‘Half of who you are’: Parent and child reflections on the emotional experiences of reversing familial language shift

Aims and objectives: The study explores the emotional experiences of a mother and son during the re-introduction of the heritage language at age 6, following a 2-year hiatus. In particular, it seeks to highlight the affordances of a collaborative, emotionally sensitive approach to family language policy, with parent and child jointly incorporating and acknowledging aspects of the child’s identity development. Methodology: Adopting an autoethnographic approach, mother and son kept a joint research diary where critical incidents (as chosen by either one or both) were written down as verbatim as possible, and subsequently reflected on together, with reflections again entered into the diary. Data and analysis: Data were coded according to various emotions (frustration, pride, joy, love, guilt, etc.) as well as a code linked to identity development. These were then thematically analysed. Findings and conclusion: Parent–child collaboration and facilitation of child agency have a positive impact on a child’s attitude towards learning the heritage language. Jointly sharing and reflecting on the emotional journey of growing up bilingual and bringing up a bilingual child affords both parent and child an insight into each other’s concerns, charting a path towards a collaborative approach to family language policy development. Originality: The study is the first officially documented long-term study focusing on the reversal of familial language shift which provides emotional and language-related data from both parent and child, as part of a complex tapestry of emotion, identity, language, and parent–child relationship. Significance and Implications: Recommendations are made for children to be more actively involved in family language policy research, taking into account their emotional links to, and experiences with, various aspects of language use. In addition, a focus on emotions can assist in a better understanding of the links between language and identity, giving children and parents the space to reflect on and articulate their different points of view, shaping family language policy.


Introduction and background
The onset of school can have a negative impact on heritage language development, both as a result of additional time spent in an environment where the societal and 'standard' languages lead to competing language ideologies (see e.g., Farr, 2011), and the added impact of school requirements and performance criteria (Department for Education and Skills, 2007) that may threaten heritage language use in the home (Kabuto, 2010). The autoethnographic study presented here follows the first 2 months of re-introducing the heritage language from age 6, following such a commonly described home language hiatus which coincided with the school start age (age 4 in the United Kingdom). Specifically, it follows the emotional experiences of both mother and child, through keeping and reflecting on a joint research diary.

A word on context
Since this research project actively involved a child as co-researcher, the paper frequently returns to the question of the role of children in multilingual research. However, this is not the core context of this paper, which is fully explored in another publication (Little and Little, 2022). This paper, instead, aims to provide enough methodological context to the study and family background to make the focus -the emotions experienced and negotiated by Toby and his mother, myself, during the reintroduction of German, the heritage language -accessible to the outside reader.
To understand the context of the research, understanding family background and family language policies is important (De Houwer, 2007). Up until school start at the age of 4, Toby experienced 1 day a week of majority (but not exclusively) German with me, his mother. Weekends and evenings were in a dual language (but English-dominant) environment with his monolingual English-speaking father and me, one day a week was spent with his English-speaking grandmother, and 3 days a week were spent at a monolingual English-speaking nursery. The family language policy was fluid -due to his father not speaking German, family conversations were often in English, typically interspersed with short German interchanges between Toby and me.
Toby's request to 'take a break' from speaking German stemmed from his confusion over the fact that I was not only the only parent speaking the heritage language, but also the one helping him to learn to read in English, and taking him through the school phonics system, that is, meeting school success criteria (Kabuto, 2010). Aware of the likely consequences linked to language attrition, but unwilling to make language a potential cause of anxiety (Sevinç, 2020) and adopting the dual identity of teacher/mother (Okita, 2002), I agreed to a language hiatus, which lasted 2 years. During this time, the use of German in the home was reduced to absolute basics, such as the occasional 'Gute Nacht' ['good night']. At age 6, however, Toby decided that his 'break' had been long enough: he felt he had learnt to read in English to his satisfaction, and wanted to re-start speaking German. He suggested that we could 'do research together', and, following an exploration of his understanding of this concept (see Little and Little, 2022 for a full discussion), we agreed to note down key moments in our joint endeavour, and to reflect on our experiences of 'resurrecting' the heritage language. While the overall study took 2.5 years in total, this article focuses on the first 2 months, since they were the most emotionally challenging from both our perspectives. As such, the study adds significantly to the under-researched area of reversing language shift at the familial level, by offering a firsthand reflective, qualitative account of emotions and experiences related to reversing language shift from the perspectives of both mother and child, particularly in terms of emotion and identity construction, and the impact this had on family language policy.
The study addresses the following research questions: What are the experiences, emotions, and attitudes of a mother and child working together to reverse familial language shift?
What are the implications of these emotions on family language policy?
Being an autoethnographic study, the research does not generalize; however, it closely inspects the context of heritage language maintenance from an angle which has been largely unexplored, specifically providing in-depth engagement with a young child's voice, views, and emotional experiences, making a significant contribution to the argument of including children in family language policy development.

Terminology
This study uses the term 'heritage language', despite Blackledge et al. (2008) pointing out that the term 'heritage' implies a certain sense of continuity-of context, culture, or values, for example. In the case of our family, we might speak of a 'disrupted heritage', since, as Bourdieu (2000) argues, heritage and inheritance are not inevitable, like any inheritance, it may be rejected (Little, 2019), either permanently, or, in Toby's case, temporarily. The term 'heritage language' has been adopted here to illustrate a familial link, a connection-or the potential of one, through shared language, which, in our case, is only 'inherited' through one side of the family, since T'.s father is monolingual in English.

Heritage language, identity, and emotion
While Bourdieu (1977) viewed the individual identity as predominantly a product of history, recent thinking about identity has focused on agency, individuality, and fluidity. Choudry (2010), for example, argues that children from multiple backgrounds are more likely to develop a flexible identity, which changes, develops, and adapts according to social context, language, and so on. Similarly, Clark and Gieve (2006) highlight multiple identities occupied by the individual, particularly in relation to language and register. In Little (2020a), I introduce a framework to conceptualise heritage language family identities, exploring attitudes towards the heritage language in terms of emotional and pragmatic orientations on the one hand, and essential/peripheral orientations on the other. This framework allows for fluctuations over time, as certain cultures and languages may be foregrounded at certain points, according to festivities, milestones such as schooling, holidays, and other considerations. The framework functions as a focal point for discussions among heritage language family members who struggle to understand each other's point of view, as it seeks to enable those who make strong emotive links between the heritage language and identity to express any innate need they have for their child to grow up speaking the heritage language. Emotional aspects of linguistics research have only fairly recently begun to include multilingual perspectives (Pavlenko, 2009), with much of the work taking place over the past two decades (see e.g., Dewaele & Pavlenko, 2001). Dewaele (2008Dewaele ( , 2013 continually and strategically explores the relationship between emotion and language in a variety of contexts, and found that expressing love can have different meaning and emotional attachment among multilingual speakers, dependent on both personal experience and cultural connotations (Dewaele, 2008). Kouritzin (2000) speaks very openly about her struggles as a multilingual parent, going so far as referring to her identity as being 'erased' (p. 314). Norton (2013) similarly identifies considerable emotional connections between language and identity, while Kwon's (2017) participants view language as a bridge to older generations. This paper presents an additional view, focusing particularly on parent-child interactions, and links to identity construction. Pavlenko (2009) makes a strong case for paying careful attention not only to how emotions are expressed through language and vocabulary choices, but also how language acquisition itself is an emotional endeavour, linked to notions of belonging and identity (Norton, 2013).
To clarify terminology, this paper uses the term 'emotions' throughout, with specific reference to emotions being socio-cultural in nature (Mesquita & Walker, 2003), and open to interpretation. Importantly, emotions may be conscious or subconscious, and are typically explored through reflections and communication, a core component of this paper's research methodology.
Child agency, emotion, and family language policy King (2016) points out that the most recent wave of research into family language policy and heritage language learning finally begins to acknowledge child agency and identity formation, and asks for research to continue to drive these aspects further, particularly with longitudinal studies. More recently, Smith-Christmas (2020) continues to develop the considerations linking child agency to family language policy, including the exploration of language choice as agentive. She challenges the simplistic notion of a child's minority language use as 'compliance' (Smith-Christmas, 2022) and contextualises family language policy within the field of childhood studies, arguing for a more holistic understanding of the dynamics and fluidity of parent-child interaction in heritage language development. In Wilson's (2021) work, parental views on translanguaging practices are explored, highlighting both varying levels of openness to the mixing of languages, and insightful family observations of family language practices. In Gharibi and Mirhavedi's (2021) study, parents showed strong language ideologies towards the maintenance of the heritage language, employing various strategies to encourage, and in some instances force, the child to use the heritage language. This study builds on previous work, not only by providing detailed data from parent-child interactions, but also by offering joint reflections on these interactions, focusing on the child as an active partner in heritage language development, and adding their voice and emotional experiences as an important component of the research discourse.
Curdt-Christiansen and Huang (2020) explore the role of family language policy in relation to the emotional factors of family life, including the strengthening of familial ties through heritage language development and use. Their reference to the imbuement of specific words and phrases with emotional, personal, and affective meaning is strongly represented in this study, especially in terms of the recognition of these personal connections from the child's perspective. Fishman (1964Fishman ( , 1991 identified the academic field around language maintenance and language shift to be focused on what happens to language use 'when populations differing in language are in contact with each other' (p. 32). While Fishman is concerned with populations, rather than individuals, his work has contributed significantly to research in generational language use and language loss, including at family level (Little, 2020b;Verhaeghe et al., 2022). Families, in the absence of contact with larger communities of speakers, function as a microcosm, treading a balance between languages. Busch (2021) warns against classifying language repertoire as a tool or fact, instead advocating for its recognition as an 'emotionally and bodily lived experience of language' (p. 191), thus making it important to consider emotion not only from the perspective of language repertoire, but also of language shift. Tannenbaum (2012) describes the emotional concerns that may lead to the maintenance of the heritage language as a 'coping or defence mechanism' to strengthen intergenerational relationships and potentially 'protect the integrity of the family system' (p. 62). Indeed, Verhaeghe et al. (2022) identified significant emotions linked to language shift and attrition, especially in parents who aim to maintain the heritage language to provide an emotional connection between children and grandparents.

Language shift and attrition as emotional concerns
While the terminology around a language shift describes the interplay between the two or more languages present, terms such as attrition and forgetting (Ecke, 2004) focus on the language that is being 'lost' or 'forgotten', typically associated with an increased input-level of the dominant language, such as the beginning of school. Brown (2011), in a US context, highlights the great societal and authoritarian pressure on secondgeneration immigrant children to learn English, arguing that 'a feeling of belonging at school will only come with learning English and simultaneously losing their [heritage language]' (31). Such binary thinking is at odds with current figures in England, which show that more than one in five children of primary school age speak more than one language (Department for Education, 2020). Home languages were included in England's national language review for the first time in 2016 (Tinsley & Board, 2016), showing that many schools struggle to acknowledge or support families with different home languages, beyond English language teaching. Kabuto (2010) indicates that the school start adds pressures and tests linked solely to one of the various languages the child might speak, thus shifting the focus in the child's mind (Wilson, 2021). Where the child in Kabuto's study, Emma, declared she didn't 'want to be Japanese anymore' (103), Toby, in this study, chose the beginning of formal schooling to demand a 'break' from German, since he struggled with one parent being both the key parent for heritage language development, and the key parent for meeting school requirements.

Methodological and ethical concerns
Ever since the United Nations Convention on the Rights of the Child (UNCRC, UNICEF, 1989), which focuses on respectful ethical practices in relation to supporting children, have researchers explored the inherent epistemological and methodological implications of the charter. Article 12, specifically, states that 'the views of the child [should be] given due weight in accordance with the age and maturity of the child' (UNICEF, 1989).
Researchers researching their own family context is a practice well established in the field of heritage languages (see e.g., Kabuto, 2010;Kouritzin, 2000;Ronjat, 1913). This study, however, intended to move beyond the child as an 'observee' or 'informant', instead aiming to establish a collaborative, metacognitive 'safe space' in which both child and parent could take a step back and seek to explore and articulate emotions linked to emergent family language policies. Creating knowledge in the context of this study involved understanding the parent-child relationship, thus any epistemological stance could only involve a collaborative, co-constructed process of discovery. Conducting a joint autoethnography with a young child had further impact on the writing process. With autoethnographical writing already embracing a wide variety of writing styles (see e.g., Wyatt and Adams' (2014) edited work), adopting a more narrative, accessible writing style for this paper was an important consideration of the co-produced elements of the research, to ensure Toby could read and member-check the findings. Since this study was originally instigated at Toby's request, and a publication with him as co-author is already in print (Little and Little, 2022), this paper acknowledges his agency in choosing not to use a pseudonym.
The ethics application was completed jointly, and concerns such as withdrawing, changing one's mind, and so on were discussed at the outset, as well as frequently throughout the study (Little and Little, 2022). The only break occurred between receiving reviews for the jointly authored paper, and making these revisions, as Toby struggled with the critique, and asked for a few months off. Mayall (2008) argues that generational issues ought to be a major focus in childhood research, specifically, that research with children is dominated by adults trying to understand children. In her work, she compares her own interviews with children to interviews conducted by the child's mother (briefed by the researcher), stating that the mother's understanding and involvement 'provides an enabling context' (119). Children are both knowledgeable and thinkers (Bruner, 1996), and as such, this study sought to identify a way to create a 'fusion of horizons' (Gadamer, 1989) that facilitated mutual understanding, i.e., an environment where parent and child can meet in an attempt to shed pre-supposed roles: the parent admitting that they might not have all the answers, the child adopting the role of expert in relation to their own context and experiences.

Research methods and the role of the research diary
The main data collection consisted of writing down exchanges between us, shortly after they happened, as verbatim as possible. Due to the long-term and naturalistic nature of the study, this method was preferred to audio-or video-recording, instead adopting an approach linked to the critical incident technique (Flanagan, 1954), where we recognised the importance of naturally occurring events as they happened, and then moved to write them down as quickly and as verbatim as possible. Entries were instigated by either one of us, but typically physically written down by me, due to Toby's age, and checked with those present for accuracy. Most individual exchanges were short, and as such, the process of member-checking allowed for an immediate process of validating research notes and ensuring they were as accurate as memory permitted. While a critical incident-approach is obviously biased towards events which we decided to be note-worthy, thus highlighting a definitive subjective influence (Cypress, 2017), we balanced these diary entries with subsequent reflections (see below), which allowed for an additional form of validation, specifically when it came to exploring the emotions we both felt at the time of the exchange, making the reflections a particularly vital part of this paper.
While some entries were linguistic in nature, most had a specific emotional attachment (see codes below). To explore these emotions, we re-visited each entry approximately 2 days later, reflecting and seeking to understand each other's emotional reactions, and noting down these reflections, sometimes in conversational format, sometimes as reflective 'monologues'. The exchanges we chose to include could thus be described as critical incidents, with both of us acknowledging that the term 'critical' depends on personal interpretation within contexts (Tripp, 1993), making our reflections a vital part of the data collection process, providing both context and rigour through triangulation, allowing for member-checking among ourselves and facilitating an exploration of our personal perceptions of 'truth' (Griffiths, 2008).
The research diary thus constructed consists of 83 pages (25,450 words), covering 2.5 years of verbatim exchanges and follow-on reflections. This article focuses on the first two months of data collection, to highlight initial critical incidents and key moments relating to Toby's language use and identity development, and specific emotional links, including our mutual frustration related to how the re-introduction of German suddenly impinged on our well-established relationship.

Data analysis
The first 12 pages of the diary cover the first 2 months of data collection (6,534 words), a disproportionate amount, in recognition of the disproportionate number of critical incidents which occurred in the early stages of the project. In analysing the data, the coding framework included both a priori codes and inductive codes (Willis, 2010), with a priori codes determining whether entries were of linguistic interest, or had an emotional focus. Inductive codes then further split these categories according to aspects of either a particular linguistic point ('grammar', 'pronunciation', and 'literacy'), or linked to specific emotions or character traits for either one of us ('frustration', 'anxiety', 'pride', 'joy/happiness', 'love', 'sadness', 'perseverance', 'guilt', 'anger'). In this framework, entries could be assigned to multiple codes, and tagged according to which one of us had the relevant emotions. Codes were assigned jointly over a series of 'coding sessions' during a 1-week holiday period, involving copious amounts of hot chocolate. Finally, an over-arching code tagged all entries Toby and I considered to be part of his 'identity'. Of particular importance for the following analysis section is that, through our coding, we discovered that Toby identified any engagement with language play as a positive experience (coding as 'pride' and or 'joy'), while also adding the 'identity' code. This highlights the importance of involving children directly in the meaning-making from data. Due to critical incidents being imbued with meaning by the participants (Flanagan, 1954), analysis was mainly according to content (Sproule, 2010), the study was less interested in observing patterns, and more focused on exploring the emotional context within which the reversal of familial language shift took place.

Limitations
Drawing on the views and opinions of one mother and her son doubtlessly presents numerous concerns and limitations, some of which have been discussed above. Addressing the issue of representativeness, Yin (2016) argues that within qualitative research, a sample is not necessarily deemed to be representative of a larger population, instead, it is intended to maximise the information which can be gained from one context. Working with Toby on the study for 2.5 years has undoubtedly affected his ability to discuss, share his views, and examine his emotions, in a manner which may not necessarily make him comparable to other children, and it is difficult to unpick where and how these views might have been shaped through parental input, however well-intended. As Miles and Huberman (1994) point out, researcher choices influence the research process, whether the researcher is aware of it or not. As an example, the excerpts presented in this paper were chosen partly based on length, showing interaction and negotiation, and partly due to being deemed to be particularly illustrative of the emotions and codes highlighted below. As a result of this, researcher bias is certainly present here, as in any paper where quotes are chosen as part of the writing process. With the study being a part of our lives for 2.5 years, and subsequently shaping our way of being with each other (Little and Little, 2022), we present the findings as part of our lived and examined experience.

Language and frustration
Only 1 week into the keeping of the diary, common terms and phrases such as 'Ich hab' Hunger' [I'm hungry] returned to daily conversations. In the diary, I reflect that it feels like archaeologythe phrases most recently lost are the ones most quickly recovered. For Toby, these early days were particularly frustrating, as he is forced to consciously experience just how much of the language he has lost. Negotiating this frustration can be painstaking work, and two weeks after Toby's decision to restart his engagement with German, we had the first exchange that we both regarded as truly 'critical' (Tripp, 1993). As the whole family was driving in the car, I was trying to ascertain whether Toby would need a snack for a performance he had the following Wednesday. The transcript gives the languages (German and English) as in the exchange, with figurative [ fig.] and literal [lit.] translations where necessary, and actions/additional information provided in {} brackets and *emphasis*.

Me:
Du, T.? [Fig: Hey The conversation continued in this vein for a further 5 minutes, before we, both exasperated, had established whether a snack would be needed. When Toby and I reflected on our exchange 2 days later, to discuss what had happened. Toby explained his frustration of battling on in German, when the situation could have been resolved in a matter of seconds in English, a frustration shared by many bilingual children who feel there is extra work linked to being bilingual (Sevinç, 2016;Little, 2020a), and that being bilingual is not necessarily positive (Ortega, 2018). Together, we realised that, had Toby waited for me to finish the sentence, the addition of the word 'snack' would have helped him discern the meaning, but also, that he could have done more to pay attention to the context (there was only one show, and he knew it was on Wednesday). Rather than me helping him with this, asking 'Wann ist deine Show?' ['When is your show?'], we both got caught up in the minutiae of practising the days of the week. We both agreed that my husband's interjection did little to help diffuse the situation, but Toby also said he felt proud when we managed to conclude the discussion without me lapsing into English. Him being more resilient than I had anticipated is a theme that occurs throughout our exchanges (see below), supporting Mayall's (2008) view that adults' attempts to understand children can only ever be a flawed process, unless children are involved directly in the meaning making.
In other exchanges, it was me who got frustrated, rather than Toby. In an exchange 3 weeks after the one above, we drove back from a plant nursery. Toby was interested in alpine plants, and asked why their leaves are so different, leading to the following exchange: In this conversation, I ultimately gave up and continued the explanation in English, aware that each attempt to clarify led to new difficulties and questions. Nevertheless, in our joint reflection, I expressed my guilt and frustration at not persevering with German, likening it to 'clipping Toby's wings' (Little and Little, 2022). In this reflection, Toby took the lead, expressing reassurance and stating that, as his German would improve, our conversations would get longer and more complex.
Such critical incidents, and our subsequent reflective discussions, led to early changes in family language policy, particularly how new words or barriers to communication were handled. If context was not sufficient to work out new vocabulary, several points of checking might occur, both as translations offered by Toby [Is Kirche [church] church?], and by requests for help [Was ist Ananas?-What is pineapple?] Explanations or translations would be provided spontaneously based on what seemed to be the most straightforward and helpful response. If a response was likely to be long and complicated [such as the response to Toby's question 'Was ist Autismus?'-What is autism], the explanation would typically be provided in English, foregrounding understanding over language development. Sometimes, a show-and-tell or geographical explanation was possible -the pineapple in the example above was identified as being 'neben den Bananen' [next to the bananas].
One frustration that is worthy of particular mention is linked to Toby's reaction to our discussion on 'modelling' (Lanza, 1997) versus 'correction', and his concern to use 'correct' language. When we discussed 'standard' language during one of our joint reflections, about a month into the research, Toby explained that he wanted to learn 'correct' German, which he seemed to loosely categorise as the German he encountered in books, i.e., 'standard' German. I explained that I did not feel comfortable 'correcting' his language use, and suggested the standard practice of 'modelling' instead. Toby reacted angrily, stating: But that's not fair. It means I have to listen and try and work out whether I did something wrong. I have to be a detective all the time, and I can't just talk. That's stupid! Toby made it clear that he perceived modelling as deceitful, and not congruent with a collaborative approach to family language policy. We took Toby's emotions as an opportunity to discuss his ideas on alternative approaches. While it could not be truthfully said that all modelling stopped-sometimes it was part of the natural flow of conversation -standard practice became me asking him if he wanted to know the 'standard' grammar or vocabulary, with him deciding spontaneously whether he was interested or not-one example of this is provided below.

Language play and joy as identity
In English, Toby was a competent language user, enjoying language play, finding rhymes, dual meanings, playing word games, etc. As his confidence in using German grew, he quickly sought to make playing with language part of his linguistic repertoire, and there are several examples where he either takes pleasure in a linguistic discovery, or in playing with words. Four weeks into our diary, the following exchange provides an example, as we sit outside watching birds:
It turned out homonyms had been covered in school the day before, and Toby's realisation that he could spot one in German made him both happy and proud. Creative use of language continued later the same morning, installing a bat box in the garden. As I lifted Toby (who was wearing a bat costume) in the air, I shouted: The word play here derives from a 'Fledermaus' [bat] literally being a 'flying' or 'winged' mouse. Toby has a cuddly toy called Kuschelhund [cuddle-dog]. Both these examples show that he is beginning to develop the confidence necessary to play with and explore the language, and to incorporate language play in German into his existing character. When we reflected on this a few days later, Toby explained: I love playing with language. When you speak two languages, but you can only play games [i.e. make language jokes, be playful with language] in one of them, then that's sad, it's like you can't control it, and you can only say half of what you mean, and other people only understand half of who you are. It's fun to make up words and to make them mean something. But I like making up words in both languages, and then you are the only one who understands -it's like we have a secret language. (Aged 6 years, 9 months) This reflective excerpt re-iterates Toby's need to be competent in the language to feel comfortable that it expresses his identity -while he did not feel his identity itself was compromised, he believed that not being able to express exactly what was in his mind would impact on his interactions with others: his sense of fun, his inventiveness, and his playfulness were not clear to outsiders, hampered by language barriers. Considering Toby's emotions as part of our family language policy, and acknowledging his growing confidence, helped him establish his identity as a multilingual. In fact, the re-emerging bilingualism gave way to a new branch of language playfulness and joy, namely by mixing the two languages, showing his emergent 'creative multilingual voice' (Cenoz & Gorter, 2013, p. 250). This would result in word creations like 'ich starviere' (using the English 'to starve' as a basis, but adding a German verb ending in the first person singular). When Toby first used the term 3 weeks into the diary, I burst out laughing, and although the standard phrase [Ich hab' Hunger -I'm hungry] appears in the diary in the very first week, 'ich starviere' remained as a family-specific neologism. The encouragement of translanguaged family neologisms is in direct contrast to some of the families highlighted in Gharibi and Mirvahedi (2021), and, to a lesser extent, Wilson (2021), where some parents would refuse to respond to the societal language, or strictly encourage heritage language use. For us, Toby's 'creative multilingual voice' was inextricably linked to his sense of joy and playfulness, and extending his opportunity to show wit and a sense of fun through translanguaging was an important enabling factor. As such, our family language policy actively celebrated multilingual family neologisms, acknowledging the growing skill it took to create them, and valuing them in their own right, rather than viewing them as 'mistakes', or a mere transitionary phase on the way to 'standard' language.

Language and love
Toby's desire to focus on 'correct' language is something I return to in the conclusion, but nowhere was our differing focus more apparent than in expressions of love. Five weeks into the project, Toby had taken to saying: For me, the emotional connection changed as Toby repeated the sentence in standard German, even though he instigated the sentiment. For me, it became a question of authenticity and 'voice', telling him how to express his love was encroaching on his identity. When we discussed this incident together, though, Toby expressed a very different view, reiterating his earlier dislike of modelling and sharing a sense of betrayal when I did not 'correct' him, in an echo of Dewaele's (2008) statement that '[t]he one thing that nobody would wish to get wrong is a declaration of love' (1753). For him, it was important to know that he could trust me to let him know about any language-related 'mistakes', regardless of content. Although for both of us, emotive content of conversations was particularly important, Toby did not share my worry of authenticity -to him, what mattered was to be able to count on me to steer him through any 'mistakes'.
Three weeks after the exchange described above, at the end of the two-month period covered in this paper, expressions of love recurred in the research diary, as Toby became determined to say Ich hab' dich mehr lieb als alles auf der Welt! [I love you more than anything in the world.] Following on from the previously discussed entry, he still struggled with Welt/world, and the very similarsounding 'als/alles' [than/anything]. The use of the dative (die Welt -auf der Welt) also remained confusing. Ich hab' dich lieb [I love you] was a staple phrase, and initially, his version was Ich hab' dich mehr lieb as alles in die [sic: auf der] world', with 'in die world' being a direct copy from his previous effort above. My attitude towards his efforts on expressing emotion had not changed, as I noted in the diary: He asks me if he gets it 'right', but in saying that it isn't, I feel like I'm criticising the sentiment as well as the language.
Instead of 'correcting' or 'modelling', I looked for a more playful way to help him. So when he said Ich hab' dich mehr lieb als alles auf der world, and he asked me to tell him whether this is 'correct', I made a joke out of it and tackled him onto the bed, tickling him and shouting Welt. . .... Welt!! Welt!!! in mock exasperation. He laughed and repeated Welt several times, before saying the full sentence in standard German. Reflecting on this moment a few days later, Toby explained that knowing he got it 'right' was an important moment of self-efficacy and control for him.
The playfulness of the incident further helped Toby get over his occupation with 'correct' German as 'standard' German, which ultimately helped us to adopt a more carefree and flexible approach to our conversations, less encumbered by 'corrections', and, ultimately, more communicative and spontaneous.

Concluding remarks
As Okita (2002) points out, being a heritage language parent can be emotionally draining work, and it can also be lonely, particularly when the heritage language maintenance falls on the shoulders of just one parent. This paper argues that children, instead of being solely at the receiving end of a parent's heritage language efforts, have a substantial role as collaborators, partners, and instigators, siding with King's (2016) call for more studies which explore child agency and identity formation, and Smith-Christmas' (2022) call for a closer link between childhood studies and family language policy. It is important to recognise that the responsibility and agency related to heritage language maintenance do not just lie with the parent -instead, children have both the right and the ability to share and reflect on their attitudes, and to guide family language policy. However, children's views and reflections remain frequently excluded, at least until children are much older (see e.g., Okita, 2002).
Both negative and positive emotions can be utilised to guide family language policy development. In Toby's case, his frustrations linked to modelling (Lanza, 1997) and his preoccupation with 'correct' language were explored through a jointly reflective, communicative approach can assist with a 'fusion of horizons' (Gadamer, 1989), which we have come to call an 'un/familiar space' (Little and Little, 2022). Through this collaborative exploration, negative emotions were minimised as part of the joint development of family language practices, which enabled the diffusion of frustrating or angering experiences. On the other hand, Toby's sense of joy when playing with language, creating family neologisms which borrowed from both his languages, and his excitement when he established links between different words, were an important part of Toby's sense of identity, and required space within the family language policy. Through joint reflection, the active creation of space for joyful, playful, and creative language use allowed us to reframe the idea of 'mistakes' into the more positive 'creative multilingual voice' (Cenoz & Gorter, 2013), which helped to establish Toby's sense of identity, since being able to 'play' in both languages enabled him to express his whole self, rather than just 'half', as he explained. While, arguably, the difference between a 'mistake' and a 'play on words' is intent, our reflections helped us both to query notions of standardised language, and view language development as fluid and evolving. In acknowledging this, children can share and lead on how their multilingual development is linked to their sense of self. Examining the role of emotions as part of family language policy can be an empowering opportunity for children to take the lead, and offer many opportunities for future research which foregrounds individual family voices and emotional experiences of all involved.
Our experiences make an important contribution to the research into heritage language learning, both at content and at methodological level. By providing an analysis of the first 2 months of our efforts to resurrect the heritage language, we offer insights into the emotional toll such efforts take, and, crucially, we do so from both the parent's and the child's perspective, highlighting how collaboratively developed strategies increase agency and investment. The analysis also provides insights into the importance of linking identity with heritage language (Pavlenko, 2009) use even during low competence levels, to provide an emotional, motivational link to the language, and a sense of self-efficacy. Tying children's language use to their developing sense of identity -acknowledging their emotions, preferences, etc. may ultimately provide stronger emotional links to the heritage language than focusing on the heritage language as a means to strengthen familial ties (Tannenbaum, 2012), especially in families where the heritage language is not a vital component for family communication. In such families, encouraging the child to lead family language policy and work with their developing identity to collaboratively explore the heritage language may encourage the child to build their own, independently strong, emotional ties to the heritage language, rather than viewing it simply as part of their 'inheritance' (Little, 2019). As an autoethnographic study, this paper claims neither representativeness nor generalisability (see methodology). Nevertheless, the findings highlight that a 'disrupted heritage' in heritage language development is not necessarily a sign of unavoidable, permanent language shift, and that working with children to develop collaborative methods to 'resurrect' the heritage language, working against 'forgetting' (Ecke, 2004).

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

Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.