Majority Language Influence and Heritage Language Maintenance in a Small Transnational Community: Hungarian-Hebrew Families in Israel

Abstract

In a globalised and interconnected world, transnational families must navigate heritage language (HL) practices within dominant majority languages (ML), often with limited institutional support. Focusing on a small and understudied community of Hungarian-speaking transnational families in Israel, this study explores how HL development is maintained and negotiated within the framework of family language policy in a dynamic multilingual environment. Fifteen Hungarian-speaking parents from bilingual Hungarian-Hebrew families participated in semi-structured sociolinguistic interviews conducted in Hungarian. A mixed-methods approach was used to analyze the interview data. Quantitative analysis was used to identify the distribution and relative frequency of language use across families. At the same time, qualitative analyses show how parental ideologies and strategies relate to HL development. The findings show that while HL input remains central in parental speech, children frequently respond using both HL and ML, indicating a dynamic bilingual repertoire and a translanguaging orientation. Overall, HL development is negotiated, maintained through cultural and emotional ties, flexible bilingual practices and dynamic family language policies.

1. Introduction

In a globalised and interconnected world, families are no longer limited to a single geographic or cultural context. Ongoing migration trends have introduced transnational families. The members of these families live across national borders, yet they are socially, emotionally and linguistically connected (Bryceson & Vuorela, 2002; Levitt & Glick Schiller, 2004; Vertovec, 2009). Within these families, language plays a central role in maintaining intergenerational ties, identity and a sense of belonging.

In host societies, transnational families are often a part of heritage-language (HL) communities that differ in size, density, and institutional support. While much existing research on HL maintenance has focused on large and well-established immigrant groups, less is known about smaller and weakly institutionalised HL communities, where opportunities for language transmission are limited, and family practices become especially central.

This study focuses on Hungarian-speaking families in Israel and explores how HL use is maintained and negotiated within a Hebrew-dominated environment influenced by English (Rosenhouse, 2012). The analysis adopts the family language policy (FLP) framework to examine HL maintenance in the family domain. Transnationalism and HL research provide the broader sociolinguistic context in which family practices take place (García & Wei, 2014; K. A. King et al., 2008). Within this framework, the study treats code-switching (CS) as an analytical tool for examining structural and directional patterns of HL–ML use, while translanguaging frames these patterns as part of speakers’ unified linguistic repertoire. Together, these approaches allow an integrated examination of parental language ideologies, interactional practices, and bilingual language use across multiple levels of FLP.

The Hungarian-speaking community in Israel is a relatively small and understudied group compared to other immigrant groups, such as English-, Russian-, or French-speaking communities (Armon-Lotem et al., 2021; Ben-Rafael, 1994; Meir et al., 2021; Rosenhouse, 2012). The Hungarian-speaking community in Israel had 13,616 members in 1948, and 24,143 people arrived from Hungary between 1948 and 1960 (CBS, 2010). Early Hungarian migrations were connected to the establishment of the state, followed by post-World War II immigration. Later waves (4706 people immigrated between 1960 and 1989) were tied to post-communist transitions of the 1990s and to economic mobility across Europe around 2008. In the period 2000–2024, 2150 people immigrated from Hungary to Israel, according to statistics from the Central Bureau of Statistics (CBS, 2025). However, these figures do not reflect the exact size of the community, as Hungarian speakers may also have immigrated from neighboring countries and been recorded under those countries of origin. Researchers such as Rosenhouse (2012) and Spolsky and Shohamy (1999) estimate that 60,000–100,000 individuals in Israel have Hungarian origins. However, the number of speakers with active Hungarian proficiency is considerably lower and difficult to quantify. Hungarian currently holds no official minority-language status in Israel and receives minimal institutional support. There is no Hungarian foreign language instruction at Israeli universities, and formal HL provision is primarily limited to bottom-up community initiatives, such as weekend and heritage schools and cultural programs organized by the Kishon EduCenter, which operates depending on parental involvement.

Early generations of Hungarian immigrants experienced rapid linguistic assimilation, with Hebrew replacing Hungarian as the primary language of daily life (Ben-Rafael, 1994). The loss of intergenerational transmission is reflected in the dominance of Hebrew as the majority language (ML) and the limited institutional support for maintaining Hungarian (Rosenhouse, 2012). Studies of Hungarian-speaking communities across the diaspora have noted their non-cohesive and heterogeneous characteristics, resulting from different migration waves and varying degrees of cultural integration (Fenyvesi, 1995–1996; Bartha, 1995; Kontra, 1990; Szilágyi & Szécsi, 2020). More recent migration waves, particularly those arriving after 2000, show a different linguistic and ideological profile. These families are typically multilingual, with prior exposure to English and other European languages, and often express cosmopolitan and mobile identities (Hatoss, 2006, 2016; Szilágyi & Szécsi, 2020). Having grown up within the European Union’s multilingual policy framework, where multilingual competence is both a cultural norm and an institutional requirement (Multilingualism, 2009; Eurobarometer, 2023), these families bring distinct language ideologies into the Israeli sociolinguistic environment. At the same time, they maintain emotional, linguistic, and cultural links with Hungary while integrating into Israeli society (Bilgory-Fazakas & Armon-Lotem, 2025).

Taken together, these characteristics position Hungarian in Israel as a particularly informative case for examining HL maintenance in small communities, where family-level practices interact with limited community support and strong majority language pressure. The present study investigates how HL development is maintained and negotiated in an understudied transnational setting, focusing on HL resilience, family bilingualism, and identity negotiation processes. Using the Hungarian-speaking community in Israel as a case study, it further explores how families shape and adapt their FLPs in response to dominant Hebrew and limited institutional support.

2. Literature Review

The theoretical background is organised to reflect the study’s analytical structure, moving from macro-level perspectives on transnational families and HL maintenance to FLP and bilingual interactional practices.

2.1. Transnational Families

Bryceson and Vuorela (2002) define transnational families as those whose members are motivated to maintain close and active relationships despite geographic distance. These families may live apart for a long time or permanently, yet they maintain connections across borders. While some migrant communities consider family as a nuclear unit of parents and children, others maintain an extended family system, including grandparents and other kin who play vital roles in care, everyday communication and intergenerational support (Baldassar et al., 2016; Bryceson, 2019; Poeze & Mazzucato, 2014).

Vertovec’s (2009) research supports the same definition of transnational families. Within these contexts, linguistic practices reflect ongoing negotiations between assimilation and cultural and linguistic continuity. The linguistic practices of these families are not static. They evolve and are dynamic, as HL use is constantly negotiated and shaped by everyday interactions and changing family circumstances (Curdt-Christiansen, 2016; K. King & Lanza, 2019). Language ideologies, management strategies, and practices may shift over time in response to changes in the family and linguistic proficiency among family members (Hirsch & Lee, 2018). Studies on Hungarian-speaking families show that FLP develops within changing linguistic environments, where mobility, transnational ties, and shifting conditions influence HL practices over time (Hatoss, 2024), consistent with earlier research highlighting the role of family relationships in HL transmission (Bryceson & Vuorela, 2002; Tannenbaum & Howie, 2002).

These features of transnational families align very well with the characteristics of the Hungarian-speaking community in Israel used in this study to explore HL maintenance and vulnerability.

2.2. Heritage Language Maintenance and Vulnerability

The well-known and most often cited definition of a heritage speaker comes from Valdés, who describes them as ‘individuals raised in homes where a language other than English is spoken and who are to some degree bilingual in English and the heritage language’ (Valdés, 2000, p. 375). Although the definition originates in an English-dominant society, any ML can replace English. Two key elements are central: the HL is acquired first, but it might not progress to age-appropriate levels due to an early shift to the ML. As a result, HL speakers show variation in HL abilities (Polinsky & Montrul, 2013, p. 133). Heritage speakers are early bilinguals who acquire a minority language in the home but go through a shift in dominance due to exposure to the ML (Rothman, 2009). From this perspective, developmental trajectories, changes in input, and shifts in dominance characterise heritage bilingualism (Kupisch & Rothman, 2018; Polinsky & Montrul, 2013).

HL speakers may be simultaneous bilinguals, acquiring the HL in natural and spontaneous situations parallel to the ML. Or sequential bilinguals, acquiring HL first before shifting to the ML through institutionalised schooling. This process often leads HL to become the weaker language by adulthood (Polinsky & Montrul, 2013, p. 133). Even though early exposure to HL would seem to be the most practical and effortless way to achieve bilingual proficiency, research indicates that not all children living in bilingual environments become bilingual and maintain HL proficiency as they grow older (e.g., De Houwer, 2007; Verdon et al., 2014). Polinsky and Montrul (2013) found that HL is shaped by reduced input, incomplete acquisition, and sociopolitical pressures. From an FLP perspective, exposure alone is insufficient, as HL development also depends on how languages are ideologically valued and used in everyday family interactions within the home (Otwinowska et al., 2021; Zabrodskaja et al., 2023).

Classic frameworks of language shift highlight that HL vitality depends not only on the family’s linguistic choices but on the broader sociocultural and institutional conditions (Fishman, 1991; Carreira & Kagan, 2011). Most recent perspectives stress variation and resilience. Kupisch and Rothman (2018) emphasise the heterogeneity of HL outcomes. They argue that HL bilingualism can be seen as a spectrum of linguistic competence shaped by input and socialisation, shifting the attention from what is missing to what is maintained. Contemporary FLP research similarly defines FLP as dynamic and adaptive rather than fixed (Zabrodskaja et al., 2023; Karpava et al., 2025). At the same time, studies also demonstrate that HLs can remain resilient when families frame them as resources of identity and belonging (He, 2010; Hornberger & Wang, 2008) and cultural identity (Bilgory-Fazakas & Armon-Lotem, 2025). Furthermore, positive HL ideologies, supported through practices and flexible management strategies—including translanguaging—may strengthen intergenerational transmission despite ML dominance (Karpava et al., 2025).

2.3. Family Language Policy

Family Language Policy (FLP) provides a framework for understanding how families navigate the transmission and maintenance of HL under the pressure of ML dominance. Building on Spolsky’s (2004) model of ideology, management and practice, FLP emphasises that what families intend to do and what they actually do are not always the same. Parental ideologies guide language choices and everyday practices. Children’s agency and sociocultural pressures may also shape them. K. A. King et al. (2008) highlight the dynamic nature of FLP. Although parents may intend to enforce strict HL rules, family strategies are negotiated due to children’s resistance, peer influence, or a desire to be included in wider social networks, leading to compromise (Schwartz, 2010). Children do not simply adapt to parental policies. They actively negotiate or resist them. Therefore, they shape the FLP and influence the family’s linguistic trajectory (Lanza, 2004). Extended family members, particularly grandparents, also play a role in maintaining HL use, linking family strategies to intergenerational continuity (Tannenbaum & Howie, 2002; De Houwer, 2021). FLP is particularly relevant for HL research because it provides a framework for explaining why HL outcomes vary across families with similar linguistic resources, by connecting everyday interactions to broader ideological and structural conditions.

Over the past decade, FLP research has examined how ideology, practice and management are expressed in everyday family life, linking them to child and family outcomes (such as socio-emotional and cognitive) rather than treating language input as the only predictor (Hollebeke et al., 2023).

Curdt-Christiansen (2013) extended the FLP to include sociopolitical influences, suggesting that family policies are shaped not only by household choices but also by school, community networks, and state-level ideologies. Therefore, families who attempt to maintain HLs negotiate within the wider system of power. Examining transnational families, Hirsch and Lee (2018) showed how FLP changes across migration contexts through negotiation and adjustment. Building on the dynamic view, Seals (2023) introduces the concept of aspirational FLP, demonstrating how language policies shift with family goals and life stages, and explicitly framing FLP as evolving rather than fixed.

Research on FLP has identified several parental strategies for managing bilingual interactions at home. One is the well-known One Parent—One Language (OPOL) approach, in which each parent consistently uses a different language with the child (De Houwer, 1999; Lanza, 2004). Another is the Time and Place approach, in which language is assigned to specific routines, settings, or activities (Barron-Hauwaert, 2004). Parents may also use the correction and reinforcement strategies, in which they reinforce HL use and model responses when children switch to the ML (Lanza, 2004; Yamamoto, 2001). In contrast, some families adopt flexible bilingual practices, in which parents strategically mix languages, accepting children’s translanguaging as part of natural bilingual development (Wei, 2012; García & Wei, 2014). These strategies change over time as families respond to children’s agency, changing circumstances, and social pressure (Hirsch & Lee, 2018; Karpava et al., 2025).

In this study, FLP provides a framework for analysing how Hungarian-speaking parents in Israel manage HL input and whether they enforce consistent HL use, allow situational flexibility, or accept bilingual practices.

2.4. Parental Language Input

Fishman (1991) emphasised that efficient HL transmission could be predicted by the parents’ use of HL in the home. De Houwer (2007) investigated the connection between parental input patterns and rates of HL transmission. She found that patterns of parental input influenced the rate of children’s bilingual use. A recent longitudinal study on HL transmission in Australia has further supported De Houwer’s notion. The study showed that personal factors, such as the parental HL use, or the physical presence of grandparent(s), and environmental factors like parental impression of the support from the educational environment, or first and second-generation immigration status, played important roles in the maintenance of a HL (Verdon et al., 2014). In transnational families, HL input also includes communication with relatives abroad and frequent visits, which further diversify children’s exposure to HL (De Houwer, 2021; Szilágyi & Szécsi, 2020; Bilgory-Fazakas & Armon-Lotem, 2025).

Previous research suggests that in small, geographically dispersed HL communities, maintenance is often supported not only by HL at home but also by external cultural reinforcement. Regular visits to the home country, sustained contact with extended family members, and literacy practices offer HL input that could compensate for limited community presence (Bartha, 1995; Szilágyi & Szécsi, 2020).

Everyday family routines, such as shared mealtime, offer opportunities to see how family members use the HL and ML. Blum-Kulka’s (1997) work on family dinner table talk shows that the dining table is much more than a place to eat. It is a social space where stories are told, turns are negotiated, and children learn how conversations work in real life. As she writes, when “dinner is constructed as an intergenerationally shared social speech event, it can serve as a pragmatic socialization context in which children learn how to become competent conversational partners in intergenerational multiparty talk” (Blum-Kulka, 1997, p. 10). Snow (2006) reinforces this perspective, emphasising that mealtimes are not just about eating routines but also interactional sites where language is used, negotiated, and serves as a platform for socialisation. Most recently, this view has been supported by a cross-cultural study of Bohn et al. (2024). They found that mealtime is a universal context for parent–child interactions across societies, which shape the quality and quantity of language input. In bilingual households, parental language choices at the dinner table reflect wider family language ideologies and serve as a key site for HL development (Hu, 2019).

2.5. Bilingual Language Use: Code-Switching (CS) and Translanguaging

Parental strategies and ideologies around HL use often include decisions about bilingual language use (Lanza, 2004; Wei, 2018). Code-switching (CS) has long been studied in the literature, and scholars differentiate between structural and interactional approaches. The structural models emphasise that CS follows predictable linguistic patterns and is driven by grammatical rules (Kovács, 2001; Muysken, 2013; Matras, 2009; Otheguy et al., 2015). In contrast, interactional models focus on the social and pragmatic features of CS, viewing it as a strategic resource for constructing identities and signalling group memberships and social relationships (Wei, 2005, 2012; Dewaele & Wei, 2014).

As children typically acquire their language(s) through adults’ input, several studies have investigated parental code switching (CS). For example, children code-switch at rates similar to those of their parents, suggesting that parental CSs serve as an important model (Genesee et al., 1995). Some studies indicate that the frequency of parental CS can negatively affect a child’s lexical development (Byers-Heinlein, 2013). In contrast, others argue that it does not affect the child’s lexical development (Bail et al., 2015).

Early on, CS was stigmatised (Myers-Scotton, 1993; Zentella, 1997), and this negative attitude led parents to restrict its use. However, more recent studies (Fuller, 2009; Dewaele & Wei, 2014) show that positive attitudes toward CS correlate with more frequent and strategic use of CS. Fuller’s (2009) study showed that if CS was considered positively, it became a marker of bilingual identity. At the level of language use, CS tends to follow structural and functional patterns. Therefore, intra-sentential switches occur most frequently in the nominal domain, as nouns and noun phrases exhibit greater semantic autonomy and are less restricted morphosyntactically (Kovács, 2001; Muysken, 2013). Similar tendencies were reported in studies of Hungarian contact varieties (Hungarian-German, Hungarian-English) (Bartha, 1995; Fenyvesi, 1995–1996). These insertions often reflect the sociocultural environment of the ML. Items referring to institutions, bureaucratic processes, education, and local infrastructure are recalled in the ML, especially when HL equivalents are unavailable or pragmatically inadequate (Matras, 2009). Lexical choices rooted in everyday cultural practices may also function as markers of belonging and identity in bilingual interaction (Guzzo, 2014).

Rather than viewing bilingual speech as “switching” of two different codes, Garcia and Li Wei think of translanguaging as the process by which “bilinguals have one linguistic repertoire from which they select features strategically to communicate effectively” (García & Wei, 2014, p. 13). This perspective emphasises the dynamic use of linguistic repertoires to express identity. It also challenges the idea that HL, ML and other languages exist as separate systems. Otheguy et al. (2015) argue that translanguaging highlights bilingual creativity and autonomy in the use of complete linguistic systems, “refers to the act of deploying all of the speaker’s lexical and structural resources freely” (p. 297). In family and community settings, translanguaging offers a framework for how parents and children co-construct communication through the fluid use of multiple languages. It reframes “mixed” input not as inconsistent but as dynamic and pragmatic, which supports inclusion and bilingual development (Wei, 2018).

In the present study, CS is used as an analytical tool for examining patterns between HL and ML, including directionality and structural integration. Translanguaging, in contrast, frames the interpretation of these patterns by highlighting how speakers use their full linguistic repertoires in everyday family interactions.

2.6. The Current Study

The objective of the current study is to explore how HL development is maintained and negotiated within a small and understudied transnational community by focusing on HL resilience, family bilingualism, and identity negotiations. Using the Hungarian-speaking community in Israel as a test case, the research examines how families adapt their FLPs under conditions of strong ML dominance and limited institutional support. Therefore, the following research questions are addressed:

RQ1. What characterises parental language input in Hungarian-Hebrew bilingual families in terms of language choices, directionality of interaction, and multilingual repertoire use during family mealtime?

RQ2. How does ML influence the use of HL across community-level, interactional and ideological dimensions of FLP?

RQ3. What structural and functional patterns of CS emerge in parental HL speech, and how do these relate to speakers’ bilingual ideologies and translanguaging practices?

Each RQ corresponds to a specific analytical level examined in the results section, allowing a structured examination of parental input patterns (RQ1), ML influence (RQ2), and bilingual language use, including CS and translanguaging practices (RQ3).

H1. 

While many parents are expected to report an intention to maintain HL at home, consistent HL input will be constrained by the ML and by variation in parental and children’s HL use (K. A. King et al., 2008; Schwartz, 2010).

H2. 

At the community level, parents are expected to report weak social ties due to the small community size and limited institutional support within the Hungarian-speaking community, resulting in reduced opportunities for HL use beyond the family domain and increased pressure to shift toward ML in public domains. In the absence of institutional support, families are expected to rely primarily on external strategies, such as frequent visits to the country of origin and maintaining contact with grandparents. (Bilgory-Fazakas & Armon-Lotem, 2025).

H3. 

CS is expected to occur with varying occurrence patterns and distributions, reflecting parental strategies and domain-specific communicative needs (Byers-Heinlein, 2013; Bail et al., 2015). The dominant CS type is expected to be intra-sentential switching (Muysken, 2013), particularly insertions into nominal groups (Kovács, 2001; Matras, 2009). These insertions are predicted to be morphologically and syntactically integrated into HL, showing a high degree of bilingual proficiency and structural integration. Crosslinguistic influence is expected in all interviews, but its extent will vary (Polinsky & Montrul, 2013).

3. Materials and Methods

3.1. Research Design

3.1.1. Participants

Participants were recruited via Hungarian-speaking Facebook groups and news portals (e.g., Israeli Hungarians—Izraeli Magyarok, Új Kelet Online) using the snowball sampling method to complete an online questionnaire examining family language use, ideologies and practices (see Supplementary Material S3 for more information). At the end of the questionnaire, they were asked whether they were willing to take part in a follow-up interview. Only parents who expressed a willingness and had at least one child under the age of 12 were eligible for inclusion in the qualitative phase. This combination of initial snowball-based recruitment followed by purposive selection of interview participants is widely used in sociolinguistic and FLP research (K. A. King et al., 2008; Curdt-Christiansen, 2009). Accordingly, fifteen Hungarian-speaking parents (fourteen mothers and one father) from bilingual Hungarian-Hebrew families were selected for this study. The participants’ ages ranged from 35 to 48 years (M = 42, SD = 4.14), and they had lived in Israel for an average of 13 years (SD = 3.48) (see Table S1, Supplementary Material S1 for more information). They all immigrated to Israel after 2000 and had at least one child under 12 born in Israel. Most (n = 13) were born in Hungary, while two were born elsewhere (Russia and Slovakia) but raised in Hungary. They all nominated Hungarian as their mother tongue. The majority of the participants (n = 13) studied at a higher educational institute in Hungary. Most have mid- to high-socio-economic status occupations (e.g., teachers, managers, designers, photographers, pharmacists), but three positions do not (e.g., assistant, cosmetician). The number of children under 12 is twenty-nine, and the average age of the children present during the interviews is 7.33 (SD = 1.95).

All families were functional multilinguals; they regularly use more than one language in everyday communication without implying balanced proficiency or native-like competence across languages (Bilgory-Fazakas & Armon-Lotem, 2025). Grandparents and extended family members who lived in Hungary motivated parents to transmit and maintain the HL for a pragmatic reason that grandparents would not be able to interact with the children if they did not acquire the HL (Bilgory-Fazakas & Armon-Lotem, 2025).

3.1.2. Data Collection

The present study is a part of a broader mixed-method project on HL maintenance among Hungarian-speaking families in Israel (Bilgory-Fazakas & Armon-Lotem, 2025). The qualitative design is based on semi-structured sociolinguistic interviews. The interviews were conducted by the first author, a bilingual researcher, between 2023 and 2024, either online via Zoom or face-to-face, depending on participants’ preferences. The interviews were conducted in Hungarian, although other languages (Hebrew and English) were accepted when preferred by participants. During the interviews, the youngest child from each family was present. Participants were encouraged to speak freely about their experiences and could interact with their children at any time. Children spontaneously participated in the interaction by responding to questions, commenting on parental statements, or correcting reported language use within the family domain. All parents signed informed consent forms for themselves and provided separate consent for their children prior to participation.

The interview protocol covered several core thematic areas, including participants’ migration and linguistic background, family language ideologies and everyday language practices. Participants were asked about language use within the household, decision-making processes regarding language choice, and strategies for managing the HL and ML across domains. Particular attention was given to language practices during family mealtimes. Parents were asked to describe typical language choices, interlocutors, and the directionality of language use (parent–parent, parent–child, child–parent and sibling interaction) during family meals, and to provide concrete examples from everyday routines. The interviews also addressed children’s social networks, transnational ties with the home country, and experiences of language use in institutional and public settings (see Supplementary Material S2 for the full protocol).

Interviews lasted between 10 and 33 min (M = 15.53, SD = 7.32). Variation in the interview length reflected differences in participants’ level of elaboration and conversational style; however, all interviews followed the same semi-structured protocol and addressed the same core topics. All interviews were recorded and transcribed verbatim for subsequent analysis. Participants were informed about the broader aim of the research, including its focus on FLP. All participants provided informed consent, and the study was approved by the Human Subjects Institutional Review Board (HSIRB) at Bar-Ilan University (approval number 231125801).

3.1.3. Data Analysis

All interviews were transcribed verbatim and coded in Excel. Each utterance was categorised according to speaker role (participant, child) and the language(s) used (Hungarian, Hebrew, English).

To explore how everyday practices shape language choice, a Sankey diagram was created from coded reports of language-use patterns among family members to provide a descriptive visualisation of family language-use patterns and their relations to language practices during mealtimes.

Thematic analyses (TA) (Braun & Clarke, 2006) were used to identify and report themes within the qualitative data. An inductive approach was employed, allowing themes to emerge directly from the interview data. During the initial coding stage, recurring references to language beliefs, reported practices, and management strategies were identified. These codes were then grouped into broader thematic categories reflecting the analytical structure of the Results section. The thematic framework compromised the following domains: (1) parental ideologies, including beliefs about HL importance and attitudes toward bilingualism and CS; (2) language management strategies, such as corrective practices and transnational reinforcement; (3) interactional language practices, including child agency and patterns of language choice and (4) structural and functional features of bilingual language use, such as distributional patterns, and morphosyntactic integration of ML-origin insertions within HL speech. CS was examined as part of how parents position themselves toward bilingual practices in interviews, in line with FLP research (Curdt-Christiansen, 2009; K. A. King et al., 2008; Schwartz, 2010) that views such talk as a space for articulating language ideologies.

All instances of Hebrew-origin lexical items that occurred during the interviews were identified as relevant tokens and entered into an Excel document for linguistic categorisations and coding. Those items listed in the Hungarian National Corpus (HNC) (in English: http://mnsz.nytud.hu/index_eng.html, accessed on 27 November 2025) were excluded (e.g., falafel, hummus, bat mitzvah, etc.), as these are established loanwords in contemporary Hungarian. All items were marked with the respondent’s identification code, listed according to the word class categories, and by the type of position of their occurrence and their functional categories. They were also tagged by word-class categories (nouns, verbs, adjectives, etc.) and discourse markers were also counted. Phrases that were code-switched (composed of more than one lexical item, e.g., a noun or adjectival phrases) were counted separately only if both elements of the constituent were inserted from Hebrew. Twenty-one instances from English were identified and listed separately from Hebrew. CS instance counts were calculated at the participant level and used in the analyses as indicators of interview-based CS production. All lexical items were coded at the participant level, while results are reported at the group level to highlight structural and functional patterns (the distributional data are presented in Table S6, Supplementary Material S1).

Exploratory statistical analyses were also conducted to examine relationships between key variables. Given the small sample size (N = 15), analyses were limited to non-parametric procedures (Spearman’s correlations and chi-square tests). Statistical findings are interpreted as tendencies within the sample.

4. Results

4.1. Characteristics of Parental Language Input

This section examines parental language input and bilingual language use within families to address RQ1. Figure 1 visualises reported language use across families. Each flow shows how participants (P), their partners (Pp), and their children (C) use Hungarian (Hu), Hebrew (Heb), English (Eng), or combinations of these languages in family interactions. The thickness of the bands reflects the dominance of each reported language choice.

The Sankey diagram is based on participants’ responses to the interview question, “Which languages do you use at home?” When aspects such as language directionality were not mentioned spontaneously, targeted follow-up questions were used to clarify parent–parent and parent–child interactions. Responses were coded for reported language choice and language combinations, which served as the basis of the colour categories shown in the diagram. Therefore, the colours represent different languages or language pairs, as it follows: Hebrew (Heb) = blue, English (Eng) = red, Hungarian (Hu) = green, Hebrew and English (Heb & Eng) = purple, Hungarian and Hebrew (Hu & Heb) = teal, Hungarian, Hebrew and English (Hu, Heb & Eng) = brown. Table S3 in Supplementary Material S1 summarises the distribution of reported language use patterns across family interaction types and provides the quantitative basis for the Sankey diagram.

Family input patterns revealed diverse and multilingual repertoires among parents, reflecting their international backgrounds and hybrid spousal language use. Five distinct language patterns were identified in parent-parent interaction (P↔P). Mixed Hebrew–English use was the most frequent pattern (6/15 families; 40%), followed by Hebrew-only interaction (5/15; 33.3%) and English-only (2/15; 13.3%). The Hungarian-only use was rare (1/15; 6.7%), while multilingual language use (Hu, Heb & Eng) was also reported by one family (6.7%) (See Table S3 in Supplementary Material S1).

In the HL-to-child (P → C) stage, parental input was predominantly bilingual, with combined Hungarian-Hebrew use accounting for 60.0% of families (9/15). Hungarian-only input was reported by 26.7% of parents (4/15), while multilingual input occurred in a smaller proportion of families (2/15; 13.3%). These patterns indicate that, although the HL remains central in parental input, it is frequently combined with the ML in everyday interactions.

Children’s responses (C → P) show greater variability and bilingual orientation. The strongest flows in the Sankey diagram correspond to the Hungarian-Hebrew use as the most frequent response pattern (6/15, 40.0%), followed by Hungarian-only responses (5/15, 33.3%) and Hebrew-only responses (3/15; 20.0%). Multilingual language use was rare, occurring in only one family (6.7%). Hebrew-only responses did not exceed either bilingual or Hungarian-only responses as reported, suggesting that the ML does not replace the HL but rather coexists with it in children’s interactional practices.

Language use during mealtimes similarly reflects multilingual negotiations. Multilingual language use was most reported (6/15 families; 40.0%), followed by Hebrew-only use (5/15; 33.3%), and bilingual Hungarian-Hebrew interaction (3/15; 20.0%). Hungarian-only use during mealtime was limited (1/15; 6.7%) (for details, see Table S3 in Supplementary Material S1). These distributions highlight mealtimes as a particularly dynamic interactional space in which multiple linguistic resources are mobilised. Overall, the quantitative distributions visualised in the Sankey diagram illustrate a bilingual environment characterised by consistent variation rather than a linear shift toward the ML. The negotiated and context-dependent nature of mealtime language use is illustrated in the following interview excerpts:

“I spoke Hungarian—that was never a question. My husband spoke Hebrew, but often English as well, since English is our shared language.”

(P2)

“When we are all together, it is only Hebrew. And when I am with them, then it is only Hungarian. For example, dinner and talking about how your day was—that is in Hebrew.”

(P7)

Language use across family interactions reflects ongoing negotiation between the HL and the ML, with bilingual and multilingual language use forming a stable component of everyday family communication.

4.2. The Impacts of ML on HL Variation

Addressing RQ2, this section examines the role of the ML in shaping variation in HL. The analysis is organised across four interrelated levels: the community level (social networks and institutional environments), the interactional level (parental language strategies), the ideological level (attitudes toward CS and interview-based CS production), and the structural level (patterns of CS within HL speech).

4.2.1. Community-Level Factors: Expectations, Integration and Family Ideologies

The pattern emerging at the community level concerns the influence of perceived pressure to integrate into the ML-speaking environment. The descriptive distributions revealed variation in both perceived integration pressure and children’s language choice with the HL-speaking parents (Table S4; Figure S1, Supplementary Material S1). Quantitative findings indicate a descriptive tendency between integration pressure and children’s language choice with the HL-speaking parents. A Spearman correlation revealed a moderate positive trend (rs = 0.50, p = 0.058), suggesting that families reporting stronger perceived pressure to integrate tended to report that their children addressed them in the ML. To further explore this pattern, the categories were grouped into a 2 × 2 table comparing families reporting high versus low or no integration pressure, and children’s language choice, with the latter categorised as mainly ML, HL, or mixed (Table S4, Supplementary Material S1). The chi-square test showed the same directional pattern (χ²(1) = 1.88, p = 0.17) and is therefore presented as supplementary descriptive evidence. Given the small sample size, the chi-square analysis is interpreted as descriptive evidence pointing toward an association between community expectations and reduced HL use within the family. Families with little or no integration pressure tended to report a higher level of HL importance. This relationship aligns with parents’ broader language ideologies: HL importance was significantly associated with lower felt integration pressure (r_s = 0.56, p = 0.030). Parents who felt socially secure or less pressured to conform to the ML norm tended to place greater importance on HL. Those who perceived stronger expectations toward ML use often reported greater difficulty maintaining HL.

Another significant relationship was revealed between HL ideology and explicit management. HL importance was associated with clear and explicit FLP rules (r_s = 0.58, p = 0.023), suggesting that the value of HL is more represented through family management. Although the descriptive data suggested a general trend in which participants who valued HL more tended to show a higher level of networking, this association was not statistically robust (r_s = 0.317, p = 0.248).

4.2.2. Interactional Level: Parental Strategies and Child Language Use

At the interactional level, the data reveal variation in how families manage the use of HL and ML. Twelve participants reported adapting their language choices in response to social pressures. For example, they described using ML when interacting in community or school-related settings to ensure mutual understanding. As one parent noted, “When we are in a community setting, then not always. Then I often answer them (children) in Hebrew” (P2). Similarly, another explained that the increased exposure to ML through daycare and peers gradually reduced opportunities for exclusive HL use at home: “Because then, you know, we started to be more and more in a Hebrew-speaking environment, mainly because of the children and their friends. Until then, it was still okay, and when he started daycare, he was still little and not yet fully speaking. And then in daycare, his Hebrew became stronger. Just because of the circumstances, it was hard for me to keep speaking only Hungarian with him in certain situations, and then we used Hebrew, so that others could also understand” (P12). These examples illustrate how institutional environments indirectly shape parental language practices and reduce the space available for HL use.

A small group of participants (n = 3) reported maintaining strict FLP, in which they use HL regardless of context or child responses. One commented, “But usually, if he happens to address me in Hebrew, then he corrects himself afterwards” (P10), while another stated that “Always Hungarian” (P13). Such practices reflect an orientation toward strict HL maintenance, relying on correction and consistency to reinforce expected language norms.

Figure 2 illustrates the number of families reporting each strategy used to promote HL development.

Visits to Hungary and literacy-related activities emerged as the most frequent strategies. Participants reported extended stays in Hungary and regular contact with HL-speaking relatives, which provided intensive and natural HL exposure that was associated with increased HL use.

“When we are in Hungary, we basically speak Hungarian, and during those times the language also becomes stronger.”

(P1)

“With the grandmother, there is Zoom—Zoom calls—and they also come to visit sometimes. So when they are here, and when they are not here but present this way, that also plays a role.”

(P7)

Literacy-related strategies included reading books in the HL and watching cartoons or films. These activities are structured, time-limited, part of the family routine, and do not require constant monitoring of language.

“My rule is that everything should be done as much as possible in Hungarian. For example, I am not willing to read to them in Hebrew—I only read in Hungarian.”

(P11)

“And then in the evening, there is bedtime story time, and I usually do the bedtime story, and that always happens in Hungarian.”

(P1)

Playdates were also reported as arranged social meetings between children and other HL-speaking peers as opportunities for children to use the HL naturally with peers, without direct adult correction or instruction.

“She has Hungarian friends. A. has B. and C. there. But she has not met with C. for a long time, unlike with B, whom they meet often.”

(P1)

“Similarly, there are other partly Hungarian families where the children also speak Hungarian very well, so he can practice the language with them.”

(P10)

Some noted occasional recasting (repeating the child’s ML utterance in the HL) but added that it required sustained attention and was difficult to maintain consistently.

“I say almost everything twice, once in Hungarian and then in Hebrew. I feel like I am a tape recorder. Ever since I have been raising my child, I have been saying everything twice—I repeat almost everything.”

(P8)

Similarly, a small number of participants reported enforcing HL-only at home because of an ML-speaking partner, leading to increased ML use in everyday family interactions.

“If my husband is there, then the mealtime conversation is in Hebrew. If he is not there, then it is in Hungarian. But if he asks for a glass of water, he will ask me in Hungarian. So he automatically addresses me in Hungarian.”

(P11)

“Well, at the moment, the common language for the three of us is Hebrew. He speaks Hebrew with his dad and Hungarian with me.”

(P10)

4.2.3. Ideological and Structural Levels: Attitudes Toward Switching to ML and Their Occurrence

At the ideological level, a Spearman correlation analysis was conducted to examine the relationship between participants’ attitudes toward CS and their actual use of CS during interviews. Attitudes toward CS were measured using a questionnaire (see Section 3.1.2. Data Collection) on a five-point Likert scale (1 = very negative, 5 = very positive), while CS occurrence was operationalised as the total number of CS instances each participant produced during the interviews. The analysis revealed a moderately strong positive association between the two variables (r_s = 0.581, p = 0.023; Table S5, Supplementary Material S1), indicating that participants who expressed more positive attitudes toward CS tended to produce more CS instances during the interview (Figure 3).

At the structural level, CS appeared as a regular but not dominant feature of HL speech. Across interviews, participants produced between 390 and 3015 words (M = 1116, SD = 683), with CS instance counts ranging from 2 to 23 (M = 8.7, SD = 6.1). Participants used multiple linguistic resources within the same discourse. CS occurred predominantly through Hebrew-origin lexical items inserted into the Hungarian speech, reflecting systematic bilingual resource use rather than alternation. Functionally, CS most commonly served as a reference to institutional terminology, culturally embedded concepts, and locally specific practices, as well as for discourse-related functions in the interviews.

4.3. Structural Features and Functional Motivations of Hebrew Insertions

Based on the coding procedures outlined in the Methodology section, this subsection addresses RQ3 by examining the structural features and functional motivations of Hebrew-origin identified in the interview data. It focuses on the distribution of insertions across grammatical categories, their position within Hungarian-dominant utterances, and the communicative functions in the interaction.

A total of 126 Hebrew-origin insertions were identified. The majority were intra-sentential; the largest number appeared in the nominal group, including single nouns (83), proper nouns (15), and noun phrases (13); verbs (9) ranked next, while the smallest groups of word classes are the adverbs (2), discourse markers (2) and numerals (2) (Figure 4). The findings cannot be compared with other research on Hungarian–Hebrew language contact. Still, they are consistent with prior studies on different language contact situations regarding the parts-of-speech hierarchy of CSs (Muysken, 2013, p. 199).

4.3.1. The Nominal Group and Its Functions

Using institutional terminology in the L2 and inserting it into L1 (Matras, 2009), most of the proper nouns (n = 15) included names of companies (e.g., health insurance companies: Maccabi, Clalit, Leumit), and names of offices and institutes (Misrad Hapnim, Misrad HaKlita, Sochnut, etc.).

[1] A Miszrád Hápnimba voltunk.

ɒ misraːd hɒpnimbɒ voltunk

The Misrad Hapnim—ILLATIVE be 1PL PAST

We were in the Ministry of Interior.

Many items referred to Israeli culture, and the participants’ personal history of immigration and experiences (n = 16). They are used because there are no equivalents in Hungarian, or, if there are, they do not have the same semantic meaning—a pattern consistent with the gap-filling or cultural loan theory (Matras, 2009).

[2] és akkor végigcsináltam a rendes alíját

eːʃ ˈɒkːor ˈveːɡiɡt͡ʃiˌnaːltɒm ɒ ˈrɛndɛʃ ˈɒlijɒt

And then I complete-1 Sg Past the proper Aliyah ACC

And then I completed the proper immigration.

Another dominant group refers to local cuisines (n = 43). These items are single nouns and mostly occur in the nominative case (20) or in the accusative case (20). They can be seen as a way for the speaker to fill cultural gaps (e.g., no translation equivalent in Hungarian; see example [4]) or to reaffirm his identity culturally (e.g., example [4]) (Guzzo, 2014, p. 163).

[3] Hát a schitzel és a krumplipüré.

haːt ɒ ʃnitsɛl eːʃ ɒ krumplipur̪eː

Well the schnitzel—NOMINATIVE and the mashed potatoes.

Well, the schnitzel and the mashed potatoes.

[4] a mallavahot, ezt is szoktam csinálni

eː ɒ malːavɒhot ɛst iʃ ˈsoktam ˈt͡ʃinaːlni

And the malawach -ACCUSATIVE, I also usually make—1 Sg Past this.

And the malawach, I also usually make this.

CS also referred to participants’ places of residence or schooling, which could have been challenging to translate or could not convey the same meaning.

[5] Itt a Schoonában

Itt ɒ ʃunaːban

Here the neighbourhood—INESSIVE

Here in the neighbourhood.

Hebrew words are integrated into the Hungarian matrix language using Hungarian morphological markings. In all the examples above, the Hebrew lexical items remain uninflected internally but receive Hungarian suffixes, indicating that speakers treat these items as syntactic constituents of Hungarian. The [1] example shows that Hebrew proper nouns may enter Hungarian syntax as complete noun phrases and take spatial cases, just like any Hungarian place name. As -t marks objects in Hungarian, both Hebrew insertions in examples [2] and [4], despite being semantically foreign, receive it, indicating syntactic integration. The productive use of the Hungarian locative suffix is shown in example [5], which uses a Hebrew lexical root that is phonologically adapted—the vowel-lengthening rule—and treated as a Hungarian stem. The data show that Hebrew nouns are morphosyntactically assimilated into Hungarian, in which Hungarian provides the grammatical frame, with case, definiteness and agreement expressed on Hebrew nouns.

4.3.2. The Verbs and Their Functions

Hungarian has two productive derivational suffixes to derive verbs from nouns by either the suffix -z or -l; the distribution of the two suffixes is phonologically conditioned: internet-ez ‘to use the Internet’, szév-el ‘to save’ (Kiefer, 2010, p. 719). There are 9 identified items in the data, of which 7 are derivational blends in which the Hebrew root עליה (in English pronunciation, aliya) is ‘Hungarianised’ (e.g., example [6]).

The Hungarian-derivational morpheme ‘z’ is attached to the end of the word to ‘make it’ a Hungarian verb. However, the suffix is preceded by the vowel-lengthening phonological rule, meaning the stem final vowels a/e are regularly lengthened before suffixes (Kiefer, 2010, p. 717) [7].

[6] 2014-ben alijáztunk

2014-ben ɒlija:ztunk

2014 INESSIVE aliya—á—z—1PL PAST

We immigrated to Israel in 2014.

[7] Sokat ginázunk.

ʃokat ɡinaːsunk

Many park—á—z—1PL PRESENT

We are a lot in the park.

Overall, the predominance of intra-sentential insertions—particularly in the nominal group and the morphologically integrated verb forms—signals systematic use of bilingual resources.

5. Discussion

5.1. Characteristics of Parental Language Input

RQ1 examined how HL, ML and additional languages are used across interactional directions within the family. The findings indicate that HL remains a central component of parental input; however, it is rarely used in isolation. Instead, HL input co-occurs with the ML and, in some cases, with English. This pattern reflects the multilingual realities of transnational households rather than a linear shift toward ML dominance. Children’s responses were most frequently bilingual, which indicates that they do not simply shift toward the ML. Instead, children use both repertoires when interacting with their HL-speaking parent. This is aligned with previous research that emphasised children’s agency in shaping their families’ overall language use (Lanza, 2004; K. A. King et al., 2008; Schwartz, 2010). It also points to variation and resilience in HL development (Kupisch & Rothman, 2018).

Shared family spaces—particularly mealtimes—emerged as dynamic interactional sites where linguistic boundaries were fluid. The co-occurrence of HL, ML and English in these contexts suggests that family communication operates through integrated multilingual practices rather than strict language separation. This observation aligns with translanguaging perspectives in which bilingual speakers mobilise their full linguistic repertoire using HL, ML and English (García & Wei, 2014; Otheguy et al., 2015).

Overall, the findings indicate that negotiation and adaptability characterise parental language input in this community, and flexible interactional practices support HL maintenance.

5.2. The Impact of ML on HL Variation

RQ2 focuses on the impact of the majority language on HL use, examining how societal pressures, parental ideologies and interactional strategies interact within FLP. At the community level, the findings show that perceived pressure to integrate into the ML-speaking environment is associated with reduced HL use in parent–child interaction. Families who experienced lower integration pressure tended to attribute greater importance to HL, suggesting that sociolinguistic security supports ideological commitment to HL maintenance. This relationship aligns with research demonstrating that HL development is shaped not only by family decisions but also through social spaces that legitimize HL use (Fishman, 1991; Carreira & Kagan, 2011).

At the interactional level, most participants reported adapting their language use to ML-speaking contexts, especially in public or institutional spaces. This was also found in previous research, which shows that external pressure—such as the school environment and peer groups—can limit opportunities for sustained HL input (Schwartz, 2010; Curdt-Christiansen, 2013). However, a smaller subgroup of participants who maintain HL regardless of context was also reported in studies that described strong parental ideology and intentional language management in the development of HL over time (De Houwer, 1999; Lanza, 2004).

The strategies participants used further reflect the earlier-mentioned dynamic negotiation. The most frequent practices—visiting the home country, literacy activities—are consistent with the notion that HL development in small communities is often supported by external cultural reinforcement rather than by strict in-home language rules (Bartha, 1995; Szilágyi & Szécsi, 2020). This suggests that families rely on lived experiences, cultural continuity, and transnational ties to maintain HL, rather than constant correction or enforcement.

Attitudes toward CS at the ideological and structural levels played a meaningful role. The positive association between favourable CS attitudes and higher CS occurrence confirms that speakers’ beliefs share their linguistic behaviour (Fuller, 2009; Dewaele & Wei, 2014). Variation in the frequency of switching to the ML suggests that ML influence is present, but not to the same degree across participants. It is also consistent with prior findings of Byers-Heinlein (2013) and Bail et al. (2015).

Together, these findings show how HL variation emerges through the interaction of societal pressure, parental ideologies, and everyday practice as components of FLP.

5.3. Structural Patterns of CS and Translanguaging Practices

RQ3 addressed the linguistic outcomes of bilingual family interactions by examining the structural patterns and communicative functions of ML-origin insertions within the HL-dominant speech.

The analysis of ML-origin insertions in HL speech shows how bilingual parents in this community strategically use multiple linguistic resources when interacting with their children. Although CS was observed across all interviews, it did not occur randomly.

The dominance of intra-sentential insertions, especially in the nominal group, aligns with Muysken’s (2013) research, as it results from their semantic flexibility and lower morphological constraints. The functional motivations of these insertions reinforce their systematic nature. ML proper nouns referring to institutions, bureaucratic processes and local infrastructure (e.g., Misrad Hapnim, Sochnut, Maccabi) when direct HL equivalents were either not available or inappropriate. These cases are described as cultural or situational loans (Matras, 2009). Similarly, lexical items related to local cuisine and everyday life are used to mark identity. This resonates with Guzzo’s (2014) findings that food-related CS represents belonging and cultural attachment.

Verb insertions displayed an additional level of integration. Hebrew lexical roots, adapted through Hungarian verbal derivational morphology (e.g., alijázni, ginázni), suggest deep structural integration into HL. This degree of morphological integration suggests high bilingual proficiency and flexible cross-linguistic processing, which aligns with Kupisch and Rothman’s (2018) study on heritage speaker competence. Although the coding procedure classified these items as CSs, their functional nature points toward a translanguaging orientation. Participants used a unified linguistic repertoire for meaning making, expressing identity and sociocultural position (García & Wei, 2014; Otheguy et al., 2015; Wei, 2018). Therefore, these insertions are pragmatic and used to create meaning and should not be reduced to a binary HL-ML contrast.

Taken together, the findings indicate that these insertions are part of a dynamic bilingual communication. They enable participants to navigate their transnational identities, maintain cultural continuity and build a shared bilingual language use within their families.

5.4. Theoretical Contributions and Synthesis

The findings demonstrate that families use cultural and emotional connections to the heritage country and flexible language practices to maintain HL in the absence of institutional support. These results support theories of transnationalism (Bryceson & Vuorela, 2002; Vertovec, 2009) by showing that HL development is not only a linguistic process but also a transnational practice supported by family ties across borders. In this way, HL becomes a tool through which families negotiate belonging in the host and home societies. The study also contributes to research on FLP by showing how ideologies, management and practices are constantly negotiated in multilingual contexts. Although parents expressed a strong ideological commitment to HL, realities of the ML-speaking environment and children’s agency shape their practices. These findings highlight adaptation and resilience in dynamic FLP models (Schwartz, 2010; Curdt-Christiansen, 2013). These families maintain HL through external reinforcements and flexible bilingual practices that allow HL to coexist with ML. Furthermore, the structural integration of Hebrew insertions indicates that bilingual parents show a flexible and productive linguistic repertoire. These findings align with translanguaging perspectives (García & Wei, 2014; Otheguy et al., 2015; Wei, 2018), indicating that mixed input can be a resource for meaning-making, identity construction, and intergenerational continuity.

6. Limitation

The study is limited by the selection of participants. The sample size consists of highly educated, post-2000 migrants, mainly mothers, with a strong positive orientation toward multilingualism and active ties to the home country. The predominance of mothers in the sample reflects a well-documented pattern in FLP research showing that mothers are more likely to participate due to their central role in everyday caregiving, educational mediation and daily language interaction within the home (De Houwer, 2007; Curdt-Christiansen, 2009).

Although the presence of children during interviews created a more naturalistic setting, it may also have influenced parents’ linguistic choices. In addition, the findings may be affected by the observer’s paradox (Labov, 1972), as participants’ awareness of being recorded can shape language behaviour, including CS patterns and language choice.

While the coding of CS and lexical insertions was systematic, the spoken data offer only a partial window into the multilingual repertoires that participants use across different domains.

These analyses are interpreted cautiously and are intended to complement the qualitative findings rather than support generalisation beyond the present sample.

7. Conclusions

This study aims to contribute new insights into HL development within a small and understudied immigrant group: Hungarian-speaking transnational families in Israel. Based on semi-structured sociolinguistic interviews, the study examined how families negotiate the use of HL in ML-dominant environments. While this paper is somewhat ambitious in covering language attitudes, language maintenance and shift and bilingual practices, it offers a useful overview of the Hungarian-Hebrew families in Israel, and the variety of FLPs adopted in this small community.

Overall, the findings highlight HL development as a dynamic and relational practice shaped by family interaction, social positioning and transnational experience. By adopting an FLP perspective, the analysis demonstrates how ideologies, management strategies and everyday practices remain in continuous negotiation. The findings underscore the importance of flexibility in bilingual family life, that HL maintenance can be sustained through adaptive multilingual practices and demonstrate that HL development is not uniform but negotiated.

Although the findings cannot be generalised beyond the present study, they point to emerging tendencies in how transnational families use linguistic, cultural, and emotional resources to support intergenerational HL maintenance.

It extends and deepens the work begun by Bilgory-Fazakas and Armon-Lotem (2025) by moving from questionnaire data to in-depth interviews that allow closer examination of bilingual practices. Future work will investigate the language use of the heritage speakers, namely, the children of the families participating in the current study. That phase will use a storytelling assessment based on visual prompts (MAIN), in which participants will be asked to construct narratives in response to a series of photographs. These photo-elicited stories will be analyzed to explore language choice, narrative structure, and how children draw on their linguistic repertoires when constructing meaning.