A Feminist Relational Discourse Analysis of mothers’ voiced accounts of the “duty to protect” children from fatness and fatphobia

Research has highlighted damaging contradictions in the responsibilisation of mothers over children's health, at once held responsible for tackling “childhood obesity” while being cautious not to encourage children to become obsessive with their bodies. While research has highlighted discourses of blame and elucidated mothers’ experiences, less is known about how mothers negotiate discourse in their voiced accounts. Utilising Feminist Relational Discourse Analysis, this study analysed interviews with 12 mothers in England to explore their experiences of a nationally mandated BMI screening programme in schools and how discourses shape their voices and experiences. In negotiating complex and contradictory discourses of motherhood and fatness, participants expressed a “duty to protect” their children from both fatness and fatphobia. Negotiating these responsibilities left mothers feeling guilt at their personal “failure” to protect their children from one or both harms. Mothers did not take up these discourses unproblematically; they resisted them, yet felt constrained by “expert knowledges” of fatness and motherhood that had clear consequences in responsibilising mothers for the “harm” of fatness. This analysis calls attention to how dominant discourses function personally and politically to responsibilise mothers for the harm caused by state-sanctioned fatphobia.

Moreover, fatness is regarded as a "virtual confessor" to irresponsible lifestyles (Murray, 2009) and, in the context of scarcity and inequality (Harrison, 2012), fatness is constructed as an immoral choice (Mulder et al., 2014). With both the USA (Dickman et al., 2017) and UK (Marmot et al., 2020) subjected to extreme inequality and deprivation, discrimination targeted towards fat people, who are constructed as failing to consume responsibly, is profoundly prevalent. Such discourses are sanctioned within government policy (Gillborn et al., 2020). These discourses do not implicate all fat people equally; discourses position working-class and racialised people as disproportionately "obese" communities that threaten the nation's health (Herndon, 2005).
BMI itself was derived from Adolphe Quetelet's measurements of what he understood to be "the average man", his work suggesting that these were white, cisgender men (Harrison, 2021). Furthermore, "obesity" discourse draws on white and middle-class moralising discourses of the "healthy body" (Azzarito, 2009;Murray, 2009), reinforced in raced and gendered ways by "media environments that privilege whiteness and thinness as quintessential womanhood" (Gentles-Pert, 2020, p. 319). These discourses, propped up by BMI, construct whiteness as the invisible, unexamined norm against which all bodies should be compared. The disproportionate marking of racialised people as failing to meet these white standards potentially "reclaim[s] race as a biological category through fatness" (Azzarito, 2009, p. 185), contributing to scientific racism and dangerously challenging data establishing that race does not have a genetic basis (Saini, 2019). To this end, blame for "obesity" and its purported risks are individualised to already marginalised groups and shifted from socially and politically located inequalities (Boero, 2009) in favour of endorsing measures and discourses that uphold the racist status quo (Strings, 2019).
Moreover, Campos (2006, pp. 38-9) refers to the current "fat panic" as "having very little to do with science, and everything to do with the economic and professional motivations of obesity researchers, eating disordered thinking, and anxieties about class, race, and social overconsumption in general". In exploring these motivations, Glassner (1999, p. 35) argues that particular risks are highlighted and emphasised by governments and researchers "either because they offend the basic moral principles of society or because they enable criticism of disliked groups". The "epidemic" status awarded to "obesity" works in this way to position fat bodies, particularly those of marginalised groups, as threatening and dangerous (Boero, 2007).

The responsibilisation of mothers
With the "healthy body" constructed through moralising discourses (Murray, 2009), an increase in children marked as "obese" is regarded as a risk to the future economic health of society (Harrison, 2012) and societal morals and norms (Coveney, 2008). Thereby, in recent years, "obesity" policy and research have increasingly targeted children, with the actions and influence of mothers a frequent target of research (e.g., Gross et al., 2014;Hughes et al., 2016). For years, feminist research has demonstrated the classed, raced, and gendered dimensions of conceptualisations of "good" and "bad" motherhood (Reppond & Bullock, 2020) and elucidated how mothers carry the burden of a social obligation to raise healthy and well-adjusted children (Rose, 1999), responsibilised for children's various outcomes and their actions interrogated from before conception (Parker & Pausé, 2017). Governmentality scholars have highlighted this responsibilisation of mothers as a "biopolitical project" in service to neoliberalism (Parker & Pausé, 2019), emphasising individual responsibility and justifying interventions for "problem bodies" (Rabinow & Rose, 2006). Therefore, discourses holding mothers responsible for children's pathologised bodies serve as a biopolitical tool, ensuring mothers govern their own behaviour and that of their children to meet idealised standards of health. Indeed, policy documents related to the implementation of the NCMP construct active engagement in the programme as the moral duty of responsible citizens and good parents (Gillborn et al., 2020), and racialised mothers of fat children in particular are more likely to fall under child welfare scrutiny, their children removed from the care of the "failing mother" and placed in the hands of the state (LeBesco, 2011;Zivkovic et al., 2010).
Mothers' responsibilities related to children's eating behaviours and weight outcomes are often contradictory, encouraged to monitor children to prevent "obesity" (e.g., Twarog et al., 2016) while being cautious not to encourage children, particularly daughters, to become obsessive with their bodies and food consumption (Vander Ven & Vander Ven, 2003). Critical scholars have highlighted these damaging contradictions, calling attention to patriarchal cultural norms that construct idealised bodies (Woolhouse & Day, 2015) and position women's and girls' bodies as in constant need of improvement in ways that harm mothers and daughters alike (Walsh & Rinaldi, 2018). Dominant discourses responsibilising mothers for the harm caused by constructed body ideals shift responsibility to individual mothers (Parker & Pausé, 2019), obscuring the role of government and policy in legitimising fatphobia as a response to "regimes of truth" about "obesity" (Gillborn et al., 2020).
Research with children has demonstrated how the NCMP reproduces a fear of fatness as an abnormality (Evans & Colls, 2009). Mothers are "incorporated as … actual or potential all[ies] into pedagogic programmes" (Rose, 1999, p. 182), and policy documents related to implementing this programme construct active engagement and cooperation as the moral duty of parents (Gillborn et al., 2020). Research has elucidated parents' opinions of the NCMP, such their rejection of the programme's pathologisation of their children (Gainsbury & Dowling, 2018). Critical research has further revealed the impact of "obesity" discourse on fat women implicated by it, such as the stressful experiences of fat pregnant women who are considered a risk to their future children (e.g., McPhail et al., 2016) and shamed into unhealthful behaviours (Parker & Pausé, 2017). There is space here, then, to consider discourse and voice together in order to understand how the women implicated by the discourses in the NCMP negotiate them in their spoken accounts, and how their experiences are shaped in multiple ways within the realms of available discourse. To this end, the current study applied Feminist Relational Discourse Analysis (FRDA; Thompson et al., 2018) to analyse discourse and voice in mothers' accounts of the NCMP. The aims of the study, therefore, were to 1) explore the discourses mothers drew on in their narratives of the NCMP and 2) explore how mothers' voices and experiences are shaped by these discourses.

Language
It is important to define the language used throughout this paper. Firstly, while discussing issues disproportionately concerning mothers and daughters, we do not assign biological essentialism to these categories. We view women and mothers as anyone who identifies with these categories and define daughters as any children who identify with the categories or experiences of girlhood and/or daughterhood. All mothers, women, daughters, and girlswhether cisgender, transgender, or gender non-conformingare harmed by patriarchal and fatphobic structures, albeit in sometimes different ways. For example, transgender women and girls in particular experience greater body dissatisfaction and disordered eating (Parker & Harriger, 2020), likely exacerbated by the arbitrary BMI limits placed on access to gender-affirming healthcare (Brownstone et al., 2021). We recognise the limitations of this gendered language, but seek to name those who are harmed by oppressive systems and structures in order to challenge this harm.
Furthermore, we reject medicalised and pathologising labels based on inaccurate measures of health such as "overweight" and "obese." Instead, in-line with language endorsed by fat activists (e.g., Cooper, 2016), we refer to those who are pathologised for their nonconformity to standardised health measures as "fat." Due to the dominant nature of "obesity" discourse and "common sense" use of BMI labels, this sometimes results in a disconnect between the language used by study participants and us as authors. When reflecting participants' words, we use the same terminology as participants (e.g., "obese") while our analysis is reflected by our use of non-medicalised language.

Participants and recruitment
The study utilised semi-structured interviews with 12 women whose children were recently invited to be measured through the NCMP. We recruited participants from across England through posters in community centres, online posts on Facebook and Twitter, and via snowball method. Informed consent was obtained and, to protect participant anonymity, participants and any additional names mentioned in interviews were given pseudonyms. The researchers' university ethics committee granted ethical approval for this study.
Twelve interviews were held in total with women who self-identified their ethnicity as "White British" (n = 7), "mixed-race: White and Black Caribbean" (n = 1), "Black African" (n = 1), "Black British" (n = 1), "British Indian" (n = 1), and one woman who declined to disclose her ethnicity. Through recruitment we did not define categories of motherhood or womanhood on behalf of participants; anyone self-identifying as a woman and a mother was invited to take part. Participants were not asked to disclose their gender or any biological relationship with their children. Furthermore, fathers were notably excluded from this study; this was in order for the study to reflect the voices of those most often implicated by discourses of children's bodies and health; namely women and mothers. Participants' children were aged between 4 and 13.
While not directly measured, interviews identified that participants' children received a range of BMI results, but no participants reported a child who had been marked as "underweight." Although data on the women's and (if relevant) their partners' occupations were collected, we did not define social class on behalf of participants due to the failure of standard operationalised measures to touch on the complexities of social class and how it is perceived and experienced (Day et al., 2020). However, during analysis, we used Holt and Griffin's (2005) guidance to look for subtle and indirect classcoded discussions.

Analytical approach
Discourses govern how phenomena can be constructed, studied, and understood and are fraught with power (Foucault, 1969(Foucault, /2002. Therefore, discourse analytic research facilitates understanding of the multiple and contrasting ways of speaking about particular phenomena, how these ways of speaking produce and shape knowledge and truth, and what purposes or interests these constructions serve (Arribas-Ayllon & Walkerdine, 2008). However, feminist researchers highlight how discourse analysis can be problematic when seeking to understand individual voices and experiences. For example, the voices of individual women are often lost when focussing on the overarching discourses that inform and underpin their narrative accounts (Saukko, 2010, as cited in Thompson et al., 2018). Furthermore, postmodern rejections of "truth" can lead to women's material experiences and disadvantages not being regarded as "true," which is counter to the goals of feminism. To this end, FRDA (Thompson et al., 2018) aims to identify the discursive realms through which experiences are made and given meaning, and individuals' embodied experiences within these realms.
The centring of voice within FRDA allows us to understand how spoken accounts of the self are intimately connected to discourse and can provide powerful counternarratives to dominant ways of understanding (Thompson et al., 2018). Moreover, analysing voice in relation to discourse can assist in understanding how lived experiences are actively structured, mediated, and negotiated within overarching discourses (Saukko, 2010, as cited in Thompson et al., 2018) and allows us to understand how discourses "hit" and "bruise" us (Ahmed, 2017, p. 30, as cited in Thompson et al., 2018).

Procedure
Data collection and analysis was conducted primarily by the first author (SG). Interviews lasted between 45 min to 90 min, during which participants were asked questions about their experiences of NCMP, including: how they first felt when they heard about the programme; whether they talked about the measurements with their children or other family members; and how they felt about the results and the "hints and tips" they had received. Interviews were transcribed verbatim and analysed using Thompson et al. (2018) FRDA guidelines to identify the discourses that shape participants' voices and locate participants' experiences in relation to these. FRDA involves two analytical phases that explore how experiences and discourses come together through voice (Thompson et al., 2018). The first phase aids in identifying the broad discourses that participants draw upon, and the second traces participants' voices through these discursive patterns.
In phase one, SG listened to interview recordings multiples times to build familiarity with the data and note down emerging voices. SG then split transcribed data into "chunks" and assigned descriptive codes which were used to generate in-vivo themes and recurring patterns of meaning within participants' talk. Six final themes were generated across the 12 interviews, including "being weighed and measured" and "responsibilities." Each theme was analysed to understand how they were discursively constructed within the data set by looking at quotes to identify the meaningful sets of statements and assumptions about that theme, which were multiple and contradictory. For example, concerning the theme "being weighed and measured", discourses included "measurements are necessary to challenge obesity" and "measurements are harmful to health", representing contrasting discourses at play around this theme. Finally, these themes were discussed between all authors to identify the overarching discursive patterns at play, named The physical and psychosocial harm of "obesity" and Negotiating mothers' responsibility. Finally, we consulted theory and research to make sense of these discursive patterns, including consulting research to understand the functions of these discursive patterns in broader social, historical, and ideological contexts (Thompson et al., 2018).
The second phase of analysis aimed to analyse emergent voices within discourses. First, SG generated "I Poems" for each participant by collating quotes pertaining to each discourse. Each statement made in the first-person tense was underlined, including every "I" statement and accompanying verbs. Statements beginning with "you" were also included when the participant was arguably discussing their own experience (e.g., Stephanie: "you see stuff happen with [your children] and you think back, well, did I do that?"). These underlined first-person statements were kept and the rest of the data deleted to create an I Poem of each discourse for each participant. These I Poems allow the researcher to hear and identify each participant's first-person voice concerning the discourses at play, allowing us to listen to how the participant narrates their own self and experiences (Gilligan et al., 2006, as cited in Thompson et al., 2018).
After creating I Poems, SG listened for "contrapuntal voices" to identify contrasting voices and stories in participants' first-person accounts. Gilligan et al. (2006, as cited in Thompson et al., 2018) state that identifying these voices allows us to recognise the multiple sides of stories being told, capturing the personal in relation to the political. What emerges here is a multi-layered account of individual voices as each participant contemplates the realms of the discourses within which they are situated (Thompson et al., 2018). For example, we heard contrasting voices of guilt and defiance within I Poems. Finally, all three authors constructed a theoretical account of the analysis, tying together the findings of the research aims in relation to individuals and considering the functions both personally and politically of the identified discursive negotiations (Thompson et al., 2018).

Positionality
In analysing the voices and experiences of other women, it is important for us to consider our positions as active participants in the production of knowledge. The first author (SG) is childfree and non-fat, limiting the extent to which the first author can empathise with mothers implicated in discourses of blame related to their children's bodies. Particularly in research of this nature, the bodies of the researcher and those being researched are brought into direct interaction (Throsby & Evans, 2013), and it is possible that participants initially assumed the interviewer was supportive of dominant knowledges of "obesity" by virtue of being a non-fat researcher investigating mothers' experiences of this programme. Despite this, we were reassured that participants who felt strongly opposed to the programme seemed comfortable to share these oppositions from the start. SG was also pleased to share some sense of familiarity with participants, offered in small but meaningful ways in which participants invited SG to share in their thoughts and experiences and through which SG offered reassurance when participants challenged dominant knowledges of "obesity." BR and MW are both mothers who are strongly opposed to the NCMP and opted their own children out of the programme when invited to take part. BR and MW therefore identified with much of what the participants said, and this undoubtedly shaped our reading and interpretation of the data when we together discussed the discursive patterns being reproduced and challenged and the personal and political functions of these discursive negotiations.

Results
The findings of data analysis are presented in two subsections, corresponding to the two phases of analysis. First, overarching discursive patterns in participants' talk are outlined using extracts from the data to illustrate how these patterns appeared in participants' talk. Second, participants' emergent voices are analysed using I Poem extracts to demonstrate how mothers negotiated these discourses in their experiences.

Phase 1: Discursive patterns
The physical and psychosocial harm of "obesity". The first discursive pattern identified in interviews with mothers relates to dominant representations of "obesity" as harmful to physical health and psychosocial wellbeing; to this end, this discursive pattern constructs weight-monitoring as a valuable harm prevention tool. Many participants reproduced constructions of "obesity" as inherently dangerous conditions that warrant governmental tracking and intervention: I do agree to a point that more and more children are obese and health-wise, later on in life, it's gonna be harder for them to get … if they continue on the path that they're on it will be harder for them, especially medically, just to sort of get over stuff. Obviously, cos if they're obese now then obviously when they get later in life and they don't change and they keep getting bigger then. (Stephanie, White British) I think it's … it's a touchy subject for a lot of people, isn't it? I think that it's quite, erm, it's quite a tricky one. … I can see that something needs to be done about childhood obesity, obviously, and obesity generally, and healthy living. (Pandora, White British) These extracts reflect the dominant discourse of "obesity" that multiple dangers of "obesity" are a given, framed as a matter of common sense where these "grotesque discourses" are taken as scientific and objective facts (Foucault, 1963(Foucault, /1989. Stephanie draws on assumptions of "obese" children becoming "obese" adults, and the number of children marked as "obese" is positioned as a measurement of the health of society. Therefore, it follows that children warrant governmental and parental tracking and intervention through the NCMP, with children positioned as the "apocalyptic demography" (Robertson, 1997) who must be protected to protect the future of social order (Holmer-Nadesan, 2005).
In line with the discourses (re)produced in NCMP materials and documents (Woolhouse & Day, 2015), fatness was reproduced as detrimental to social elements of individuals' lives: Bullying as well … People that are obese are given a hard time, really given a hard time, and sometimes some of these people have health conditions which make them obese so, you know, they're not at fault. But I think that we really look down on, as a society we do look down on people, and you know, that's something that a child may have to put up with in the future. (Sonika, British Indian) As a parent, I quite agree with [the programme] that it's a good thing, especially for the children, … they really need to get to have a normal healthy weight. … Then they can even have confidence and trust in theirself, love themselves, learn to appreciate themselves, you know.
[If children are overweight] that confidence might not be there, and may even lead to … [it] might make some of them to have low self-esteem, like not really appreciating themselves. (Rayowa, Black African) In the above extracts, both weight stigma as expressed by others and as felt by oneself are positioned as inevitable responses to fatness; it follows, then, that a "healthy weight" will protect children's mental wellbeing. It is worth mentioning that both Sonika and Rayowa's children had been marked as "normal weight" through the NCMP; therefore, while they reflect on issues that fat children may face, they do not necessarily do so from a position of concern for their own children.
These extracts echo discourses of fatness (re)produced through policy that position successful weight loss as a preventative measure to avoid bullying or low self-esteem (Gillborn et al., 2020). To this end, avoiding fatphobia by successfully conforming to normalised standards of health is necessary and can be achieved via weight-monitoring, while fatphobia is a reasonable and expected response to abnormal bodies. Sonika also draws on binary constructions of the "deserving" versus the "undeserving" fat, not dissimilar to those of the undeserving poor (Himmelfarb, 1992). In both constructions, fatness is framed as an abnormality whether caused by illness or personal failure. For those whose fatness is seen as within their control, fatphobia is positioned as an inevitable response; meanwhile, those who are fat because of factors outside of their control deserve to be protected from ridicule.
In constructing fatness as the cause of issues related to body image and self-esteem, this discourse frames weight-monitoring as essential towards protecting children from this stigma. Individualising blame in this way obscures the harm caused by pathologising discourses of fatness granted legitimacy through public health campaigns (Catling & Malson, 2012) and interventions, including the NCMP (Evans & Colls, 2009). Indeed, government policy positions "normal weight" as a method of protection from bullying and low self-esteem, establishing this as an "expert knowledge" (Gillborn et al., 2020) that is evidently drawn on by mothers making sense of the programme.
Mothers seem to reproduce this discourse in the interests of protecting a "promise of happiness" (Ahmed, 2017) for their children in line with the assumption that a fat life is an unhappy life. To this end, mothers reproduce stigmatising discourses of fatness in ways that are cloaked in the desire to do "what is best" in the interests of children's happiness. This relates to the second discursive pattern identified in participants' talk, in which mothers negotiate the responsibility, and desire, to protect their children from harm.
Negotiating mothers' responsibility. The second discursive pattern identified in participants' talk around the NCMP was related to mothers' responsibility to protect their children from harm, including the harm of weight-monitoring and weight-talk. This represents resistance to dominance discourses, instead positioning weight-monitoring as something children should be protected from: We have a big thing in this house that we don't talk about weight in any negative way at all. You just are who you are, if that makes sense. So, weight, and being weighed, and dieting, and anything like that is pretty much not ever gonna happen in this house. These extracts resist weight-monitoring as the most effective intervention (Gillborn et al., 2020) by re-presenting is as harmful and to be kept out of the family home and out of children's thoughts and worries. This represents resistance, contradiction, and complexity; while many mothers drew on dominant discourses to frame the measurement programme in general as reasonable and potentially helpful, in the context of their own children's wellbeing, mothers constructed such monitoring as potentially harmful: I'm always on a diet. I'm very, very careful never ever to mention that in front of Imogen. … If I was that way inclined, I could be really, really strict, only giving her certain things, but then I think that could promote her trying to steal it and trying to get things cos she's not allowed it … that could lead to her always being on a diet … or, again, binge eating on things because she doesn't think that she's allowed it. (Shannon, White British) You kind of scrutinise everything you do with them … But then you think to yourself, I can't [talk to daughters about their NCMP BMI results] cos I'm gonna give them a body complex. I'm making it worse because I'm now saying to them that [their weight] is an issue, and it's not. … And then over the years you see stuff happen with them and they get their own body complex issues … and then you think back, "well, did I do that? Did I, did I trigger that?" Because I got these letters and then I tried to stop you from eating that and that, [or] I tried to fatten you sort of, d'ya know what I mean? (Stephanie, White British) These extracts challenge the efficacy of weight-monitoring while drawing on a dominant discourse of mothers' actions as highly influential in daughters' eating behaviours and body insecurities (Vander Ven & Vander Ven, 2003;Woolhouse & Day, 2015). Through this discourse, strictness and leniency with food and eating are positioned at opposite ends of a spectrum of mothers' behaviour, within which mothers must achieve the right balance to avoid causing harm to their children. Therefore, monitoring children becomes a process of ensuring children do not become fat (Stephanie: "I got these [NCMP BMI result] letters and then I tried to stop you from eating that and that") while protecting them from body image issues that might accompany the pathologisation of their bodies (Shannon: "I could be really, really strict … [but] that could lead to her always being on a diet"). This balancing act represents a "calibrated femininity" (Cairns & Johnston, 2015), a process of continual adjustment to meet an idealised standard of the "good mother" who appears knowledgeable and adequately health-conscious but not overly obsessive and controlling.
In addition, while both extracts position mothers' actions as influential in their children's development of obsessive monitoring or "body complexes," Shannon positions mothers' complicated relationships with dieting as an inheritance they may pass onto their daughters. While Shannon acknowledges the heritable nature of body management, saying she is cautious to hide diet-talk from her daughter, the discourse reproduced in this extract individualises blame towards mothers as those who have passed down such concerns at the expense of recognising how fatphobia operates in ways that harm mothers and daughters alike. For instance, Walsh and Rinaldi (2018, p. 98) have highlighted how fatphobia "manifest[s] in and on women's bodies, and [is] enacted through everyday interactions within both maternal relationships and wider social contexts affecting mother and daughter, women old and young".
Interestingly, in mothers' negotiations of these discursive patterns, there appear to be oppositional discourses through which the prevention of "obesity" is constructed as a shared parental responsibility (Mirana: "parents need to know in case they're putting their child at risk"), while the protection of children from the harm of weight stigma is constructed as a private responsibility that mothers must attend to alone (Julia: "We have a big thing in this house that we don't talk about weight"; Pandora: "I don't want to get into conversations with them about weight"). Therefore, weight interventions are considered important insofar as preventing or tackling "obesity," yet they hold considerable potential for harm through the confirmation of children's bodies as pathological. With these conflicting responsibilities and discursive patterns highlighted, the next stage of analysis seeks to situate participants' voices within these patterns and understand how they negotiate these in their spoken accounts.
Phase 2: Voices and I poems Voices: The duty to protect. Having outlined the discursive patterns emerging from participants' talk about the NCMP, the following phase of data analysis involved identifying tones and contrapuntal voices in relation to these discourses (Thompson et al., 2018) to demonstrate how women's experiences are shaped in multiple, often conflicting ways within the realms of available discourses. As demonstrated below, participants' voices were marked by contrasting tones of defiance, confidence, frustration, and guilt. The main narrative pattern to emerge in the participants' talk was "the duty to protect" their children, both from fatness and fatphobia. Below, I Poem extracts will be presented in relation to each discursive pattern to explore how participants negotiated these discourses.
I poems: The physical and psychosocial harm of "obesity". The first discursive pattern represented "obesity" as harmful to both physical health and psychosocial wellbeing; therefore, weight-monitoring is a valuable tool in harm prevention. However, participants did not simply reproduce this discourse unproblematically. The I Poems below show that participants voiced the impact of this discursive pattern in complex ways: In this I Poem, Karen sympathises with some of the programme's methods for engaging children in "health behaviours". However, recalling her own weight-monitoring experience after giving birth to her daughter, she speaks anxiously about how this constant monitoring "consumed" her, expressing frustration that the worry she experienced occupied so much of her maternity leave. Karen narrated this experience as a personal battle, expressing resentment at the impact these measurements had on her. Reflecting on the NCMP as feeling like "a turning on its head" of what she went through with her children's rigid weight-monitoring, Karen conveys sadness and frustration that her son Adam's "overweight" result has made her feel pressured to monitor his weight. Therefore, this I Poem demonstrates that, rather than simply reproducing the discourse of weightmonitoring and protection from harm unproblematically, this discourse has "hit" and "bruised" Karen as she expresses her feeling of being "back to that position" she describes as all-consuming and anxiety-inducing.
Towards the end, Karen speaks with an irritated yet humoured tone, laughing as she asserts that Adam is "perfectly fine" but expressing that his "overweight" result has led her back to these concerns about monitoring him. Despite her resistance to constructions of BMI labels as meaningfully related to health, Karen's anxious and concerned response is nevertheless informed by a pathologising discourse of "obesity". Karen's doubt is arguably related to the dominant discursive position of the state as holding the only rational, "objective truth" about children's health in opposition to the emotional and irrational judgements of the public (Cook et al., 2004). However, Karen does not reproduce this discourse of the state's authority on health uncritically; she parodies it, using humour whilst expressing concern over what it means for her son's health. Therefore, while this I Poem sheds light on the pervasiveness of these dominant discourses of fatness and responsibility, it also shows how Karen negotiates these discourses with conflict, frustration, and even subversion through humour.

I poem: Mirana
I don't like classifying kids as, like, overweight or underweight.
But then, at the same time, I think parents need to know in case they're putting their child at risk.
When I were younger, and just being overweight, I had to go see a dietician and be weighed like every month, And if I'd not lost weight, I could tell that the dietician were, like, telling my mum and dad off.
But then, I were always sat there, so I felt like they were telling me off as well.
I think, now, I'm like, almost anxious about getting into that position with my kids.
I think I've got my own issues with it, So soon as she were like labelled nearly overweight, I were like, "woah", I don't want her to be like I was.
Mirana talks about the complexities of weight-monitoring as a tool in the "fight against obesity". Multiple voices shape her negotiation of this discourse. First, Mirana bluntly asserts that monitoring is essential so that parents are aware of the potential harm caused to their children; therefore, this is a responsibility shared by all parents. However, Mirana feels conflicted; she uncomfortably conveys her concern with the potential consequences of labelling children's bodies as pathological. Later, she laughs as she reflects in a defeated tone on her own unhelpful and painful experience of weight-monitoring.
Mirana speaks of her desire to protect her daughter characterised by two conflicting concerns, both related to ensuring that her daughter achieves a "healthy weight"; first, to protect her from the painful experiences she had as a child marked as "overweight," and second, from the health risks associated with "obesity" in the first place. She anxiously reflects on her own unhelpful and painful experiences of weight-monitoring while wrestling with the potentiality that such monitoring might help to protect her children from this harm. Mirana narrates her duty to protect her children from fatness itself and the social consequences of living in a fat body; she positions the prevention of "obesity" as a responsibility she shares with other parents, while the fear that her daughter will experience weight stigma is an anxiety that she faces alone.
I poems: Negotiating mothers' responsibility. The second discursive pattern identified was related to motherhood and mothers' responsibility for protecting their children from harm, be that in the form of fatness or fatphobia. In the below I Poems, we demonstrate how participants negotiated and positioned themselves within this discourse in their voiced accounts.

I poem: Pandora
It did make me think, maybe, should I be restricting her diet in some way? I don't feel like I really need to up her activity levels in any way, In this I Poem, Pandora talks about her responsibility for her daughter's health and wellbeing, but does not reproduce this discourse in a straightforward manner. Pandora expresses doubt over her actions as a mother, with her daughter's "overweight" result leading her to wonder whether she should restrict her diet or "up her activity levels". Pandora grapples with a discourse of mothers as the guards over their children's bodies, speaking in a pensive and thoughtful tone in considering whether her actions have influenced her daughter's weight. She engages with this discourse with reluctance and doubt. This I Poem, then, demonstrates a reluctance to engage in her daughter's weight-monitoring but, at the same time, a questioning of this reluctance; Pandora's resistance is not straightforward, and is instead peppered with self-doubt. Later in this extract, Pandora talks about her desire not to discuss weight and dieting with her daughter for fear of it becoming an issue that might then cause harm. Pandora here negotiates her two responsibilities to protect, first, her daughter's body image, and second, her physical health. Pandora negotiates these responsibilities in ways that position them as being in tension and therefore potentially incompatible, therefore "hitting" Pandora in a way that causes discomfort, doubt, and indecision. Pandora goes on to express confidence in her daughter's general health, more assertively resisting the dominant discourse of her responsibility as a mother to intervene and "hassle" her daughter about her diet. Pandora's reproduction of this discourse is therefore shaped by multiple contrasting voices that speak to the complexity of her experiences of resistance against her daughter's weight-monitoring. You're the parent, You're the mum and you're with them, And you think you've harmed them.
You think you've done them some damage.
Over the years you see stuff happen with them and they get their own body complex issues, And then you think back, well, did I do that?
Did I trigger that?
You think you've put them in harm's way.
You're there to protect them, And you feel like you've done them wrong.
Even though you know that what you've done is right, it's just the way they are, You're still, you're devastated.
You just feel like you've failed them.
Although Stephanie often speaks in the second person, she is arguably speaking of her own experience. She speaks of mothers' duty to protect their children from harm, with a sense of obligation demonstrated directly through notions of "the mum" who is "there to protect them". However, Stephanie speaks in a tired and defeated voice. In relation to the discourse, she speaks of her "failure" to protect her children with assertions that even though she has done her best as a mother, her children's bodies are "just the way they are". There is a sense of impossibility here; Stephanie seems to feel as though protecting her children from harm is impossible.
This I Poem extract demonstrates how the discourse of mothers' responsibility has hurt Stephanie as a mother and caused conflict and feelings of failure, despite her statement that "what [she has] done is right". Stephanie speaks of the overwhelming devastation, grief, fear, and shame she has experienced within this discursive realm.

Conclusion
This study applied FRDA (Thompson et al., 2018) to understand how discourses shape voiced accounts and how participants actively locate and relocate themselves within discursive realms. We identified two discursive patterns in participants' talk, related to 1) the harms of "obesity" and 2) mothers' responsibilities to protect their children from harm. While participants expressed doubt over the meaning associated with BMI results, they often conceptualised these as objectively true; "overweight" or "obese" were constructed as causes for concern over children's health and their success as mothers, judged by how well they protected their children from harm. This arguably demonstrates how such "regimes of truth" (Foucault, 1975(Foucault, /1995 cannot be resisted straightforwardly and without consequence on their position as mothers. These "expert knowledges" are crucial to governmentality. Resisting these knowledges results in accusations of being an "obesity denier" (e.g., Katz, 2015), serving to encourage mothers' compliance with constructed norms of health knowing that failure to comply leaves them ostracised as bad mothers and irresponsible citizens.
The most striking narrative pattern present throughout participants' accounts was "the duty to protect" children from both fatness and fatphobia, constructed in opposing ways. Participants actively located and relocated themselves within the realms of available discourses, constructing themselves at times as both "good" and "bad" mothers. These contrasting positions correspond to the contradictions in dominant discourses of mothers' roles in "obesity" and "disordered eating". Firstly, "obesity" is positioned as a health epidemic that all mothers have a shared moral duty to prevent (Boero, 2009;Gillborn et al., 2020). On the other hand, mothers are aware that such damaging constructions of fatness may result in children experiencing issues with body image and "disordered eating" (Catling & Malson, 2012); yet this fatphobia was regarded as an inevitable cog in the system that fat people should expect to experience due to their non-conformity with normalised constructions of health. Therefore, mothers constructed their responsibility to deal with the fatphobia their children may experience as a personal battle to be dealt with alone. To this end, a combination of these discourses holds that mothers' efforts to protect their children from one harm may expose them to another, and mothers voiced guilt and anxiety in their narratives of the fatphobia they wished, but sometimes felt unable, to protect their children from. Many participants drew upon dominant discourses of "obesity" and blame, their concerns caught up in the criteria of health and happiness shaped by the "expert knowledges" informing health policy (Rose, 1999). These discourses "hit and bruised" them profoundly.
Successfully protecting children from both of these harmsfatness and fatphobiawas positioned as unachievable and mothers told of the pain inflicted by trying to manage the weight of these expectations. Mothers were aware that children might regard sudden lifestyle changes, in-line with the NCMP's advice, as confirmation that their bodies are problematic; thereby, efforts to protect children from "obesity" contradict the duty to protect children from fatphobia. Conversely, encouraging children to appreciate their bodies, which have been constructed as pathological through government policy, left mothers in fear for their children's present and future health. While the reproduction of these discourses in government policy serves to encourage compliance with constructed health norms, negotiating these opposing responsibilities leaves mothers vulnerable to being constructed as forever failing. Through analysis of participants' voices and experiences, it is evident that several participants experienced this harm.
The guilt experienced by mothers in response to these discourses functions as a strategy of governmentality, reinforcing self-governance and compliance with normalised body ideals and constructions of "good motherhood" (Parker & Pausé, 2019) by keeping mothers in a state of fear about their children's health. These feelings of guilt, fear, and failure, and the interventions that foster them, are not conducive to improving children's physical health but do hold real implications for the wellbeing of mothers and children and their relationships with each other. The maintenance and reproduction of these pathologising discourses is likely to continue to harm children's body image (Catling & Malson, 2012;Evans & Colls, 2009), while at the same times blaming mothers for these harms and responsibilising them for prevention and amelioration.
Though the current study analysed discourse and voice concerning a UK government policy, policies and interventions built on these assumptions about fatness have been rolled out in schools elsewhere (e.g., Vander Schee & Boyle, 2010) and the discourses identified here are drawn upon globally (e.g., Abou-Rizk & Rail, 2015). To this end, the harm of these discourses stretches beyond the UK and the programme investigated in this study. Our participants' narratives add to the body of prior research indicating that weight-monitoring and the discourses legitimised through such policies have the potential to harm children's self-image and wellbeing (Catling & Malson, 2012;Evans & Colls, 2009). Furthermore, the current research calls attention to how these discourses hurt mothers.
It is essential to recognise that "obesity" interventions construct a "healthy" BMI as achievable if individuals successfully follow state health guidance (Gillborn et al., 2020). Although this expectation does not reflect the reality of the complex relationships between bodies and health (Gard & Wright, 2005;Guthman, 2009), women are discursively positioned as bad mothers (Halse, 2009) when their children fail to meet this expectation. In addition, mothers are harmed by sexist and individualising systems that foster these discourses; as Walsh and Rinaldi (2018, p. 97) argue, "fat shaming can be inscribed upon the mother's body in ways that become a daughter's inheritance". Discourses responsibilising mothers for the reproduction of these harmful ideas facilitate further governance and control under the guise of protecting the nation's health, while state-sanctioned programmes that legitimise these harmful discourses go unchallenged.
Though mothers in this study were critical of weight-monitoring and aware of the potential harms it poses to children's wellbeing, it is not possible for mothers to singlehandedly break down and resist the state-sanctioned discourses and powerful expert knowledges that legitimise fatphobia. If the health and wellbeing of children is a public health priority, policy and programmes such as the NCMP that pathologise bodies and responsibilise mothers for the harm of fatphobia should be scrapped. Indeed, fatphobia is a systemic and structural harm; challenging this requires wider interrogation of the knowledges privileged in our understandings and the motivations, wider impacts, and desired outcomes of government interventions.