The gendered migrant experience: a study of family language policy (FLP) amongst mothers and daughters in the Somali community, Bristol

ABSTRACT This article adopts a gendered take on Family Language Policy (FLP) by questioning the way that gender impacts on the issues faced by refugee woman during and after flight. For this reason, the ethnographically informed research addresses the concerns and experiences of mothers and daughters in the Somali community in Bristol, one of the fastest growing communities in the city but one that remains a ‘neglected social group … everywhere present but in many ways invisible’ (Wallace & Kahin [2017]. Somali parents and schooling in Britain (p. 1). UCL Institute of Education Press.), with little known about their experiences on or after arrival (Warfa et al. [2006]. Post-migration geographical mobility, mental health and health service utilisation among Somali refugees in the UK: A qualitative study. Health & Place, 12(4), 503–515.). The study of FLP not only contributes to our understanding of the processes of language shift and change, it also sheds light on broader language policy issues at societal levels. Analysis suggests that it is principally mothers who take on the demanding, yet invisible work of FLP in the home and that mis-matched fluency between mothers and daughters results in a fracturing of family relations with the potential for long-term emotional repercussions. The findings have implications for educational and public sector organisations involving immigrant communities.

This article adopts a gendered take on Family Language Policy (FLP) by questioning the way that gender impacts on the issues faced by refugee woman during and after flight. For this reason, the ethnographically informed research addresses the concerns and experiences of mothers and daughters in the Somali community in Bristol, one of the fastest growing communities in the city but one that remains a 'neglected social group … everywhere present but in many ways invisible' (Wallace & Kahin, 2017, p. 1), with little known about their experiences on or after arrival (Warfa et al., 2006).

Family language policy
Whilst language policy has traditionally focused on public and institutional contexts, with less attention being given to the intimate sphere of the home and family 1 more recent work King et al., 2008;King & Fogle, 2006Smith-Christmas, 2016) has firmly established the field of family language policy (FLP), concerned with what Luykx (2003) refers to as the 'ecology of the family.' FLP adopts a holistic approach that seeks to understand the top-down, macro forces, and bottom-up, micro influences that shape the 'explicit and overt as well as implicit and covert language planning' (Curdt-Christiansen, 2018, p. 420) that takes place in family homes. FLP also acknowledges the, at times, 'informal and unplanned ' (O'Rourke & Nandi, 2019, p. 4) nature of FLP with grassroot practices becoming de-facto policy. At the microlevel, research addresses the dynamic and complex practices and planning decisions made and contested by family members in the home (e.g. Smith-Christmas, 2014).
Whilst research is increasingly interested in how families are constructed through multilingual language practices (King & Lanza, 2019), a more critical question would be the extent to which language plays a role in the destablising or breaking down of family ties and connections. Whilst families can, through everyday talk, forge the ties that make a family (Tannen et al., 2007), family members can also contest, negotiate and challenge the values and beliefs. This research will view FLP as being shaped by experiencesthe experiences of different generations and individuals within families have a significant implication on policy, practice and research (Hua & Wei, 2016).
FLP also acknowledges that however tightly knit, families do not live in a vacuum, isolated from the larger sociocultural environment. On the contrary, they constantly interact with others in sociolinguistic, sociocultural, socioeconomic, and sociopolitical contexts (Spolsky, 2012) and therefore, practices within the home are shaped by societal developments on local, national and global scales. FLP makes 'visible the relationships between private domains and public spheres' (Curdt-Christiansen, 2018, p. 421). The complexity of this relationship is central to the field of FLP but is of particular interest to this research given that the relationship and contrast between private domains and public spheres will be particularly stark in refugee/migrant homes and families. This paper takes up the call, from King (2016, p. 2) for the inclusion of different family types, such as displaced families, and the addition of more diverse language constellations into the field of FLP and seeks to add to the existing literature on FLP within refugee families (Bezcioglu-Goktolga & Yagmur, 2018;Gafaranga, 2011;Navarro & Macalister, 2016;Purkarthofer & Bordal Steien, 2019;Revis, 2020). Whilst these families' situations may be comparable to those of socio-economically advantaged migrants in some aspects, Revis (2020) points out that 'their situations tend to be more delicate and vulnerable' (p. 41) not least because as the victims of forced migration many families arrive with limited financial resources, lack of host-language proficiency and limited access to employment and support.

Gendered FLP
It has become common place to conceive of gender as something that is 'done' and 'performed' (Butler, 1990) and consequently, academic research has reached the conclusion that gender is rooted in and jointly constructed in particular communities of practice (Eckert & McConnell-Ginet, 1992;Tannen, 2008), with differences in ideologies and meanings of gender and gendered linguistic practices having been identified across several speech communities around the world (See Pavlenko et al., 2001).
FLP research has yet to fully address how 'language and gender ideologies intersect' (Piller & Gerber, 2021, p. 12). Tannen suggests that 'family interaction is a continuing negotiation of gender identities and roles' (Tannen, 2008, p. 200) and Wright (2020) reminds us that a focus on families and how they talk about, use, negotiate and form language policies can bring to the foreground gender roles within the family language processes. This research further focusses on understanding the 'gendered processes and social relations contained within experiences of forced migration' (Hopkins, 2010, p. 520) with an acknowledgement that embodied in roles, relations, and hierarchies, gender is a core organising principle of social relations and opportunities (Boyd, 1999).
This research seeks to understand the gendered dynamics amongst the Somali community in Bristol. 2 Whilst there has been an increased focus on refugees, including women, in research and policy, Hopkins (2010) argues that the way that gender impacts on the issues faced by refugee woman during and after flight, her responses to situations and strategies for resolution and the outcomes of flight and resettlement, requires further investigation. This research seeks to do just that by understanding the gender dimension of migrationhow do resettlement and transnational connections produce, reproduce and alter gender, cultural and ethnonational identities? In other words, given the complex social, cultural and political processes of migration in what 'multiple ways … do women and girls … actively and passively … [accommodate and resist]' (Kirk, 2010, p. 161) forces which both empower and entrap them (Marchand & Runyan, 2000). Additionally, Bezcioglu-Goktolga and Yagmur (2018) remind us that FLP of second-generation immigrant families is an under-researched topic. The novelty of this research is its elaboration of the understanding of the second generation by combining their experiences with those of their mothers. This will shed light on the intercultural differences are emerging between generations within the UK Somali community (Sporton & Valentine, 2017). What linguistic choices are made? What language experiences are had? and how does this impact on language learning trajectories and family relations?
Role of the mother Tannen (2008) suggests that mothers play a key role in families (See Braunshausen, 1928 for an early discussion of different roles in the family), functioning as part of 'dyads' (Gross, 1951) with other family members. This paper seeks to explore one such dyad, the dynamics between mother and daughter and will examine the ways in which family interaction differs across gendered lines and the means by which it is negotiated between 'caregivers and children in tandem' (Smith-Christmas, 2018, p. 131).
By taking an active role in their children's socialisation, mothers become crucial to language maintenance or shift within the family (Kayam & Hirsh, 2012;Nakamura, 2016;Tuominen, 1999), playing a 'more active role than fathers in their management strategies' (Bezcioglu-Goktolga & Yagmur, 2018, p. 55) but research acknowledges that this remains largely un-noticed (Okita, 2001). To date, research has yet to fully address the relationship between mothers and their children, something this paper seeks to redress principally because, as (Fogle, 2013) points out, children perceive themselves as agents in shaping FLP. Additionally, if we accept that bilingualism and multilingualism are, for the purpose of this research, best seen as experiences; the experiences of different generations and individuals within families have a significant implication on policy, practice and research (Hua & Wei, 2016).

Emotion and FLP
King (2016) characterises what she terms (pp. 2-3) the 'fourth' (and current) phase of FLP research as involving 'questions that examine language competence not just as an outcome, but as a means through which adults and children define themselves, their family roles, and family life.' This is an important reminder that the decisions and the conversations that happen at home have an impact far more important, in many ways, than language competencyfeelings about yourself, your family, your value and your role.
Whilst there has been a growing interest in the intersection of language and emotion (e.g. Dewaele, 2004;Pavlenko, 2006), Smith-Christmas (2018) emphasises the lack of research into the affective component of FLP -FLP is often affected by and affects emotional issues and psychological dimensions, but these are seldom acknowledged as central in the analysis of such policies. Tannenbaum (2012) outlines the impact of forced migration on individuals and familiesit is In this regard, FLP is inherently different from broad national and social policies in the input and effect of emotional issues and psychological dimensions. Past and present experiences and hopes and worries about the future are a core aspect of family life and ignoring this aspect misses the core concern of many families.
Heritage language maintenance has been found to support closer intergenerational relationships, parent-adolescent discussion, and greater family solidarity (Farr et al., 2018;Tannenbaum & Howie, 2002;Tseng & Fuligni, 2000). Data from Smolicz (1981Smolicz ( , 1992, who explored language as a core value in a range of contexts and minority groups, pointed out the emotional centrality of language maintenance in the family on the one hand, and of acquiring a new language on the other. Smolicz et al. (2001) indicated that a shared family language was linked with a closer family bond with parents identifying that this prevented a possible disintegration of family life, something the work of Guardado (2002) also confirms in the Canadian Hispanic community. By contrast, reduced heritage language proficiency in children has been shown to have negative repercussions on family well-being including tensions in parentchild relationships, reduced intergenerational communication, and feelings of alienation from the wider cultural community (Oh & Fuligni, 2010;Tseng & Fuligni, 2000). Issues of language anxiety have been especially dealt with in the work of Sevinç (2020) where there is a detailed discussion of the 'contagion of anxiety, spreading from one generation to the next' (p. 84)a language anxiety principally focused on the linguistic ability of younger generations and their future prospects and triggered by 'sociolinguistic and emotional pressure on normative standards, cultural values, beliefs, and practices, such as the tension between home language maintenance and shift (Sevinç, 2016).
In summary, this paper adopts a broad and multi-faceted understanding of FLP, engaging with the 'explicit and overt as well as implicit and covert language planning' (Curdt-Christiansen, 2018, p. 420) that takes place in family homes, the influence of wider external influences as well as the gender dynamics, and the challenges, obstacles and frustrations of family language maintenance.

Research study
This paper reports on research 3 carried out and inspired by an outreach project co-constructed by researchers from a British University and members of the Somali community in Bristol between May 2018 and August 2019.
Ninety-one languages are used in Bristol, with Somali reported to be the third most frequently used, behind English and Polish (Bristol City Council, 2012). The Somali population in Bristol is estimated to be between 8,300 and 10,000 (Bristol City Council, 2014) and is concentrated in some of the most deprived wards of the inner and eastern areas of the city (Insight, Performance and Intelligence Service, 2019).
Somali migration to Bristol can be divided into three distinct phases (Sporton & Valentine, 2017) with the first arrivals to Bristol being economic migrants from the British Protectorate of Somaliland in the early twentieth Century, employed mostly by the Merchant/Royal Navy in the seafaring trade (Olden, 1999). The late 1980s saw the arrival of many Somalis fleeing political persecution and civil war (Sheikh & Healy, 2009) and since then, many families have arrived in the UK from other European countries (Van Liempt, 2011) as part of the secondary or 'two step migration' (Wallace & Kahin, 2017, p. 82). Thus, the current UK Somali population not only comprises of different migratory journeys (Rutter, 2004) but is also characterised by continuous movement within the diaspora (Kapteijns & Arman, 2004). However, the Somali community, despite being one of the fastest growing communities in Bristol, remains a 'border-line community of migration' (Bhabha, 1994, p. 12), a 'neglected social group … everywhere present but in many ways invisible' (Wallace & Kahin, 2017, p. 1), with little known about their experiences on arrival (Warfa et al., 2006). This paper, in line with Curdt-Christiansen (2013) adopts an ethnographically grounded approach to the study of FLP in order to gain an insight into the language ideologies held by members of the Bristol Somali community in order to better understand how FLP is established, negotiated and implemented. I carried out participant observations of weekly conversational classes and excursions and spent time in the homes of many of the Somali women, observing the language practices of mothers and daughters. These observations resulted in 24 sets of fieldnotes, representing approximately 72 h of fieldwork. These fieldnotes are treated as 'primary data sets' (Blackledge & Creese, 2010), and therefore 'evidentiary … and used equally alongside other data' (Blackledge & Creese, 2010, p. 79) rather than merely 'aides memoir' as Emerson et al. (1995, p. 23) argue. In addition to these, a series of ethnographic chats (Selleck, 2017) were run, through the medium of English, with both mothers and daughters. These ethnographic chats provided an additional 165 min of data which was subsequently transcribed. Whilst these ethnographic chats were run primarily through English, the relaxed format allowed for informal translation and conversational support to be given by other participants. Additionally, some use of technology was apparent with mothers translating key words where needed. 4 Whilst the data presented within this article were elicited primarily through the ethnographic chats, data from fieldnotes also informs the analysis. 5 The observations evidenced 'ideologies in action' (Jaffe, 1999), what people actually do whilst the more structured question-based ethnographic methods were deployed to elicit evaluative discourse and key ideological stances as well as an analysis of reported language practices. It was hoped that the combination of methods would allow for a richer and more nuanced understanding of the research context as well as allowing for a deeper understanding of the complex cultural and ideological pressures on Somali women in Bristol.
Extract 2 -Sahra and Talibah 6 (Somali Mothers) S: In the house just Somali (laughs) I decide just Somali (.) it is important to keep the Somali (.) this is my rule as mother (.) but my children they always try speak English (.) they can speak Somali with a little help (.) I teach them the Somali words but always they say that English is easy (.) easier T: Yes (.) the children (.) for English is easy (.) when I'm talking my daughter (.) I'm talking in Somali but she (.) S: They mix hey (.) English and Somali (.) but we have to teach our children to speak Somali language (.) it's important as their mum (.) yeah (.) it's how they understand (.) because we don't speak English very well and this is our important role to keep Somali (.) a reminder of home (.) of home (.) Somali language is being Somali Extract 3 -Aisha (16years old) My mum is the main person to push Somali (.) to be honest my dad is just as bad as me In both extracts, it emerges that the role of FLP is largely taken up by mothersthey carry out the emotionally demanding yet invisible task, largely by themselves, of negotiating and managing the complex language needs and capabilities of their children, who span, in many cases, a wide range of years with multiple migratory trajectories. This task is made more difficult by the apparent lack of support from fathers who are represented as problematic partners and parents, who cannot be relied upon to keep their part of bilingual parenting arrangements and whose language practices with their children need to be policed by mothers. That said, this practice is broadly in line with the cultural expectations placed on the Somali mothers. In Somalia, a woman's relationship to her own and her husband's family and her identity and capacity as a woman, wife and mother, are mediated by traditional rules of kinship which set norms for appropriate gender-related and age-related behaviour (Kapteijns, 1993). Somali society is inherently patriarchal, with girls expected to be obedient and passive and men, traditionally, playing 'little part in the domestic sphere' (Wallace & Kahin, 2017, p. 15).
For the Somali mothers bringing up their children in Bristol, the desire to maintain cultural identity is a significant motivating factor in transmitting their heritage language to their children (See Guardado, 2002). They see their role, as a mother, to care and nurture a sense of Somali in their children, to in their words keep a connection, 'a reminder of home' (Extract 2, line 13). Language, for these mothers, is the medium through which they hope children will develop their own sense of Somaliness and develop attachment and belonging to Somalia (Extract 2, line 14). Language can thus be seen to function as an 'anchor point' (Ramsden & Ridge, 2013, p. 241), a 'thread of continuity and security extending into their new lives' (Ramsden & Ridge, 2013, p. 241). Correspondingly, parents saw loss of the Somali language as a loss of Somali culture and a loss of connection with the place many parents still viewed as 'home' (See also Extract 8). Clyne (2003) argues that refugees may be more likely than migrants to uphold cultural traditions. The reason, he contends, is that they are usually forced to leave their country because of a 'temporary' plight and may well be longing to return once the situation has stabilised.
What emerges in the above extract is that at the level of general policy, the mothers and the daughters don't appear to differ in the belief that it is important to learn both Somali and English; where they differ is in how to manage the bilingual repertoire of the younger generation via the family language policy and via the less structured ideologies of language that govern practice.
Whilst the mothers discuss the 'mixing' of Somali and English by the younger generation, the implication of the comment 'we have to teach our children to speak Somali langauge' is that the version of Somali that is aspired to should be void of any English borrowings; in other words 'just Somali.' In this sense, the mothers seem to adopt a sense of 'separate bilingualism,' the traditional notion of keeping languages separate and exercising a choice between them (Blackledge & Creese, 2010). The young Somali women seem to align more closely with a sense of 'flexible bilingualism' and other related concepts such as translanguaging (García, 2009) and heteroglossia (Bailey, 2007), all of which point to 'an approach to bilingualism that is centred, not on languages as has often been the case, but on the practices of bilinguals' (García, 2009, p. 140), where these practices are presumed to include fluid movements of various sorts between language codes. 7 An important element of FLP is that discussions also account for the efforts taken by parents to support their children's heritage language maintenance. Through ethnographic observations, it was noted that mother's expend considerable time looking to promote and encourage engagement with Somali culture. Whilst, at the time of carrying out the research, there was no formal community language education for Somali in Bristol, mothers sought out opportunities for their children to engage in informal peer-to-peer Somali conversation through initiatives like holiday homework camps. An emphasis was placed, my both mothers and daughters, on return trips to Somali. These were seen as a way to foster cultural awareness and to highlight the importance of maintaining language competency in Somali.
That said, it has been argued that in taking on this role of passing on traditional values and protecting the ethnic language, mothers become confined to socialising inside their own ethnic community and thus sheltered from the influence of the host language (e.g Joudi Kadri, 2009, p. 95), something that we saw mentioned in Extract 1, above.
That said, like many immigrant families (see King, 2000) the pro-community language ideologies held by the Somali mothers do not directly translate into community language maintenance. The Somali women report (in Extract 2 and more explicitly discussed in Extract 4, below) that they encounter problems establishing the language rules; in practice, transmitting the heritage language to children becomes a daily struggle, requiring effort to consistently speak the language and vigilance to ensure it is spoken back.
Extract 4-Kameela (Somali mother) My children refuse to speak Somali at home (.) they only want to use English (.) not only between themselves but also to us (.) perhaps that's why my English is better now (laughs)they don't even like Somali food (.) instead they choose to eat fried food (.) pizza (.) it is what their friends are eating (.) I understand but it makes me sad (.) it all starts when they start school (.) they go from being Somali children in the family to Bristol children Despite an emphasis on speaking the heritage language in the home environment, the mother here reports that once children start mainstream school, they often struggle to see its relevance to their life and actively resist their parents' (or mothers) efforts to maintain it (See Donghui & Slaughter-Defoe, 2009;Ren & Hu, 2013;Seloni & Sarfati, 2013), with these pro-minority language ideologies coming into conflict with the majority view (Dauenhauer & Dauenhauer, 1998). There is a strong sense that the children go from being Somali children to citizens of Bristol, no longer do the mothers perceive themselves as sole mitigators, guardians and curators of their children's experience of the world, instead the school and the peer pressure that this exposes young people to, comes into play with fairly dramatic and long-lasting effects on the family dynamic. The refusal to not only speak their heritage language but also to engage in other culturally desirable activities, is seen here to be a constant source of relationship tension and upset (see Fillmore, 1991 for further discussion of this). Moreover, the children seem to take significant agency by introducing English into the home domain, in some cases even influencing other family members to use it (Line 3), and thus initiating language shift. In this sense, the parents surrender to external forces by ceasing, or at least limiting, the cultural and linguistic practices of their heritage language and, possibly unaware of the conversational mechanisms by which language shift occurs, are unknowingly playing a part in the demise of the minority (and home) language (see Smith-Christmas, 2016). This highlights a contraction between parents' reported linguistic ideologies and the 'practiced' family language policy (Mirvahedi, 2021) and reminds us of the sometimes ad-hoc, 'defacto, informal and unplanned' nature of FLP (O'Rourke & Nandi, 2019, p. 4).
Therefore, the children themselves contribute to the process of forming the language policy around them, something also confirmed by Mirvahedi (2021) who discussed the particular importance of older siblings who have more experience of the outside world. Therefore, the willing participation of young people in adult-initiated practices cannot be assumed (Paugh, 2005;Rindstedt & Aronsson, 2002). So how then do young people perceive the family language policy? On what basis do they resist or affirm home-based ideologies and practices?

The gendered turn
Research suggests that home and school experiences amongst immigrant families vary across intersecting gender and race boundaries (Lopez, 2003a(Lopez, , 2003b in ways that may impact on bilingual proficiencies (Confresi, 1999;Hondagneu-Sotelo, 1994;Toro-Morn & Alicea, 2003;Urciuoli, 1991). Research indicates that girls and boys in immigrant families are socialised differently across gender lines and encouraged to pursue different paths in life (Portes & Rumbaut, 2001). Consequently, research has suggested that girls are more likely to maintain bilingual proficiency than boys (Lutz, 2006;Portes & Hao, 2002;Stevens, 1986;Veltman, 1981). These arguments are reflected in the data, with Somali daughters self-identifying as the 'eldest,' 'first-born' daughter with the suggestion that this label brings with it stricter rules and responsibilities than would be expected of their brothers, particularly in relation to the use of Somali. Comments such as 'I'm the eldest daughter of course, I have to speak Somali' or 'I'm the eldest girl, naag weeyn, (translating roughly as 'great women') I try to teach my siblings Somali' were common. With these large multi-generational families, the older daughters seem to play an important role in the early patterns of language use amongst younger siblings, which again gives a nod towards the role that young people play in dictating and shaping FLP. Given that older children tend to receive more overall input in the home or minority language than their younger siblings and therefore achieve higher levels of competency in the language (Dumanig et al., 2013) it is perhaps unsurprising that we see older daughters reporting that they adopt a role, as their mothers also do, of passing on the Somali language and culture but of course it raises important questions around the gender-specific challenges facing immigrant women.
In the following extract, the girls make gender explicit and portend its troubles, something Langellier (2010) also identified in her research with Somali refugees in America.
Extract 5 -Zahi (16years old) and H: My brother totally gets away with not speaking Somali (.) it's like all ok for him the oldest boy (.) all his friends are white British (.) he listens to English music and stuff (.) but for us girls its like about who we are seen to be (.) Z: oh god yeah its like who is she going to marry (.) a nice Somali boy (.) she needs Somali (.) and I think I agree (.) the older I get the more I realise I need to work on my Somali (.) I've missed out The girls suggest that they feel the burden to linguistically conform more than their brothers who apparently 'get away with not speaking Somali,' and who are encouraged, it seems on the basis of this extract, to integrate into the host society. The girls perceive that they should keep up appearances, look to behave like 'good,' traditional Somali girls, with this defined, and noted in fieldnotes as someone who works hard at school, dresses modestly, speaks Somali, shows respect for elder members of the community and takes on responsibility for younger members of the family. Whilst in some ways the girls appear to accept this pressure (Lines 7-8), suggesting that they feel they have 'missed out' on Somali, the use of reported speech in Line 5 also indicates that the girls have a degree of disidentification with Somali culture and appear to reject its patriarchal investments and values that reverberate through family, clan, and tradition, something that has been identified in scholarship and activism (Abdi, 2007;Kapteijns, 1999;Korn & Eichhorst, 2006). Thus, the Somali girls here narrate a 'bewildering entanglement of cultural differences layered with diasporic tensions that deploy feminism against traditional tribal practices' (Langellier, 2010, p. 69) and the result is that the girls appear challenged to sort out what to reject and what to preserve of their culture. On the one hand, the girls are attempting to affirm their own place within British society but against the backdrop of the local Somali community, and Somalia, using language to create cohesion and Somaliness. In many ways then, the girls' language decisions engage a trifocal gaze, one that is 'focused on the dynamic present and another that is looking forward' (Jain & Wee, 2018, p. 112), but with the added complexity that they must also adhere to traditional expectations and values. Hyndman and de Alwis (2003) note from their research in Sri Lanka that the impact of war has been 'both disabling and enabling for women and men … destabiliz [ing] the sexual division of labour, resulting in the redefinition of women's roles in society' (Hyndman & de Alwis, 2003, p. 223) The same may be said for the Somali women in Bristol, they are faced with challenges and opportunities to redefine what it means to be a Somali woman in the city.

The emotive turn
The data is full of accounts and reports from parents and children that parent-child disfluency in a common language is associated with several challenges in the parent-child relationship something neither party is content with.
Extract 6 -Talibah and Faduma (Somali mothers) T: Yeah I no like this (.) it is easy now for English because she have more friend and every day is busy for her (.) for text message and F: Oh yes text message T: talking and (.) F: School and T: When she is talking nothing I am (.) no understanding what she said (.) "bla bla bla"[imitates her daughter talking very fast] F: ahh very quick T: yeah (laughs) she is very quick for me ahh (laughter) you are mum (.) imagine how you feel in this situation (.) hey Extract 7 -Nadifa (Somali Mother) I feel like a bad mum because I can't even take my children to the doctor Extract 8-Talibah (Somali Mother) My daughter needs my help at school and with her body but she don't speak all the Somali words and I no have all the English words I need (.) Somalia means nothing to my daughter, she has not lived there and she has no connection with the peoples there (.) she sees Somali like anybody else (.) she just wants to fit in (.) I must learn English so I can proper mum again Extract 9 -Bilan (14years old) There are times when I feel I don't know my mum (.) I can't just talk to her (.) I need to plan my chat (.) look up words or talk with my dad around (.) and why would I want to do that (.) especially some chats (.) I just get on with my own stuff(.) make my own decisions Extract 10-Aamlin, (Somali mother) M: well I translate using my phone (.) when my children no understand me I go on the phone yeah to help (.) I feel so sad that this is the way I speak with her (.) we do not have a (.) a bond There is clear evidence of changing dynamics in the family unit with reports from both mothers and daughters of the absence of meaningful interactions particularly around academic and personal issues. This seems to result in mothers and daughters feeling isolated from each other with signs that they are increasingly growing apart, emotionally. Although alienation between parents and children is not uncommon, as a by-product of adolescent development (Tseng & Fuligni, 2000) it seems to be more pronounced in the Somali families in this study, with children entering a world that their mothers quite literally cannot understand.
There is an awareness amongst mothers that Somali culture is challenged in resettlement but inter-generational tensions around language are more complex than communication. In Extract 8 it becomes clear that tensions exist around the sense and strength of Somali identity that is fostered in second-generation children. Tensions appear to partly concern creating children's Somali identity and are partly associated with parents' own desire to maintain long-term ties with Somalia. Somalia is an 'oral society' (Wallace & Kahin, 2017) and any perceived breakdown in communication between mothers and daughters is particularly poignant for Somali women who, if their children cannot speak Somali, lose the ability to pass on histories and experiences considered so vital in Somali culture (Andrzejewski & Lewis, 1994). The reality, however, is that the relationship between language and ethnicity is complex; ethnicity is just one component of identity, and it may be highlighted depending on a given situation. Le Page and Tabouret-Keller (1985), in their seminal work, remind us that individuals choose to use language (or not) to 'reveal both their personal identity and their search for social roles ' (1985, p. 14). For the Somali girls there is no single way to mark their ethnicitythey are, instead, at the forefront of carving out the parameters of a new sense Somaliness and of womenhood with indicators of belonging like dress, religion and language coming to hold new values.
As family life becomes dominated by English, participants' ability to cope is mediated by the strategies they use to manage the impact of change. In some instances, language brokers are drawn on, principally fathers, (Extract 9) but also the wider, extended family, and unsurprisingly technology also features (Extract 9 and Extract 10). Whilst it is beyond the scope of this research to address the digitally mediated nature of family communication, Lanza and Lexander (2019, p. 245) conclude that greater research emphasis should be placed on digital as well as non-digital family language practice. Palviainen and Kędra (2020) argue that this would allow researchers to address critical questions such as 'how language choice and multilingual practices are shaped and multilingual discourses constructed through multiple devices' (p. 107). It would also help us to understand whether technology offers the possibility of sustaining the multilingual family in contexts of migration.
Through these strategies, mothers and daughters report having limited exchanges in the home, 'constrained but not ruptured intergenerational communication' (Portes & Rumbaut, 2001, p. 144) but what resonates is the sense of low self-esteem and the feelings of inadequacy as a result. Mothers, through their lack of English, are left on the side lines looking in at family life and young girls, through their lack of Somali, are left dealing with important adolescent milestones alone.
Whilst there is a desire amongst mothers to learn English, to amongst other things, facilitate meaningful interaction with their children (Extract 8), it is also beneficial for these mothers' emotional well-being to be able to speak their first language among themselves and with their children (see Souza, 2008). The vulnerabilities of refugee and asylum-seeking women become clear with the fracturing of mother/daughter interaction having a more significant impact on some mothers' own sense of wellbeing. Mothers position themselves as 'bad mothers' (Extract 8) who are 'failing' to meet expectations, expectations that seem to not only come from within the community, where women are expected to embody all the reminiscences of the country of origin but that are also more generally felt by mothers (See Meeussen & Van Laar, 2018). In Extract 8, Talibha expresses the desire to become 'a proper mum again' by regaining or acquiring shared fluency with her children. There is a strong sense of isolation from their own children something that Bilan (Extract 9) also shares.

Discussion
Through their complex accounts, their memories and lived language experiences, we can understand mothers and daughters manifold positions with regards to family language policy. The ethnographically informed research presented here has highlighted the essential interplay between micro family language practices and macro political policy decisions at the societal level. Untangling and understanding FLP allows for important insights into the everyday processes of language use and communicative practices and can thus lead to better policies to support language maintenance. This paper has addressed the complexities of the female refugee/migrant experience, as seen through the relationship between mothers and daughters. The mothers and daughters represented in this paper have articulated the challenge of not only resettlement and orientation, but also the process of learning and adjusting to what it is to be a Somali woman, mother and daughter in Bristol. We see an almost continuous tension between mothers and daughters, each with their different values and expectations, with the home becoming a hotbed for this angst; the home is anything but a sanctuary for the community language. Mothers and daughters in this study reported that at times there was a fracturing of the communication between mothers and daughters, with limited parent-child fluency in a common language associated with several challenges.
It emerged in the discussions of the data that the work of family language planning fell primarily to the mothers and elder daughters, with fathers represented as problematic partners and parents who cannot be relied upon to fulfil their role in bilingual parenting arrangements and whose own language practices need to be policed by mothers. Additionally, daughters reported that they were subject to differing rules and expectations with pressure to uphold customary norms for gender propriety. Both arguments underline the relevance of a gender-specific division of language 'work' that locates women in the private sphere of the family, housework and care (Davis & Greenstein, 2009;Lopata, 1993).
Contradictory values were placed on language by the Somali mothers and daughters. On the one hand, the community-specific benefits of learning and using Somali are foregrounded (as opposed to any perceived economic benefits), such as the ability to communicate with grandparents and extended family, or the ability to appreciate the literature of a particular language. The Somali language is positioned as a 'gift' (Piller & Gerber, 2021, p. 622) from parents to their children. Yet equally English is considered an asset and the sole requirement for full participation in British life, work and education. This is further compounded by the practices of the state education system in the British context; bilingual language development is regarded primarily as parental responsibility (e.g. Nicholas, 2015;Pacini-Ketchabaw & Armstrong De Almeida, 2006) rather than one that is supported by the state education system and this in turn, places additional burdens on already disenfranchised refugee mothers. The inadequacy of the British state education system in supporting heritage language maintenance in the UK is widely discussed (see Carruthers & Nandi, 2021 for an up-to-date discussion of the issues). So, whilst community languages are a vibrant part of everyday life in the UK, they remain undervalued in terms of cultural capital, community cohesion and economic prosperity (Ayres-Bennett & Carruthers, 2018) and language learning is therefore restricted to outside mainstream schooling, sometimes through supplementary or community schools or through other voluntary organisationsall of which suffer from limited funding opportunities.
In sum, whilst the work of Barakos and Selleck (2019), King and Fogle (2006) and Selleck (2020) amongst others, suggest that parents elsewhere see bilingualism as an investment and an asset, the mothers represented here see only the challenges and obstacles posed by bilingualism; bilingual parenting has begun to add yet another dimension to parental anxieties.
Yet despite this, Somali mothers express both a desire to learn English and a frustration over their perceived inability to do so. Learning English is perceived as a way to gain symbolic capital (Pavlenko, 2002) which would allow women to reconfigure and regain control over the family dynamics disrupted by migration. There is a perception that it would provide resources to them to perform their roles as mothers, which have been negatively impacted by being a migrant in a foreign linguistic and social context. So, whilst great improvements have been made in terms of social inclusion within the Somali community and despite being one of the fastest growing communities in Bristol, the Somali community remains a 'border-line community of migration' (Bhabha, 1994, p. 12), marginalised and disenfranchised, everywhere present but in many ways invisible' (Wallace & Kahin, 2017, p. 1). More needs to be done to enable the Somali community to organise itself socio-economically and politically in order to assert a more powerful voice in present-day Britain. This would allow them, as a community, to move away from the somewhat 'laisser-faire style of language planning'' (Piller, 2001, p. 66), to take the initiative and become agents of change.
Notes 1. Although note that Fishman was an early proponent of the importance of the family in language transmission and maintenance (Fishman, 1991).
2. Ninety-one languages are used in Bristol, with Somali reported to be the third most frequently used, behind English and Polish (Bristol City Council, 2012). The Somali population in Bristol is estimated to be between 8,300 and 10,000 (Bristol City Council, 2014) and is concentrated in some of the most deprived wards of the inner and eastern areas of the city (Insight, Performance and Intelligence Service, 2019). 3. This research meets the ethical guidelines as set out by UNIVERSITY. The research was given approval by the X department ethics committee. 4. The ethnographic chats took place at the end of the outreach initiative, where the focus had been on equipping the Somali mothers with English language skills and above all confidence to use their English. Across the board, the women's English had developed considerably as had their confidence. 5. Whilst the ethnography presented here draws on all of the aforementioned data sources I have had to make choices about what data to present and how best to present that data. In this sense this ethnography, like any other, is partial and restricted and therefore not presented as comprehensive. The examples provided as data extracts are illuminating moments, highlighting key elements of the unfolding story. Most significantly, they show ideological values that are salient to participants, and hence to the research aims. 6. Note that all names have been anonymised.
Transcription conventions -(.) -An untimed, short pause (3.0) -A timed pause, in seconds Speech -Transcribed speech (text) -Commentary [text] -Clarification "speech" -Voiced speech 7. Flexibility, as an ideology, does not entail that, in terms of practice there will be no detectable boundaries between languages at the level of use, only that this boundary is less firmly policed and respected.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Notes on contributor
Dr Charlotte Selleck is a senior lecturer at the University of the West of England. She is also a member of the Bristol Centre for Linguistics. Her research interests remain in the field of sociolinguistics, including sociolinguistic theory, the sociolinguistics of Wales and interactional sociolinguistics. She also has a keen interest in linguistic ethnography.