A corpus-assisted discourse study of parental concerns regarding multilingual child-rearing

Abstract

Many parents have concerns about raising their children with multiple languages. However, there is a paucity of previous research regarding parental concerns about multilingual child-rearing, particularly in multilingual societies. We address this gap with a corpus-assisted discourse study of parental concerns regarding multilingual child-rearing in Quebec, Canada. We created two corpora comprising 641 parents’ responses to an open-ended survey question regarding their main concerns about raising multilingual infants/toddlers (English corpus: 12,502 words, French corpus: 9,360 words). We examined frequencies, collocations, concordance lines, and longer segments to investigate the nature and strength of different concerns. Our results revealed that two previously-attested concern types — cognition concerns and exposure-fluency concerns — were most prominent. The results also provided more nuanced insights into the nature of these concerns. Moreover, the results revealed two additional concern types: concerns regarding trilingual/heritage language transmission and concerns about the effect of multilingualism on children’s identity and social/emotional well-being. These had not previously been attested. Our research makes a theoretical contribution by advancing knowledge about parental concerns and how they can contribute to the study of family language policies. Additionally, our findings may serve as the basis for improving support for multilingual families.

1. Introduction

Children who grow up to be proficient users of multiple languages reap both economic and social benefits (Seward et al., 2021; Tannenbaum, 2005). Many parents are aware of these potential benefits and, as a result, are motivated to raise their children to be competent multilinguals (Byers-Heinlein & Lew-Williams, 2013). However, some parents also have concerns about their children’s ability to achieve this goal and about the challenges their children may face along the way (King & Fogle, 2006a; Piller & Gerber, 2018). Parental concerns have been defined as feelings of sympathy and care for their child (Algarvio & Leal, 2016), and fear and uncertainty regarding their child’s future (King & Fogle, 2006a, b). They may influence parents’ decisions regarding their language use with their children (Braun & Cline, 2010). Parents’ language use with their children, in turn, plays a key role in their multilingual development (De Houwer, 1999). However, research investigating parental concerns is limited. Parental language-related concerns have primarily been investigated for their ability to identify early language delay (e.g. O’Toole et al., 2016; Chung et al., 2011), and they have largely focused on only one type of concern: language development. Moreover, these concerns are often treated categorically in that parents are either deemed concerned or not (e.g. Miękisz et al., 2016). Studies of parental concerns about multilingual childrearing have focused almost exclusively on parents raising multilingual children in predominantly monolingual societies. In these studies, parents have expressed a mixture of optimism about children’s ability to handle multilingual exposure and concern that children may be delayed compared to monolingual peers (e.g., King & Fogle, 2006b; Piller & Gerber, 2018). In societies where multilingualism is widespread at the community level, comparisons with monolingual peers may be less prevalent, but parents in such societies might have other concerns.

The current study takes place in the Canadian province of Quebec. French is the sole official language in Quebec and is spoken by the majority of the province’s population, but there is also a sizable English-speaking minority officially recognized under Canadian law. Both French and English effectively function as societal languages in large parts of the province, and both are held in high esteem (Kircher, 2014, 2016). The province is also home to a large and growing number of heritage language (HL) speakers (Statistics Canada, 2023) – that is, speakers of languages that are not the dominant language in a given social context. Approximately 1 of 6 Quebec children under 4 is raised with multiple languages in the home (Schott et al., 2022).

Parental concerns in this context have rarely been directly studied; however, in one study, mothers raising children in Montreal (Quebec’s urban center) in linguistically mixed families with Japanese as a HL were interviewed (Tsushimi & Guardado, 2019). These mothers expressed concerns centered on ‘the potential negative effects on their children’s linguistic, cognitive, and academic capabilities as a result of their multilingual lives’ (Tsushimi & Guardado, 2019, p. 319). For example, mothers worried about the negative impact of their use of the HL on children’s societal language development. More recently, we conducted two studies investigating concerns in the Quebec context, which included a wider sample: parents raising young children with more than one language at home. In the first study, we conducted focus groups and interviews with 27 parents living in Montreal (Ballinger et al., 2022). Thematic analysis of the resulting qualitative data revealed that participants generally expressed weak concerns regarding multilingual child-rearing. However, comparatively stronger concerns were noted among parents raising their children with HLs in addition to English and/or French. These parents were concerned they would not be able to support their children’s development in the HLs sufficiently and/or that their children might later refuse to speak the HLs, potentially leading to conflict or impeded communication within the family. The findings from this study motivated a subsequent questionnaire-based study among parents throughout Quebec, focused specifically on parents’ concerns regarding their children’s multilingual development (Quirk et al., 2023). In that study, we administered the Parental Concerns regarding Multilingual Development Questionnaire (PCMD) to 821 parents raising children aged 0-4 years with multiple languages in the home.² Factor analysis of Likert-scale ratings of 11 statements regarding concerns about their children’s multilingual development revealed two factors: (1) cognition concerns – which center on the potential for multilingual exposure to lead to cognitive difficulties in children, such as confusion and language delay, and (2) exposure/fluency concerns – which relate to the quality and quantity of children’s language exposure, and whether their language environment supports their attainment of fluency in their languages. These factors accounted for 66% of the variance in parents’ responses, indicating their importance – but also leaving a sizable portion of the variance unaccounted for, suggesting that other concern types exist.

Building on this previous research, the current study contributes to our understanding of parental concerns regarding multilingual development and other aspects of multilingual child-rearing by exploring other potential concern types and shedding more light on the concern types revealed in our earlier study. It does so through a corpus-assisted discourse study of qualitative data obtained with the only open-ended question in the aforementioned PCMD questionnaire, which we had also used to elicit the quantitative data (Quirk et al.,2023). The open-ended question tapped parents’ main concerns about raising their children with multiple languages, and we examined parents’ responses to address the following research questions:

RQ1: What is the nature of parents’ concerns regarding multilingual child-rearing?

To answer this question, we investigated whether the two previously-attested concern types revealed in the quantitative analyses would also emerge from the qualitative data elicited by the open-ended question – and whether additional concern types would emerge.

RQ2: What are the strongest concerns that parents have regarding multilingual child-rearing?

To answer this question, we investigated whether certain concerns would feature more prominently than others in the qualitative data, which would suggest that these concerns are particularly strong.

Regarding RQ1, we predicted that parents’ open-ended responses would indeed reveal the previously-attested concern types as well as additional concern types. Regarding RQ2, we predicted that some concerns would figure more prominently than others in parents’ responses. However, given the paucity of research on concerns, we did not make specific predictions about what additional concern types would emerge, nor about the nature of the differences in prominence across concern types.³

Addressing these questions is important for both practical and theoretical reasons. In terms of practical significance, information about parents’ concerns is urgently needed to provide high-quality support to the large number of families raising their children multilingually. For example, this information could inform the development of tailored parenting resources that specifically address parents’ strongest concerns. This support is critical for the well-being of caregivers and children, as well as for supporting HL maintenance since parents’ concerns influence their decisions about intergenerational language transmission (Braun & Cline, 2010). Furthermore, our focus on a lesser-studied social context, namely a multilingual society, builds the empirical basis for developing resources that are adapted specifically to families in such contexts, which will in turn make these resources more effective (Fibla et al., 2022). In terms of theoretical significance, this study takes important first steps towards building a theory of parental concerns regarding multilingual child-rearing. Parental concerns have been investigated indirectly through related concepts such as parental expectations (e.g. Curdt-Christiansen, 2009) and expressions of stress and tension related to multilingual child rearing (e.g. Okita, 2002); however, these studies do not systematically investigate the construct of concerns as a distinct type of belief. Addressing this gap by integrating concerns into a theoretical framework for multilingual child-reading is important because parents’ concerns influence their language use with their children, and parents’ language use in turn plays a key role in shaping children’s multilingual development. In this study, we argue for the integration of parental concerns into one particular framework – namely, family language policy (FLP). In the following section, we elaborate on this. We summarize the existing literature which touches on parental concerns regarding multilingual child-rearing, and we discuss how such concerns fit into the proposed theoretical framework, arguing that concerns are an essential yet unrecognized and understudied aspect of family language policy.

2. Theoretical Framework

FLP refers to “explicit and overt planning in relation to language use within the home and among family members” (King et al., 2008, p. 1). In recent years, FLP has emerged as an effective lens through which multilingual child-rearing can be studied. It draws on and contributes to two distinct fields of study: child language acquisition and language policy (King et al., 2008). Based on the latter, the “inner core” of FLP includes caregivers’ language beliefs, language practices, and language management (Curdt-Christiansen & Huang, 2020, p. 175). The focus here is on language beliefs.

Previous research has investigated parents’ language beliefs from different perspectives: their language ideologies (e.g. Curdt-Christiansen, 2009), their attitudes towards the specific languages they are transmitting to their children (e.g. Kircher, 2022), and their attitudes towards childhood multilingualism (Kircher et al., 2022, 2024).⁴ Some studies have also examined attitudes thematically related to concerns, like the belief that speaking two languages at home can confuse children (e.g. Kang, 2015). However, concerns are arguably different from such attitudes. As noted above, parental concerns are characterized by fear and uncertainty about children’s future (King & Fogle, 2006a, b). A distinguishing feature of parental concerns, thus, appears to be that they are linked to situations and future outcomes specific to the parent’s own situation (and that of their children). In contrast, related constructs such as attitudes are more general and can be positive or negative. Given this distinct nature, we argue that concerns should be integrated into the framework of FLP as part of the language beliefs component, as a separate belief type to be studied alongside language ideologies and attitudes. This study begins to do so by building on the aforementioned research by Quirk et al. (2023), which provided important initial insights regarding the nature and structure of parents’ concerns about their children’s multilingual development. Specifically, this study provides more comprehensive and nuanced insights into the underlying structure and sheds light on further concern types. This new information advances our understanding of the construct of concerns, thereby setting the groundwork for its integration into theories of FLP. This constitutes the principal theoretical contribution of the study.

3. Methodology

3.1. Data Collection

The qualitative data analyzed in this study were obtained via the PCMD’s only open-ended question, item 50: In your own words, could you describe your main concern(s) about raising your child with more than one language? The PCMD, the pre-registration for this study, the datasets, and supplementary materials are on the OSF: https://osf.io/bhsf4/.

From September 2020 to February 2021, the PCMD was distributed online via databases of Quebec-based parents interested in research, mailing lists, traditional and social media, and snowball sampling. Parents could choose to respond in either French or English. Item 50 yielded data from 641 parents raising children aged 0–4 years with multiple languages in the home. Notably, 59% of participants were raising their children with at least one HL, and 31% with three or more languages.

3.2. Data Analysis

The data elicited with item 50 were used to create two corpora: an English corpus (12,502 words, based on 375 participants’ responses) and a French corpus (9,360 words, based on 266 participants’ responses). We used the corpus-linguistic software AntConc (Anthony, 2019) to conduct a study that broadly falls under the remit of corpus-assisted discourse studies (Kircher & Fox, 2021; Kircher & Kutlu, 2022; Vessey, 2016). Corpus-linguistic analysis can effectively reveal discursive patterns in large amounts of data (Baker, 2010). Specifically, analyzing frequencies provides important insights because “words that are repeated … have a particular function within the society producing the texts” (Vessey, 2016, p. 5). We thus began with a frequency analysis, using deductive coding to determine frequently-occurring terms pertaining to the already established concern types: cognition concerns and exposure/fluency concerns. Then, we employed inductive coding to establish semantic categories pertaining to additional concern types. Subsequently, we performed a collocation analysis to establish the words that co-occur frequently with the following lemmas: CONCERN* and WORRY* as search terms in the English corpus, and PRÉOCCUPER* and INQUIÉTER* (equivalent to concern and worry) in the French corpus. We employed the standard window of five words to the left and right of the search term and looked at their Mutual Information (MI) scores, which use a logarithmic scale to express “the ratio between the frequency of the collocation and the frequency of random co-occurrence of the two words in combination” (Gablasova et al., 2017, p. 163). As is customary, MI scores above 3 were considered significant (e.g., Vessey, 2017). To obtain more nuanced insights into parents’ concerns, we complemented these quantitative corpus-linguistic procedures with qualitative analyses. Specifically, we explored concordance lines, which “present an individual lexical item within its ‘co-text’ across numerous texts” (Vessey, 2016, p. 6). Finally, we examined larger segments (entire responses). Examining frequencies, collocations, concordance lines, and larger text segments enabled us to address RQ1. Comparing frequencies relating to the different concern types enabled us to address RQ2.

4. Results

4.1. Frequencies

The frequency analysis revealed the presence of not only the previously-attested concern types, cognition concerns and exposure/fluency concerns, but also two additional concern types, which we will refer to as HL/trilingualism concerns and identity/well-being concerns. The words pertaining to each concern type are shown in Table 1, along with their frequencies, both absolute and normalized to words per ten thousand (wptt), to facilitate comparisons.

Table 1:

Categorization, with absolute frequencies and relative frequencies in words per ten thousand (wptt) across corpora

Category	Category-related words in the English corpus	Category abs. freq.	Category rel. freq. (wptt)	Category-related words in the French corpus	Category abs. freq.	Category rel. freq. (wptt)
Cognition	delay(-s)(-ed) (43); school (32); confuse(-ion)(-ed) (22); read(-ing) (21); development (16); write(-ing) (15); slow(-ly) (5); overwhelm (3); handle (2); distinguish (2)	161	128.8	(re)tard (35); difficile/difficulté(s) (30); développe(-ment)(-er) (19); école (15); distinguer/distinction (9); mélange (9); écrire (8); scolaire(s) (8); écrit(-ure) (4); confus(-ion) (5)	142	151.7
Exposure/Fluency	exposure(-ed) (44); teach(-ing) (24); fluent(-cy) (21); words (20); proficient(-cy) (11); proper(-ly) (8); grammar (6); skills (5); mistakes (4); sentences (4); master (4); balance(-ed) (3); completely (2); knowledge (2)	158	126.4	exposé(e)(s)/exposition/exposer (25); mots (17); maîtrise(-er) (-é) (17); niveau (11); suffisamment/suffisant(e) (11); vocabulaire (11); phrases (7); couramment (7); occasions (5); perde (5); erreurs (2)	118	126.1
HL/Trilingualism	Spanish (25); Italian (10); three/3/third (language(s)) (10); Arabic (4); Chinese (4); ; Portuguese (2); German (2); trilingual (1)	58	46.4	espagnol(e) (28); trois / 3 (langues) (6); arabe (5); portugais (5); khmer (4); dari (3); turc (3); trilingue (2); euskara (1); minoritaire (1)	58	62.0
Identity/ Well-being	feel (9); comfortable / comfort (7); accent (5); peers (5); anglophone (5); groups (5); culture(s) (5); confident (5); friends (4); identify(-ity) (3); bullying (1); barrier (1)	55	44.0	frustrante/frustration(s)/frustrer (6); culture(-l)(-s) (4); discrimination (2); émotions (1)	13	13.9

As noted above, cognition concerns center on the potential for multilingual exposure to lead to cognitive difficulties in children, such as confusion and language delay. In the open-ended responses, the presence of this concern type is reflected, for example, in parents’ repeated use of the lemmas delay*/retard*,⁵ confusion*/confusion* and distinguer* (distinguish), which indicates that parents were concerned their children might experience language delay and/or not be able to tell their languages apart. Notably, some frequently-used words in the open-ended responses provided more nuanced information than we had been able to gather from the quantitative analyses of the closed items in Quirk et al. (2023). For example, the frequent mention of school/école indicated that parents were particularly concerned about children’s multilingual exposure detrimentally affecting their performance once they start school.

As explained above, exposure/fluency concerns relate to the quality and quantity of children’s language exposure and whether the children’s language environment supports their attainment of fluency in their languages. In the open-ended responses, the presence of this concern type is reflected, for example, in parents’ repeated use of the lemmas exposure*/exposition* and master*/maîtriser*, indicating their worry that children’s language exposure would fail to support mastery of their languages. Again, the frequently-used words provided additional detail about the concerns tapped by the previous quantitative analysis in Quirk et al. (2023). For instance, parents’ reference to teach and homework indicated a concern about their ability to support their children in their school work. In the earlier quantitative analysis of concerns, parents were found to be moderately concerned about their ability to support their children’s language development. Thus, this finding sheds light on a specific aspect of that concern. This is explored further in our analysis of longer text segments (below).

The first of the additional concern types revealed by the qualitative data, HL/trilingualism concerns, is thematically related to exposure/fluency concerns. However, while the latter relates to exposure/fluency at a general level, the former is specific to the challenges of children acquiring a third and/or HL. The most frequent word we found in this category was Spanish/espagnol, which reflects the most frequently-transmitted HL by the parents in this study. Several other HLs were mentioned repeatedly, including Italian, Arabic/arabe, and Portuguese/portugais. Additionally, parents repeatedly wrote three/trois and third (language) to express concerns specific to trilingual acquisition. This indicates that parents’ concerns focused on languages their children are acquiring that are not spoken widely in the community and/or on difficulties associated with acquiring three or more languages.

The second additional concern type revealed by the qualitative data, identity/well-being, relates to parental concerns about the effect of multilingualism on children’s identity and their emotional or social well-being. The presence of this concern type is reflected, for example, in participants’ frequent use of terms such as peers, friends, and groups in the English corpus, and the lemma culture*/culture* in both corpora, indicating that parents worried about how multilingualism might impact their children’s ability to form social bonds and develop a sense of belonging. Relatedly, words such as bullying in the English corpus and discrimination (discrimination) in the French corpus suggested parents were concerned about their children being targeted by such acts due to their multilingualism. Words in this category also referred to potential effects of multilingualism on children’s emotional health. Such words included the term confident and the lemmas FRUSTR* (frustrate/frustration/frustrating) and ÉMOTIONS* (emotions), which indicated concern that children would suffer from lower self-confidence and more frustration due to their multilingual exposure. For example, some parents were worried that their children would find it difficult to communicate with teachers, peers, and non-family members in one of their languages, which could lower their self-confidence and cause frustration.

Figure 1 summarizes the frequency findings in both corpora combined.⁶ It shows that words expressing cognition concerns and exposure/fluency concerns were used with similar frequency to one another – and much more frequently than words relating to HL/trilingualism concerns and identity/well-being concerns.

Figure 1: Word frequencies by semantic grouping.

4.2. Collocations

Our search terms for the collocation analysis in the English corpus were the lemmas CONCERN* and WORRY*, which appeared 157 and 106 times, respectively. The search terms in the French corpus were the corresponding lemmas INQUIÉTER* and PRÉOCCUPER*, which appeared 89 and 59 times, respectively. The collocation results were in line with the frequency findings, providing evidence for all four concern categories: (1) cognition concerns – e.g. delays (MI=5.70) and retard (delay, MI=4.33); (2) exposure/fluency concerns – e.g. exposing (MI=7.81) and apprenne (learn, MI=5.74); (3) HL/trilingualism concerns – e.g. Spanish (MI=3.64) and japonais (Japanese, MI=6.33); and (4) children’s identity and emotional/social well-being – e.g. communication (MI=7.08; no significant collocates in the French corpus). A table containing all significant collocations by concern type can be found in the supplementary materials on OSF.

4.3. Concordance lines and larger text segments

The qualitative analyses of concordance lines and selected longer segments from parents’ responses lend support to the findings from the quantitative corpus-linguistic procedures – and they provide more detailed and nuanced information about parents’ different concerns.

The concordance lines for delay (Figure 2) shed light on the specific nature of cognition concerns. For example, in the context of concerns about language delay, parents also mentioned other cognitive difficulties, such as stuttering and the need to see a speech therapist. Moreover, parents showed concern that code-switching could cause delays or confusion. This is seen, for example, in the concordance line mentioning the use of “frenglish” (Figure 2). The following segment also exemplifies this:

Figure 2:

Selected concordance lines for delay

(1) Confusion and initial communication delay are my main concerns. At home we use the 3 languages with no specific differentiation, and sometimes we use 2 languages in the same sentence.

Here, we also see another recurring theme: namely, that delays occur initially and would only be fleeting. Parents frequently used phrases such as “initial” or “short-term” when discussing their concerns about delays – and this sentiment is often cited as a reason why this concern is only minor, as in the following segment:

(2) Un retard de language est mon plus grand préoccupation, mais je ne m'inquiète pas beaucoup. Il va se retraper au fur et à mesure. (Language delay is my biggest concern, but I don’t worry much. With time, he’ll catch up.)⁷

Nonetheless, the data indicated that even slight concerns about language delay affected parents’ behavior. One participant even stated that this concern dissuaded her and her partner from using one of their mother tongues, as shown in this segment:

(3) I was slightly concerned about short-term language delays. That's one of the reasons that we didn't add a third "mother-tongue" – I wasn't sure that my clinical side could handle the possibility of language delay, even if I was prepared for it and knew that it wouldn't be an issue in the long-term.

Thus, while parents may “know” that the delays will eventually go away, the possibility of a delay in the early stages of development was still viewed as a negative consequence to be avoided. Indeed, the conflict between “logical” thoughts/knowledge about their children’s multilingual development and “irrational” fears and concerns was a recurrent theme in parents’ mentions of cognition concerns. For example:

(4) Je SAIS que le bilinguisme vient naturellement aux jeunes enfants, mais je ne peux m'empêcher de trouver cela mélangeant pour elle. J'ai peur que mes inquiétudes (irrationnelles) influencent négativement la façon dont je lui apprends à parler. (I KNOW that bilingualism comes naturally to young children, but I can’t help but think it’s confusing for her. I worry that my (irrational) concerns negatively impact the way that I teach her to speak.)

The concordance lines for enough (Figure 3) shed light on the nature of parents’ exposure/fluency concerns. These included, for instance, concerns about parents’ ability to support their child’s language development due to their own lack of language proficiency. Examination of longer segments revealed an additional detail – namely that some parents also worried that their own status as (perceived) outsiders may have detrimental effects on their children, as illustrated by this segment:

Figure 3:

Selected concordance lines for enough

(5) Will be hard to help with homework, and probably embarrassing to them when I speak in front of their friends/others.

Another detail that emerged from the concordance lines for enough was that parents worried in particular about the effect that multilingual exposure might have on their children after school entry. As noted above, this had already been indicated by the frequent mention of school/école (Table 1). Additionally, as seen in the segment below, some parents worried that problems at school would then translate into emotional or social difficulties for their child:

(6) Not being fluent enough by the time they reach school, as was the experience with my eldest; while she is learning French at a faster rate, she also experiences lack of confidence, lack of feeling understood and gravitates to other English children as she feels like an outside in the French school system.

Another aspect of parents’ exposure/fluency concerns was seen in the concordance lines for accent (Figure 4): parents were particularly concerned about their children’s pronunciation skills, worrying that their children might not attain “proper” and “authentic” accents. As with concerns about difficulties at school, these concerns were also linked to identity/well-being concerns, as exemplified by this response from a parent who expressed their desire for their child to be welcomed by the school language community:

Figure 4:

Selected concordance lines for accent

(7) I worry that my children will have an accent in English and not be fully accepted by the Anglophone community.

In addition to the concern that non-native pronunciation could lead to a lack of acceptance by one particular language community, examination of longer segments revealed that parents worried about children’s multilingual and multicultural backgrounds precluding them from finding a sense of belonging in any community. This is illustrated by this segment from a participant who had experienced this herself:

(8) There were times I didn't know certain expressions in Spanish nor English because I was always jumping back and forth between different identities / cultures / groups. I am also concerned that my children will be judged or experience they don't belong to any groups. I have experienced rejection from Latino groups because they see me as Canadian and vice versa.

The concordance lines for friends (Figure 5) also shed light on nuances in the identity/well-being concerns and links between these and exposure/fluency concerns. For example, parents worried about the lack of friends from one of their language/cultural groups and difficulties making and communicating with friends at daycare. As in the concordance lines for delay, one finds mentions of children’s ability to keep up with monolingual peers. Inspection of the larger segments additionally revealed that some parents worried about their relationship with their child suffering due to their child’s multilingualism, as in this segment:

Figure 5:

Selected concordance lines for friends

(9) It makes me sad to think that we might never have a shared "best" language and that our communication will be impeded as a result.

The concordance lines for enough (Figure 3) revealed parents’ concern for both community language and HL acquisition, as shown in segment 10:

(10) Since I am the only person who interacts with my child in French on a daily-level, I worry that I am not able to teach him the language sufficiently for him to be comfortable entering French-language school in the future.

However, parents frequently expressed stronger concern for the HL, as evinced by segment 11:

(11) Pour l'anglais et le français je n'ai pas d'inquiétude puisque les opportunités de parler et d'entendre ces langues sont abondantes au Québec. Pour le japonais, je m'inquiète que les opportunités ne soient pas suffisantes à l'extérieur de la maison et qu'il perde l'intérêt. (As for English and French, I’m not worried because there are abundant opportunities to hear and speak these languages in Quebec. As for Japanese, I worry that the opportunities out of the home will be insufficient and that he’ll lose interest.)

Finally, the concordance lines for three/3/third (Figure 6) revealed repeated mention of concerns related to children’s developing trilingualism, such as difficulty attaining sufficient exposure and achieving fluency in three or more languages. Examination of longer segments further revealed parents’ worry about children not attaining fluency in any of their languages, as illustrated here:

Figure 6:

Selected concordance lines for 3/three/third

(12) Il en résulte qu'il [mon conjoint] ne maîtrise pleinement aucune de ces 3 langues. Je suis donc inquiète que la même chose se produise avec notre enfant. (As a result, he [my partner] has not fully mastered any of his three languages. So I’m worried that the same thing will happen with our child.)

Finally, as with other concern types, the longer segments revealed a link between HL/trilingualism concerns and identity/well-being concerns. For instance, in the following segment, one parent worried about strained relationships between HL-speaking parents and children:

(13) Son refus de parler espagnol vient souvent du fait qu'elle comprend moins bien et ça la frustre, elle préfère que je lui parle en français. (Her refusal to speak Spanish often comes from the fact that she understands it less well, and that frustrates her. She prefers that I speak to her in French.)

5. Discussion

In this study, we investigated the nature of parents’ concerns about multilingual child-rearing (RQ1) and the relative prominence of their concern types (RQ2). For RQ1, our results affirmed a central finding from the earlier quantitative study among the same parents, Quirk et al. (2023): namely that parents had concerns about the effect of multilingualism on children’s cognition and concerns regarding children’s exposure to and fluency attainment in their languages. Parents’ concerns for children’s cognition related to mixing in the input, confusion, and – most frequently – language delay. The most prominent exposure-fluency concerns included parents’ potential inability to support children’s language development and children’s (resulting) failure to attain fluency in one/all of their languages.

In addition, the analysis of parents’ open-ended responses provided novel insights into the nature of these concern types and built on related work in FLP research. For instance, for cognition concerns, parents repeatedly mentioned conflict between what they “know” about multilingual development, namely that it usually proceeds without significant difficulty and is generally successful, and “irrational” feelings and anxiety about the consequences of multilingualism on their child’s cognitive development. These sentiments were frequently expressed in the context of concerns related to language delay. This is perhaps surprising, given that prior work reports highly positive attitudes regarding the cognitive benefits conferred by childhood multilingualism (Kircher et al., 2022, 2024; Piller & Gerber, 2018). Our results suggest that such attitudes are not incompatible with concerns regarding cognitive difficulties incurred by multilingualism. Rather, parents seemed to view the cognitive benefits of multilingualism as being both orthogonal to and compensatory for the challenges their children might face in terms of language development. Under the view that concerns constitute a construct that is separate from attitudes, this finding is more readily interpretable: parents’ attitudes about the abstract, and in some cases long-term, benefits of multilingualism are distinct from their concerns about their and their child’s more immediate experiences and outcomes.

Thus, our findings support our aforementioned argument that concerns should be integrated into the framework of FLP as a separate construct. For example, parental concerns could readily be integrated into current models of FLP such as the Dynamic Model (Curd-Christiansen, 2009; Curdt-Christiansen & Huang, 2020) as a separate type of belief. In the Dynamic Model, families’ beliefs are influenced by external factors, including sociolinguistic, socio-economic, and socio-cultural factors, and by internal factors, such as families’ emotions and identity, parents’ impact beliefs, and families’ culture. These relationships are echoed in many of the longer segments expressing parental concerns. For example, we found that parents’ concerns were sensitive to the status of the languages being transmitted in society, with concerns being concentrated on HLs (segment 11). We also found that parents’ concerns reflected aspects of educational policy in the province, specifically, the exclusive use of French in the vast majority of public schools (segments 6 and 10). Concerns also fit nicely into the beliefs component of the Dynamic Model in that they are influenced by family-level factors, for example, harmonious parent-child relationships (segment 9) and parents’ impact beliefs, including their knowledge of multilingual development (segment 4) and their view of their own role in supporting children’s language development (segment 10). Importantly, while concerns are shaped by some of the same factors as other types of beliefs, such as attitudes, notably by one’s past experiences (e.g. Kircher & Fox, 2019), we argue that they are distinct in that they have as a focus the present and future experiences of individual children/parents. Thus, even when thematically related, attitudes and concerns may diverge, as has been reported by us and others regarding the relationship between multilingualism and cognitive development.

One aim of this study was to view concerns in a new context —a society where multilingualism is the norm rather than an exception. Our findings, particularly for cognition concerns, pointed more to similarities than differences. For example, parents raising young multilinguals in more monolingual societies also expressed both the belief that multilingualism confers cognitive advantages and the concern that their children would not keep pace with monolingual peers (e.g. Piller & Gerber, 2018). Thus, contrary to expectation, widespread multilingualism at the societal level (as it exists in Quebec) did not appear to render monolingual comparisons less salient. Notably, however, even in Quebec, children raised multilingually in the home, and thus exposed to multiple languages before school entry, are still a minority (16.7%; Schott et al., 2022). Possibly this minority status caused parents to fear that their children would perform below average at school due to their multilingualism. While this remains to be ascertained, it leads us to another theme in parents’ concerns revealed in the qualitative data: namely, concerns related to school readiness.

The frequency analysis revealed that both cognition and exposure-fluency concerns were linked to school-related concerns. In Quebec, all children can attend public school in French while access to English instruction is restricted to certain groups, and public school options in HLs are unavailable. Consequently, many children in Quebec will start school in their non-dominant language. This situation was mentioned repeatedly as a source of concern. Specifically, parents were concerned that cognitive issues incurred by children’s multilingualism could lead to difficulties at school, including frustration, teacher prejudice, and an inability to keep up with classmates. Parents also expressed concern that they would be unable to fully support their children at school due to their own language skills, which could impede their ability to communicate with teachers and support their children’s homework. Thus, these parents had school-related concerns centered on their own abilities and their children’s abilities. In line with this finding, in a related study conducted among these same parents, we found a strong desire for resources that will prepare children for school – even though they are not yet school-age (Quirk et al., in prep.). Our findings also suggest that some parents felt a lack of comprehension and support from teachers and schools in Quebec, and this applied in particular to parents who did not grow up with the school language(s) themselves. Indeed, in a previous study, Mandarin-speaking parents reported being less engaged than English-speaking parents with their children’s French-language public schools in Montreal (Riches & Curdt-Christiansen, 2010) – where, as previously noted, French-English bilingualism is widespread. Mothers of Japanese-acquiring children in Montreal also expressed concern regarding the potential negative impact of multilingualism on their children’s academic lives (Tsushimi & Guardado, 2019). This is in line with a theme that emerges from prior studies with the participants from the present study (Ahooja et al., 2022; Kircher et al., 2022, 2024; Quirk et al., 2023): namely, that the experiences of HL-transmitting parents differ from those of parents transmitting English and French. We return to this point below.

For RQ1, in addition to cognition concerns and exposure/fluency concerns, our analyses also revealed two further concern types: HL/trilingualism concerns and identity/well-being concerns. While some such concerns have emerged from previous research (e.g. Tsushima & Guardado, 2019), these types have not been explicitly identified and labelled in prior work, thus constituting a novel contribution of this study. HL/trilingualism concerns, although conceptually linked with exposure-fluency concerns, are distinct in that they focus on the challenges particular to acquiring a HL and/or three or more languages. At times, parents themselves made this distinction explicit by stating that their concern was concentrated on a particular HL, and not French or English. The themes that emerged in this concern type were lack of opportunity and motivation for children to use a HL/third language, the particular difficulty of maintaining a HL/third language after school entry, and potential parent-child conflict. The distinction in concerns about French-English bilingualism on the one hand and multilingualism involving HL(s) is likely linked to the socio-political context of Quebec; specifically, that French is the official and dominant language of the province, and English is the largest and only language minority to receive state support in the form of public schooling in their language. Parents in this context may feel pressure to privilege the transmission of English and French, the two societal languages, effectively necessitating trilingual transmission in order to maintain their HLs. Indeed, in earlier work with mothers raising children with Japanese as a HL, such pressure pushed some mothers to reduce or stop HL use with children (Tsushima & Guardado, 2019). Despite all mothers expressing positive attitudes towards HL transmission, the authors concluded that these pressures made the multilingual society of Montreal non-conducive to HL transmission. Interestingly, this pattern contrasts with findings for HL-transmitting parents in an English-dominant, and hence less multilingual, part of Canada. In a study of Russian-transmitting parents in Saskatchewan, Makarova and colleagues (2019) reported that parents showed no concern for the effect of bilingualism on children’s English development. This points to the possibility that trilingualism concerns may have a synergistic relationship with HL concerns.

Notably, the aforementioned higher status of the English-speaking minority with respect to other minority language communities in Quebec has been found to influence parental beliefs and practices in an earlier study of Chinese- and English-transmitting families in Montreal (Curdt-Christiansen, 2009). In this study, Chinese-transmitting parents put more emphasis on HL-transmission as a means of establishing children’s identity than English-speaking parents. This could, in turn, lead to stronger concerns about their children’s HL development, given its key role in establishing a HL identity. Also, Chinese-speaking parents reported engaging less with their children’s French-language schools, possibly due to their lower French proficiency relative to the English-transmitting parents. The perception of language proficiency as a barrier to supporting children at school was a general theme in parents' concerns, both HL-transmitting and non-HL-transmitting, but future research might explore how such concerns differ across HL-transmitting and non-HL-transmitting parents. Concerns might be stronger in HL-transmitting parents if, as in Curdt-Christiansens’ study, they also have lower French proficiency. Insights into the experiences of HL-transmitting families in Quebec such as these are urgently needed given that a HL is spoken in almost 60% of multilingual households in Quebec, and the number of HL-speaking families is expected to grow in coming years (Statistics Canada, 2023).

Identity/well-being concerns, as noted above, were often linked to HL/trilingual transmission: many parents wondered if their children would find it difficult to identify with HL speaker groups since they have few HL-speaking peers and limited exposure to the associated culture. However, these concerns were also expressed by parents raising children with French and English only, for example, whether their children’s accents in French or English might influence how they are perceived by others, including whether they would “pass” for native speakers or not. Thus, despite holding very positive views of childhood multilingualism, some parents may still adhere to monolingual norms and hold a negative view of common aspects of multilingualism, such as cross-linguistic transfer. Interestingly, in a recent study of parents raising multilingual children mostly in Montreal, parents encouraged children to develop pluralistic identities rather than pushing them to adopt monolingual identities (Le Gall & Meintel, 2015). An interesting possibility is that identity/well-being concerns arise more in contexts where pluralistic identities are not as commonplace, for example, outside of the urban centers of Quebec. Our sample was also skewed towards families living in the Montreal metropolitan area (72%), which could explain the relative infrequency of this concern type relative to other concern types. Future research focusing on multilingual families in less multilingual contexts could test this possibility.

For RQ2, our results also shed light on the relative prominence of the different concern types. Specifically, we found cognition concerns and exposure/fluency concerns to be parents’ most prominent concerns about multilingual child-rearing. This finding is in line with earlier work (King & Fogle, 2006b), thus constituting another similarity with concerns of the parents in the current study and those raising multilingual children in monolingual societies. In our prior quantitative study of concerns, cognition concerns were weaker than exposure-fluency concerns (Quirk et al., 2023). By contrast, here, we found that these concern types were equally prominent in parents’ open-ended responses. One possible reason for this is that parents may well have been less concerned about their children’s cognition but felt compelled to clarify their thoughts about language delay in the open-ended item, given that this concern was linked to mixed feelings. That is, parents simultaneously expressed the belief that their child will benefit cognitively from multilingual exposure and the fear that this exposure could incur a language delay and other cognitive difficulties for their child. The other two concern types — HL/trilingualism concerns and identity/well-being concerns — were relatively infrequent. The infrequency of HL/trilingualism concerns is somewhat surprising, given that over half of parents were transmitting a HL. One possibility is that, due to their thematic closeness, HL/trilingualism concerns were subsumed in more general mentions of concerns about exposure and fluency. The infrequency of identity/well-being concerns, as noted above, may be linked to specific characteristics of our sample, namely the dominance of families living in highly multilingual and multicultural areas of Quebec. Another possibility is that these concerns are generally less common or weaker than other concern types, a possibility that could be tested in future research.

6. Conclusions

This study contributes to an improved understanding of parental concerns regarding multilingual child-rearing by affirming the existence of two previously-attested concern types – namely cognition concerns and exposure/fluency concerns – and by providing more nuanced insights into the nature of those concern types. This study also revealed two new concern types – namely, HL/trilingualism concerns and identity/well-being concerns. The presence of HL/trilingualism concerns highlights a distinction which has been found to impact these same parents’ attitudes towards childhood multilingualism (Kircher et al., 2022, 2024) and their engagement with resources to support their children’s multilingual development (Ahooja et al., 2022). This suggests that the status and number of languages being transmitted play an important role in shaping the experiences and beliefs of multilingual families in a multilingual context such as Quebec. Future studies should seek to clarify (1) how HL transmission and trilingualism shape concerns independent of one another, (2) the underlying reasons for differences in experiences and beliefs relative to parents transmitting only societal languages, and (3) whether this generalizes to other sociolinguistic contexts.

An important theoretical contribution of this study is the generation of knowledge regarding the multi-faceted nature of parental concerns regarding multilingual child-rearing, demonstrating that such concerns are distinct from the language attitudes that are typically studied as part of the beliefs component in FLP research. This finding points to the need to establish concerns as a separate focus of inquiry in order to obtain a more coherent and comprehensive understanding of parental beliefs regarding multilingual child-rearing.

In terms of practical implications, our findings could serve to improve support for parents raising young children multilingually, both in terms of determining how parents should be supported as well as which groups of parents would benefit most from that support. For example, our results indicate that even parents raising children in a multilingual society like Quebec have concerns about language delay and difficulties at school. Though these parents frequently encounter models of successful multilingual acquisition, they still view challenges for their children acquiring multiple languages in early childhood, and they may struggle with decisions about language use in and outside of the home. Given the role of these decisions in shaping early language environments and, in turn, children’s multilingual outcomes, this study constitutes a basis for the development of support for both parents and children alike.

7. Strengths, Limitations, and Future Directions

A notable strength of our study was its large sample and the inclusion of parents transmitting a variety of language combinations. However, not all groups were represented, and in particular our sample included very few parents transmitting Indigenous heritage languages. Furthermore, we did not consider data from parents who had opted not to raise their children multilingually. The concerns of these parents may differ from those of our participants; thus, future research should make targeted efforts to recruit them.

Another strength of our study was that it gathered quantitative and qualitative data from the same sample, thereby yielding much more nuanced insights. However, the fact that parents responded to our open-ended question after completing the closed items, which referred to cognition and exposure-fluency concerns, may have primed them to mention these concern types in their open-ended responses. Future work could test this possibility by enquiring about parents’ main concerns without any prior questioning/discussion of specific concern types.

Lastly, the survey format allowed us to collect large amounts of data, but at the expense of our ability to follow up on unanticipated findings. Thus, it will be important for future studies to build on our findings using complementary methods, such as focus groups and interviews. For example, our finding that parents’ attitudes and concerns regarding childhood multilingualism were, at times, incongruent – along with the insights that our study provided on the nature and dimensionality of concerns as a separate construct – motivate possible future studies of concerns directly alongside attitudes.

Future quantitative analyses might also determine whether the additional concerns we found, HL/trilingualism concerns and identity/well-being concerns, indeed constitute separate concern types. Since the current study and related research among the same parents have found HL/trilingualism-specific differences in the experiences and beliefs of parents (Ahooja et al., 2022; Ballinger et al., 2022; Kircher et al., 2022, 2023), it is possible that – rather than constituting a distinct concern type – HL/trilingualism constitutes an overarching characteristic that links to (and generally intensifies) all types of concerns.

Finally, an important next step in research on concerns will be to test the hypothesis that concerns about multilingual childrearing are a distinct type of parental belief. This could be done through investigations of such concerns alongside other parental belief types, such as parents’ attitudes toward their languages and towards child multilingualism, to test if and how they behave independently from these constructs. Because of their individualized, more concrete nature and their orientation to future experiences, we believe that their addition to existing models of FLP will capture additional nuances in how families develop their FLP and its link to multilingual children’s linguistic and socio-emotional development.