Highlights
-
• Young monolinguals and bilinguals did not differ in their inhibition, switching and updating of information.
-
• Monolinguals and bilinguals show distinct correlation patterns between language and executive control.
-
• Contrary to bilinguals, monolinguals with lower bilingualism scores show better inhibition abilities.
-
• Dual-language contexts might have opposite influences for monolinguals and bilinguals.
1. Introduction
In recent years, the claim that bilinguals exhibit superior cognitive abilities over monolinguals in tasks involving executive control – a concept known as the bilingual advantage (Bialystok, Reference Bialystok2017; Poarch & Krott, Reference Poarch and Krott2019; Takahesu Tabori et al., Reference Takahesu Tabori, Mech and Atagi2018) – has begun to diminish (Antoniou, Reference Antoniou2019; Nichols et al., Reference Nichols, Wild, Stojanoski, Battista and Owen2020. Recent studies in young bilingual adults (Donnelly et al., Reference Donnelly, Brooks and Homer2019; Lehtonen et al., Reference Lehtonen, Soveri, Laine, Järvenpää, de Bruin and Antfolk2018; Mukadam et al., Reference Mukadam, Sommerlad and Livingston2017; Ware et al., Reference Ware, Kirkovski and Lum2020) have indicated that the hypothesis advocating for enhanced higher order cognitive abilities due to bilingualism is not as strong as initially suggested by earlier studies. Furthermore, this deflating phenomenon is even reaching the strongholds of the, until recently, developmental phases in which the effect of bilingualism on executive control was relatively undisputed, namely, children and older adults (Dick et al., Reference Dick, Garcia, Pruden, Thompson, Hawes, Sutherland, Riedel, Laird and Gonzalez2019; Papageorgiou et al., Reference Papageorgiou, Bright, Tomas and Filippi2019, respectively). These recent null results have led to a focused investigation into the particular language interaction context under which bilingualism might exert benefits in nonlinguistic cognitive task performance. As a result, a growing body of research is exploring how diverse and unique language experiences may distinctly influence executive control (Jiao et al., Reference Jiao, Grundy, Liu and Chen2020; Sanchez-Azanza et al., Reference Sanchez-Azanza, López-Penadés, Aguilar-Mediavilla and Adrover-Roig2020). Hence, to grasp the nuanced impacts of bilingualism on executive control, it is proposed to move beyond a binary view of bilingualism and instead consider it as a spectrum, thus acknowledging the wide range of language experiences and proficiency levels (Dash et al., Reference Dash, Joanette and Ansaldo2022a, Reference Dash, Joanette and Ansaldo2022b; Luk & Bialystok, Reference Luk and Bialystok2013). Considering the recent challenges to the bilingual advantage hypothesis and the diverse findings across various age groups, the next logical step is to delve into bilingual language experiences in greater detail.
The adaptive control hypothesis is a theoretical perspective that seeks to account for the diversity of findings in the study of bilingualism and cognition (Green & Abutalebi, Reference Green and Abutalebi2013). This hypothesis posits that bilinguals’ language control abilities adapt to their specific interactional contexts, indicating that the cognitive benefits of bilingualism are more nuanced and dependent on these contexts than previously thought. Green and Abutalebi (Reference Green and Abutalebi2013) detailed three specific interactional contexts for bilinguals: the single-language context (environments in which languages are used separately in different contexts), the dense code-switching context (environments where a mixture of both languages is spoken) and the dual-language context (environments where both languages are commonly used and are almost interchangeable).
Consequently, bilinguals with different conversational habits are likely to show varied executive control benefits (Beatty-Martínez et al., Reference Beatty-Martínez, Navarro-Torres, Dussias, Bajo, Guzzardo Tamargo and Kroll2020), shaped by these interactional contexts –single-language, dense code-switching and dual-language. Therefore, based on the adaptive control hypothesis, bilingualism would show an impact on tasks tapping into inhibition (i.e., suppression of irrelevant information) because bilinguals must inhibit the nontarget language to avoid cross-language interference (Abutalebi & Green, Reference Abutalebi and Green2008; Bialystok, Reference Bialystok2015). Moreover, bilingualism would affect the performance of bilinguals in tasks tapping into switching (i.e., task engagement and disengagement) because they find themselves in the need of shifting languages back and forth depending on the context and to whom they are speaking to (Prior & Macwhinney, Reference Prior and Macwhinney2010). Finally, since bilinguals must be on constant alert to monitor in which language other people are talking to them (Costa et al., Reference Costa, Hernández, Costa-Faidella and Sebastián-Gallés2009), bilingualism affects their updating abilities (i.e., goal maintenance, conflict monitoring and manipulation of information in working memory).
Even though these purported effects seem to be common to all bilinguals to a greater or lesser degree, it has been hypothesized that the dual-language context is the most demanding interactional environment (Crespo et al., Reference Crespo, Gross and Kaushanskaya2019;Green & Abutalebi, Reference Green and Abutalebi2013; Hartanto & Yang, Reference Hartanto and Yang2016), potentially affecting a broader set of executive processes, such as inhibition, switching and updating (Miyake & Friedman, Reference Miyake and Friedman2012), than the other two contexts (Green & Abutalebi, Reference Green and Abutalebi2013). In this sense, some authors have suggested that studies asserting bilingual advantage in executive control typically involve bilingual individuals facing large contextual demands (i.e., dual-language context), whereas studies less likely to capture this phenomenon are those with bilingual participants with minimal contextual demands (Lehtonen et al., Reference Lehtonen, Soveri, Laine, Järvenpää, de Bruin and Antfolk2018; Verreyt et al., Reference Verreyt, Woumans, Vandelanotte, Szmalec and Duyck2016). Given that bilingualism is increasingly recognized as a continuous domain affected by idiosyncratic experiences and not as a static entity (Dash et al., Reference Dash, Joanette and Ansaldo2022a; DeLuca et al., Reference DeLuca, Rothman, Bialystok and Pliatsikas2020) with a high dependence on contextual and individual variations (Green & Abutalebi, Reference Green and Abutalebi2013; Hartanto & Yang, Reference Hartanto and Yang2016), the same should be acknowledged for monolinguals.
In this sense, one aspect that remains unstudied and could be a source of confounding factors leading to some of the inconsistent results in the field involves the linguistic profile and language experience of monolinguals (Castro et al., Reference Castro, Wodniecka and Timmer2022). As Antoniou (Reference Antoniou2019) noted, research on bilingualism often places significant emphasis on the linguistic profile of the bilingual samples, yet comparatively less consideration is given to their monolingual counterparts. Monolingualism has traditionally been treated as a uniform category, yet previous research suggests that monolingual speakers also experience varying degrees of linguistic exposure and use (Castro et al., Reference Castro, Wodniecka and Timmer2022; Luk, Reference Luk2015). Factors such as living in multilingual environments, prior foreign language learning or passive exposure to other languages can shape their linguistic experiences and cognitive engagement with language (Bice & Kroll, Reference Bice and Kroll2019; Ellis, Reference Ellis2008). Research indicates that even without active bilingualism, the linguistic environment can influence monolinguals’ language processing (Bice & Kroll, Reference Bice and Kroll2019) and potentially other cognitive functions (Castro et al., Reference Castro, Wodniecka and Timmer2022). Exploring this variability is crucial for accurately characterizing monolinguals and understanding their role in comparisons with bilingual populations. Initially developed to explain how bilinguals’ language control abilities adapt to their specific interactional contexts (Green & Abutalebi, Reference Green and Abutalebi2013), the adaptive control hypothesis provides a valuable framework to examine the broader influence of linguistic context on cognitive processes. Extending this perspective, we propose to consider how the principles of the adaptive control hypothesis, particularly the emphasis on contextual demands and adaptive control, might relate to monolinguals exposed to the rich linguistic diversity of dual-language environments. This leads us to predict that monolinguals in such contexts, despite not actively switching between languages, may exhibit enhanced cognitive control functions due to the constant presence and processing (even passively) of multiple languages in their environment.
1.1. Purpose of the study
We aim to explore a novel extension of this hypothesis by testing whether monolinguals immersed in a dual-language context experience cognitive benefits from the linguistic context to an extent comparable to bilinguals who are also exposed to the same environment. Additionally, it explores the relation between their distinct language abilities and their executive control performance. To this end, monolingual and bilingual young adults completed a battery of cognitive tasks and bilingualism measures to address the following research questions:
Research question 1: Do monolingual and bilingual individuals from dual-language contexts show differences in performance on the three subcomponents of executive control, specifically inhibition (measured by the Stroop task), switching (measured by the Trail Making Test) and updating (measured by the Digit Span)?
Within the adaptive control hypothesis (Green & Abutalebi, Reference Green and Abutalebi2013), monolinguals in a bilingual environment, especially in a dual-language context, are also exposed to the common use of both languages to some extent, even though they are not able to produce and/or understand one of them. Consequently, monolinguals might be subjected, to some degree, to the potential benefits bilinguals enjoy from sharing the linguistic context. This would theoretically result in the absence or, at least, the blurring of context-specific benefits (namely, bilingual advantages in the case of bilinguals alone) when comparing monolinguals and bilinguals from the same dual-language contexts. The most straightforward way to verify this assumption may be to compare the cognitive outcomes of bilinguals and monolinguals that are immersed in a dual-language context, since it is assumed to be the most demanding context and where the effect of the environment should be more noticeable for bilinguals.
Thus, we speculate that, in this specific setting, a context-specific benefit would emerge in both inhibition and switching for bilinguals since monolinguals are not sufficiently proficient in a second language that is not completely unfamiliar to them. That is, the reduced likelihood of cross-language activation and interference (Blom et al., Reference Blom, Küntay, Messer, Verhagen and Leseman2014), along with the lack of possibility to fluently switch between languages, do not make it likely to produce long-term effects on these subcomponents of executive control – inhibition and switching. In other words, the absence of the need to actively inhibit another language and the lack of active practice in switching between languages means that substantial changes in inhibition (understood as control of internal language competence) and cognitive flexibility are not expected in monolinguals. In contrast, for the updating component of executive control, we might anticipate similar performance between monolinguals and bilinguals in their working memory capabilities when immersed in a dual-language context. This is because monolinguals, akin to bilinguals (Costa et al., Reference Costa, Hernández, Costa-Faidella and Sebastián-Gallés2009; Liu et al., Reference Liu, Li, Zuo, Wang, Guo and Schwieter2022), would have to constantly monitor potentially ambiguous linguistic situations (i.e., conflicting linguistic cues) to appropriately adapt their conduct in such shared environment. That is, similarly to how bilinguals in a dual-language context develop a greater monitoring capacity to manage the coexistence of the two languages, monolinguals immersed in such an environment may potentially develop a greater sensitivity to linguistic cues and a better ability to monitor them. In this context, studies have shown that online language comprehension is closely associated to the individual’s ability to update information in working memory (Kim et al., Reference Kim, Oines and Miyake2018; Meyer, Reference Meyer2018). Moreover, conflict monitoring adjustments have been related to the working memory capacity of the updating abilities, a core component of executive control (Engle & Kane, Reference Engle and Kane2004; Weldon et al., Reference Weldon, Mushlin, Kim and Sohn2013).
Thus, we hypothesized bilinguals to outperform monolinguals in those measures tapping into inhibition (i.e., Stroop task) and switching (i.e., Trail Making Test), but not in those tasks that tap into updating (i.e., Digit span) for the reasons mentioned earlier (Hypothesis 1). Following this line of inquiry, the second research question focuses on quantifying the relationship between specific facets of bilingual language experience with the various aspects of executive control in both bilinguals and monolinguals.
Research question 2: To what extent are different measures of bilingualism associated with different subcomponents of executive control among both bilingual and monolingual individuals?
Other context-related components may also influence the executive control of monolinguals in a dual-language context, namely, the self-reported proficiency and percentage of second language (L2) exposure and/or use. In this vein, research has shown that higher self-reported L2 proficiency can lead to greater cognitive advantages (Giovannoli et al., Reference Giovannoli, Martella, Federico, Pirchio and Casagrande2020). Furthermore, the daily percentage of L2 use, which reflects the frequency of using the L2 in daily life, has been linked to variations in executive functioning. Individuals who use an L2 more frequently tend to exhibit greater cognitive benefits in terms of executive control, such as better task-switching abilities and improved attentional control (Gullifer et al., Reference Gullifer, Chai, Whitford, Pivneva, Baum, Klein and Titone2018). These findings underscore the dynamic relations between language proficiency and executive control, highlighting the potential cognitive advantages associated with bilingualism and the extent of L2 use in monolingual individuals residing in dual-language contexts. Thus, we expected the objectively tested degree of bilingualism (i.e., bilingualism score), the self-reported proficiency in L2 and the daily percentage of L2 use of both monolinguals and bilinguals to positively correlate with the subcomponents of executive control (Hypothesis 2). However, while for bilinguals these correlations were expected for all subcomponents (albeit we expect them to be more evident for the inhibition and switching outcomes), for monolinguals we hypothesized a direct association between the degree of bilingualism and the inhibition and switching mechanisms, but not with updating (Hypothesis 3).
Given the critical need for methodological precision in bilingualism research, underscored by previous review studies (Dash et al., Reference Dash, Joanette and Ansaldo2022b; Grundy, Reference Grundy2020), the current study implements stringent methodological standards to bridge gaps in the existing literature. In contrast to several studies in which bilinguals and monolinguals are either immigrants, recruited from various countries or regions within the same country (Bialystok et al., Reference Bialystok, Klein, Craik and Viswanathan2004, Reference Bialystok, Barac, Blaye and Poulin-Dubois2010; Coderre & van Heuven, Reference Coderre and van Heuven2014; de Abreu et al., Reference de Abreu, Cruz-Santos, Tourinho, Martin and Bialystok2012; Hernández et al., Reference Hernández, Martin, Barceló and Costa2013; Kalia et al., Reference Kalia, Wilbourn and Ghio2014; Paap & Greenberg, Reference Paap and Greenberg2013; Sanchez-Azanza et al., Reference Sanchez-Azanza, López-Penadés and Adrover-Roig2020; Zhou & Krott, Reference Zhou and Krott2016), all the participants in the current study were from the same geographical area and none of them were immigrants from other countries. By using this approach, we intended to ensure that the situation and the immersion of the participants in the same background were as similar as possible in our sample to draw better conclusions about the impact of bilingualism (de Bruin et al., Reference de Bruin, Bak and Della Sala2015). Hence, monolingual participants were also immersed in the dual-language context, even though they did not consider themselves as bilinguals and their L2 proficiency and usage significantly differed from that of the bilingual participants. This method enables the control of potentially confounding sociocultural and demographic factors and ensures a more homogeneous study group based on interactional context, thereby facilitating a clearer understanding of the impact of bilingualism. Another contributing factor to inconsistencies in the bilingual advantage hypothesis is the variation in cognitive tasks used across studies, making direct comparisons between studies challenging due to the fundamentally different nature of the cognitive tasks under study. Seldom have multiple cognitive tasks been employed in the same group within a single study (Bak et al., Reference Bak, Nissan, Allerhand and Deary2014, 2022a). To address this issue, the current study had the same participants perform three distinct cognitive tasks, each designed to measure the impact of bilingualism on different subcomponents of executive control, thereby ensuring a more accurate and comprehensive assessment.
2. Method
2.1. Participants
A sample of 56 healthy young adult participants from Montréal (Canada) took part in this study. Our sample size estimation was informed by foundational studies that underpin the adaptive control hypothesis, including Bialystok et al. (Reference Bialystok, Craik, Green and Gollan2009), Bialystok and DePape (Reference Bialystok and DePape2009) and Abutalebi et al. (Reference Abutalebi, Della Rosa, Green, Hernandez, Scifo, Keim, Cappa and Costa2012). These studies reported mostly large effect sizes in executive function task comparisons between monolinguals and bilinguals. Based on these precedents, which report effect sizes in the medium to large range (Cohen’s d = 0.5–0.8), and assuming a significance level of .05, power of 0.80 and an expected large effect size, the required sample size was estimated at 29 participants per group. Our final sample from Montréal closely matches this requirement.
Montréal is a diverse city where French is the official language; however, English is widely used and understood throughout. In Montréal, monolingualism, whether in English or French, is less common among the youth. Among workers in the Montréal census metropolitan area, 80% are bilingual and 28% of them have achieved trilingual proficiency. Bilinguals from Montréal use both languages independent of their location, especially in informal contexts, such as on the street and in coffee houses (Lamarre et al., Reference Lamarre, Paquette, Kahn and Ambrosi2002), and they code switch frequently (Kellert, Reference Kellert2023). Therefore, Montréal is considered a highly bilingual city (Gullifer & Titone, Reference Gullifer and Titone2020), standing, in line with the adaptive control hypothesis, as a prime example of a dual-language (or a dense code switching) context where interactions unfold seamlessly in both English and French.
Although Montréal is predominantly bilingual, monolingualism is still present, particularly among certain youth populations who, despite being passively exposed to multiple languages, report minimal active use or proficiency beyond their dominant language. Within this context, their presence can be understood in light of sociocultural and educational factors. Previous research has shown that attitudes toward bilingualism, neighborhood composition and patterns of family language use can contribute to reduced engagement in bilingual practices (Lamarre, Reference Lamarre2013). Moreover, differences in educational opportunities and personal motivation to study and use a second language also play a significant role in shaping individual linguistic trajectories (Mady, Reference Mady2010). These factors help explain why a subset of individuals continue to identify and function as monolinguals despite living in a predominantly bilingual environment. Here, we define monolingualism as Ellis (Reference Ellis2008) proposed: a monolingual individual is one who does not have access to more than one linguistic code as a means of proficient social communication. Nevertheless, even under this definition, note that the languages’ proficiency of monolinguals lies on a continuum, as does that of bilinguals. Thus, to better capture this reality, in this study, we refer to these individuals as functional monolinguals, emphasizing that while they may have some degree of passive exposure or limited knowledge of an L2, their linguistic repertoire is effectively restricted to a single language for social communication (Bice & Kroll, Reference Bice and Kroll2019). In the present sample, half of the participants were self-rated functional monolinguals (18 females; 22.2 ± 3 years old; 15.2 ± 2 years of education), while the remaining 28 participants were bilinguals (15 females; 22.8 ± 3 years old; 16.4 ± 2 years of education). As Table 1 shows, both groups were similar in terms of age, and bilinguals received slightly more formal education.
Table 1. Sample descriptive statistics
Note. L1: first language; L2: second language; BAT, Bilingual Aphasia Test (part C).
a Scale from 0 (no knowledge) to 30 (perfect knowledge).
b This measure was only collected from 18 (of 28) participants in each group.
c Scale from 0 (no knowledge) to 42 (perfect knowledge).
2.2. Linguistic background
All participants were administered both a subjective and an objective performance-based language background questionnaire.
2.2.1. Self-reported language assessment
The Language Experience and Proficiency Questionnaire (LEAP-Q; Marian et al., Reference Marian, Blumenfeld and Kaushanskaya2007) was administered to assess both the first (L1) and second language (L2) proficiencies and the percentage of daily language use for all participants. Participants were asked to subjectively report their proficiency in speaking, understanding, and writing in both their L1 and L2, using scales from 0 (no knowledge) to 10 (perfect knowledge). Bilingual participants also reported their age of L2 acquisition. As expected, monolinguals rated their L2 proficiency below the cut-off of 7 points in each of the 0–10 self-reported scales (see Table 1), a threshold previously used by Marian et al. (Reference Marian, Blumenfeld and Kaushanskaya2007) to distinguish individuals with limited second language proficiency from more proficient bilinguals. Monolinguals and bilinguals considered themselves as such, and all bilinguals acquired their L2 early. All participants estimated the percentage of daily time they used each of the languages, which amounted to <3% in the monolingual group and over 42% in the bilingual group. Both monolinguals and bilinguals reported either French (35% and 50%, respectively) or English (65% and 50%, respectively) as their first language (L1), and bilinguals spoke the other language (i.e., French or English) as their second language (L2). Preliminary t tests showed no significant differences in executive control scores (e.g., Stroop, TMT, digit span) between French-L1 and English-L1 participants, whether monolingual (ps ≥ .081) or bilingual (ps ≥ .062), indicating that first language did not affect cognitive task performance.
2.2.2. Objective language assessment
Given that all participants lived in Montréal (i.e., a dual-language interactional context), they all had some degree of knowledge of both languages, even though monolingual participants self-identified as monolinguals. Thus, we also objectively tested the proficiency in L2 via the Part C of the Bilingual Aphasia Test (BAT; Paradis & Libben, Reference Paradis and Libben2014), which is intended to assess the languages of bilingual or multilingual individuals using a culturally and linguistically equivalent test adapted to each culture, and to quantify language proficiency among neurologically healthy speakers. It consisted of three parts: in the first part, the participant was asked to choose from a list the translated words from their L2 into their L1 (5 words: 5 points awarded) and from their L1 into their L2 (5 words: maximum 5 points awarded). The second part was similar to the first one, but this time the participant had to produce the equivalent translation without alternatives (10 words from L1 into L2, and vice versa: 20 points awarded). The final part consisted of sentences with different grammatical structures that the participant was asked to translate into English and into French (6 sentences from L1 into L2, and vice versa: 12 points awarded), respectively. Participants were given one point for each correct answer, and the total objectively assessed level of bilingualism (which we will refer to as “bilingualism score”) ranged from 0 to 42. As can be consulted in Table 1, bilinguals and monolinguals were equivalent in terms of L1 proficiency (regardless of their L1), and the bilingual group showed a much higher proficiency in L2, as compared to the monolingual group. The bilingualism score, calculated for both monolingual and bilingual groups, evaluates the impact of a dual-language context by measuring translation skills between languages, thereby assessing the influence of bilingualism across both groups.
All participants were right-handed, as shown by the Edinburgh Handedness Inventory (Oldfield, Reference Oldfield1971) scores. No participant reported psychiatric or pharmacological treatment at the time of testing, and none presented noncorrected visual or auditory deficits. All participants provided written informed consent. This study was approved by the ethics committee of the Centre de Recherche de l’Institut Universitaire de Gériatrie de Montréal.
2.3. Executive control tasks
The selection of tasks was guided by theoretical, empirical and practical considerations, with the aim of capturing distinct components of executive functioning while ensuring ecological validity and feasibility of administration. Specifically, these tasks are widely used in clinical and neuropsychological settings and in bilingualism research, facilitating comparison with previous studies (e.g., Gold et al., Reference Gold, Kim, Johnson, Kryscio and Smith2013; Salvatierra & Rosselli, Reference Salvatierra and Rosselli2011). The selected tasks also allow for the examination of domain-specific effects of bilingual experience, as proposed by frameworks emphasizing the multifaceted nature of executive control (Miyake et al., Reference Miyake and Friedman2012; Bialystok, Reference Bialystok2015). Furthermore, we included both verbal and nonverbal tasks to account for potential modality effects, as some studies suggest that bilingual advantages may differ depending on the verbal load of the task (Bialystok et al., Reference Bialystok, Craik and Luk2008; Paap et al., Reference Paap and Greenberg2013). Finally, their psychometric robustness and established normative data make them suitable for diverse populations (Strauss et al., Reference Strauss, Sherman and Spreen2006).
2.3.1. Inhibition
The Stroop test (Stroop, Reference Stroop1935) was administered to evaluate interference control. This task consisted of three parts, two of which intend to be different baselines of the third conflicting condition. The three parts were administered in the participants’ L1. In the first part (Color), participants were asked to name the colors (red, green and blue) of the pseudo-randomly 24 dots presented in the sheet as quickly as possible. If the participant finished all the columns before 45 seconds, the participant returned to the first column and started again. Upon commission of an error, the test leader said “no” signaling a mistake, and the participant was to correct the error and to continue without stopping. In the second part (Word), the participant was asked to name aloud the colors in which common words (e.g., “over,” “hard” or “when”) were written. The layout of the page was the same as in the first part. The third part (Color–Word) was similar to the second, except for the fact that instead of common words, the words were color names and participants were requested to name the color of the ink (red, green and blue), where the ink color was always incongruent with the color word that the participant should try to ignore (e.g., blue, red, green). The dependent variables were the number of items read by the subject during a 45-second interval and the number of errors for each part. Additionally, the Stroop effect (Scarpina & Tagini, Reference Scarpina and Tagini2017) was calculated as the difference between the Color and the Color–Word conditions, for both the number of words uttered in 45 seconds and the number of errors. Note that higher scores on this derived measure indicate worse inhibition.
2.3.2. Switching
The Trail making test (TMT) parts A and B (Reitan, Reference Reitan1958) were used to measure, respectively, processing speed and both switching and updating (Sánchez-Cubillo et al., Reference Sánchez-Cubillo, Periáñez, Adrover-Roig, Rodríguez-Sánchez, Ríos-Lago, Tirapu and Barceló2009). In the TMT-A, participants were instructed to connect 25 numbered circles (1–25) in ascending order. TMT- B included both numbers (1–13) and letters (A–L), and the participant was asked to connect the circles in ascending order alternating between the numbers and the letters (e.g., 1-A-2-B-3-C, etc.). Participants were encouraged to connect the circles as fast as possible without lifting the pen from the paper. Errors were pointed out immediately to allow for correction. The dependent variables collected were the number of seconds required to complete the task for both parts (A and B), as well as the number of errors. Additionally, the derived score B–A was calculated and used as dependent variables. This derived measure seems to provide a purer indicator of task-switching ability than TMT-B, since it is relatively absent of the influence of processing speed and updating (Sánchez-Cubillo et al., Reference Sánchez-Cubillo, Periáñez, Adrover-Roig, Rodríguez-Sánchez, Ríos-Lago, Tirapu and Barceló2009). Note that higher scores on this derived measure indicate worse switching.
2.3.3. Updating
Updating capacity was assessed by using the Digit span test from the Wechsler Adult Intelligence Scale (WAIS-III; Wechsler, Reference Wechsler1997). In this task, increasingly longer strings of numbers were read to the participant in their L1. Participants were asked to repeat the numbers in the same order (Forward) or in the reverse order (Backward). There were two sets of digits for each string length. Both the total number of correctly remembered and the span score were collected as dependent variables for both conditions’ sequences. While the Forward condition would be more related to short-term memory capacity, monitoring and goal maintenance, the Backward condition seems to reflect the manipulation of information in working memory (Unsworth & Engle, Reference Unsworth and Engle2006). Note that higher scores on all these measures indicate better updating.
2.4. Data analysis
To address Research Question 1 (Executive control comparison between bilinguals and monolinguals), a series of analyses of covariance (ANCOVAs) were performed to compare the performance between both bilinguals and monolinguals across tasks and conditions while statistically controlling for the effect of the distinct number of years of Education (i.e., the covariate) among groups (Hypothesis 1). Note that the Stroop task analyses were conducted with 54 participants (instead of the 56 of the other tasks), due to the loss of data regarding two bilingual participants. In any case, half of the sample was composed of monolinguals and half of bilinguals. Prior to conducting the ANCOVAs, assumptions were verified. In particular, residuals were approximately normally distributed for each category of the independent variable, as visually confirmed through Q-Q plots. Moreover, Levene’s tests indicated homogeneity of variances for all executive control measures (Stroop, ps > .800; Digit span, ps > .663; TMT, ps > .103).
However, since the reliability of this procedure has been questioned regarding its use to control for the disparity of pre-existing group differences not attributable to the linguistic background (Paap et al., Reference Paap, Johnson and Sawi2014), we additionally tested the participants’ performance on executive tasks by means of a subsample. In this vein, we matched participants from both groups who reported the same years of education, and we obtained a subset of the original sample composed of 34 participants (18 participants for the monolinguals and 16 for the bilinguals). Both groups showed similar age, t(34) = 0.25, p = .804, and years of education, t(34) = 0, p = 1, in the subset. Moreover, when the dependent variables between the original and the subset samples for both bilinguals and functional monolinguals were compared, none of the specific scores (including all executive control tasks and conditions) reached significance (all ts < 1, all ps > .356).
Additionally, we aimed to further understand the potential relation between specific linguistic variables and executive control measures within each group, Research Question 2 (Correlation patterns between linguistic variables and executive control indices). Thus, the objective bilingualism score, the self-reported level of proficiency and the percentage of L2 daily use of both groups were separately correlated with each relevant condition or derived score of every task to explore the correlation patterns (Hypothesis 2). Finally, we tested whether the patterns of correlations among groups were different for these correlations using the Fisher’s Z test (Hypothesis 3). Finally, correlations of all relevant measures for all participants were calculated. Given that our correlational analyses were theoretically driven and prespecified based on established research questions, we did not apply corrections for multiple comparisons (e.g., Bonferroni or false discovery rate). This approach aligns with methodological recommendations cautioning against routine multiplicity adjustments in hypothesis-led studies, as such corrections can inflate Type II error rates and reduce statistical power, potentially masking meaningful effects (Hooper, Reference Hooper2025; Rothman, Reference Rothman1990). To preserve transparency, uncorrected p values are reported with appropriate caveats, and findings must be interpreted requiring cautious consideration.
Statistical analyses were conducted using JASP (Version 0.18; JASP Team, 2024). A significance level of p < .05 was used for all analyses.
3. Results
The descriptive statistics of the main tasks for the original sample can be consulted in Table 2.
Table 2. Descriptive statistics of the main study variables for the whole sample in each group
3.1. Research Question 1: Executive control comparison between bilinguals and functional monolinguals
3.1.1. Inhibition
Regarding in the Stroop task (see Figure 1A), the ANCOVA results showed that bilinguals and monolinguals’ performance was similar for each condition concerning the number of items correctly uttered in 45 seconds, Word: F(1, 50) = 1.17, p = .284, η2 = .022 (Group × Education: F < 1), Color: F(1, 50) = 0.22, p = .638, η2 = .004 (Group × Education: F < 1), Color–Word: F(1, 50) = 1.01, p = .319, η2 = .018 (Group × Education: F < 1). Moreover, the Stroop effect was not significant, F(1, 50) = 0.14, p = .709, η2 = .003 (Group × Education: F < 1).
Complementarily, t-test results on the matched subset revealed the same results for the number of utterances in 45 seconds, Word: t(32) = −1.37, p = .180, Color: t(32) = −0.24, p = .815, Color–Word: t(32) = −1.21, p = .234. Furthermore, the Stroop effect remained nonsignificant, t(32) = 0.85, p = .403. Since the number of errors was negligible (on average, less than one error for each condition), we did not examine this outcome (see Table 2).
3.1.2. Switching
Concerning the TMT (see Figure 1B), the ANCOVA results showed that the performance of both language groups was similar for each condition in both the time, A: F(1, 52) < 1, p = .616, η2 = .005 (Group × Education: F(1, 52) = 3.3, p = .075, η2 = .054), B: F(1, 52) = 1.69, p = .199, η2 = .031 (Group × Education: <1), B–A: F(1, 52) = 2.7, p = .106, η2 = .048 (Group × Education: F < 1) and the accuracy, A: F(1, 52) < 1, p = .636, η2 = .004 (Group × Education: F(1, 52) = 1.6, p = .207, η2 = .03), B: F(1, 52) < 1, p = .719, η2 = .002 (Group × Education: F(1, 52) = 2.3, p = .131, η2 = .042).
Complementarily, t-test results on the matched subset revealed the same results for both the time, A: t(34) < −1, p = .713, B: t(34) < 1, p = .335, B–A: t(34) = 1.31, p = .198 and the accuracy, A: t(34) < 1, p = .560, B: t(34) < 1, p = .411.
3.1.3. Updating
As for the Digits task (see Figure 1C), the ANCOVA results showed that the performance of the bilinguals and the monolinguals was similar for both the total number of correct sequences, Forward: F(1, 52) = 0.35, p = .559, η2 = .006 (Group × Education: F < 1), Backward: F(1, 53) = 3.03, p = .088, η2 = .054 (Group × Education: F(1, 52) = 2, p = .166, η2 = .035), and the total span, Forward: F(1, 52) = 0.42, p = .520, η2 = .008 (Group × Education: F < 1), Backward: F(1, 52) = 3.17, p = .081, η2 = .056 (Group × Education: F < 1).
Complementarily, t-test results on the matched subset revealed the same results for both the total number of correct sequences, Forward: t(34) = 0.08, p = .941, Backward: t(34) = −1.88, p = .069, and the span, Forward: t(34) = 0, p = 1, Backward: t(34) = −1.55, p = .131.
3.2. Research Question 2: Correlation patterns between linguistic variables and executive control indices
3.2.1. Group comparisons
Since the number of errors committed by participants, on average, was negligible (see Table 2), we decided to disregard these variables and focus only on measures with more variability to conduct the correlation analyses. As can be seen in Table 3 (bilingualism score and self-reported proficiency in L2) and Supplementary Material 1 (daily percentage of L2 use), the correlational patterns between the linguistic variables (the objective bilingualism score and self-assessed proficiency and use scores) and executive control measures (mainly inhibition and updating of information) were distinct between the two groups.
Table 3. Group patterns of correlation and comparison (Fisher’s Z) between the objectively evaluated bilingualism (BAT), the self-reported proficiency in L2 and the most relevant executive control measures of the study
Note: BAT, Bilingual Aphasia Test (part C).
a The data of all 54 participants were available.
b Only 36 participants’ data were available due to data loss TMT: Trail Making Test.
*Asterisk indicates significant relations (p < .05) for reading easiness. Bold depitcs a significant difference in correlation patterns between groups for reading easiness.
All Stroop and TMT measures are based on time (RT), and the Digits variables display the span measure.
In particular, while lower bilingualism scores and a less frequent percentage of L2 daily use in functional monolinguals was associated to a better inhibitory ability (note that higher scores indicate better performance on this task condition; see Figure 2), in bilinguals the reverse pattern is observed (Table 3 and Supplementary Material S1). Similarly, a lower self-reported proficiency in L2 in monolinguals was related to a better ability to maintain their objective in memory and to monitor for conflict, opposite to bilinguals (Table 3).
Furthermore, it is noteworthy that both groups differed in the way the objective bilingualism score is related with their inhibition ability. As previously stated, in the case of monolinguals, a lower degree of bilingualism was associated with their higher task-specific inhibition. In contrast, the opposite pattern was found for bilinguals, whose inhibition ability seems to be boosted more broadly, and in relation with other more general executive mechanisms (but not inhibition specific) by their bilingualism score. No other correlation pattern differed between groups.
3.2.2. Whole sample
On a separate note, and to supplement the section and the shared body of literature, we have decided to provide the overall patterns of correlations between the main variables in this study in the Supplementary Material 2. Ideally, this information will be useful in the future to extract reliable conclusions about the task specificity of the bilingual advantage, which, besides the advances (Ware et al., Reference Ware, Kirkovski and Lum2020), seems to be in question (Paap & Greenberg, Reference Paap and Greenberg2013).
In this vein, the correlation pattern calculated for all participants shows a picture in which several of the measures of the separate tasks are correlated between each other (see Supplementary Material S2). Nevertheless, it must be noted that the most relevant indices of the inhibition and switching tasks seem to correlate with each other (i.e., both Stroop Color–Word and Stroop effect with both TMT B and TMT B–A; see Supplementary Material S2), yet none is related with the updating markers.
Interestingly, when examining the various language measures across all participants altogether, they display correlations exclusively with the span of the Digits Backward task. Specifically, higher scores in the objective bilingualism score, L2 proficiency and the percentage of daily L2 use are associated with better updating abilities in general. However, other measures were close to be correlated with the linguistic characteristics of the participants (ps < .1; see Supplementary Material S2).
In sum, monolinguals with lower levels of bilingualism and L2 use exhibited better inhibition and monitoring abilities, whereas bilinguals showed the opposite pattern. When analyzing all participants together, language measures correlated exclusively with updating ability (Digits Backward), indicating a relation between overall bilingualism and working memory updating.
4. Discussion
The current study seeks to determine if monolinguals immersed in a dual-language context experience context-specific benefits in executive control (what has been widely known as the bilingual advantage for bilingual individuals) when compared to bilinguals living in the same environment, guided by the framework of the adaptive control hypothesis (Green & Abutalebi, Reference Green and Abutalebi2013). This hypothesis originates from the assumption that both monolinguals and bilinguals in Montréal are regularly exposed to two languages, characterized by dense code switching and high language entropy (Gullifer & Titone, Reference Gullifer and Titone2020). In this regard, we recognize that in a bilingual context such as Montréal, it is unusual for there to be no exposure to L2, which is why we have classified the group of people who self-identified as monolinguals as functional monolinguals. Overall, our goal is to challenge strict dichotomies in bilingualism research and highlight the importance of taking a more nuanced approach to classifying language users when studying the bilingual advantage (Bice & Kroll, Reference Bice and Kroll2019; Surrain & Luk, Reference Surrain and Luk2019).
The main finding of the current study is the absence of a bilingual advantage when comparing bilinguals to monolinguals immersed in a dense code-switching interactional context, across all subcomponents of executive control – inhibition, switching and updating. Second, we observed that monolinguals and bilinguals exhibit distinctive patterns of association between measures of bilingualism and subcomponents of executive control, specifically inhibition and updating. These findings are discussed in detail next.
4.1. Research Question 1: Executive control comparison between bilinguals and functional monolinguals
Previous works have reported that bilinguals immersed in dual-language contexts show smaller switch costs and better conflict resolution abilities than bilinguals immersed in single-language contexts (Hartanto & Yang, Reference Hartanto and Yang2020; Ooi et al., Reference Ooi, Goh, Sorace and Bak2018). Based on the association between greater language entropy and better task-switching performance (van den Berg et al., Reference van den Berg, Brouwer, Tienkamp, Verhagen and Keijzer2022), we anticipated that young bilinguals from a dense code-switching environment and a high between-languages entropy (French–English) would outperform their monolingual counterparts in aspects of executive control, particularly inhibition and task switching, but not in information updating. Along these lines, monolinguals are not sufficiently proficient in the L2 to undergo a substantial cross-linguistic activation or to shift between languages (Blom et al., Reference Blom, Küntay, Messer, Verhagen and Leseman2014). Despite this, they must frequently monitor for relevant linguistic information (i.e., being spoken in the language they master) in dense code-switching contexts. Hence, we anticipated that monolinguals would develop updating abilities to a level comparable to that of bilinguals (Costa et al., Reference Costa, Hernández, Costa-Faidella and Sebastián-Gallés2009; Kim et al., Reference Kim, Oines and Miyake2018; Weldon et al., Reference Weldon, Mushlin, Kim and Sohn2013). However, they would not practice additional inhibition- or switching related abilities as the latter, at least not to the degree in which the demands placed over the linguistic and executive control mechanisms are expected to be noticeable. The results of this study contradict the initial expectations, showing that both groups performed similarly across all executive control tasks, as illustrated in Figure 1. This similarity in performance was observed both in the original sample, with years of education considered as a covariate, and in the comparison of education and age-matched functional monolingual and bilingual groups. Although these findings are against our main hypotheses (regarding inhibition and switching), it is in line with the conclusions of several current studies that call into question the bilingual advantage (Antoniou, Reference Antoniou2019; Hilchey et al., Reference Hilchey, Saint-Aubin and Klein2015; Lehtonen et al., Reference Lehtonen, Soveri, Laine, Järvenpää, de Bruin and Antfolk2018; Sanchez-Azanza et al., Reference Sanchez-Azanza, López-Penadés and Adrover-Roig2020).
Figure 1. Graphical results of the three different tasks. (A) Stroop task: number of items uttered in 45 seconds for each condition and the Stroop effect. (B) TMT: direct scores (A and B) and the derived B–A score, scale values represent time in seconds. (C) Digits: Forward and Backward, including both the direct and span scores.
Figure 2. Scatterplots representing the associations between the bilingualism score (Bilingual Aphasia Test, part C) and both the (A) Stroop Color–Word and (B) Stroop Effect measures with the complete sample of bilinguals and monolinguals.
Specifically, bilinguals and functional monolinguals performed similarly in each condition in terms of the inhibition task. These results concur with previous research, which also found no differences between these groups in terms of their performance on the latency of inhibition measures, such as the Stroop tasks (Antón et al., Reference Antón, Carreiras and Duñabeitia2019; Blumenfeld & Marian, Reference Blumenfeld and Marian2014; Duñabeitia et al., Reference Duñabeitia, Hernández, Antón, Macizo, Estévez, Fuentes and Carreiras2014). However, it is important to mention that some studies have reported a bilingual advantage in the performance of the Stroop paradigm, with better inhibition skills in bilinguals compared to monolinguals (Kousaie & Phillips, Reference Kousaie and Phillips2017; Prior & Gollan, Reference Prior and Gollan2011). In terms of differences in task switching or alternating attention, the present results showed that the performance of both groups was similar in each condition in terms of time and accuracy in the TMT-derived score. Again, these results are consistent with previous studies that have found no significant differences in task-switching ability between bilinguals and monolinguals when using the TMT (Lehtonen et al., Reference Lehtonen, Soveri, Laine, Järvenpää, de Bruin and Antfolk2018; but cf. Ware et al., Reference Ware, Kirkovski and Lum2020 for results in favor of, and Massa et al., Reference Massa, Köpke and El Yagoubi2020 against the bilingual advantage). Regarding the results on updating, again, the performance of bilinguals and functional monolinguals was similar both for the total number of correct sequences and for the total reach in the forward and backward parts of the digits task. Although some studies have reported a possible bilingual advantage in working memory (Bialystok et al., Reference Bialystok, Craik and Luk2008; Morales et al., Reference Morales, Calvo and Bialystok2013), our findings are consistent with previous research that has found similarities in working memory updating abilities between bilinguals and monolinguals (Sanchez-Azanza et al., Reference Sanchez-Azanza, López-Penadés and Adrover-Roig2020).
Although differences between monolinguals and bilinguals are not always apparent, they have been related to factors associated with the bilingual experience, such as the type of immersion context and the degree of exposure to the L2, the proficiency in L2, the linguistic similarity between both languages, the cognitive load associated with the constant management of parallel language activation and competition and the sociocultural context in which the bilingual experience takes place (Chung-Fat-Yim et al., Reference Chung-Fat-Yim, Poarch, Comishen and Bialystok2021; Green & Abutalebi, Reference Green and Abutalebi2013; Hartanto & Yang, Reference Hartanto and Yang2016; Incera & McLennan, Reference Incera and McLennan2018; Jiao et al., Reference Jiao, Grundy, Liu and Chen2020; Ooi et al., Reference Ooi, Goh, Sorace and Bak2018; Sanchez-Azanza et al., Reference Sanchez-Azanza, López-Penadés, Aguilar-Mediavilla and Adrover-Roig2020; Schwartz & Kroll, Reference Schwartz and Kroll2006). In this study, it is possible that, by sharing the same dense code-switching environment with a large language entropy, the performance between bilinguals and functional monolinguals in executive control has been levelled. Thus, we speculate that monolinguals living in such a culturally rich and linguistically complex environment where languages are switched on a very regular basis may have accumulated some cognitive advantages even without being bilingual. Future research could delve into a comparative analysis of monolingual individuals, as well as bilinguals, hailing from diverse linguistic contexts to provide solid support for our findings, shedding light on the interplay between language use and executive control.
4.2. Research Question 2: Correlation patterns between linguistic variables and executive control indices
Based on previous research, we expected that the objective degree of bilingualism (bilingualism score from BAT), the subjective self-reported L2 proficiency and usage, to correlate with subcomponents of executive control. These correlations were anticipated in both groups for inhibition and switching, and exclusively in the bilingual group for updating. However, our hypotheses 2 and 3 were not completely satisfied, as the majority of the expected outcomes were not verified. Despite these unmet hypotheses, notable findings from our study show that functional monolinguals and bilinguals present distinctive associations between both their objective degree of bilingualism, the self-reported L2 proficiency, the daily percentage of L2 usage and their inhibitory capacity. Furthermore, an opposite pattern was found for functional monolinguals and bilinguals in terms of the association between their self-reported proficiency in L2 and their short-term memory. Thus, the observed correlational patterns suggest that the ability to monitor and suppress irrelevant information is of a more general nature in functional monolinguals, and in the case of bilinguals it appears to be task specific. Interestingly, functional monolinguals showed that the lower their bilingualism score, the lower their self-reported proficiency in L2 and the lower their daily percentage of L2 use, the better their capacity for inhibition and selective attention to relevant information (Color–Word condition of the Stroop test). These correlations were not significant for bilinguals, which explains the clear difference in association patterns between linguistic and cognitive executive tasks, as revealed by the Fisher tests. Building on this, Linck et al. (Reference Linck, Hoshino and Kroll2008) showed in their Study 2 that the ability to suppress information between bilinguals immersed in the context of their L1 and those immersed in the context of their L2 was similar. These results suggest that immersion in a specific language context does not necessarily produce an improvement in inhibitory control. In the case of this study, and as compared to Linck et al. (Reference Linck, Hoshino and Kroll2008), we speculate that functional monolinguals and bilinguals may use their inhibition related cognitive resources differently, hence displaying a different correlation pattern, despite sharing the same linguistic context. For example, functional monolinguals may focus on suppressing the influence of other languages, whereas bilinguals may need to control the activation of both languages in a more balanced way. This difference in the functional allocation of cognitive resources could explain the different relations between bilingualism and inhibitory control observed in the two groups.
We also found that the lower the measures of bilingualism (bilingualism score, the self-reported L2 proficiency and daily usage; Table 3 and Supplementary Material 1) in functional monolinguals, the better their ability to maintain objectives in memory and to monitor conflict (Digits forward). This pattern of execution by functional monolinguals, in which the ability to update might moderate the inhibition of irrelevant information, is in line with previous literature (Duell et al., Reference Duell, Icenogle, Silva, Chein, Steinberg, Banich, Di Guinta, Dodge, Fanti, Lansford, Oburu, Pastorelli, Skinner, Sorbring, Tapanya, Uribe Tirado, Alampay, Al-Hassan, Takash and Chaudhary2018). On the other hand, a clear and substantial relation was found in the case of the bilingual group between the bilingualism score (Table 3) and the Stroop effect: the higher the bilingualism score, the better the inhibition capacity of bilinguals in this task (note that lower Stroop effect indicate a smaller difference between conditions, which implies better overall processing). That is, it seems that bilinguals (unlike monolinguals) have fewer differences in efficiency between automatic (reading words) and controlled (naming color words) information processing. In summary, the different correlation patterns seem to show that, in a dual-language context, monolinguals with a low knowledge and daily use of an L2 would show performances resembling the Bilingual Executive Processing Advantage (BEPA), as this group shows a paradoxical association between language skills and updating, which plays an important role in general cognitive functioning, as it is an underlying process that bridges working memory, attention and executive control. While bilinguals seem more inclined to a processing style more associated with the Bilingual Inhibitory Control Advantage (BICA) shown by differences in Stroop effect suggesting a more efficient processing style with less interference from irrelevant information (Hannaway et al., Reference Hannaway, Opitz and Sauseng2019; Hilchey & Klein, Reference Hilchey and Klein2011).
Furthermore, previous literature generally concurs that a higher degree of bilingualism is usually associated with the ability to inhibit irrelevant information (Carlson & Meltzoff, Reference Carlson and Meltzoff2008; Incera & McLennan, Reference Incera and McLennan2018; Sabourin & Vinerte, Reference Sabourin and Vinerte2015; Tse & Altarriba, Reference Tse and Altarriba2012), and with updating contextual information (as shown by the correlation results including all participants; Antón et al., Reference Antón, Carreiras and Duñabeitia2019; Monnier et al., Reference Monnier, Boiché, Armandon, Baudoin and Bellocchi2022). However, other studies have found that neither the inhibitory control nor the updating abilities of bilinguals are always associated with the degree of bilingualism (Kałamała et al., Reference Kałamała, Szewczyk, Chuderski, Senderecka and Wodniecka2020; Lehtonen et al., Reference Lehtonen, Soveri, Laine, Järvenpää, de Bruin and Antfolk2018). In general, the different pattern of correlations of bilinguals in this study seems to go in the same direction as most previous literature, indicating that the degree of bilingualism favors better overall processing of conflicting information.
However, to our knowledge, the present results in the case of monolinguals and the association between the bilingualism score, the self-reported proficiency in L2 and daily percentage of L2 use and inhibition are novel. It is possible that the distinct correlation patterns are a result of functional monolinguals having a stable profile, as they lack the need to switch between languages in dual-language contexts (in terms of the stability–flexibility dilemma, Dreisbach & Fröber, Reference Dreisbach and Fröber2019). In brief, this dilemma states that two opposing strategies compete to adjust goal-directed behavior: stability, in which cognitive processes that maintain the current task representation avoiding irrelevant intrusions (i.e., inhibition); and flexibility, in which cognitive processes allow to shift efficiently to another task representation (i.e., switching). The balance between these strategies is dependent of both the individuals and the context, and it can be biased or directed toward stability or flexibility, thus triggering a trade-off. Additionally, particular features of a given environment, such as a dual-language context, might contribute to the persistence of a bias or style favoring either stability or flexibility (Hommel & Colzato, Reference Hommel and Colzato2017). In this sense, Sanchez-Azanza et al. (Reference Sanchez-Azanza, López-Penadés, Aguilar-Mediavilla and Adrover-Roig2020) found, using latent variables, that bilinguals immersed in dual-language contexts seemed to be more inclined to flexibility mode than to stability mode. This study describes how the language-switching context, in terms of linguistic input and conversational demands among bilinguals has led them to become unbalanced toward flexibility (which seems to favor switching), potentially resulting in a trade-off with stability (which could favor inhibition). Thus, the opposite pattern might apply in the case of monolinguals immersed in a dual-language context. Since they do not need to switch between languages because of a lack of or a very low proficiency in L2, they develop a predisposition for a style related to stability (Musslick et al., Reference Musslick, Bizyaeva, Agaron, Leonard and Cohen2019; Sanchez-Azanza et al., Reference Sanchez-Azanza, López-Penadés, Aguilar-Mediavilla and Adrover-Roig2020; Serrien & O’Regan, Reference Serrien and O’Regan2019), in line with our correlation results. This stable style would probably be associated with a response pattern more compatible with the BEPA, predicting more general executive control benefits instead of inhibition specific gains (Dreisbach & Fröber, Reference Dreisbach and Fröber2019). Thus, we speculate that monolinguals with poorer knowledge of an L2 may be required to restrain the influence of other languages in their day-to-day interactions to prevent unwanted interference from codes they are not able to use adequately. That is, unlike bilinguals, who need to inhibit internal information from their second language, functional monolinguals immersed in a bilingual environment would undertake a similar process of suppressing external linguistic information and monitoring their use by others. In part, this could explain the correlation patterns in inhibition and updating found in monolinguals (as they would have to monitor their environment to actively mitigate linguistic influences). Although these findings do not establish a clear context-specific benefit, as ruled out through group comparisons, they do suggest that monolinguals and bilinguals process irrelevant information differently to adapt to the demands of their shared environment. Therefore, it is essential to consider the sociolinguistic aspects of environmental factors in monolingual and bilingual populations when interpreting the results from various bilingual research. For instance, Dash et al. (Reference Dash, Berroir, Ghazi-Saidi, Adrover-Roig and Ansaldo2021) also corroborated the distinction in operational strategies between monolingual and bilingual elderly populations through the dual mode of cognitive control. Consequently, this study substantiates that performance disparities might be rooted in strategic differences between these populations concerning the subcomponents of executive control.
An alternative explanation for the observed distinct associations between lower L2 use and better inhibition in monolinguals is that these individuals may rely on compensatory cognitive strategies. Monolinguals immersed in a dual-language environment, such as Montréal, might develop enhanced attentional control to manage frequent incidental exposure to a second language without the need for active language switching. This compensatory mechanism differs from bilinguals’ reliance on flexible language switching, which modulates their executive control balance between stability and flexibility (Ng & Yang, Reference Ng and Yang2022). This notion is also supported by structural brain studies showing compensatory neural adaptations in response to environmental cognitive demands (Torres et al., Reference Torres, Rosselli, Loewenstein, Lang, Vélez-Uribe, Arruda, Conniff, Curiel, Greig, Barker, Rodriguez, Adjouadi, Vaillancourt, Bauer and Duara2022). In contrast, while the stability–flexibility dilemma explains bilinguals’ inclination toward cognitive flexibility – favoring switching over inhibition – monolinguals may adopt a more stable control mode accompanied by compensatory attentional strategies to cope with environmental demands (Marton et al., Reference Marton, Goral, Campanelli, Yoon and Obler2017; Sanchez-Azanza et al., Reference Sanchez-Azanza, López-Penadés, Aguilar-Mediavilla and Adrover-Roig2020). Recent evidence from neurocognitive and neuroimaging literature supports that such compensatory mechanisms reflect increased cognitive reserve, where bilingualism strengthens executive control networks and delays cognitive decline despite greater brain atrophy (Costumero et al., Reference Costumero, Marin-Marin, Calabria, Belloch, Escudero, Baquero and Ávila2020). This dynamic suggests that the lack of a global bilingual advantage in dense code-switching contexts may arise from different adaptive control strategies rather than an absence of benefit.
4.3. Limitations
However, we believe it is relevant to consider some limitations that could influence the generalizability of the present findings. First, the conclusions of this research are based on a relatively small sample size. Previous studies have suggested that a larger sample size may increase statistical power and reduce the likelihood of false positive or false negative results in relation to bilingualism and executive control (Button et al., Reference Button, Ioannidis, Mokrysz, Nosek, Flint, Robinson and Munafò2013; Paap et al., Reference Paap, Johnson and Sawi2014). Although our sample size was determined based on power estimations from foundational studies reporting medium-to-large effects (e.g., Bialystok et al., Reference Bialystok, Craik, Green and Gollan2009; Bialystok & DePape, Reference Bialystok and DePape2009), we acknowledge that more recent meta-analyses suggest that bilingual advantages are typically small-to-medium in magnitude (Lehtonen et al., Reference Lehtonen, Soveri, Laine, Järvenpää, de Bruin and Antfolk2018). Thus, while our study was adequately powered to detect effects of the size reported in earlier work, it may be underpowered to capture subtler differences. Accordingly, null findings should be interpreted with caution. Nonetheless, our results remain informative within the current literature, and future replication with larger samples will be valuable to further validate these observations. Furthermore, it is important to consider that the tasks used in this research involved paper-and-pencil measures. Although it appears that performance on computer-based tasks and traditional paper-and-pencil tests correlates moderately to strongly (Hartle et al., Reference Hartle, Mendes-Santos, Barbosa, Balboni and Charchat-Fichman2021), this may have limited the accuracy of the cognitive assessments to some extent (Schatz & Browndyke, Reference Schatz and Browndyke2002). In any case, our choice of the Stroop, TMT and Digit Span was intentional, as these tasks are widely used in clinical contexts and remain robust and ecologically valid measures of executive control (Arbuthnott & Frank, 2000; Strauss et al., Reference Strauss, Sherman and Spreen2006). While we acknowledge that computer-based paradigms such as the Flanker, Simon or n-back could provide greater construct specificity, the tasks we employed do not diminish in value; rather, they offer a complementary approach. Future research would benefit from integrating both clinically grounded and experimental paradigms to provide a more comprehensive assessment of executive functions. Furthermore, despite having assessed the degree of bilingualism through objective measures (not just self-reported), this study would have benefited from more ecological measures that better capture the linguistic patterns of both bilinguals and monolinguals in their everyday lives. In this line, we also recognize that the inclusion of another monolingual control group immersed in a single-language context would improve the generalizability of the present findings. Finally, several of the present findings are derived from correlational analyses that do not allow strong causal conclusions regarding the relation between language experience and executive control. Although significant uncorrected associations are observed, these should not be regarded as direct evidence for causality. Moreover, the strict application of multiplicity corrections might reduce false positives while increasing false negatives, especially with modest sample sizes. Given our theory-driven analytical plan, we report uncorrected p values alongside transparent caveats, allowing us to balance inferential rigor with the potential for discovery. This approach facilitates meaningful interpretation while acknowledging the limitations inherent to correlational designs. Future longitudinal and experimental studies with larger samples should seek to replicate these associations with corrections applied, and to elucidate the underlying mechanisms (Maxwell & Cole, Reference Maxwell and Cole2007).
4.4. Conclusion
Overall, this study illustrates that in a shared dual-language context, bilinguals and functional monolinguals perform similarly on executive control tasks (Sanchez-Azanza et al., Reference Sanchez-Azanza, López-Penadés and Adrover-Roig2020). Furthermore, our findings emphasize the importance of considering the context in which monolinguals are evaluated in research on bilingualism and executive abilities, as associations between linguistic and cognitive variables might be paradoxical. Thus, the study of executive performance in monolinguals provides a valuable comparison to better understand the effects of bilingualism and its contexts on cognitive control, as the different patterns of correlations seem to demonstrate. Monolinguals and bilinguals may allocate cognitive resources differently for language control, resulting in distinct patterns of correlation with inhibitory control. In dual-language contexts, monolinguals with low L2 exposure and proficiency may exhibit characteristics of the Bilingual Executive Processing Advantage (BEPA), reflecting a more stable and response style. In contrast, bilinguals tend to show patterns consistent with the Bilingual Inhibitory Control Advantage (BICA), indicating more efficient interference management. These effects are more pronounced with higher bilingualism scores, which appear to support a more flexible and adaptive cognitive processing mode.
These findings underline the need for a more comprehensive and contextualized approach in research on bilingualism and executive control. Following the recommendations of other studies (Czapka et al., Reference Czapka, Wotschack, Klassert and Festman2020; De Cat et al., Reference De Cat, Kašćelan, Prévost, Serratrice, Tuller and Unsworth2023; Marian & Hayakawa, Reference Marian and Hayakawa2021; Samuel et al., Reference Samuel, Roehr-Brackin, Pak and Kim2018; Wagner et al., Reference Wagner, Bialystok and Grundy2022), future research ought to adopt more markedly methodological approaches that ensure a better understanding of the phenomenon of bilingualism and the purported bilingual advantage. In sum, the lack of between-group differences, together with the opposite correlation patterns between linguistic and executive measures in functional monolinguals and bilinguals underlines the need to approach the study of bilingualism from new methodological perspectives that consider not only the particularities of bilingual participants, but also those of monolinguals. In doing so, a more complete and refined understanding of the relation between bilingualism and executive control shall be developed.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/S1366728925100886.
Data availability statement
Data supporting the findings of this study are openly available at https://osf.io/dnb9x/
Acknowledgments
This work was supported by FRSQ (Fonds de Recherche en Santé du Québec, Grant nr. RBIQ09). We also thank all participants for their valuable time.
Competing interests
The authors declare none.
Open access
