Do bilingual and monolingual preschoolers acquire false belief understanding similarly? The role of executive functioning and language

Abstract

Bilingual children often show advanced executive functioning (EF) and false belief (FB) understanding compared to monolinguals. The latter has been attributed to their enhanced inhibitory control EF, although this has only been examined in a single study which did not confirm this hypothesis. The current study examined the relation of EF and language proficiency on FB reasoning in bilingual and monolingual preschoolers to answer two questions: (1) Are there differences in bilinguals’ and monolinguals’ FB, language proficiency, and EF? If so, (2) is there a differential role for language proficiency and EF in predicting FB reasoning in these two groups? Thirty-two Spanish–English bilinguals and 33 English monolinguals (three to five years old) were compared. While monolinguals outperformed bilinguals on language proficiency, after controlling for this, bilinguals outperformed monolinguals on FB reasoning, and marginally on EF. General language ability was related to FB performance in both groups, while short-term memory and inhibitory control predicted FB only for monolinguals.

Keywords

Bilingual development bilingualism executive function false belief theory of mind

During early development, children develop an understanding of mental states reflecting the emergence of a theory of mind (ToM), which refers to an understanding of what the mind is. One aspect of ToM development is the ability to conceptualize the contents of the minds of others as different from one’s own, by recognizing others’ false beliefs (FBs) (Wellman, Cross, & Watson, 2001). An example of a FB reasoning task is the Sally–Anne task about the displacement of an object. In this task, Sally has a basket and Anne has a box. Sally places a marble in her basket and goes away. Anne then moves the marble to her box. When Sally returns, the children are asked where will Sally look for the marble. To answer correctly, children need to ignore their own correct knowledge, and assess what Sally knows by conceptualizing Sally’s mind and knowledge as different from the children’s own (Baron-Cohen, Leslie, & Frith, 1985; Wimmer & Perner, 1983).

Research has focused extensively on the role of executive functioning (EF) and language proficiency on the development of FB understanding. EF is an umbrella term for a variety of functions related to the management of one’s own cognitive functions and behavioral responses such as directed attention, inhibition of impulses or inhibitory control, planning, categorization, and working memory (Diamond, 2013).

An overwhelming number of studies have confirmed the relationship between EF (particularly inhibitory control) and language proficiency on the development of FB understanding (for a meta-analysis on each topic respectively see Devine & Hughes, 2014; Milligan, Astington, & Dack, 2007). These studies, however, have examined the relationship between FB understanding and language proficiency and EF almost exclusively in monolingual children. Recently, studies of bilingual preschool children and adults demonstrated that bilinguals have an advantage over monolinguals in FB reasoning (e.g., Goetz, 2003; Kovács, 2009; Nguyen & Astington, 2014; Rubio-Fernández & Glucksberg, 2012). For example, Goetz (2003) compared Chinese–English bilinguals to both English and Chinese monolinguals on FB reasoning. Bilinguals outperformed both groups of monolinguals, and the two monolingual groups’ performance did not differ from one another in spite of coming from two vastly different linguistic and cultural backgrounds. In a similar fashion, Kovács (2009) found that three-year-old Hungarian–Romanian bilinguals passed FB reasoning tasks more often than Romanian monolinguals of similar age.

This FB advantage for bilingual children has been attributed to their enhanced EF (particularly inhibitory control), where they also often outperform monolinguals (e.g., Goetz, 2003; Kovács, 2009). Studies of monolingual children indicate that EF is strongly related to FB reasoning (Carlson & Moses, 2001; Devine & Hughes, 2014; Hughes & Ensor, 2007), although this relation is not always found (Montgomery & Koeltzow, 2010).

The reasoning behind this suggestion is a three-part argument. The first part stems from the inhibitory control component of the FB reasoning tasks. Inhibitory control is the main EF component thought to be involved in FB reasoning tasks, and the one referred to most often in the current work. Accordingly, in order to answer the FB questions about the displacement of an object as laid out before, children need to inhibit the preponderance of their own correct knowledge, as well as the attentional pull of the true location of the object. Instead, they must choose the erroneous expectation of another – a false belief – and point to an empty object location. Part two of the argument is that bilinguals have enhanced EF inhibitory control abilities due to the need to hold and manage dual representations for linguistic information. This includes having to constantly inhibit one set of linguistic representations in favor of another set (Bialystok, 2001; Bialystok, Craik, Klein, & Viswanathan, 2004). Finally, this enhanced EF in bilinguals boosts their performance on the FB reasoning tasks, due to the EF component in the FB tasks as described above.

The EF bilingual advantage has been demonstrated in multiple studies. For example, Bialystok and Viswanathan (2009) compared eight-year-old Canadian monolingual children to Canadian bilingual children who spoke English and one of a variety of other languages, and also to bilingual children in India who also spoke English and one of two other languages. Results indicated that bilingual children from both Canada and India outperformed monolinguals on EF tasks requiring inhibitory control and switching abilities. Significant bilingual–monolingual differences in four- to five-year-old preschoolers’ EF performance have been reported as well (Martin-Rhee & Bialystok, 2008). In this study bilingual preschoolers outperformed monolinguals only when the task required the children to ignore a misleading command such as in the Simon Says task, and not when they only had to refrain from executing a salient response.

Similarly, Carlson and Meltzoff (2008) tested Spanish–English bilinguals, English monolinguals, and English speakers enrolled in a second-language immersion kindergarten on a variety of EF tasks. After controlling for marked differences in socio-economic status (SES), Spanish–English bilinguals performed better than English monolinguals and the language immersion group on conflict tasks requiring cognitive flexibility (e.g., sorting cards according to different rules), but not on delay tasks (e.g., waiting to open a present). An EF advantage for bilinguals has also been reported in toddlers (Poulin-Dubois, Blaye, Coutya, & Bialystok, 2011) and infants (Kovács & Mehler, 2009). Differences in memory flexibility have also been found between bilingual and monolingual children aged 1;6 (Brito & Barr, 2012).

While bilinguals have typically been found to have an advantage on EF tasks involving inhibitory control, this is not always the case. An accumulating body of studies has indeed failed to replicate the bilingual advantage in EF tasks. For instance, Nguyen and Astington (2014) reported that French–English bilingual preschoolers did not differ from monolingual preschoolers on EF tasks involving inhibitory control or working memory. Morton and Harper (2007) also failed to find cognitive control differences between bilingual and monolingual children of identical SES. Similarly, bilingual young adults tested on a variety of EF tasks such as anti-saccade tasks, Simon tasks, flanker tasks, and color-shape switching tasks, showed no advantage on any of the tasks, even when participants were matched on demographics and fluency (Paap & Greenberg, 2013). However, the proposed relationship between enhanced EF in bilinguals and a resulting boost in FB understanding has only been tested in a single study of bilingual preschoolers (Nguyen & Astington, 2014). Contrary to expectations, this study did not find a relationship between FB reasoning and inhibitory control.

In addition to EF, a relationship between language proficiency and the development of FB reasoning has been widely reported in monolingual children. There is some debate though regarding whether general language ability, comprised of both semantic and grammatical abilities, is sufficient, or whether the specific complementation syntactic structure is required to understand FBs (de Villiers, 2005; Slade & Ruffman, 2005). Complement structures involving mental state verbs allow the embedding of one proposition in another. For instance, in the sentence ‘The man thinks that it is raining but it is really the sprinkler’, the presence of the mental state verb think makes the truthfulness of the sentence independent of the real state of the world. Thus, sentential tensed complements allow a distinction between what a person thinks and what is actually true, which is the core of FB understanding. In addition, processing of this kind of sentence requires dual representations, which is another hallmark of FB reasoning; simultaneously representing both clauses, and both an FB and true belief. Research supporting both the general language proficiency (e.g., Cheung et al., 2004; Farrar, Lee, Cho, Tamargo, & Seung, 2013) and the complementation positions (de Villiers & Pyers, 2002) has been obtained.

Language proficiency has been a contentious topic in which to compare bilinguals and monolinguals, as bilingual children often score lower on language proficiency tests compared to monolinguals when tested in only one language (e.g., Bialystok & Viswanathan, 2009; Carlson & Meltzoff, 2008). This difference was corroborated in a meta-analysis by Bialystok, Luk, Peets, and Yang (2010) where bilinguals’ receptive vocabulary was smaller compared to that of monolinguals across studies. As with EF, studies of bilingual children and the relation of language proficiency to their FB performance is limited. Recently, it was reported that vocabulary proficiency in both Spanish and English was associated with bilingual children’s FB performance (Gordon, 2016). Whether grammar and complementation is related to FB reasoning in bilingual children remains unknown.

Current study

Of interest in the current study is whether bilinguals show enhanced FB reasoning, and whether language ability and EF play similar or different roles in FB reasoning for bilingual and monolingual children. These issues are examined by comparing preschool Spanish–English bilinguals to English monolinguals. Spanish is the most common second language in the United States (U.S. Census Bureau, 2009), and therefore the most relevant and accessible population of bilinguals for researchers in the United States. Two issues were explored. First, and of primary interest, is whether the relationship between FB reasoning to both EF (inhibitory control and cognitive flexibility) and language proficiency (vocabulary, grammar, and complementation) is different for monolingual and bilingual preschoolers. Of secondary interest is whether the bilingual FB advantage reported in prior studies was present in our sample, and if so, could this difference in performance be explained by the aforementioned differential relationships.

Method

Participants

Thirty-two Spanish–English bilingual and 33 English monolingual children aged between 3;4 (40 months) and 5;5 (65 months) (bilingual M = 50.51, SD = 6.93, 37–64.5 months; monolingual M = 49.52, SD = 6.42, 39.8–65 months) were tested. Bilingual children were identified by their parents as being fluent in both Spanish and English, and regularly interacting with speakers of both these languages. The majority of the bilingual children (61%) had been exposed to both languages since birth, and all had been exposed to their non-dominant language for at least one year. These values are presented in Table 1, and were obtained through language questionnaires that the parents filled out along with the consent forms.

Table 1.

Bilingual children’s ‘Parent/guardian language questionnaire’ responses.

	% English only	% Spanish only	% Both
Languages child speaks (>1 year)	_	_	100
Child’s preferred language	81.3	12.5	6.3
Home	19.4	19.4	61.3
Spoken at preschool	77.4	3.2	19.4
Mother’s native language	16.1	64.5	19.4
Father’s native language	12.9	67.7	19.4
Other’s native language	7.1	92.9	_
Language mother talks to child	19.4	22.6	58.1
Language father talks to child	25.8	25.8	48.4
Language other talks to child	_	71.4	28.6
Reading language at home	45.5	9.1	45.5
Language exposure at birth	19.4	19.4	61.3

Measures

False belief

Unexpected content task

A closed crayon box was presented to the child while the experimenter (E) asked: ‘What do you think is inside the box?’ Upon receiving an answer from the child, E revealed that the box actually contained candles (Perner, Leekman, & Wimmer, 1987). E then asked the child: ‘What is inside the box?’ After receiving an answer from the child and putting the candles back in the box, E asked a representational change question: ‘When you first saw the box, all closed up like this (closing box and pointing to the closed top), what did you think was inside?’ E then brought out a puppet named Ernie and asked the child a false belief question: ‘Ernie has never looked inside the box. What does Ernie think is inside the box?’ Finally, E asked the control question: ‘What is really inside the box?’ The child received 1 or 0 for each of the two belief questions for a total of two possible points, if he or she also answered the control question correctly, otherwise no points were awarded.

Unexpected location task

E enacted the following story in front of the child while narrating it: ‘Big Bird is putting his toy car inside the box. Now he is going outside to play. This is Ernie’s friend the Puppy. Puppy is taking the toy car out of the box and putting it inside this bag. Now the Puppy is going home.’ E then asked a false belief question, ‘When Big Bird comes back inside, where will he look for the car?’ E then showed Big Bird where the toy car was, and asked the child, ‘Before I showed him, where did he think the toy car was?’ The child received 1 or 0 for the last two questions for a total of two points if she also answered the control questions correctly as to where the object is right now and where was it placed originally, otherwise the points were eliminated (see Wimmer & Perner, 1983).

Object disappearance, low inhibitory control demand FB task

E narrated the following story using custom made illustrations (as described in Wellman et al., 2001). ‘Maxi is putting his chocolate inside the box. Now he is going outside to play. Maxi’s mom takes the chocolate out of the box and eats it. Now she leaves. When Maxi comes back inside, will he look for the chocolate?’ The child received 1 or 0 for the last false belief question, and another point for the false belief question ‘Now the mom tells Maxi that she ate the chocolate. Before she told him that, would he have looked for the chocolate?’ for a total of two points if she also answered the control questions correctly as to where the object is right now and where was it placed originally, otherwise the points were eliminated.

Appearance-reality, object identity task

E introduced a sponge that looks like a rock and said: ‘Look at what I have here, what is it?’ After the child’s response, E said: ‘Here, you can touch it.’ At this moment the child realized the object was actually soft and spongy. E then asked the child a series of reality questions and belief questions: ‘What does this look to your eyes right now? 1.What is it really? (Appearance reality). 2. When you first saw this, before you touched it, what did you think it was? (representational change). Elmo has never touched this before. What does Elmo think this is?’ (false belief). The child received 1 or 0 for the last three numbered questions, if she also answered the control question ‘What does this look like to your eyes right now?’ correctly. If the child failed that question no points were awarded for the reality question ‘What is it really?’

Appearance-reality: object property task

E introduced a white fish cutout and asked: ‘When you look at this fish, what color is it?’ E then placed a red filter over the fish and asked: ‘When you look at this fish with your eyes right now, what color does it look?’ While the filter remained over the fish, E asked: ‘What color is this fish really and truly?’ and ‘Before I put this plastic over it, what color did you say it was?’ The child received 1 or 0 for the last two questions for a total of two points.

An FB reasoning composite was created with the child’s scores on all of these measures for a total of 11 points.

Executive functioning

The EF tasks were selected based on two criteria: (1) those tasks that have been shown to differ between monolingual and bilingual preschoolers and (2) those that are frequently related to FB understanding in monolingual preschoolers (Carlson & Moses, 2001).

Day/night Stroop-like task

The child was asked to identify a card depicting a sun and one depicting a moon and stars (Gerstadt, Hong, & Diamond, 1994). If identified correctly as ‘night’ and ‘day’, the child was given the following instructions: ‘When you see this card (pointing to the night card), I want you to say “day”, and when you see this one (pointing to the day card), I want you to say “night”.’ Understanding was confirmed by giving the child one practice trial with each card. The instructions and practice trials were repeated until the child answered both practice questions correctly. The child was then given 16 trials in a fixed random order in which she was asked to name each card according to the rule. The final score was the sum of times the child named the card correctly according to the rule.

Dimensional change card sort task

E introduced two containers, one labeled with a picture of a red rabbit another with a blue boat (Frye, Zelazo, & Palfai, 1995). E instructed the child to play the shape game, in which ‘the rabbits go with the rabbits and the boats go with the boats.’ E demonstrated this rule with a blue rabbit and one depicting a red boat, and asked the child to sort a set of cards without feedback. E then instructed the child to switch to the color game in which ‘the blues go with the blues and the reds go with the reds,’ using another set of cards. The child’s score was computed out of four possible points based on the correct sorting of the red boat (twice) and blue rabbit (twice) in the second set.

Bear/dragon Simon Says task

E asked the child to imitate 10 actions such as ‘stick out your tongue’ or ‘touch your ears’ (Reed, Pien, & Rothbart, 1984). E then introduced two characters and the rules of the task: ‘This is the nice bear, when he talks to us, we will do what he tells us to do. This is the naughty dragon, when he talks to us we won’t listen to him. If he tells us to do something, we won’t do it.’ E then ran practice trials with each puppet until the child effectively followed at least one bear command, and resisted at least one dragon command. When the task began, five bear trials and five dragon trials of different actions were given in alternating order. The child received one point for ignoring each of the dragon commands, and one point for following each of the bear commands for a total of 10 points.

In addition to raw scores, an EF composite score was created by transforming each task score into a z-score due to the widely differing number of trials per task compared to our other measures. A composite was created by adding the z-scores.

Memory

A short-term memory task was introduced to serve as a general cognitive ability control measure in place of an intelligence quotient test (Bialystok, 2010).

Short-term memory digit span task

E read single digit lists at a rate of one per second starting with two digits (as reported in Bialystok, 2010). The child was asked to repeat the numbers in the same order. One digit was added after each second trial until the child was unable to reproduce both trials of the same number of digits. The final score consisted of the last number of digits in which the child was able to reproduce at least one sequence.

Language

Language tests were provided in English to English monolinguals, and to Spanish–English bilinguals in their identified dominant language (see below for the identification procedure).

Clinical Evaluation of Language Fundamentals (CELF)

Children’s general language proficiency was measured using the CELF’s three core language subtests: Expressive Vocabulary, Sentence Structure, and Word Structure, in English or Spanish as appropriate (Wiig, Secord, & Semel, 2004). For the Expressive Vocabulary, children were shown a series of pictures and were asked to name each one. In the Sentence Structure trial, the child was read a series of sentences and was asked to point to the picture that represents the sentence out of four related drawings. For the Word Structure trial, the child was asked to use the progressive –ing, prepositions, pronouns, plurals, etc., by for example telling the child ‘Here is one cat’ while showing them a corresponding picture, and asking them to follow suit with a picture showing two cats. Total CELF scores and subtests scores were obtained.

Receptive One Word Picture Vocabulary Test (ROWPVT)

Children’s receptive vocabulary was measured using the English or Spanish–English bilingual versions of the ROWPVT as appropriate (Gardner, 2000b). For this test, children were presented with a series of four drawings. The experimenter instructed the child to point to the picture that matched the word that the experimenter said.

Complementation task

Comprehension for complementation was measured in the following manner: E read to the child 12 sentences with embedded complements, accompanied with illustrative pictures (de Villiers & Pyers, 2002). E then asked a memory question, e.g., ‘She thought the girl was reading a book, but she was really playing cards’; the memory question was, ‘What did she think?’ The child’s score was a total of 12 possible points. Half the sentences contained a mental state verb (think) and half contained a communication verb (say). Different verbs were used to separate the semantics of complementation (think condition) from the grammar of complementation (say condition). Total scores and subscores of mental and communication complementation were calculated.

Procedure

Parents were asked to fill out a demographic questionnaire as well as a language questionnaire that asked about the child’s language preference and exposure to different languages. All parents were compensated for their child’s participation with a $20 gift card. The testing was conducted in quiet corners at the children’s preschools in play-like interactions. Testing was conducted over two sessions each lasting around 30 minutes, to prevent fatigue. A procedure was in place to discontinue the session if the child became distracted or uncooperative, but this was not necessary. The two sessions took place around a week apart.

All measures were given in English to English monolinguals, and to bilinguals in the language selected by the parents as the child’s dominant language in the parent questionnaire. The language dominance identified by the parents of the bilingual children was corroborated by asking the children’s teacher for their own dominance assessment, followed by directly asking the child for their preference. The testing sessions began with a language test (the CELF) in the dominant language, further giving the experimenter an opportunity to confirm dominance. The tasks were given to bilinguals in their dominant language in order to assure the bilinguals’ performance was not hindered by their non-dominance in their second language. The principal investigator, who is a native Spanish speaker and fluent in English, did the translation of all testing and consent materials, as well as conducted all the testing sessions. The testing sessions began with a language test involving pointing to make the child familiar with the experimenter before he/she was asked to speak. Testing then proceeded with FB and EF tasks in alternating order. The tasks were given to all children in a fixed sequence.¹

Results

In preliminary analyses, we examined demographic differences between the monolingual and bilingual groups. Mothers of monolingual children had on average a higher education level than mothers of bilingual children (t (62) = 3.85, p < .001; monolingual M = 5.63; bilingual M = 4.63) with our scale ranging from 0 = ‘Less than High School’ to 6 = ‘Graduate Degree.’ We included maternal level of education as a demographic control variable on our correlation and regression analyses. There were no significant differences between the groups on age and gender distribution or significant gender differences on any of the dependent measures below. In addition to these standard demographic considerations and controls, we inspected the data for potential differences between those children who had been bilinguals since birth (n = 19), versus those who had not (n = 13). Importantly, all children in the study had been exposed to their non-dominant language for at least one year. Preliminary results indicated no significant differences between cradle versus non-cradle bilinguals on our main variables of interest (FB, EF, and language ability), and thus were included in the same bilingual group in all further analyses. This is likely due to the fact that this is a preschool sample. Because of their young ages, the children are learning both languages simultaneously, in spite of earlier exposure to one language versus the other. Indeed, exposure to a second language before the age of five is considered dual language learning (Castro, García, & Markos, 2013).

Differences in false belief reasoning, executive functioning, and language proficiency in bilingual and monolingual children

Means and standard deviations of the primary dependent measures are presented in Table 2. The first question addressed was whether monolingual and bilingual children differed on FB reasoning, EF, and language abilities. Comparisons of monolingual and bilingual children were conducted on all the measures using t-tests. As hypothesized, there were significant differences between bilinguals and monolinguals on language ability as indicated by the CELF (t (62) = 5.11, p < .001), the ROWPVT (t (62) = 4.93, p < .001), and complementation comprehension, (t (63) = 2.08, p < .05). Monolinguals outperformed bilinguals on each of these language measures.

Table 2.

Means, standard deviations, estimated means, and standard errors for bilinguals’ and monolinguals’ false belief, executive function, language, and short-term memory measures.

	False belief	EF^b	CELF	ROWPVT	Complements	STM
		Bilinguals (n = 32)
M	6.22	−.25	46.00	44.71	7.69	2.32
SD	(2.42)	(2.15)	(14.43)	(13.55)	(3.74)	(.83)
Estimated M^a	7.29	.49
SE	(.41)	(.34)
		Monolinguals (n = 33)
M	7.09	.24	62.45	59.18	9.55	2.59
SD	(2.70)	(1.72)	(10.96)	(9.75)	(3.47)	(.67)
Estimated M^a	6.04	−.47
SE	(.40)	(.32)

Note: Standard deviations and standard errors in parentheses.

Estimated least squares means and standard errors obtained from an ANCOVA controlling for language ability using the CELF and ROWPVT.

Composite created by adding z-transformed scores from the individual tasks.

EF = Executive functioning.

CELF = Clinical Evaluation of Language Fundamentals – Core Language subscale.

ROWPVT = Receptive One Word Picture Vocabulary Test.

STM = Short-term memory.

In contrast to the hypotheses, there were no raw differences between monolingual and bilingual children on the EF measures as shown in Table 2, all ts (63), ps > .05). Untransformed ranges and means for the individual tasks are presented in Table 3. There were no differences on the short-term memory (STM) measure, t (63) = 1.21, p > .05. Of particular interest was whether there were differences between bilingual and monolingual children on the FB reasoning measures. However, there were no performance differences on the FB reasoning composite.

Table 3.

Untransformed minimum, maximum, mean, and standard deviation of EF tasks.

	Minimum	Maximum	Mean	SD
		Bilingual (n = 32)
Day/Night	3	16	12.59	3.68
Bear/Dragon	0	10	8.03	2.51
Card sort	0	4	2.75	1.65
		Monolinguals (n = 33)
Day/Night	2	16	11.94	3.58
Bear/Dragon	6	10	8.97	1.47
Card sort	0	4	3.09	1.42

To further explore whether there are differences between the monolingual and bilingual children on EF and FB reasoning, we conducted a multiple analysis of covariance (MANCOVA) controlling for overall language ability as assessed by the CELF and ROWPVT since the groups differed on these language scores. The MANCOVA indicated that there was a significant main effect of bilingualism status on FB task performance at F (1, 64) = 4.02, p = .049, η_p² = .063 and a marginally significant main effect on EF at F (1, 64) = 3.52, p = .06 η_p² = .055 with bilinguals outperforming monolinguals. These results indicate that after controlling for the marked differences on language ability between these groups, bilinguals displayed an advantage over monolinguals on FB related reasoning and EF (see Table 2).

Predictors of false belief reasoning in monolingual and bilingual children

Of primary interest was the relation of language proficiency and EF to FB reasoning in the monolingual and bilingual groups. The correlations for the two groups were examined separately. As seen in Table 4, for the monolingual group, after controlling for age and maternal education which differed between the groups, FB reasoning overall correlated with the CELF, receptive vocabulary, complementation, and STM. In addition, while the EF composite measure was not related to FB, the Bear/Dragon measure of inhibitory control was associated with FB in monolinguals. For the bilingual preschoolers only the CELF (general language ability) was found to be correlated to FB reasoning. Neither complementation, nor STM nor EF were related to FB reasoning. Thus, prior speculation that the advantage that bilingual preschoolers have shown in FB reasoning could be attributed to their EF skills was not supported.

Table 4.

Bilingual (lower diagonal) and monolingual (upper diagonal) partial correlations controlling for age and maternal education, between false belief, executive function, and language measures.

	FB	EF	DCCS	B/D	D/N	STM	CELF	ROWPVT	Comp.
Bilinguals	Monolinguals
FB	–	.24	.23	.32*	−.017	.59*	.59**	.35*	.39**
EF	.05	–	.78***	.61***	.54**	.37*	.50**	.29	.50**
DCCS	.19	.60***	–	.55**	.01	.24	.41**	.25	.36**
B/D	−.07	.73***	.28	–	−.14	.35*	.57**	.39**	.57**
D/N	−.03	.42**	−.19	−.04	–	.16	−.01	.10	.13
STM	.11	.06	.32*	−.04	−.22	–	.58***	.38*	.56**
CELF	.45**	.35*	.31	.45**	−.13	.26	–	.57**	.63***
ROWPVT	.15	.16	.36**	.04	.01	.22	.51**	–	.43*
Comp.	.18	.25	.41**	.00	.06*	.24	.31*	.50**	–

FB = False belief.

EF = Executive functioning.

DCCS = Dimensional card sort.

B/D = Bear/Dragon.

D/N = Day/Night.

STM = Short-term memory.

CELF = Clinical Evaluation of Language Fundamentals – Core Language subscale.

ROWPVT = Receptive One Word Picture Vocabulary Test.

Comp. = Complementation.

p < .10; **p < .05; ***p < .001.

To further explore these relations, we conducted a hierarchical regression analysis with the FB reasoning composite as the outcome measure. Block 1 included demographic variables (age and maternal level of education). Block 2 included language status (monolingual or bilingual) and general language ability (CELF) because it was the one measure significantly correlated with FB reasoning for both groups. Block 3 included interaction terms of language status by [variable] for only those measures that showed a differential relation to FB for the monolingual and bilingual populations in the correlation analyses described above. Specifically, interaction terms were created for: language status × STM, language status × complementation, and language status × Bear/Dragon EF measure. In this regression, age and general language ability were significant predictors of FB reasoning. Once all the variables in the model were included, language status was also a significant predictor. Finally, of the interaction terms only language status × Bear/Dragon EF measure was significantly related to FB (see Table 5). These results suggest that while we see a natural progression with age, general language ability is related to FB reasoning in both groups. In turn, EF function as reflected in the Bear/Dragon measure only has an effect for the monolingual group.

Table 5.

Summary of hierarchical regression analysis for variables predicting FB (N = 65).

Variable	B	SE B	β
Step 1 (R² =.35, p < .001)
AgeMother’s education	.20.31	.03.23	.57** .14
Step 2 (∆ R² =. 50, p < .05)
AgeMother’s education	.13–.03	.04.24	.37* –.013
Language status	.54	.60	.11
CELF	.09	.22	.50**
Step 3 (∆ R² =.54, ns)
AgeMother’s education	.14.001	.04.25	.39** .001
Language status	4.23	2.01	.82*
CELF	.11	.025	−.63**
Lang. status × STM	−.47	.44	−.24
Lang. status × complementation	.033	.10	.06
Lang. status × B/D	.31	.15	−.53*

Note: B/D = Bear/Dragon.

p < .05; **p < .01.

Discussion

There were two main objectives to the present inquiry. First we wanted to determine the presence of performance differences in our sample of bilinguals and monolinguals on FB reasoning, EF, and language ability. In a complex interaction among related factors, although bilinguals underperform on language proficiency measures compared to monolinguals (e.g., Bialystok & Viswanathan, 2009), they often score higher on EF measures (Bialystok, 2010; Carlson & Meltzoff, 2008; Kovács & Mehler, 2009; Martin-Rhee & Bialystok, 2008). Similarly, bilinguals have been shown to outperform monolinguals on FB reasoning measures, although this has been limited to a handful of studies (Goetz, 2003; Kovács, 2009; Nguyen & Astington, 2014).

The primary goal was to determine the ways in which EF and language proficiency (known predictors of FB in monolinguals) were related to FB in bilingual and monolingual preschoolers. In particular, we were interested in the possibility that bilinguals’ previously reported advantage in EF might be responsible for their advantage in FB reasoning. Is it the case that inhibiting one language for another facilitates the process of inhibiting one’s knowledge in favor of the knowledge of another? This is not what we found in our sample. In turn, we examined the role of language proficiency as a related factor for both groups, while features of EF were only related to FB reasoning for the monolingual sample. This asymmetry in the relation between FB and EF for the two groups (corroborated by the interaction in the regression analysis) flips on its head the prevailing hypothesis on the source of the bilingual advantage in FB as emanating from an advantage in EF. In a related study by Fan, Liberman, Keysar, and Kinzler (2015), the authors found that both bilingual and bilingually or multilingually exposed children outperformed monolinguals on a perspective taking social communication task. Importantly, performance on this task was unrelated to EF abilities as measured by the Dimensional Change Card Sort task. The authors concluded that enhanced social perspective taking in bilinguals is not driven by EF, but rather by the exposure to a multilingual environment. If indeed it is the case that bilinguals and monolinguals arrive at being able to reason about FBs in different ways and related to different factors, this forces a re-evaluation of our understanding of FB development from a theoretical standpoint, as will be discussed below.

Bilingual–monolingual differences

As expected, monolinguals significantly outperformed bilinguals on language ability using general language (CELF), receptive vocabulary, and memory for complementation tasks. After controlling for these language differences, bilinguals performed better than monolinguals on the FB tasks. Similarly, a study of English–French bilinguals found an FB advantage after controlling for language ability (Nguyen & Astington, 2014). This finding indicates that growing up in a bilingual environment has the potential to enhance FB related reasoning. We argue that bilingualism influences social cognition in positive ways.

Similarly, we predicted based on previous findings that bilinguals would perform better than monolinguals on EF inhibitory control measures (see Bialystok, 2010; Carlson & Meltzoff, 2008; Kovács & Mehler, 2009; Martin-Rhee & Bialystok, 2008). Only a marginal difference between the language groups was found for the composite EF measure, but not for the individual EF measures (see also Nguyen & Astington, 2014). It is possible that this difference was only marginal due to our relatively small sample size.

We argue that precisely because of the sizable complexity in acquiring two different languages at the same time and the limited access to language learning sustaining biases such as the mutual exclusivity bias (Davidson, Jergovic, Imami, & Theodos, 1997), bilinguals rely more extensively on socio-cognitive cues of the language environment. This particular language learning that they must negotiate through social cues, in turn, facilitates their construction of a ToM. Moreover, this framing of a bilingual’s construction of ToM as a result of their linguistic challenge speaks to the competence versus performance debate of the development of ToM. Is it the case that the competence is there since infancy but children are only able to perform on it once their EF abilities are sufficiently developed? Or does the competence itself emerge at that age? Our detection of a potential bilingual advantage influenced by linguistic factors and not by EF and cognitive mechanisms such STM suggest support for the competence account of FB reasoning from a constructivist perspective.

False belief reasoning relationships

The primary objective was to examine the relation of EF, and language proficiency to FB reasoning for both the monolingual and bilingual groups. In the regression analysis, general language (CELF) predicted FB reasoning for both groups while the interaction for language status × Bear/Dragon EF was significant, suggesting a differential effect for both groups. Thus, even though the mean performance of bilingual children was lower than the monolingual children on language measures, the bilingual children’s relative performance on FB reasoning tasks was predicted by their language proficiency in their dominant language. For monolingual children, while the relation to language proficiency is also strong, other cognitive factors such as STM and some aspects of EF are involved.

The particular aspects of language related to FB reasoning were also examined. In the correlations all the language measures were linked to FB in monolinguals, whereas for bilinguals only the CELF measure was related. In addition, once we control for general language ability using the CELF, the significant correlation between FB and complementation goes away for monolinguals. These findings contradict previous reports arguing for the unique role of complementation comprehension beyond general language (de Villiers, 2005; de Villiers & Pyers, 2002; Milligan et al., 2007). Instead, our findings support the claim that general language (composed of general grammar and semantic abilities) is sufficient for facilitating FB reasoning in both monolinguals as well as bilinguals (Cheung, 2006; Farrar, Benigno, Tompkins, & Gage, 2017; Slade & Ruffman, 2005).

The results also indicate that EF measures of inhibitory control did not account for bilinguals’ enhanced FB reasoning as prior studies have hypothesized but not examined (e.g., Goetz, 2003; Kovács, 2009). While inhibitory control (Bear/Dragon task) was related to monolingual children’s FB understanding, it was not related to bilinguals’ understanding. Interestingly, STM was related to performance in the monolinguals group but not the bilingual group. Some previous studies have shown that working memory, as measured by a backward digit span task, was related to FB reasoning in both monolinguals and bilinguals (Nguyen & Astington, 2014; Tardif et al., 2007; although see Slade & Ruffman, 2005). However, we used a forward rather than a backward digit span test. This forward digit span STM measure was intended to serve as a general cognitive ability control measure in place of an intelligence quotient test (Bialystok, 2010), and not as another EF measure. Together the results indicate both similarities and differences in the emergence of FB understanding in monolingual and bilingual populations.

We acknowledge some limitations of the current study. As discussed, a working memory backward digit span task would have been an appropriate addition to the EF measures. A larger sample size would have also afforded us more power in the analyses, such as in the near significance for a bilingual advantage in the EF composite. Another possible explanation for this latter finding, however, is that the bilingual sample was overwhelmingly dominant in the majority language (81.3%), thus more often inhibiting their non-dominant language. Finally, the need to expand this line of research is clear to us in order to draw further theoretical conclusions regarding the specific mechanisms related to the development of FB reasoning in bilinguals.

Conclusions

As far as we know, this is the first study to systematically and simultaneously examine the relative contribution of different language proficiency measures and EF measures in bilingual children’s FB understanding. These results suggest both common and different predictors of FB for the two groups. Specifically, while general language ability was related to FB reasoning for both groups, other cognitive measures such as STM and inhibitory control were only related for monolingual children.

Footnotes

Acknowledgements

This work was completed as part of the first author’s thesis submitted to the University of Florida. We want to thank Scott A. Miller and Linda Lombardino for their input, as well as the anonymous reviewers for their thoughtful commentary on the manuscript. This work would not have been possible without the undergraduate research assistants at the Cognitive Development Lab at the University of Florida, the preschool administrators who opened their doors to our research team, and the parents and children who volunteered.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The research was supported by the Jacquelin Goldman Graduate Student Research Award.

Notes

References

Baron-Cohen

Leslie

A. M.

Frith

(1985). Does the autistic child have a theory of mind? Cognition, 21, 37–46.

Bialystok

(2001). Bilingualism in development: Language, literacy, and cognition. Cambridge, UK: Cambridge University Press.

Bialystok

(2010). Global-local and trail-making tasks by monolingual and bilingual children: Beyond inhibition. Developmental Psychology, 46, 93–105.

Bialystok

Craik

F. I.

Klein

Viswanathan

(2004). Bilingualism, aging, and cognitive control: Evidence from the Simon task. Psychology and Aging, 19, 290–303.

Bialystok

Luk

Peets

K. F.

Yang

(2010). Receptive vocabulary differences in monolingual and bilingual children. Bilingualism: Language and Cognition, 13, 525–531.

Bialystok

Viswanathan

(2009). Components of executive control with advantages for bilingual children in two cultures. Cognition, 112, 494–500.

Brito

Barr

(2012). Influence of bilingualism on memory generalization during infancy. Developmental Science, 15, 812–816.

Carlson

S. M.

Meltzoff

A. N.

(2008). Bilingual experience and executive functioning in young children. Developmental Science, 11, 282–298.

Carlson

S. M.

Moses

L. J.

(2001). Individual differences in inhibitory control and children’s theory of mind. Child Development, 72, 1032–1053.

10.

Castro

D. C.

García

E. E.

Markos

A. M.

(2013). Dual language learners: Research informing policy. Chapel Hill: The University of North Carolina, Frank Porter Graham Child Development Institute.

11.

Cheung

(2006). False belief and language comprehension in Cantonese-speaking children. Journal of Experimental Child Psychology, 95, 79–98.

12.

Cheung

Hsuan-Chih

Creed

Wang

S. P.

(2004). Relative roles of general and complementation language and theory of mind development: Evidence from Cantonese and English. Child Development, 75, 1155–1170.

13.

Davidson

Jergovic

Imami

Theodos

(1997). Monolingual and bilingual children’s use of the mutual exclusivity constraint. Journal of Child Language, 24, 3–24.

14.

de Villiers

J. G

. (2005). Can language acquisition give children a point of view? In Astington

J. W.

Baird

J. A.

(Eds.), Why language matters for theory of mind (pp. 186–219). New York, NY: Oxford University Press.

15.

de Villiers

J. G.

Pyers

J. E.

(2002). Complements to cognition: A longitudinal study of the relationship between complex syntax and false-belief-understanding. Cognitive Development, 17, 1037–1060.

16.

Devine

R. T.

Hughes

(2014). Relations between false belief understanding and executive function in early childhood: A meta-analysis. Child Development, 85, 1777–1794.

17.

Diamond

(2013). Executive functions. Annual Review of Psychology, 64, 135–168.

18.

Fan

S. P.

Liberman

Keysar

Kinzler

K. D.

(2015). The exposure advantage: Early exposure to a multilingual environment promotes effective communication. Psychological Science, 26, 1090–1097.

19.

Farrar

M. J.

Benigno

Tompkins

Gage

(2017). Are there different pathways to explicit false belief understanding? General language and complementation in typical and atypical populations. Cognitive Development, 43, 49–66.

20.

Farrar

M. J.

Lee

Cho

Tamargo

Seung

(2013). Language and false belief in Korean-speaking and English-speaking children. Cognitive Development, 28, 209–221.

21.

Frye

Zelazo

P. D.

Palfai

(1995). Theory of mind and rule based reasoning. Cognitive Development, 10, 483–527.

22.

Gardner

(2000a). Expressive One-Word Picture Vocabulary Test. Novato, CA: Academic Therapy Publications.

23.

Gardner

(2000b). Receptive One-Word Picture Vocabulary Test. Novato, CA: Academic Therapy Publications.

24.

Gerstadt

C. L.

Hong

Y. J.

Diamond

(1994). The relationship between cognition and action: Performance of children 3.5-7 years old on a Stroop-like day-night test. Cognition, 53, 129–153.

25.

Goetz

(2003). The effects of bilingualism on theory of mind development. Bilingualism: Language and Cognition, 6, 1–15.

26.

Gordon

K. R.

(2016). High proficiency across two languages is related to better mental state reasoning for bilingual children. Journal of Child Language, 43, 407–424.

27.

Hughes

Ensor

(2007). Executive functions and theory of mind: Predictive relations from ages 2 to 4. Developmental Psychology, 43, 1447–1459.

28.

Kovács

A. M.

(2009). Early bilingualism enhances mechanisms of false-belief reasoning. Developmental Science, 12, 48–54.

29.

Kovács

A. M.

Mehler

(2009). Cognitive gains in 7-month-old bilingual infants. Proceedings of the National Academy of Sciences, 106, 6556–6560.

30.

Martin-Rhee

M. M.

Bialystok

(2008). The development of two types of inhibitory control in monolingual and bilingual children. Bilingualism: Language and Cognition, 11, 81–93.

31.

Milligan

Astington

J. W.

Dack

(2007). Language and theory of mind: Meta-analysis of the relation between language ability and false-belief understanding. Child Development, 78, 622–646.

32.

Montgomery

D. E.

Koeltzow

T. E.

(2010). A review of the day–night task: The Stroop paradigm and interference control in young children. Developmental Review, 30, 308–330.

33.

Morton

J. B.

Harper

S. N.

(2007). What did Simon say? Revisiting the bilingual advantage. Developmental Science, 10, 719–726.

34.

Nguyen

Astington

(2014). Reassessing the bilingual advantage in theory of mind and its cognitive underpinnings. Bilingualism: Language and Cognition, 17, 396–409.

35.

Paap

K. R.

Greenberg

Z. I.

(2013). There is no coherent evidence for a bilingual advantage in executive processing. Cognitive Psychology, 66, 232–258.

36.

Perner

Leekman

Wimmer

(1987). 3-year-olds’ difficulty with false-belief: The case for a conceptual deficit. British Journal of Developmental Psychology, 5, 125–137.

37.

Poulin-Dubois

Blaye

Coutya

Bialystok

(2011). The effects of bilingualism on toddlers’ executive functioning. Journal of Experimental Child Psychology, 108, 567–579.

38.

Reed

Pien

D. L.

Rothbart

M. K.

(1984). Inhibitory self-control in preschool children. Merrill-Palmer Quarterly, 30, 131–147.

39.

Rubio-Fernández

Glucksberg

(2012). Reasoning about other people’s beliefs: Bilinguals have an advantage. Journal of Experimental Psychology: Learning, Memory, and Cognition, 38, 211–217.

40.

Slade

Ruffman

(2005). How language does (and does not) relate to theory of mind: A longitudinal study of syntax, semantics, working memory, and false belief. British Journal of Developmental Psychology, 23, 117–141.

41.

Tardif

Kaciroti

(2007). Language and false belief: Evidence for general, not specific, effects in Cantonese-Speaking preschoolers. Developmental Psychology, 43, 318–340.

42.

United States Census Bureau. (2009). Selected social characteristics in the United States: 2007 (Archived from the original on April 25, 2009). Retrieved from http://www.chamberorganizer.com/millvalleycc/docs/Social_Characteristics.pdf

43.

Wellman

Cross

Watson

(2001). Meta-analysis of theory of mind: The truth about false belief. Child Development, 72, 655–684.

44.

Wiig

E. H.

Secord

W. A.

Semel

(2004). Clinical evaluation of language fundamentals (2nd ed.). San Antonio, TX: Harcourt Assessment.

45.

Wimmer

Perner

(1983). Beliefs about beliefs: Representation and constraining function of wrong beliefs in young children’s understanding of deception. Cognition, 13, 103–128.