characterized bilingual sample. In a visual-world task, participants saw pairs of objects as they heard instructions to select a target image. Instructions were either unilingual or codeswitched from one language to the other. We found that only bilinguals who use each of their languages in separate communicative contexts and who have high attention ability, show differential attention to unilingual and codeswitched speech. Bilinguals for whom codeswitching is common practice process unilingual and codeswitched speech similarly, regardless of attentional skill. Taken together, these results suggest that bilinguals recruit different language control strategies for distinct communicative purposes. The interactional context of language use critically determines attentional control engagement during language processing.

Judgments of learning in bilinguals: Does studying in a L2 hinder learning monitoring?

“…The reason why this contrast effect was only obtained when participants were first presented with L2 is not evident and might be due to a number of reasons (e.g., more interference from L1 that balances out the possible benefits of greater effort for L2); therefore, this should be a subject of further research. However, in line with many previous bilingual studies [14,116,117], these results provide evidence that the obtained L2 effects are dependent on the context in which L2 learning is achieved.…”

Baseline pupil size encodes task-related information and modulates the task-evoked response in a speech-in-noise task

“…This modelling approach assumes a polynomial behavior of the time series and it is, furthermore, unable to account for potential autocorrelation of the response (Baayen et al, 2017;van Rij et al, 2019). Recently, generalized additive mixed modeling (GAMM; Hastie & Tibshirani, 1990) has been proposed as an alternative method to analyze pupillometry data (e.g., Algermissen et al, 2019;Aydın & Uzun, 2022;Beatty-Martínez et al, 2021;Boswijk et al, 2020;Huijser et al, 2020;Lõo et al, 2016;Pandža et al, 2020). This method overcomes some of the GCA weaknesses and could be used for similar analyses to the one presented in this study.…”

“…The time increment Δt in this study was chosen to be 0.08 s, equivalent to five time-samples. A region of the recording was considered a saccade when the calculated velocity over that region was above 22°/s, as suggested by Holmqvist et al (2011). Missing data and blink portions were linearly interpolated.…”

Attending multiple languages: the relation between individual multilingual language use and attentional control

“…It is precisely the variable control demands of different communicative contexts, determined by the diversity within an interactional context, and habits of language use (e.g. switching) that were found to determine attentional engagement and state [82]. This prompts not only the need for a characterization of individual differences in language use and experience, and of the language diversity in different interactional contexts and their attentional demands, but also of language use within the different contexts.…”