Using network science to map what Montréal bilinguals talk about across languages and communicative contexts

Abstract

Recent work within the language sciences, particularly bilingualism, has sought new methods to evaluate and characterize how people differentially use language across different communicative contexts. These differences have thus far been linked to changes in cognitive control strategy, reading behavior, and brain organization. Here, we approach this issue using a novel application of Network Science to map the conversational topics that Montréal bilinguals discuss across communicative contexts (e.g., work, home, family, school, social), in their dominant vs. nondominant language. Our results demonstrate that all communicative contexts display a unique pattern in which conversational topics are discussed, but only a few communicative contexts (work and social) display a unique pattern of how many languages are used to discuss particular topics. We also demonstrate that the dominant language has greater network size, strength, and density than the non-dominant language, suggesting that more topics are used in a wider variety of contexts in this language. Lastly, using community detection to thematically group the topics in each language, we find evidence of greater specificity in the non-dominant language than the dominant language. We contend that Network Science is a valuable tool for representing complex information, such as individual differences in bilingual language use, in a rich and granular manner, that may be used to better understand brain and behavior.

Bilingualism: A Neurocognitive Exercise in Managing Uncertainty

“…Whether someone is trying to decipher multilingual signs at high speeds on the highway, order coffee in a bilingual city, or communicate academic research to multilingual peers, the people involved in these interactions bring to the table their individual levels of language knowledge, language fluency, language preferences, overt goals, and covert intentions. Bilingual environments thus have fluctuating language demands (Anderson et al, 2018;Beatty-Martinez et al, 2019;Grosjean, 2001;Gullifer et al, 2020;López, 2020;López et al, 2020;Tiv, Gullifer, et al, 2020b), which corresponds with a set of cognitive, linguistic, and social uncertainties. Individuals must resolve or adapt to these uncertainties by tuning the neurocognitive systems responsible for language and cognitive control (Abutalebi & Green, 2016;Green & Abutalebi, 2013;Green & Wei, 2014).…”

“…The notion of behavioral context is central to many usage-based theories about language and bilingualism, because people perceive and produce the various languages that they know with interlocutors in their environments (such as at home or in the workplace). This rich contextualization of language has wide-ranging consequences for language fluency, processing, representation and control, and it may also carry consequences for domain general cognitive control and underlying brain mechanisms (Adler et al, 2020;Anderson et al, 2018;Beatty-Martinez et al, 2019;Green & Abutalebi, 2013;Grosjean, 2001Grosjean, , 2016Hofweber et al, 2020;Tiv, Gullifer, et al, 2020b). To give one example, the adaptive control hypothesis (Green & Abutalebi, 2013) posits that language usage within particular INTERACTIONAL CONTEXTS will have adaptive consequences for control and brain organization, where interactional contexts consist of the "recurrent pattern of conversational exchanges within a community of speakers" (Green & Abutalebi, 2013, p. 516).…”

“…The level of language-related uncertainty can be estimated, at a basic level, using entropy measures (Eleta & Golbeck, 2014;Guzmán et al, 2017), either at the aggregate level (for a sample of participants), or as an individual difference measure (Gullifer et al, 2021;Guzmán et al, 2017). An even richer characterization can be provided by network scientific approaches (Eleta & Golbeck, 2014;Tiv, Gullifer, et al, 2020b). Here, the entities in an environment and their interrelationships are modeled as networks using graph theory, allowing for a set of measures, including language entropy, to be extracted that provide information about the fundamental structure of an interactional (or behavioral) context.…”

“…While network models of multilingual language usage have been constructed from online sources, like Twitter (e.g., Eleta & Golbeck, 2014), they have not, to our knowledge, been used to assess in-person, bilingual language usage. In a recent paper, we provide an example of how network science can be leveraged to uncover information about naturalistic language usage (Tiv, Gullifer, et al, 2020b). We surveyed individuals about the languages that they use to discuss several topics of conversation (e.g., politics, sports, moral issues, religious issues) throughout different communicative contexts (e.g., at home, at work, etc.).…”

Moving From Bilingual Traits to States: Understanding Cognition and Language Processing Through Moment-to-Moment Variation

“…For example, higher language entropy, or greater diversity in language experience, relates to greater resting-state brain connectivity between regions associated with second language speech production and executive functions (e.g., between the bilateral putamen and ACC; Gullifer et al, 2018). Another fruitful avenue is to incorporate methods from social-network analysis to further quantify the nuances of bilingual language experience (Tiv et al, 2020) and demonstrate its effect on language processing (e.g., Lev-Ari, 2018, Lev-Ari & Shao, 2017. Using an exposure-based approach to quantifying bilingualism, one study drew different conclusions and interpretations about the effects of bilingualism and age on executive function than studies that treat age and bilingualism as dichotomous (Incera & McLennan, 2018).…”