Vocabulary matters! The relationship between verbal fluency and measures of inhibitory control in monolingual and bilingual children

Highlights

• Verbal fluency and attentional control were tested in 8-year-old monolinguals and bilinguals.

• Monolinguals and bilinguals had comparable English receptive vocabulary sizes.

• Bilinguals outperformed monolinguals on tasks of verbal fluency.

• Vocabulary and attentional control predicted children’s letter verbal fluency.

• Vocabulary size is fundamental in assessing bilinguals’ verbal fluency skills.

Abstract

The role of early bilingual experience in the development of skills in the general cognitive and linguistic domains remains poorly understood. This study investigated the link between these two domains by assessing inhibitory control processes in school-aged monolingual and bilingual children with similar English receptive vocabulary size. The participants, 8-year-old monolinguals and bilinguals, completed two Verbal Fluency Tasks (VFTs), letter and category, and two measures of inhibitory control. Results showed that bilinguals outperformed monolinguals on the VFTs, but performance was similar on the inhibitory control measures approaching ceiling for both monolingual and bilingual children. Importantly, it was shown that both vocabulary proficiency and general inhibitory control skills underlie monolingual and bilingual children’s performance on VFTs. These results demonstrate that vocabulary proficiency plays a fundamental role in comparing monolingual and bilingual VFT performance. The bilingual advantage found in this study seems to have escaped previous studies that did not account for vocabulary size in populations of bilingual and monolingual school-aged children.

Introduction

Whereas a bilingual advantage has been shown across various cognitive domains (see Bialystok, 2001, Bialystok, 2015, Bialystok, 2017, and Costa & Sebastián-Gallés, 2014, for reviews), bilinguals tend to show poorer performance on a number of language processing tasks compared with their monolingual peers (Sandoval, Gollan, Ferreira, & Salmon, 2010). However, recent research has demonstrated that bilingual adults do not always show lower performance in language processing tasks compared with monolinguals (Friesen et al., 2015, Luo et al., 2010). These studies show that successful bilingual performance can be related to a number of factors, including linguistic proficiency in each of the bilingual’s languages, degree of bilingualism, and domain-general cognitive abilities such as inhibitory control. The current study specifically investigated the interplay of these factors and the ability to rely on general inhibitory control skills in language processing tasks among bilingual and monolingual school-aged children.

When compared with their monolingual peers, bilinguals demonstrate a disadvantage in lexical processing tasks when assessed in only one of their languages (see Gollan & Kroll, 2001, for a review). For instance, when administered with single-language tasks, children and adults who are proficient bilinguals tend to obtain lower receptive vocabulary scores than monolinguals (Bialystok, 2006, Bialystok and Luk, 2012, Bialystok et al., 2010). This finding is usually attributed to bilinguals’ tendency to receive less exposure to each of their languages compared with monolinguals and to spend less time using each of their languages (Byers-Heinlein, Fennell, & Werker, 2013). However, when bilinguals’ vocabulary size in the two languages is considered, no bilingual disadvantage is observed, and indeed bilinguals’ combined vocabulary sizes often surpass monolinguals’ single-language vocabulary sizes (Pearson, Fernandez, & Oller, 1993). The bilingual disadvantage is even more pronounced in tasks that rely on productive skills, especially those targeting lexical retrieval. For instance, bilinguals perform slower in picture naming tasks (Bialystok et al., 2008b, Ivanova and Costa, 2008), exhibit more tip-of-the-tongue states (Gollan & Acenas, 2004), and obtain lower verbal fluency scores (Gollan et al., 2002, Rosselli et al., 2002). Aside from differences in language proficiency, these difficulties in production tasks have been attributed to nontarget language intrusion. That is, the two languages of a bilingual are constantly active, resulting in greater demands for suppression of cross-linguistic competition during the production process (Bialystok et al., 2008a, Bialystok et al., 2008b, Rodriguez-Fornells et al., 2005). In addition, bilingual performance can be affected by a number of individual and environmental factors such as age of acquisition of each language (Valian, 2015), proficiency in each language and patterns of language use (MacLeod, Castellanos-Ryan, Parent, Jacques, & Séguin, 2017), and socioeconomic status (SES) among others (Calvo & Bialystok, 2014). Thus, it appears that bilinguals’ performance in language processing tasks is dependent simultaneously on their ability to overcome the cognitive demands of the specific task and their lexical competence in the target language.

A commonly used measure of lexical retrieval is the Verbal Fluency Task (VFT). It has been successfully used as a neuropsychological measure of efficiency in lexical retrieval in diverse typical and clinical populations (Anderson et al., 2001, Bialystok et al., 2009, Filippetti and Allegri, 2011, Friesen et al., 2015, Korkman et al., 2001, Luo et al., 2010, Malek et al., 2013, Matute et al., 2004, Ostrosky-Solis et al., 2007). This task usually comprises two conditions, category and letter, where participants are asked to produce as many words as possible either belonging to a semantic category (category VFT; e.g., list animal names) or starting with a particular letter (letter VFT; e.g., list words that start with the letter f) in a specified period of time, usually 60 s. The two conditions recruit both lexical and semantic knowledge. That is, the VFTs require participants to retrieve words from their lexicon in the case of monolinguals or from one of their lexicons in the case of bilinguals.

Aside from assessing lexical and semantic competence, the two VFTs rely on executive functioning skills, which is of particular interest for the current study. Executive functioning refers to general self-regulation and control processes that involve working memory, monitoring, and inhibitory control (Miyake et al., 2000). More precisely, the two VFTs have been shown to engage two different cognitive processes that are comprised within the definition of inhibitory control (Indefrey and Levelt, 2004, Levelt, 1999, Levelt, 2001, Perret, 1974, Robinson, 2003, Schmidt et al., 2017), which refers to the ability to inhibit responses to irrelevant information while focusing on a cognitive goal (Kochanska, Murray, & Coy, 1997). The two cognitive processes comprised within inhibitory control are response inhibition and interference suppression (Luo et al., 2010). Response inhibition assesses participants’ ability to inhibit a usual response in favor of an unusual response in the task (Bunge, Dudukovic, Thomason, Vaidya, & Gabrieli, 2002). Interference suppression, on the other hand, requires the participants to switch their attention between two competing responses during the task (Martin-Rhee & Bialystok, 2008). Following this classification, the category VFT has been proposed to rely on response inhibition skills. When performing this task, the participants are required to inhibit the production of semantic neighbors that do not fall within the category specified in the task instructions, but they are not required to suppress a competing response or to switch between two response rules (Friesen et al., 2015, Luo et al., 2010, Schmidt et al., 2017). Thus, this task is similar to lexical retrieval in everyday communication because it complies with common strategies of semantic retrieval and organization (i.e., words are stored in semantic categories in the lexicon). Conversely, the letter VFT is proposed to rely on interference suppression skills because participants are required to generate words following a strategy that is not naturally used for lexical retrieval in everyday communication and need to suppress the automatically activated semantic neighbors for the words listed during the task (Perret, 1974, Schmidt et al., 2017). It is noteworthy that these executive functioning demands may be particularly pronounced in bilingual populations because, aside from taxing inhibition and interference suppression, the VFTs also recruit more general conflict monitoring skills. That is, bilinguals are also required to monitor the use of only one of their languages during the task (Sandoval et al., 2010), which resembles the monitoring demands that bilinguals face in most everyday communicative situations (Bialystok, 2001).

Further evidence for the involvement of domain-general executive functioning abilities in the VFTs is found in neurophysiological studies. For instance, a recent meta-analysis of functional magnetic resonance imaging data on healthy monolingual participants’ performance on the letter and category VFTs showed evidence for activation in the left frontal lobe but also in the anterior cingulate gyrus (Wagner, Sebastian, Lieb, Tüscher, & Tadić, 2014). These findings provide evidence that the inhibitory control demands of this task fall beyond those related to the demands of speech processing (Snyder, Feigenson, & Thompson-Schill, 2007). Therefore, comparisons of performance on the VFTs between monolinguals and bilinguals reveal an intriguing discrepancy. That is, whereas bilingual adults and children commonly outperform monolinguals in measures of executive functioning skills, particularly inhibitory control (see Bialystok, 2015, for a review), they underperform monolinguals on the VFTs despite the fact that these tasks also recruit general executive functioning skills. Bialystok et al. (2008a) showed that the interaction between bilinguals’ executive functioning skills and vocabulary size in the target language may account for this apparent discrepancy. In their study, adult monolingual participants were compared with high- and low-receptive vocabulary groups of bilinguals on category and letter VFT performance. The bilingual adults included in this study had acquired English as their second language during childhood but had used it as their dominant language for schooling and social activities. This study showed that high-vocabulary bilinguals outperformed monolinguals in the letter VFT, whereas their performance was comparable in the category VFT. Thus, vocabulary size and the inhibitory control demands of the task both appeared to moderate bilingual performance, specifically in the letter VFT condition that incurred the interference suppression component of inhibitory control.

The use of VFTs to compare monolingual and bilingual children’s performance has been scarce and has produced mixed results. Kormi-Nouri, Moradi, Moradi, Akbari-Zardkhaneh, and Zahedian (2012) tested 1600 monolingual and bilingual 7- to 12-year-olds using a Persian version of VFTs. Participants were Turkish–Persian bilinguals, Kurdish–Persian bilinguals, and Persian monolinguals. Results showed that only the Turkish–Persian bilinguals outperformed the monolingual group in the letter VFT. In the category VTF, monolinguals outperformed both bilingual groups. However, the analyses did not account for proficiency in the language in which the tests were administered (i.e., Persian). Thus, the lack of a bilingual advantage in the Kurdish–Persian bilingual group may be due to the fact that this group was not as proficient in the target language (Persian) as the Turkish–Persian group (Friesen et al., 2015). Indeed, the method adopted by Kormi-Nouri et al. (2012) and the large number of participants may have posed difficulties in terms of accounting for variables such as age, SES, and level of bilingualism.

In a more recent study, Friesen et al. (2015) used the VFTs to compare performance in bilingual and monolingual children at the ages of 7 and 10 years as well as young and older adults. Importantly, the samples of bilingual children recruited at the two ages differed in their degrees of English proficiency. The bilingual 7-year-olds had comparable English vocabulary sizes to monolinguals, but the bilingual 10-year-olds had English vocabulary sizes that were significantly lower than their monolingual peers. The findings showed that bilingual and monolingual 7- and 10-year-olds performed similarly in both the letter and category VFTs. It appears that 10-year-olds with lower vocabulary skills were able to compensate for their lower linguistic knowledge by recruiting higher levels of inhibitory control and, thus, attained similar performance to monolinguals. Bilingual younger and older adults, however, did outperform their monolingual counterparts, but only in the letter condition. Thus, the findings show that bilinguals’ ability to engage inhibitory control in verbal tasks emerges sometime around 7 years and before 10 years of age, a bilingual advantage in VFTs can be captured around young adulthood, and it may be more evident when the bilingual and monolingual samples are matched for proficiency in the target language.

Previous research, therefore, indicates that vocabulary knowledge and inhibitory control underlie performance in measures of lexical retrieval, specifically VFTs. More advanced inhibitory control associated with bilingualism leads to an advantage in VFTs, particularly in the letter condition, which poses demands for the interference suppression component of inhibitory control, and this advantage emerges before adulthood. It is possible that as children’s executive functioning abilities are developed, bilinguals become able to recruit their inhibitory control skills to successfully perform in this task. The aim of the current study was to further investigate the interplay between linguistic proficiency and inhibitory control abilities and their relation to monolingual and bilingual children’s VFT performance, particularly at the age when a bilingual advantage is expected to emerge (i.e., before 10 years according to Friesen et al., 2015).

In this study, monolingual and bilingual children performed the letter and category conditions of the VFT and two measures of inhibitory control. Because the letter and category conditions are associated with different components of inhibitory control, whereby the letter VFT is a measure that taps into interference suppression, and the category VFT is a measure that taps into response inhibition, the two inhibitory control measures were selected to match this dissociation. The advanced version of the Dimensional Change Card Sorting task (DCCS; Zelazo, 2006, Zelazo et al., 1997) was selected as the measure of interference suppression. This task requires solving conflictive dimensions while avoiding distractions from salient but irrelevant cues (Bialystok and Martin, 2004, Bunge et al., 2002, Sorge et al., 2017). In contrast, the Day–Night version of the Stroop task (Day–Night Stroop; Diamond, Kirkham, & Amso, 2002) was selected as the measure of response inhibition because this is a task that creates conflict between two simple options of the same stimulus feature (Bunge et al., 2002). Previous studies have shown that bilingual children outperform monolingual peers only in tasks that assess the interference suppression component of inhibitory control (Bialystok and Shapero, 2005, Carlson and Meltzoff, 2008, Iluz-Cohen and Armon-Lotem, 2013, Kalashnikova and Mattock, 2014, Qu et al., 2016, Zelazo, 2006), but not in tasks that assess response inhibition (Bialystok et al., 2009, Martin-Rhee and Bialystok, 2008, Sorge et al., 2017).

It is worth emphasizing that one of the aims of the current study was to improve on previous studies by including a sample of bilingual children whose vocabulary proficiency in the language tested was comparable to that of the monolingual group. Therefore, any difference in performance between groups could not be attributed to a group difference in vocabulary proficiency. Given the comparable vocabulary size for our groups and following the findings by Friesen et al. (2015), we predicted a bilingual advantage in the letter VFT among children in this study. Furthermore, we were particularly interested in assessing the contribution of vocabulary proficiency and general cognitive skills in monolingual and bilingual children’s VFT performance. In that respect, if receptive vocabulary size ensures successful VFT performance in children at this age, it should be a significant predictor of children’s VFT scores for both groups of children and in both VFT conditions. Alternatively, if children also rely on their nonverbal inhibitory control skills, their inhibitory control scores that involve interference suppression (i.e., DCCS scores) but not response inhibition (i.e., Day–Night Stroop scores) should predict their VFT performance over and above the contribution of receptive vocabulary size. In addition, if bilingualism enhances interference suppression skills, we would expect stronger predictive power of interference suppression measures on the letter VFT for our bilingual group.