Migrant pupils as motivated agents: a qualitative longitudinal multiple-case study of Russian-speaking pupils’ future ideal selves in English primary schools

ABSTRACT The paper discusses the findings regarding future possible selves based on research with Russian-speaking migrant pupils in English state-funded primary schools at Key Stage 2 (7–11 years old). Its aim is to explain the nature and characteristics of ideal selves of primary-level migrant children. The methodology comprises a qualitative longitudinal multiple-case study with five embedded cases. Evidence derives from 63 interviews and seven months of participant observations. Drawing on seminal work in language learning motivation research, the study uses an integrated drawing-based elicitation technique. The findings challenge earlier claims that middle-childhood level children do not have a clear future possible self, showing how pupils express diverse, at times pragmatic and practical but remarkably profound, ideal selves in their L2 (second language) schools. The study expands on the concept of the ideal self and calls into question any definitive conclusions in the field of primary level children’s ideal selves and learning motivation.


Introduction
It has long been maintained in learning motivation research that primary-level children do not have clear future ideal selves (Dörnyei 2009).Perhaps this is not surprising given the sparsity of research into primary school children's motivation (Boo, Dörnyei, and Ryan 2015) with few studies exploring primary pupils' learning motivations and LL (language learning) motivations in particular (e.g.Graham et al. 2016;Robins 2012).Existing research, although somewhat peripheral to the current study, states that a clear ideal self (be it ideal L2/bi/multilingual self or ideal self) is absent among primary-level children (Zentner and Renaud 2007).It is described as having a 'fantasy' as opposed to clear goals and aspirations (Lamb 2012(Lamb : 1015)), and possible selves, it is claimed, are not linked with motivations among younger children (Macintyre, Mackinnon, and Clément 2009b: 197).
Having reached a consensus on the matter, researchers to date have tended to avoid further speculations, with one exception.Chambers (2019: 9) explored possible selves among primary age MFL (modern foreign language) learners, confirming that 10-to 12-years-old pupils' ideal selves were attuned to 'the here and now'.Applying the L2 Motivational Self System, Chambers (2019: 229) found that pupils' ideal selves were neither applicable to real life nor clear.Given the scarcity and brevity of pupils' responses, he concluded that the findings did not satisfy the conditions of ideal selves (Dörnyei 2015), comprising having an elaborated, vivid and difficult to attain but plausible future image.Chambers (2019) suggested that more studies are required using more targeted longitudinal and ethnographic methods of investigation.This may also be strengthened by the use of creative techniques in interviews with children.
In relation to future possible selves, there is in the literature some sense of primary-level children's learning goals (e.g.Rodkin et al. 2013;Wilson et al. 2016 in relation to achievement goals), which are shaped by children's selves (e.g.Dweck 1991), but they are not considered as part of ideal or multilingual selves of migrant language learners at primary level.Research into primary-level children's learning motivation does not reflect the deep ongoing experiences of migrants whereby their future and present selves are continuously re-shaped and negotiated.This is not the case in adult migration research, where often traumatic experiences of migration are found to impact substantially on desired and feared possible selves (Stevenson 2019).In his state-of-the-art article on motivational research in language education, Lamb (2017: 334) stressed a 'priority for future L2 motivation research' was to focus on young English second language (L2) learners.Thus, the characteristics of the ideal selves of primary-age migrant children are not yet understood nor have they been dealt with in depth, yet the significance of this neglected area of investigation is acknowledged.By focusing on different contexts and languages, such an exploration would highlight the 'cross-cultural' aspects impacting on their possible selves (Huang, Hsu, and Chen 2015: 37).This might then further the analysis of individuals' selves among language learners and selves in various cultures (Macintyre, Mackinnon, and Clément 2009a) emphasising the fundamental interaction between contextual factors and learners' emergent visions of possible selves (Serafini 2020).This is pertinent given the start of a global and broader multilingual movement in LL motivation research (Ushioda 2017).
Therefore, the aim of this study is to explore the nature and characteristics of the ideal selves of primary-level migrant children, thereby contributing to the studies researching possible self-guides in relation to primary-level migrant pupils.The paper draws on research with Russian-speaking migrants in English primary schools at Key Stage Two (7-11 years old, middle childhood).The methodology comprises an interpretive paradigm employing a qualitative, longitudinal multiple-case study approach.The evidence is based on 63 interviews with children using creative techniques and seven months of participant observations of in-school experiences.By using the creative technique 'My hopes and dreams', the reported hopes and dreams of pupils in their L2 schooltraced over timeembodied diverse ideal selves, which at times seemed trivial and practical but were also surprisingly profound.The ideal selves are presented through the PAP (protective, agentive and prospective) model of migrant children's ideal selves in L2 schools.With an inherent complexity and clarity, these ideal selves reflected the real-life learning issues of migrant pupils, thus appearing as much more than 'fantasies'.
The next section explains the key terminology used in this paper and briefly characterises the participants' age group.This is followed by an elaboration of the theoretical orientation.I then move to the materials and methods, the findings section and a discussion of the findings in relation to the literature, proposing a model of migrant children's future ideal self.I conclude by reiterating the theoretical and practical implications of the study.

Ideal self
Ideal selves, or desired and hoped-for future possible self-guides, are socially constructed, cognitive images, unique to every person, that originate from their desires, values, dreams, goals, and meanings (Dörnyei 2005).They also act as determiners of actual selves (Zentner and Renaud 2007: 557) and are described as a 'critical catalyst' in teacher education research (Gaines et al. 2018: 210;Lee and Schallert 2016).Future possible selves are thus linked with the present through the interrelation of both real-life and envisioned self-identities (Dörnyei 2009).In migration contexts they are found to be dynamic and continuously emergent (Stevenson 2019).
They unite motivation and self-cognition and are based on the possible-selves concept delineated by Markus andNurius (1986) andHiggins' (1987) self-discrepancy theory.The former includes the expected self, the hoped-for self, and the feared self, each of which has a different influence on motivation.Conceptualising the learning self into actual, ideal, and ought-to domains and into two standpoints (own/other), Higgins (1987: 319) suggested that the opposition of the ideal and the feared selves serves as an influential tool for motivation as it creates a certain level of displeasure or irritation.The person, then, mostly subconsciously, aims to diminish displeasure by making the ideal and ought-to selves consistent with each other.
Being an essential part of motivation (Higgins 1987), an ideal possible self is a component of the leading theory of LL motivationthe L2 Motivational Self-System (Dörnyei 2009).In the conceptualisation of LL motivation, the corresponding orientation of the ideal self is an integrative motivation (with intrinsic motivation corresponding to the L2 learning experience and extrinsic motivation to the ought-to L2 self) (Dörnyei 2009: 30).Such an explanation denotes that in the context of immersion in L2 schools (in anglophone countries) the ideal L2 self (integrative motivation) is seen as an obligation and a necessity as opposed to reflecting a 'desire to learn target language' (Dörnyei and Ushioda 2009: 22), i.e. migrant pupils' ideal L2 selves would naturally be directed towards understanding and becoming familiar with the people who speak L2 in and outside of their L2 schools.Previous research has confirmed this explanation, averring that the role of integrativeness for L2 motivation in the case of migration is substantial (Taguchi, Magid, and Papi 2009: 67).However, in relation to migrant pupils in L2 schools, this explanation seems to limit the nature of the ideal self, making it prescriptive, self-explanatory, and narrowed to L2 and more generally to LL motivation, thereby overlooking other affordances of ideal selves in the field of migration.

Participants' age group
Middle childhood, i.e. 7-11 years old, is particularly interesting for the present study since, as McAdams (2015a) and Coll and Szalacha (2004) have said, it is during middle childhood that children start to form aspirations about the future.Overall, it is a period of fundamental dynamic development (Smith and Cowie 1991) and 'profound transformation' (Rappley and Kallman 2009: 50) when children experience 'a global shift in cognition, motivation, and social behaviour', marked by a 'heightened sensitivity to the environment' (Del Giudice 2014: 193-199).I recognise that this terminology is somewhat problematic.Complex and distinctive as this formative stage inherently is, it nevertheless potentially includes children at significantly varying stages of development, i.e. early (7-8 years old) and later years (9-11 years old), which can be challenging in a research context (Collins 2005).Furthermore, even within these ages, children's behaviour is not precisely phased within motivation or language comprehension categories (Collins 1984).Although it is essential to be mindful of generalisations across age ranges within middle childhood, it is considered to be a discrete period of development (Rappley and Kallman 2009).

Theoretical orientation
Situated within a linguistic multi-competence framework (Cook 2016), the studies in learning motivation and LL motivation have expanded into research on migration immersion contexts (e.g.Harvey 2017;Nakamura 2019;Thompson 2017aThompson , 2017b)), marking a global multilingual turn in LL motivation (Ushioda 2017).As part of this framework, Henry (2017: 555) propounds a multilingual ideal self, defined as a perpetual interrelation between an individual's languages, LL and the ideal self.The ideal multilingual self thus corroborates a broader view on the ideal L2 self.
In this vein, and in an attempt to view the ideal self holistically, this paper follows McAdams' (2015a) understanding of motivationthe motivated agent line.The motivated agent line focuses on the present and future, asking 'What do I want?What do I value?' (McAdams 2015b: 260).The development of a person as a motivated agent in terms of life values, goals, plans, and dreams forms a 'motivational agenda', which conveys the steps that an individual takes to acquire the attributes of their future self (McAdams 2015a: 256).The development of the motivational agenda is idiosyncratic; it may remain stable or undergo unique shifts.Clear goals and motives start to develop during the preteen period (middle childhood) and shape pupils' everyday behaviours (314).This perspective allows one to see migrant primary-level children as motivated agents comprehensively rather than exploring their learning motivations (e.g.achievement) or their LL motivations in isolation, i.e. the ideal self is seen as being inclusive of ideal L2 self.
Within the motivated agent line, McAdams proposes to use a cognitive perspective -Ryan and Deci's (2000,2006) self-determination theoryas opposed to the dynamic, process-oriented perspectives.However, aiming to move towards a more all-encompassing process-oriented view on pupils' motivation accounting for the constant dynamics of the context and the self 'as a dynamic actor, playing a variety of roles' (Harter 2003: 611), in this study the motivated agent line is supplemented by the ecological model of agency (Biesta, Priestley, and Robinson 2015;Biesta and Tedder 2007).It complements McAdams (2015a) motivated agent line, comprising three interrelated elements: the iterational, projective and practical-evaluative agency, accounting for a possible future, informed by the past and present (Figure 1).
Agency is understood here as a fluid and unfixed participation in and interconnectedness with various contexts in the new L2 learning environment (Emirbayer and Mische 1998).Iterative agency relates to evaluations of life history and past background by an individual, framed as memories of the past (Biesta, Priestley, and Robinson 2015;Emirbayer and Mische 1998).Projective agency focuses on the future possible directions of individuals' activity, informed by children's motivational agenda (hopes, wishes, fears and values).Practical-evaluative agency focuses on the present and migrant children's attempts to examine and internalise possible future visions in relation to their L2-immersion learning experience.The ideal self closely interlinks with these dimensions, forming pupils' motivational agenda which is also shaped by the present experiences (and the actual self).Although the framework situates the future possible ideal self at the centre, it, nevertheless, does not negate other dimensions of the self (e.g.feared/ought-to self) existing in learning motivation.

Materials and methods
The findings are drawn from a larger study (Gundarina 2019).The methodology is based on an interpretive paradigm using a qualitative longitudinal multiple-case study (Yin 2014) with five embedded cases, each representing one pupil.The qualitative research umbrella is chosen to allow an in-depth investigation while simultaneously bringing to the foreground pupils' voices, their own ideas about the world, their actions, and their motives (Becker 1996); this is then strengthened by a multiple design providing different perspectives on the research questions.The cases were selected following Punch ( 2005) and use five criteria: they use Russian as their dominant language (L1); they are firstgeneration immigrants; they attend state-funded primary schools at Key Stage 2 (7-11 years old); at least one parent/carer is a native Russian-speaker; they (and their school and parents) agree to participate in the research; they migrated no longer than six years prior to the study (Gundarina 2020).The time-frame was chosen following Suarez-Orozco, Carhill, and Chuang (2011) and Este and Ngo (2011), that immigrant children had no vivid identity shift in the five years after immigration and their L2 level took seven years to reach high proficiency.This time-frame would also allow for more flexibility among potential participants, as, based on a pilot study, finding participants was challenging.Participants were recruited using data from the Department for Education, comprising the names and addresses of 516 schools in London.Having contacted 216 schools, two expressed their interest in taking part in the study with four pupils fitting the criteria.One more pupil (School B) was recruited through posting an ethically approved advertisement to a social network for Russian-speaking migrants.All pupils, their school and parents who agreed to take part and fit the criteria were recruited to take part in the study.
The empirical part of the study was conductedfrom October 2016 to May 2017in three statefunded primary schools in South, West and Central London.The evidence is based on 63 semi-structured interviews with pupils using creative techniques and seven months of participant observations.I observed each pupil for approximately one working month, for an average of 24.8 school days.The interviews were conducted twice a month with each child (average duration: 27 minutes), except for the first and the last months of the study when I conducted an introductory/ finalising interview.Sixteen interviews with the children's parents and teachers were also conducted at the start and end of the study, lasting for an average of 54 minutes.The results section primarily focuses on the data from the interviews with children, using observations and interviews with adults as supplementary data to contextualise the children's reports.During the study I aimed to take a non-hierarchical role as a friend and build good relationships with children, who were given freedom to end the interviews at any time and not respond to any questions (by saying 'skip') or take part in activities.
The research followed strict ethical standards and requirements (BERA 2018); ethical approval was granted by the affiliating institution.Informed consent was obtained from children, parents, and class teachers.All proper nouns related to the data collection used thus far and henceforth are pseudonyms (children's names) or otherwise anonymised (all other participants, school names).The interviews were conducted in children's L1 or using all their language repertoire, depending on a child's preference.The data were translated by the author.For a detailed methodological discussion of L1 use and ethical positioning in research with children in this project, see Gundarina (2020).

Drawing-based elicitation technique
In order to investigate the ideal possible selves of the pupils, I created a drawing elicitation technique, 'My hopes and dreams', in an attempt to increase children's meaningful participation and accessibility during the research project, seeing them as 'experts' in their lives (Christensen and James 2008).
The technique is informed by the motivated agent line framework, attending to the broad questions 'what do I want?' and 'what do I value?' (McAdams 2015b: 260).The essence of the sociological use and analysis of drawings is that the researchers ask children to comment on their drawings while accepting that the meaning is purposefully generated by the pupils (Davies 2015), enabling them to participate in the knowledge-making process (Fargas-Malet et al. 2010).The choice of topic for the drawings was guided by the aim of the investigation, participant characteristics (their age), and the research setting (primary schools).Children could choose what and how to draw (e.g.objects or activities, using pens/pencils).This was followed by a conversation based around the technique.The children were asked: What did you draw?What does it mean?Do you dream about anything?Do you dream anything about English or linked with the school?It should be noted, that using L1 for the interviews eliminated a potential confusion between a 'dream' as an aspiration and daydreams or night-time dreams in English as Russian delineates between future ('мечты') and nighttime dreams ('сны').

Data analysis
While being guided by the purpose of the paper (Cohen, Manion, and Morrison 2017), i.e. focussing on ideal selves of migrant pupils, the data emerged experientially during the course of the research and were unique to each pupil.The analysis comprised an NVivo software assisted iterative and multiphase thematic approach (Bazeley 2009;Cohen, Manion, and Morrison 2017;King and Horrocks 2010).Additionally, I applied the logic models analytic technique, which involved 'matching empirically observed events to theoretically predicted events' (Yin 2014: 155), i.e. based on the existing theoretical categories, such as, ideal multilingual self (Henry 2017; Ushioda 2017), thereby implementing a 'cyclical movement' between induction and deduction in the process of data analysis (Hennink, Hutter, and Bailey 2011;Patton 2011Patton : 1421)).
During the coding, or aggregation of categories (Creswell and Poth 2016), stage I read the transcripts for the first time, wrote initial codes/nodes in NVivo 11, coding the relevant parts using one to three words (e.g.'agency and voice' 'dreams and wishes' including 'wish to be good at mathematics').I highlighted an abstract transcript and a relevant part of a child's drawing and assigned a relevant code (meaningful summary) to it.The types of codes which evolved were expected codes (i.e. using logic models techniquefor example, ideal L2 selffuture dreams related to L2) and frequently occurring codes.I also included the codes that were not necessarily frequent but important for other elements (e.g.wishes for free choice and having fun) as well as 'rare and influential', i.e. unexpected codes (LeCompte & Schensul, 2010, p. 203), such as wishes to be 'the most important', including consideration of silences/no answers (Creswell, 2013).
Having completed the within-case analysis, the findings were merged into a cross-case synthesis (Stake 2013).As this is a part of a larger study, the codes relating to children's dreams and wishes were included in the 'motivated agent line' overarching theme within children's 'ideal self' node/ code.Following children's reports, there emerged 5 themes within the ideal self-node: future dreams of speaking English; dreams of agency, voice and self-worth; dreams for future occupations and exams; dreams for bi/multilingualism; and recreational events.I further outline the cases and present the results in the form of an evidence-based analytical narrative.

Overview of cases
Table 1 shows the information about the cases at the beginning of the study.All participants had minimal or no previous exposure to L2.As was clear from observations and interviews, Alisa, Yulia and Katerina had well-being and/or attainment and progress issues in learning in their L2 schools.Based on the observations and as supported by teachers' reports, Alisa and Yulia were overly shy and quiet in class, with Yulia also making slow progress.Katerina had academic achievement and behaviour issues (which are detailed elsewhere -Gundarina and Simpson 2021).In contrast, Ivan was the most successful pupil in terms of academic progress and had no reported well-being issues at the time of the study.This was similar to Rita, who had high achievement and progress, which were supported by their parents and teachers and in-class observations.

Findings
The following sections present the five main themes which emerged in the analysis: future dreams of speaking English; dreams of agency, voice and self-worth; dreams for future occupations and exams; dreams of bi/multilingualism; and recreational events.

Future dreams of speaking English
One salient dream relating to L2 proficiency occurred in Yulia's and Katerina's cases.Associated with 'being a good pupil', Yulia reported she wished to speak and learn English because, she said, 'this way I will be able to understand teachers quicker'.In another interview, Yulia's dream was 'to speak the English language!' saying that she imagined that she spoke English, 'in my dreams only!'.Later in the year, describing her drawing of her dreams, Yulia said that she dreamt of speaking English 'so that I am understood'.She felt she was not understood, 'because I am new, and they can't understand me or what I want to express' (quietly and sadder).Having spent only 2 months in the UK prior to the start of the data collection, Yulia's dream was a necessity linked with something important to her (expressing herself), that she felt she was unable to do ('speak') and did not have (being 'understood').In another interview, in her drawing Yulia included a Soviet-style assessment used in Belarus, '5+', which meant 'A+' or excellent, and the words 'Английский' -'English' (Figure 2).She further explained it meant, 'to become a good pupil (sadly).Well, I want to speak English'.She repeated this on another occasion, expressing a dream to 'become an excellent student'.In her comments Yulia seemed to equate 'being good' and 'speaking English' which reveal the way L2 proficiency conditioned a sense of self-worth because of her L2 level.
Similarly, Katerina's main dream when learning English was to have 'good langwig', as she wrote in her drawing (Figure 3), so she would be able to 'speak two languages' when she grows up.In the same drawing about learning in other subjects, Katerina emphasised her proficiency in mathematics, writing a number sentence ('100 + 1000 = 100000') which, as she explained, was something she would like to be able to solve.These illustrate Katerina's realistic (L2 competency) and distinct images of the future which she clearly communicated.Given that Katerina had low achievement and made slow progress at the time of the study, her dreams about learning were closely aligned with her actual attainment issues in learning in L2 and in mathematics.In fact, for both Katerina and Yulia, a dream for L2 proficiency reflected their real-life issues.Consistent with L2 motivational self-system theory (Dörnyei 2009) and self-discrepancy theory (Higgins 1987), these examples also foreground that the ideal selves are often directed towards something that pupils generally feel they lack in the present.
Katerina repeated this scenario over three successive months: for example, describing her drawing she said, 'Чтобы я прыгала на столах!Ха-ха-ха-ха!(Злобно)' -'So that I could jump on the desks!Haha-ha-ha! (angrily)'.Seemingly embodying her wish for voice and agency or freedom for expression (standing on tables) and her attitude towards her teacher's teaching style ('very strict'), Katerina reported dreams that reflected her in-class issues.This was supported by the later observations.At about this time in February and in January Katerina said that she had no hopes or dreams linked to English.Thus, Katerina seemed to be occupied with her needs for voice and agency in class which were more important than L2 proficiency; they were also of a more immediate nature (related to the present).Somewhat resembling Katerina's dreams about voice and agency, the first time I asked Alisa about her dreams, she expressed a wish to be the 'most important in the school': Чего бы тебе хотелось в школе?Alisa: Быть самой главной в школе.

Researcher:
What would you like to have in school?Alisa: To be the most important in school.Researcher: Eh? Alisa: The most important in school.Researcher: The most important?OK.

Researcher:
Самая главная?Alisa: Угу. Alisa: I sit in an office, there … I look at what I want … in it … and this is dreaming, resting … Researcher: Do you sit in the school office?Alisa: Yep.

Researcher:
And are these the sweets on your desk?Alisa: Yep. Researcher: You are the most important?Alisa: Yep.
As in Katerina's case, through her reports Alisa communicated her deep feelings.She dreamt of attention and felt unnoticed and possibly unimportant in her school.This was not the only dream Alisa had, as she also added, 'and I know everything in the world'.Being a recent L2 speaker in her school might have increased her language and social integration awareness (Evans and Liu 2018), revealing that potentially she felt incompetent and not worthy in her L2 school.

Dreams for future occupations and exams
Rita and Ivan both reported dreams of a future related to occupations or more distant dreams for exams, in contrast with other cases.When I asked Rita about her dreams for school, she said that the main dream was to pass the SATs (standard attainment tests) 'well' in November, January, February, March, and May and her GCSEs (General Certificate of Secondary Education) in the future.She explained that it was because of her expectation for the SATs: 'I will get a "C", maybe even a "B"'.She said she was motivated for the SATs because of their future importance in finding a job and a future 'good education', as she also explained in January, February and March.Repeatedly focusing on tests results shows how Rita's ideal self for academic tests (academically-focused ideal self, i.e. 'passing exams') was more stable and distant (since the tests were not to be taken in the very nearest future).In Dörnyei and Ushioda's (2009: 19) terms, this shows an awareness of anticipated plausibility, which comprises the student's personal understanding of 'what is possible for the self'.Ivan consistently named and described three main things: becoming a designer (or an architect), an artist, or a writer.Ivan's dreams and future plans were particularly clear whenever he tried to negotiate his learning with his mother: 'I don't need this! Mathematics!I will be an artist!' (Interview with his mother).It was also clear from the observations that Ivan's dreams of becoming a designer did not change; however, he did have 'momentary alternatives' to try.His mother pointed out that Ivan's dreams were still 'childish'; however, it was already clear 'what lies closest to his soul'.Figure 4 shows that Ivan reported a dream/hope of passing the 11+ high-stakes examination at the end of primary school, in addition to becoming an artist (represented by an easel), a computer programmer (a laptop), and an architect (a bridge).Writing in L2 and the dream of becoming a writer presented both a distinct dream and a goal; Ivan said, 'I'm working now with a helper to write a book'.I asked Ivan about this book: Excerpt: 4 Researcher: Это у вас задание учительницы такое было?Ivan: Нет. Мы сами.

Researcher:
Is this a task your teacher gave you?Ivan: No. We started it by ourselves.Physically creating books, which he compiled with a stapler, embodies a part of Ivan's dream of becoming a writer, and demonstrates that Ivan's wishes are not mere fantasies.Chambers (2019: 229) found that pupils did not express a 'concrete plan to get where they want to be'.Ivan's case showed that starting to write a book with a 'helper' exemplified his plan of becoming a writer.Overall, Ivan and Rita expressed clear dreams/wishes that were more distant/long-term than linked to a nearer future.

Dreams for bi/multilingualism
Apart from tests, I asked Rita about her dreams regarding languages in the future and she said she would like to live in Russia 'Because there I can teach children English'.Rita repeated that her dream regarding English was 'to teach somebody English' and 'maybe work in England' 'as a translator' in another interview.Further, Rita said that her dream was to 'speak English well', explaining that she wanted to be an interpreter.Rita's reports are versatile in that she considered different options in her L2 use in the future (teaching L2 in Russia or work as a translator/interpreter in England).However, these dreams only hint at her wishes to be bilingual.In an interview in March, Rita explicitly said that she would like to speak two languages in the future, 'as a Russian person who speaks English better than Russian'.Through the medium of her multilingual ideal possible self (Kim 2020), Rita's dreams here are pertinent to being bilingual exemplifying linguistic identity, i.e. being 'a Russian person' is something Rita wishes to preserve.Showing some sensible uncertainty about her future ('maybe work in England'), these dreams are realistic and applicable to real life.
In contrast with Yulia and Katerina, who, as we remember, earlier expressed wishes for L2 proficiency, Rita's dreams are directed towards her distant future (being a teacher of L2) rather than the more immediate present (speak L2 to be understood in school).This 'temporality' seems to be a salient feature of the future images: the more distant versus more immediate goals and dreams.Seeing dreams as goals is known in the literature as a temporal dimension, for instance, following Ushioda's (2014) L2 learning timeline (temporal perspective) children in this study reported both long-term and short-term goals.
Alisa expressed a full-scale take on languages: 'to speak every language in the world'.Although somewhat unspecific and one may only hypothesise if it is perceived as plausible by Alisa, this dream is an example of wishes for being multilingual or a 'diverse possible self' (Kim 2020: 12).Reflecting Henry (2017: 555), Alisa, Rita and Katerina included an ideal self, which was directed towards learning two or more languages and could be subsequently seen as an ideal bi/multilingual self.

Recreational events
The children expressed dreams about recreational, pleasurable events (Katerina, Alisa, Rita, Yulia).For example, on 14 March, Katerina reported that she did not have any dreams linked to English but she wished she was going home, where her mother would have cooked something in the garden.Yulia mentioned dreaming about 'having a house in England' and 'having more friends', wishing she could eat sweet things and bring a pencil case to school.Similarly, on one occasion, when I asked Rita about her dreams for school, she said she wanted to go to a restaurant where her parents were planning to take her the following Saturday.Alisa likewise mentioned having 'very many toys' in school.Other dreams were also related to recreationspecifically, 'free choice', food, and a computerso that, as Alisa commented, 'I could do whatever I want!'.

Discussion
The aim of this paper was to explore the nature and characteristics of the ideal possible selves of primary-level migrant children.A more integrated approach allowed a view of the ideal selves of migrant children as part of their wider motivational agenda including their both broader and language-directed ideal L2/multilingual selves.Supporting the formerly established continuously evolving and emergent nature of possible selves in adult migration contexts (Stevenson 2019), the reported ideal selves in this study are diverse and complex, including seemingly superficial dreams (pleasurable events) as well as clear goals for the future.These challenge previous claims in relation to clarity and vividness (Chambers 2019) and stability (Zentner and Renaud 2007) suggesting that, in migration context, primary-level pupils may have clear ideal selves thereby satisfying relevant conditions for an ideal self (Dörnyei 2015).While being varied and clearly expressed by pupils, their ideal selves were often embodied in imagined pleasurable future events that they would take part in or would have specific positive attributes.Lamb (2012Lamb ( : 1015) ) noted that 'younger adolescents' ideal selves are less realistic than older adolescents'.Although some ideal selves may be interpreted as 'fantasy' (Lamb 2012(Lamb : 1015)), the majority of ideal self-guides expressed by migrant pupils were self-concordantly selected and linked with real experiences.The nature and characteristics of the ideal self-reports are presented through the proposed PAP (protective, agentive and prospective) model of migrant children's ideal/future possible self in L2 schools (Figure 5).It is informed by the framework of the motivated agent line seeing pupils as motivated agents (McAdams 2015a(McAdams , 2015c(McAdams , 2015d) ) accompanied by the ecological model of agency (Biesta, Priestley, and Robinson 2015;Priestley, Biesta, and Robinson 2015), exhibiting inherent dynamism within and between the elements of the model and the immersion environment of L2 schools.
The prospective ideal self comprises academically-directed, temporally-bound (distant and stable, and more immediate and unstable) 'prospective reflections' (Urzúa andVásquez 2008: 1935), directed at the future.In relation to English (L2), ideal L2 selves (future hopes and dreams about L2 proficiency or related to L2) were expressed by the migrant pupils with L2 proficiency stages perceived as low and with issues in L2 attainment (Katerina and Yulia), corresponding to the need for competence in L2.However, their ideal L2 selves were of a more immediate nature rather than involving long-term future goals, and they were unstable.Ivan expressed a clear and stable ideal self in an L2 country, repeatedly expressing his vision for his future professions, one of which was a writer in L2.However, Ivan did not seem to report any short-term or near-future-related ideal self in learning L2 (or generally).Thus, a distinction is made between the immediate and unstable, the clearer and more distant academically focused (including ideal L2) future selves of Russian-speaking migrant pupils in L2 schools, informing the prospective dimension of migrant children's ideal selves.
The agentive dimension of the migrant children's ideal selves shapes children's wishes for selfworth and voice (freedom of expression and action), focusing on the past, present and future.An intriguing finding that those pupils who had immediate (short-term and unstable) situationspecific visions of future success related to learning (e.g. a wish for 'good langwig' related to L2 proficiencythe ideal L2 self) also expressed a need for self-worth, i.e. 'sense of worth or personal value' (Covington 1984: 4), fairness and voice (freedom of expression; the ideal self) (Katerina, Alisa, Yulia).However, those pupils in this study who persistently reported clearer visions of the more distant future related to learning (Ivan, Rita) about their lives and learning (including visions related to academic achievement or language, i.e. their ideal L2 selves), did not seem to express any ideal self in terms of self-worth, fairness or voice.Alisa expressed an ideal self in L2 schools related to having a voice and a need for self-worth, a dream of being valued (important) and competent (knowing everything) in school in general.Yulia also repeatedly expressed an ideal self related to self-worth ('being a good pupil and having good L2', i.e. she felt a 'bad' pupil because of her low L2 proficiency and achievement rather than, for example, feeling like an emergent bilingual/multicompetent pupil).Katerina imagined securing a voice in school through misbehaviour in class with her teacher directed at having fair treatment from her teacher (see Gundarina and Simpson 2021).The data suggest this need for fairness, reported as an ideal self, is in fact a precondition for voice (freedom of expression), which appears as a need in learning situations that lack perceived fairness.This echoes the concept of being himself/herself in Ushioda's (2011) study andTaylor's (2013) findings regarding the freedom of being 'themselves' leading to higher educational achievement.
Directed at present and future, the protective ideal self relates to the migrant pupils' wishes to maintain their [emergent] bilingualism (Baker and Wright 2017) or multilingualism and/or their L1, protecting, negotiating and presenting their linguistic identities (Kim 2020).In the literature, the ideal L2 self (integrative motivation) in a migration context was reported as a necessity rather than a 'desire to learn target language' (Dörnyei and Ushioda 2009: 22).In this study, such 'desire' was pertinent to children's ideal L1 selves, as part of migrant children's linguistic identity representation (Evans and Liu 2018).
The ideal L2 selves reports of children could generally be described as necessities for the pupils since the ideal L2 self did not occur if L2 competency was perceived to be already achieved.However, rather than a necessity for L2 competency, the fluctuations and integration of ideal L2 selves into ideal selves and vice versathe need to be important, to be good (ideal selves for self-worth) and have L2 proficiency (ideal L2 selves)reiterates an in-depth relationship between learning motivation and the context of an L2 school (Serafini 2020) whereby an ideal self-acts as a dynamic 'critical catalyst' enabling one to understand the present experiences (Gaines et al. 2018: 210).It is only when ideal L2 selves were explored and analysed in the context of broader ideal selves that the underlying permeating issues in learning for migrant pupils and their causes were identified, particularly relating to their essential well-being needs (self-worth, fairness, and voice).In other words, having analysed pupils' own reports of their broader ideal selves, I see the emergence of the essential issues and an absence of a vision of a more distant future only among those pupils who also reported ideal L2 selves related to L2 proficiency.Chambers (2019) found a lack of depth in the answers of FLL primary-level pupils as most responses were 'yes' or 'no ' (229).The data in my study showed the clear, yet considerably fluctuating, nature of migrant children's ideal L2 selves, though only when they were contextualised and explored alongside broader ideal selves.These suggest it is imperative that pupils' ideal L2 selves are investigated through an exploration of their broader ideal selves to understand the causes of their reports.The cause of the contrasting findings might lie in the fact that my study used a different research design; for a critique of the instrument, that was adapted in Chambers (2019) study, see Taylor (2013).Another reason may be in the specifics of these pupils' migration status, which potentially increases their awareness of language(s) and makes their visions of the future more acute and, therefore, more developed.
It follows from these that having or developing a more distant and clear future vision in LL and beyond (including academic future vision) may be a potential indicator of well-being in an L2 learning environment and might, thus, help migrant pupils to prevent or address their present needs and concerns.This also somewhat furthers a view that academically directed future selves are essential for achievement and engagement (Dörnyei and Ushioda 2009: 22;Oyserman, Johnson, and James 2011).

Conclusion
This research contributes to the existing literature on the learning motivation of migrant primarylevel children in relation to a future possible self, addressing the paucity of research into migrant pre-secondary pupils' ideal selves in England.In the context of primary children's migration the ideal self may be seen as clear and complex images and thoughts of possible futures, framed by children's lived experiences.The three-sided context-dependent and dynamic PAP model of migrant children's ideal self is proposed which re-frames and expands existing conceptualisation of the ideal self-including: prospective (academically-directed), protective (maintaining bi/multilingualism or L1) and agentive (self-worth and voice) dimensions.Although the ideal selves are images of possible futures, they are context-dependent, conditioned by migrant children's ongoing real-life experiences.These may be academic, linguistic (L2, bi/multilingual) and/or recreational, distant and stable or of a more immediate nature, and through them children may express deeper and more fundamental issues about voice, agency or self-worth.The proposed model indicates agentive and context-dependent nature of migrant children's motivations, which were linked to well-being, academic and linguistic identity representations.Given the expansion of global migration, these findings are particularly relevant across various multilingual contexts with monolingualising assimilationist ideologies (Blackledge 2001;Simpson 2015).However, the ways migrant children's future visions allow practitioners to identify and resolve the well-being issues of primary-level migrant pupils in the present as well as to understand the applicability of more distant and/or stable ideal selves requires clarification.As affordances of a qualitative genre are confined, the model may not offer an exhaustive typology of the ideal self.Neither are they as advanced as adults' ideal selves (Nakamura 2019).By challenging earlier claims, the findings do urge abstaining from any definitive conclusions and suggest that our understanding of children's future visions is only beginning to unfold.One way forward is perhaps through implementing less restrictive instruments of investigation to provide the children with more natural tools to voice themselves.

Figure 1 .
Figure1.An integrated agentive approach to exploring future self.

Figure 5 .
Figure 5.The PAP model of migrant children's ideal selves.

Table 1 .
Cases' information and number of days observed.