Excursions as corporate agents: A critical realist account of children’s agency

Recent theorising of children’s agency has focused on relational approaches. Critical realism can provide additional theoretical reinforcement as it demands a focus on relatively enduring patterns of disadvantage and potential powers. Participatory research with children and young people confirms the relevance of Archer’s conception of influence achieved by Selves, Social Actors, and Primary and Corporate Agents. In moments, children within organised collectivities set agendas and shape some circumstances that affect others. When modified by generation-sensitive insights, Archer’s framework may provide understandings of children’s individual and collective agency. These insights might also strengthen critical realist understandings of children and childhood.

A recent flurry of academic debate highlights the lack of theoretical underpinning for the concept of agency within childhood studies (Esser, 2016;Oswell, 2016;Spyrou, 2018;Stoecklin and Fattore, 2018). In many empirical studies of children and childhood the term agency is used without clear definition (Kallio and Häkli, 2013). Raithelhuber (2016) referring to Larkins (2014), suggests there is a tendency for agency to be ascribed to children without an adequate explanation of its meaning, and influence is demanded for children without indicating how this might operate. A theoretically grounded approach to understanding influence would serve to create firmer ground for our 'ontologically weak paradigm' (Oswell, 2016: 24). I suggest that critical realist relational sociology has the potential to enrich this theoretical grounding, if modified by an adequate understanding of children and childhood.
This article outlines recent understandings of the term agency in childhood studies. I summarise Archer's (1995Archer's ( , 2000Archer's ( , 2003 critical realist (CR) account of agency and indicate how her work diverges from similar relational approaches. Drawing on participatory research with children and young people, I discuss the limitations and affordances of these approaches. I suggest that a modified version of Archer's work could help address some of the challenges faced by our discipline and reground understandings of the many things that agency can mean in relation to children and childhood.

Current understandings of childhood agency
Since the 1990s, childhood studies have been influenced by Giddens' conception of agency (Oswell, 2016), defined as the capability of the individual to reflect, to decide on a course of action and to 'make a difference' to a pre -existing state of affairs or course of events (Giddens, 1984: 14 in Raithelhuber, 2016. Childhood studies authors inspired by Giddens tend, however, to ignore his analysis of the contingent and situated interaction of structure and agency in social relations, choosing to focus on either structure or agency in a way that reinforces a sense of dualism, they 'either totalise, globalise and universalise structure or individualise, localise, and particularise agency'. (Oswell, 2013: 63-64).
Recent developments in theorising agency within childhood studies move away from Giddens' approach to agency. Accounts of individual agency are rejected as these reflect majority world (Punch, 2016), liberal, rational and conscious (Esser, 2016) or essentialist (Raithelhuber, 2016) misconceptions. Relational approaches are proposed, which conceive of agency as a social (Esser, 2016) or collective achievement (Oswell, 2013), because a course of action can only be implemented (and changes in social contexts potentially achieved) through dispersed networks or assemblages of human and nonhuman actors (Oswell, 2016) or people, things and process (Raithelhuber, 2016). This can result in there being no distinction between structure and agency collapsing. Whereas for Giddens (1984: 377), 'Rule-resource sets, implicated in the institutional articulation of social systems' are structures, for Esser (2016: 15) 'Rules and regulations … may work as agents'. Esser (2016: 8) suggests that avoiding duality thinking, relational theories make visible the social by revealing 'the materiality and messiness of agency as well as its intersectionality'. Roets et al. (2013) suggest moving towards a lifeworld orientation, to focusing attention on systems as well as individual and social interactions.
An account of scale is important in childhood studies, as there is controversy about the extent to which children's actions can result in influence because children themselves are often dismissed as insignificant or incapable of informed decision-making. Esser (2016) argues that theorising scale needs to move beyond binaries in which agency is described as thin (rather than thick) when it achieves only small change, or agency is described as collaborative (rather than revolutionary) when it maintains inferior social positions. Oswell (2016) argues that in theorising scale the focus of study should be the multiplicity and overlapping nature of space and children's interactions with and through these.
An alternative CR approach to understanding agency, structure and scale within childhood is advocated by Mayall (2002) and Alderson and Yoshida (2016). They argue that children's relationship to agency does not lie at either extreme of individual or collective. They echo the relational emphasis on the importance of contexts, conceived as four planes of social being (bodies in nature, interpersonal relations, social structures and inner being). In contrast to Esser (2016), however, they argue that only humans can enact agency and that each child (and each adult) must be seen as having distinct capacities for agency. The CR approach places a greater stress on the formulation of a goal and decision-making about a course of action, as these arise from human impulses. However, human impulses do not have to be intentional or individual, as agency may be 'ambivalent, intended or inadvertent, rational or foolish, cautious or risky, compliant or resistant' (Alderson and Yoshida, 2016: 86). Structures are understood as relatively stable patterns of advantage and disadvantage in the distribution of resources, attitudes and actions that form contexts for action, which have potentially causal powers (Archer, 2000). Like Giddens, they reject any dichotomy or dualism between structure and agency as structures, reproduced by the activities of human agents, 'are both medium and outcome of the practices they recursively organise' (Giddens, 1984: 25).

Critical realism and Archer's account of agency
Critical realism (a philosophy of the social sciences) has become strongly associated with Donati's relational sociology (Donati and Archer, 2015). Donati (2016) contrasts his approach with transactional or relationist sociology (associated with, for example, Deleuze, Latour and Emirbrayer), which have more prominence in childhood studies. For the sake of clarity, the rest of this article follows Donati's use of the terms CR relational sociology and relationist sociology.
Accounts of ontology differ between CR relational sociology and relationist approaches. Relationist sociology understands social units as 'preeminently dynamic in nature, as unfolding, ongoing processes rather than as static ties among inert substances' (Emibrayer, 1997: 289). It challenges 'substantialist categories that deny the fluidityhence, the mutability-of figurational patternings' (ibid: 308). Individuals are an effect and inseparable from their relational contexts, as the 'units involved in a transaction derive their meaning, significance, and identity from the (changing) functional roles they play within that transaction' (Emirbrayer, 1997: 287). In contrast CR (Archer, 2010: 208) is concerned with focusing on 'relations between people and structures' and the positioned practices through which pre-existing properties may be triggered into causal powers, contributing to continuity or change over time. Put another way, CR criticises relationist sociology for focusing on the flat level of observable patterns of practices and interactions. Instead, it focused on relational subjects and structures which have an emergent impact on social relations; as Donati (2016: 353) explains, subjects' consciousness and its relations 'emerge through different temporal phases in which consciousness and relation … reciprocally condition each other over time'.
This focus on temporality found in CR is also present in some relationist sociology. Emirbrayer and Mische (1998: 970) describe human agency as 'the temporally constructed engagement by actors of different structural environments -the temporal-relational contexts of action'. Agency occurs through analytically distinguishable human processes of iteration (routine incorporation of past patterns of thought and action, which provide stability); projectivity (imaginative generation of possible future actions, which may change received structures); and practical evaluation (judgements about alternative actions in present situations, which are then carried out in practice) (ibid: 971). In CR, the time sequence for the conditioning and transformation of subjects parallels Archer's (1995) description of structural elaboration: conditioning (but not determining) structures and consciousnesses exist (in Time 1), structures are maintained and transformed through collective action by reflective humans in interaction with each other, social conditions and non-human resources (Time2-Time 3) and then morphostasis (reproduction) or morphogenesis (change) of structures results (Time 4). While there is a common interest in temporality, Archer (2003) takes an approach to thinking about subjects as prior to social relations, which relationists may reject as substantive. She also proposes an analytical tool for looking at the influence of individual and collective subjects comprising two individual positions (Selves and Social Actors) and two collectivities (Primary Agents and Corporate Agents). The meaning of Archer's use of these terms, which are sometimes used interchangeably in childhood studies (Mayall, 2002), is summarised below.
Selves are conscious individuals who have developed a sense of social identity through experience, emotion and reflection in a process of internal dialogue (Time 1). They exercise influence through their internal dialogue and personal decision-making (Time 2 and 3), and these choices have an impact on social identities (Time 4). This process of shaping personal identities is influenced by the natural, practical and social circumstances in which individuals are embedded. Internal dialogue is evaluative and responsive to the experiences encountered; it is not a passive process of socialisation and social identities are not fixed. Archer's (2000) suggestion that subjects have a nature and emergent generative powers (Time 1), prior to ongoing relational encounters (Time 2-3) contrasts with relationist accounts of the impossibility of taking 'pre-given units such as the individual … as ultimate starting points of sociological analysis' (Emirbrayer, 1997: 287).
Primary Agents are overlapping collectivities of subjects, occupying similar involuntary social positions associated with shared structured and pre-existing patterns of advantage or disadvantage in natural, practical and social circumstances. They are unable to articulate and organise for shared interests. Archer gives the example of immigrants and non-working youth (although these examples may be contested). While people are born into a social position, shared with others, the relative advantage of that collectivity may change over time. The uncoordinated and unstrategic actions of Primary Agents tend to result in morphostasis, but may result in aggregate effects that reconstitute or transform the environment that they experience, and this can be considered agency. They are not intrinsically passive but are 'denied an effective say, using the use of non-decision-making keeps their concerns off the agenda' (Archer, 2000: 265). Relative passivity emerges as a consequence of their relationship to Corporate Agents.
Corporate Agents are collectivities of individuals who have access to relevant resources that enable them to articulate interests, organise and engage in action. They shape, affect and transform the conditions which other collectivities experience and reconstitute. To pursue their collective goals, Corporate Agents also respond to the maintenance or change in environments which Primary Agents are generating through their agency. Primary Agents can become Corporate Agents when they gain access to resources, often through the growth of social movements. For example, while in modernity the church, state, privileged estates, leisured aristocracy might all have been considered Corporate Agents, now in high modernity, Primary Agents working through defensive associations (e.g. trade unions) and promotive social movements (e.g. civil rights) have organised around their interests and gained access to resources that enable them to become Corporate Agents.
Social Actors are individuals who have found 'a role(s) in which they feel they can invest' and whose accompanying social identity 'is expressive of who they are as persons in society' (Archer, 1995: 261). Social actorship is born of agency, as the sorts of roles to which individuals have access is affected by social relations and placement and engagement of Selves in structured environments, as Primary or Corporate Agents. As Social Actors, individuals can exercise some influence by attempting to transform the roles they fulfil. They can reflect upon the boundaries that govern expectations about their role, and they decide whether to reproduce or transform these relational patterns (Archer, 1995: 288). Archer gives the example of a mother, who as a Self in the structured environment of norms about mothering roles may transgress the expectations of that role. For example, this could be by choosing innovative forms of parenting or dedicating more time to other social roles.
I suggest that Archer's four interlinked categories are relevant for authors who claim social actorship and agency for children in opposition to pervasive notions of children's passive socialisation. However, I have concerns about Archer's (2000: 289-290) suggestion that infants 'have a long way to go'; 'Upon maturity Adam becomes both Agent and Actor'; 'minors could not complete an inner conversation' necessary to develop as selves, and that their interior dialogues are 'provisional (and highly corrigible) "dry runs" at completion'. Her understanding relies upon a reading of Piaget (Archer, 2000: 146-151), which could be described as uncritical. But, failure to critically understand childhood is not sufficient reason to reject insight from social theory outside of childhood studies. Drawing on participatory research with children and young people, the rest of this article therefore explores Archer's work to consider its relevance and to suggest how insights from children and childhood studies could reground CR theories of agency in an understanding of generation.

Methodology
A research study, Stories 2 Connect, initiated by a group of 12 young researchers (aged 13-24) in partnership with 8 adult researchers, involved 50 children and young people aged 10-24 years. Participants were recruited by approaching organisations, which young researchers defined as being part of their community (people with whom they shared an experience of using similar services). All participants had experience of social welfare or specialist education services provided by government, private and charitable sector organisations. The interviews, conducted by young and adult researchers, used open questions, sometimes supplemented with draw-write techniques (Barker and Weller, 2003). They explored children's understandings of community, important life events, experiences of disadvantage, plus relationships and resources that assisted achievement of goals and enabled challenges to be overcome. Adult researchers also conducted pre-and post-study evaluative interviews with 10 of the 12 young researchers and with 14 parents and professionals identified as supportive by young researchers. Young and adult researchers and creative writers worked with all the interview data to create sociological fiction (see www.stories2connect.org).
This article draws on the 14 interviews with adults and 34 interviews conducted with the 27 children and young people who were aged 10-17 years at the time of first interview. Seven were young researchers (5 male, 2 female; all White European; 6 disabled, 1 care-experienced, 2 young carers) and 20 were young research participants (11 male, 9 female; 14 White European, 4 Roma/Travellers, 2 South Asian; 5 disabled, 6 care-experienced, 2 young carers). The adults (all White European) included seven parents (aged 34-57 years; 2 male, 5 female) and seven professionals (aged 35-65 years; 2 male, 5 female) from education, youth work, participation, social care and international nongovernmental sectors. To protect identities, given the small sample groups, age, gender and sector are not detailed in quotes from adults. All participants lived in either northwest England or France.
Participants received information about the study and gave signed consent to participate. Parental consent was given to interview children aged under 16. Young researcher consent was given to interview parents or professionals they had mentioned. Ethical approval for the study was granted by the University of Central Lancashire.
Data analysis was multi-layered. Creative engagement through sociological fiction enabled familiarisation with the data, alongside a hybrid approach of inductive and deductive thematic analysis (Fereday and Muir-Cochrane, 2008). Using NVivo software, three researchers coded goals, challenges and participation activities described in transcripts of the interviews. Drawing on a relational approach, coding of contexts focused on interactions between people, things and processes (Raithelhuber, 2016). These codes were discussed and revised with young researchers through storytelling. The author then analysed moments of interaction in which agency was identified using Oswell's (2013) minimal definition of agency, combining perspectives from children, parents and professionals. Archer's (2000) typology was subsequently used to explore these moments.

Findings
Children taking part in activities (relationships with family and friends, education, leisure, formal participation, volunteering, paid work and unpaid informal caring) used their personal capacities, supportive relationships and material objects things to pursue goals. Engagement in all of these activities fits Oswell's (2013: 53) minimal definition of agency as 'at some basic level … the capacity to do things'.
The reproduction and transformation in structured contexts interviewees described recalls Emirbrayer and Mische's (1998) accounts of agency and relatively stable patterns of social relations. Following educational expectations, by engaging in schooling, could be seen as routine incorporation of 'past patterns of thought and action' in current activity (Emirbrayer and Mische, 1998: 971), resembling iterative agency. Transformation of personal contexts included getting a job, qualifications, positive relationships, independence or a secure home. These activities could be seen as balancing iteration and projectivity (imagining and hoping for better futures) to choose courses of action in response to currently evolving situations, resembling practical evaluation (Emirbrayer and Mische, 1998). The research participants had shared experience of challenging experiences related to education, relationships, health, discrimination, age-based transitions, living arrangements, employment, leisure and poverty. The shared and relatively stable nature of these challenging experiences suggest relationist accounts which emphasise fluidity (e.g. Powell and Dépelteau, 2013) are less relevant than Emirbrayer's (1997). Archer (2000) provides additional insight into the contrasting ways in which different individuals and groups of people experience and influence these circumstances. In some moments, activities described by interviewees involved taking on specific roles (e.g. as helpers and carers), suggesting that Archer's description of Social Actors is relevant. For example, contribution to her mother's recovery of health and return to work was described by this participant who became a carer when she was aged 12 years: she'd be sat there and she'd just randomly start crying … I just kind of grew up from there. I became the mum and my mum became the daughter. (Female, Taking on this role transformed the personal contexts of both mother and daughter but may also transgress role norms associated with childhood dependence. The experience was also significant in the formation of the child's sense of Self as she said, 'that's when I became who I was … that's why I'm so mature'. In a further example, one young researcher spoke of his roles as a campaigner for disability rights, peer mentor, sportsman and helper at his sports club. His mother suggested his helper role served to mask his physical condition, rather than to challenge discriminatory attitudes about disability: For the last two or three years (.) he's been a [sports] helper. … a lot of them don't realise he's got Down's Syndrome because all I hear is … 'He's good, isn't he? ' (Parent) Despite his role as social actor, membership of a Primary Agency collectivity (as a disabled child) meant he was born into a context of discriminatory attitudes about disability. This had an impact on the circumstances he encountered (as Self or Social Actor).
There were, however, moments that appeared to resemble Corporate Agency, when children and young people appeared to contribute to transformation of contexts that would subsequently be experienced by others. As one interviewee explained to a young researcher, membership of a campaigning group of children and adults led to their concerns being heard and subsequent regional and national improvements in services: nationally just making people listen to people. Getting the [National Health Service] better. Getting different services to become better here…making it regional. Going into London. … Challenging it and then changing it (Female, aged 16-17) While Archer's (2000) four categories of Selves, Social Actors, and Primary and Corporate Agents appear relevant lenses for understanding children's social actorship and agency, our data challenge Archer's (2000) conception of children by showing the active and reflexive capacities of children within these contexts. The challenges posed to each element of Archer's framework are therefore examined in more depth below.

Selves
Interviewees mentioned self-building (developing independence, self-confidence, communication skills, compassion, talents and capacities) and pursuing personal goals (happiness, successful relationships, education, employment, prosperity, independence or mobility). For example, one disabled young person moved from a position of lacking confidence to establishing himself as a DJ (disc jockey):  Archer's (2000Archer's ( , 2003 description of Selves who have a level of discernment (about of emotions, experiences, structured opportunities, personal competencies), are able to deliberate (about preferences and probabilities) and are able to determine a chosen course (hopes and goals and role). In the quote, there is evidence of discernment in the choice of this youth club (where 'you don't feel like you're being judged'), of deliberation ('getting to the stage') and of personal preference (if others don't like it 'tough!'). Discernment was also evident in assimilating challenging experiences to pursue new chosen courses of action. For example, one interviewee described an example of overcoming a challenge: I was in the care home. I had a lot of problems with people saying 'Ahhh you're different to us' at school. … it's been very hard. And I've always reacted, like, physically and verbally. … but now I'm having help and I'm learning different strategies … I just walk away and just ignore them. (Male, Through external conversations with supportive staff and internal dialogue, this boy assessed his circumstances, evaluated his options and chose a specific goal, related to not getting wound up. Internal deliberations are not then entirely internal, but rather an ongoing and responsive interaction. Emirbrayer and Mische's (1998) emphasis on the importance of dialogue is therefore relevant. Experiences and opportunities are also vital resources. For the boy in the care home, guidance on coping strategies from staff provided knowledge and options. For the DJ, the youth group for disabled children created a place of safety. Archer (2003: 104-105) states that Selves have causal effect, they enable us to 'modify ourselves reflexively' and also to 'modify the world as a consequence of our internal deliberations about it'. The two examples challenge Archer as they show the world might be altered by child Selves who have not yet reached Archer's suggested age of maturity (18 years), through less conflict in school, more music in the youth club and subsequent change in relationships or internal dialogues. The importance she places on patterns of advantage and disadvantage remains relevant, however, as interviewees revealed a growing lack of relevant opportunities and services being cut through austerity measures. This affected the resources available for self-building and the extent to which Selves as members of similarly positioned collectivities (Primary Agents) can access certain social roles (e.g. as DJs) and translate personal goals into achievements. Archer (2000) suggests that children are ascribed roles (as pupils or daughters) and could not be said to freely choose a social role, because they cannot competently complete internal dialogue; therefore they cannot be Social Actors. Our findings show, however, that children and young people chose and invested in multiple roles, showing regular daily, weekly or monthly commitment (as friends, participants, sportsman, helpers, advocates, campaigners, youth workers, cashiers and carers). Children and young people used these roles as ways of describing themselves. Some children experience more constraints than adults experience when choosing, transforming and balancing time in different roles. Family, community and generational contexts affected what roles were available for and chosen by young interviewees. Being a sports helper was possible due to family connections. A role as carer was taken on through choice but also due to the lack of mental health services. A role as member of youth parliament was possible through participation groups, but roles as paid politicians or electors were not possible due to laws excluding children (in their generational position as Primary Agents). Dialogue between an interviewee (male, aged 12-13), an adult researcher and a parent indicate the interplay of choice, social relations and structured patterns of expectations. The child, from a Traveller family, discussed the competing pressures around his role as pupil:

Social actors
Researcher: What made you [go to school]? Child: My Mam and [Worker from education outreach service for Travellers]. Researcher: So, did that take a lot of persuasion? Child: Yeah, a lot. I didn't get a choice. Parent: You didn't get a choice, did you? Child: It's hard to go to school because my cousins don't go to school either… because like as being like a Traveller and that it's looked down on going to school and that … because not like what you normally do.
The child described pressure to attend school from his mother and a professional working for the municipality. The mother questions his interpretation. But family, professional and generation-related expectations of attending education were in conflict with community expectations that he leave education to engage in work or sports careers, like the majority of boys in his community. While generation-related structures may be constraining, choice about being a pupil is possible, especially in a context where cousins of a similar age are not attending school. Choice of social roles for adults may also be limited, especially where adults have to engage in employment that does not express their identities in order to earn money. It is crucial to move beyond age-based assumptions of capacity and autonomy to choose and towards the conception of interdependent humans in connection with others (Moran-Ellis and Sünker, 2018). CR relational sociology does however draw attention to relevant emergent potential properties of relational subjects and structures, which can aid understanding of children's opportunities to engage in roles as Social Actors. For example, a young researcher described his mother supporting his maintenance of the role as pupil: I probably wouldn't be anywhere without Mum. … when I was about to get kicked out of primary school for having Asperger's. She basically…kicked off about that[challenging the school]. (Male, In a separate interview his mother stated he had got himself through the hardest challenges, through his own determination, what Donati (2016) might describe as a preexisting feature of his relational self. As well as the emergent power of relational Selves and dynamics within personal relationships, emergent powers of enduring structures appear to have had an influence. Economic structures informed spending on specialist mainstream educational support and teacher training, as well as expectations about children's needs to create themselves as future productive workers. Cultural norms recognising the legitimacy of a 'warrior parent' appear to have been triggered more forcefully than dominant conceptions that children should fit education systems.

Primary and corporate agents
Interviewees described their achievement of goals and roles being constrained by experiences of negative attitudes towards disability, childhood, youth and Travellers and inadequate resourcing of education and youth services. The relevance of Archer's (2000) conception of collectivities of Primary Agents is reinforced by the data; relatively disadvantaged positions were shared between certain groups of interviewees. These collectivities were sometimes intergenerational, as in Travellers of all ages sharing similar circumstances of entrenched racism. Membership of different collectivities are also sometimes age-related and overlapping. Disabled children are a collectivity disadvantaged in terms of cultural attitudes and social provision related to both childhood and being disabled. Disabled children are also members of class-based intergenerational collectivities which may afford access to relative advantage or disadvantage. While Archer (2003) notes that children are born into positions as Primary Agents, acknowledgement is needed of generation as a relevant and intersecting structure in the circumstances experienced by Primary Agents.
Primary Agents can become strategic, coordinated and agenda setting Corporate Agents through social movements which gain access to resources (Archer, 2000). Our findings demonstrate moments where this shift from Primary to Corporate Agents may be occurring, when children and adults worked together through social movements and invited participation activities (through advisory or decision-making boards, delivering training, lobbying public decision-makers and participatory action research or social action groups). For example, a young researcher describes co-delivering a training session about disabled children to social care professionals: lots of people said to me that they understood disabled, disability more … It was good to have a worker with me to help explain [technical questions about how we have improved services locally] … we're making a difference these days. (Young researcher, female, aged 17) Disabled children and young people, as Social Actors, had taken on specific roles (trainers, lobbyists) in this activity, but the worker involved also described how the training sat within coordinated action of adults and disabled children who were setting agendas and influencing circumstances: a lot of work we've done with yourselves and University of Central Lancashire that's really, really massively helped raise the profile. And for ourselves, linking in with the Council for Disabled Children and other national organisations that's helped. (Professional) These changes in understandings of disability and services provision were occurring within a collective, intergenerational process of facilitated participation. Concept of intergenerational agency (generagency) is therefore relevant. Obstruction of meetings with decision-makers was also sometimes described, suggesting Archer's (2000: 265) reference to 'non-decision-making' keeping Primary Agents' concerns 'off the agenda' also remains relevant. Rather than the achievement of corporate agency, until influence that enables direction of collective resources is secured, examples such as this might be termed temporary excursions into Corporate Agency.
The extent to which social movement and invited participation activities can enable even excursions into Corporate Agency is highly variable. A professional interviewed in France suggested that global political will is turning towards valuing the presence of children in local and national governmental planning committees. He cited examples from India and the establishment of a French national high commission on age with a dedicated advisory group of children. One young researcher questioned the value of such initiatives, citing a European event designed to lobby decision-makers which 'was a disaster … They said the wrong things, the big people … they're not doing things very well'. Invited participation groups and events can be tokenistic instruments designed more to educate children than to include them in governance (Kiili and Larkins, 2018). Similarly, social movements may not achieve change, as explored in the following subsection. Nonetheless, research suggests there are moments in which collectivities of children and adults organise to articulate interests and have an impact on public decision-making through both social protests and invited participation (Kiili and Larkins, 2018). These collectivities tend, however, to be requesting change by other resource holders rather than using resources they control to enact the changes they seek. Children's exclusion from the political and economic arena is an entrenched barrier here. Yet, these moments show what may be possible, in certain contexts over time. Archer's (1995Archer's ( , 2000 focus on Primary and Corporate Agency, Selves and Social Actors is useful for childhood studies, as it provides an account of the emergent powers of individuals and collectivities, across different scales, contexts and times. A focus on Selves and Social Actors provides a lens for exploring the influence that individual children have in their own lives, and in the lives of people around them. It resonates with Stoecklin and Fattore's (2018) findings that the capacity for reflexive monitoring and choice-making are central to children's own conceptions of agency. Archer's notion of collectivities of Primary and Corporate Agents expands the repertoire for childhood studies. These conceptions of agency provide a framework for considering how prevailing conditions of disadvantage are maintained or transformed over time and the excursions towards Corporate Agency that become possible. But Archer's framework would benefit from an intersectional and interdependent understanding of the structuring influence of generation and the emergent powers of child selves. Relational child Selves bring personal capacities into the contexts they encounter and may exercise significant influence in the roles they choose and how they transform these. The emergent powers of structures related to generation may or may not be triggered in any given moment, to enable or constrain interdependence and autonomy in the choice of roles, influence achieved through social roles and the potential for successful excursions into corporate agency, for both children and adults.

Potential contribution of and to Archer's account of agency
Archer's framework may be complemented by analytical aspects of relationist sociology to assist understanding of contexts, but a CR focus on planes of social being is also beneficial. Emirbrayer and Mische's (1998: 974) focus on agency as a dialogical process may help expand the understanding of how agency operates, as interviewees' external dialogue enabled Selves to deliberate and determine chosen goals, and collectivities to develop and organise for shared interests. But, it was evident that collective influence required much more than just dialogue. For Caetano (2015), Archer lacks understanding of the role of habitus (stable patterns of dispositions), and there was evidence of interviewees having a disposition to narrate what might be termed pro-social behaviour ('giving back', 'making a difference') and what might be termed self-help (developing positive mental reflexes and personal skills). These narratives reflect the habitus of the field of child participation and the expectations of social contribution and neoliberal self-governance that imbue invited participation opportunities (see Raby, 2014). Archer's defence is that habitus is not determining because Selves are reflexive. She might suggest that children's repetition of these narratives is a sign of their immaturity and incapacity for complete and independent internal dialogue, but the professionals interviewed also repeated these narratives and wanted to 'make a difference'. Attention to the role of habitus is therefore relevant for both children and adults. Archer (2003: 141) is criticised for not according sufficient weight to contexts (Caetano, 2015;Dépelteau, 2008) and in participatory analysis with children, a relationist focus on people, things and processes (Raithelhuber, 2016) and a lifeworld (Roets et al., 2013) focus on systems did assist identification of relevant contextual factors that were enabling or impeding children's goals. However, Archer does acknowledge the enabling and constraining nature of objective circumstances that comprised the four planes of social being (bodies in nature, interpersonal relations, social structures and inner being). A focus on these four planes helped us develop understanding of how different forms of interpersonal relations (between children, workers, parents, peers) and social structures (particularly related to discrimination and austerity) affected interviewees contexts of action.
Suggesting use of aspects of relationist thinking within Archer's framework does not mean embracing a relationist ontology, which remains overly focused on fluidity and appearance rather than causal mechanisms. Dépelteau (2008: 71) questions whether Archer's co-deterministic theory has anything useful to offer as it does not see the 'fluidity of social processes', but a key benefit of Archer's approach is to not be distracted by narratives of fluidity. Following Boltanski andThevenot (2006 in Roberts andJoseph, 2015), conceiving of the social world in terms of fluid networks and flows can be seen as a colonisation of social theory by the logic of the market. It reflects dominant ways of thinking associated with financial neoliberalism. It emphasises analysis of fluid appearances rather than underlying relatively stable conditions (Roberts and Joseph, 2015). For example, Emirbrayer (1997: 293) suggests that differences in advantages and disadvantages 'crystallize' around practices and 'Unfolding transactions, and not pre-constituted attributes, … most effectively explain equality and inequality'. Emirbrayer gives no account of how crystallising happens and no justification for why pre-constituted attributes are less important. Our data, in contrast, have shown the relevance of the CR assertion that pre-constituted attributes of relational Selves (e.g. determination) and social structures (e.g. discriminatory attitudes) are brought into the circumstances that children encounter and these have an influence on achievement of their individual goals and shared agendas.
The affordances of a generation-sensitive version of Archer's framework are perhaps best understood by reference to an example. In our study, a Traveller young man described demonstrating, with his parents and other community members, to try to gain more authorised stopping sites for his community. The organised articulation of their interests was disrupted by police attitudes and actions.
Dad: They're prosecuting [demonstrators] for putting the horses on Parliament Square. Child: Even though they had permission! In this situation, generation-sensitive CR analysis of agency demands attention to relatively enduring disadvantage and potential for change, for example, (1) how preexisting patterns of disadvantage (experienced by intergenerational collectivities of Travellers as Primary Agents) are maintained by the ongoing organised interests of Corporate Agents (perhaps governments and housing developers), (2) how access to relevant resources and opportunities for agenda setting is or can be accessed (by Travellers as Corporate Agents), (3) how pre-existing attributes (potential powers) and social roles the Traveller young man brought into this situation had any effect and (4) how patterns of disadvantage or agenda setting experienced were related to his memberships of collectivities of Travellers, men and children. 1

Conclusion
Our findings confirm that Archer's (1995Archer's ( , 2000Archer's ( , 2003 work has relevance for understanding agency within childhood, however, there is need to draw understandings of childhood sociology into her work. Given that Archer's work acknowledges the cyclical process through which consciousnesses act and develop in relational contexts, her fixed notion of childhood competence is surprising. This seems to be an uncritical reflection of dominant notions of children as not yet beings, rather than recognition that children (like adults), are involved in process of both being and becoming (Uprichard 2008).
A generation-sensitive CR account of agency is multifaceted. Children's (like adults') internal dialogues are affected by dialogue with others and habitus as well as other aspects of the contexts they encounter. Children as relational Selves bring potentially emergent powers into these contexts, but the extent to which their power brings about change is related to the specific, temporally located circumstances encountered. Children choose and transform roles as Social Actors, but they do face some constraints on their freedom to engage in certain roles. These constraints are related to generational (and other) structures. As members of collectivities (related to generational and other affiliations) children are Primary and sometimes Corporate Agents. Through these collectivities they may contribute to the maintenance (morphostasis) or transformation (morphogenesis) of patterns of social relations and structures. Potential to exercise Corporate Agency is related to age-related (and other) circumstances which affect the extent to which children can set agendas and direct the use of resources to affect the contexts which they and others experience.
Archer's analytical dualism (identifying structure and agency) and insistence on preexisting potentially emergent powers of Selves and Structures is valuable to childhood studies. It demands that we question not just the social relations between people, things, processes or systems that play out in the contexts children encounter. It also forefronts attention to the underlying potentially causal mechanisms that have personal or historical roots, the selves and the enduring shared factors that affect children's experience of advantage and disadvantage. Considering lifeworlds and relationist understandings of contexts remain useful analytical devices, including in participatory research with children and young people.
The innovative adaptation of Archer's framework outlined provides a new starting point for looking into children's descriptions of their experience and influence, adding a generation-sensitive CR account of agency to the repertoire of childhood studies. This is a provisional tool, however, as befits a CR approach. Over time, further dimensions of influence and understandings of the overlaps and interactions between these may be developed. Research could usefully focus on charting children's experience of moments of corporate agency, by identifying circumstances indicative of potential for disrupting existing patterns of disadvantage. As Archer recalls, change at this scale, tends to be slow. An essential first step is to look below and through the (sometimes) fluid appearances to recognise the emergent potential powers of children and structures, and the specific circumstances and contexts through which they do or could exercise influence.

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) disclosed receipt of the following financial support for the research, authorship, and/ or publication of this article: This work was funded by a UK Arts and Humanities Research Council, Connected Communities Grant (Grant Number AHM001539/1). PI Candice Satchwell, CoI Cath Larkins.

Note
1. For anyone interested in reflecting on these questions, a fuller account of this moment is available as sociological fiction at www.stories2conenct.org