Session 2D

In psycholinguistic and L2 studies, a key mechanism brought up to tackle the issue of accessing the right language at the right time is that of language control. Thus far, language control processes have been widely considered to be specific to the bilingual brain. However, recent studies (Declerck et al. 2020) suggest there might be more similarities than expected between bilingual and monolingual lexical access. Control mechanisms could also be at work in monolinguals, specifically when it comes to using different speech registers. 
Three experiments will be conducted to investigate whether the control processes involved in bilingual lexical access could also be implemented in a single language context to control speech registers. Our participants are French-English bilinguals presenting different degrees of bilingualism. Bilingualism will be analyzed as a continuous variable to examine the evolution of the processes with L2 proficiency. 
Our lexical decision task is aimed at exploring the similarities in the links between two languages and two speech registers of the same language. Using the masked priming technique, we will compare the effect of 4 priming stimuli on the recognition of a target word: 1) a translation equivalent, 2) an informal variant, 3) a semantic associate, 4) a control word. The comparison of the RTs in the different conditions should reveal whether the processing of a register variant is closer to the processing of a translation equivalent or to the one of a semantic associate, thereby giving us insight upon the status of speech registers in the mental lexicon. 
Our switched naming task provides an experimental setting to investigate the overlap of control mechanisms during lexical access in single and dual language context. Following the methodology of Declerck et al. (2020), participants will name pictures alternating a) French and English and b) formal and informal French. RTs will be analyzed and switch costs compared in both conditions. If switch costs are similar for the between-languages and the between-speech-registers conditions, then similar control mechanisms could be involved in single and dual language context. The results obtained by Declerck et al. (2020) partly support this hypothesis but need to be replicated.
Finally, since ecological data on language switching are scarce, we will use a picture description task to investigate both language and register switching. In the bilingual condition, participants will describe 6 illustrations a) in French, b) in English, c) alternating French and English between pictures. In the monolingual condition, the same procedure will be used with formal and informal French. Comparing the disfluencies produced in each condition should tell us whether controlling the speech register or the language of production has a similar cost in natural conditions. 
This study should contribute to the understanding of language cognition in its relation to both the context of communication and the linguistic profile of the L2 learner.


Declerck, M., Ivanova, I., Grainger, J., Duñabeitia, J.A. (2020). Are similar control processes implemented during single and dual language production? Evidence from switching between speech registers and languages. Bilingualism: Language and Cognition, 23, 694–701.

Peer interactions in daycare can contribute to young multilingual children's language development, although this depends on who they interact with and how much (Washington-Nortey et al., 2020). There is preliminary evidence that children's language background and proficiency may influence the nature of peer interactions. Specifically, earlier research, including our own teacher surveys, indicates that multilingual children may in some cases experience difficulty connecting with their peers, especially if their daycare language proficiency is low (Cekaite & Evaldsson, 2017; Dominguez & Trawick-Smith, 2018). In those cases, they might find recognition in other children with the same home language (Feng et al., 2004; Kyratzis, 2010). However, evidence is often anecdotal, and sometimes contradictory. Therefore, we aim to shed more light on the effect of children's multilingualism on peer networks and peer interactions in daycare by looking at both quantitative and qualitative aspects of interaction. Specifically, we aim to answer the following questions to what extent are language background and proficiency related to who multilingual children interact with and how much, and how do these factors affect the nature of such interactions? 



We are collecting observations of peer interactions of over 100 young children (2-4 years old), who attend approximately 15 different daycare groups in the Netherlands. These daycare groups are multilingual in composition, and diverse in terms of socioeconomic status and geographic location. Free play episodes are collected on two days in every group with video cameras and a specifically designed social network analysis app (Thieme & Thieme, 2021). We are using these data for two analyses. The first is a social network analysis of multilingual children's peer interactions: how much they interact with peers and who they interact with, while controlling for child and group characteristics (e.g., age, gender, sociability, composition of the group in terms of language backgrounds). This social network analysis is based on observation and valued exponential random graph models (Krivitsky, 2012), which to our knowledge has not previously been applied to multilingualism research. The videotaped interactions will also be analysed using a grounded theory approach (Glaser & Strauss, 1967). We will examine how language background and proficiency shape these interactions, for example in terms of home language use and peer play access strategies. The results of both analyses will help to show to what extent multilingualism impacts peer interactions in this very young age group. 


In this presentation, we will present the preliminary results, with the aim of discussing the following questions: 
The daycare groups in our study are very different from each other in terms of composition. How might this impact our analysis and interpretation of results?  
Are there any control variables that we are missing, or conversely, that should not be included? 
In which ways can we relate the results of the two methods of analysis to each other?