Navigating migrant infrastructure and gendered infrastructural violence: reflections from Brazilian women in London

Abstract This paper explores some of the institutional and theoretical silences within debates on infrastructural violence with reference to migrant women survivors of gendered violence. Drawing from feminist thinking around structural and symbolic oppression, it develops the notion of gendered infrastructural violence to help understand how migrant women survivors navigate statutory and non-statutory institutions when seeking support. Empirically, the paper elucidates how diverse Brazilian migrant women in London negotiate multiple forms of passive and active infrastructural violence played out in terms of xenophobia, discrimination and a hostile immigration environment. Such experiences can dissuade them from reporting due to actual and perceived fear of further violence being perpetrated against them. While infrastructural violence perpetrated by an oppressive racial state can exacerbate Brazilian migrant women’s suffering of direct gendered abuse, migrant and/or feminist organisations provide invaluable support and an essential protective bulwark. Yet these experiences are mediated differently depending on women’s social locations in terms of intersecting race, class, occupational and immigration status and language competencies.


Introduction
Recent thinking around infrastructure in cities has acknowledged how lack of access to services has deleterious consequences for marginalised groups where their exclusion constitutes a form of 'infrastructural violence' (Rodgers and O'Neill 2012). While the gendered dimensions of infrastructure deficiencies have been recognised, this is rarely viewed as infrastructural violence (although see Chaplin and Kalita 2017;Datta and Ahmed 2020). Furthermore, despite significant diversification of what constitutes infrastructure, service provision by state and non-governmental organisations that support migrants has seldom been conceived as such, and not from a feminist perspective (Graham and McFarlane 2015;cf Xiang and Lindquist 2014).
This paper argues that infrastructure provides a fruitful conceptual lens to examine the intersections between migrant women's experiences of gendered violence and wider structural violence as they navigate support services. With an empirical focus on Brazilian migrant women survivors of gendered violence in London, themselves a diverse group, the paper explores how exclusions from migrant infrastructure constitute 'gendered infrastructural violence' . It feeds into burgeoning debates on how structural abuses of power can lead to 'slow violence' or 'slow death' (Nixon 2011;Pain and Cahill 2020;Perry 2013;Smith 2016). This is especially potent in relation to the discriminatory actions of a racial state, as Goldberg (2002) notes; acting with impunity and/or violence against certain citizens, especially women of colour or from migrant backgrounds (Dominguez and Menjívar 2014).
The paper contributes to feminist geographical debates on the need to move narratives around gendered violence beyond the individual towards structural and intersectional oppressions (Pain 2014a(Pain , 2014bPain and Staeheli 2014), especially in relation to migrant women navigating support (Erez, Adelman, and Gregory 2009;Sokoloff 2008). These women confront multiple forms of passive and active infrastructural violence played out in terms of xenophobia and discrimination when they report gender-based violence. This can strengthen the alienation felt by women within the hostile immigration environment potentially 'incarcerating' them (Cassidy 2019); encouraging them to leave the country altogether or dissuading others from migrating. While infrastructural violence perpetrated by the state can exacerbate migrant women's suffering of direct gendered violence, migrant and/ or feminist organisations provide invaluable support and an essential protective bulwark.
The empirical base for the paper draws on research conducted with more than 200 Brazilian migrant women from diverse backgrounds and with service providers in London in collaboration with a migrant women's rights organisation with whom the authors have long-standing partnerships. The discussion begins with an outline of the conceptual debates around gendered infrastructural violence, followed by consideration of migrant infrastructure, and the nature of gendered violence against migrant women. This is followed by a contextualisation of the Brazilian community and gendered violence in the UK, and an outline of methods. The paper then examines gendered violence against Brazilian women in London. We analyse how they fear reporting and how they navigate barriers, especially when they are non-white and non-citizens in relation to gendered infrastructural violence. It shows how violence is experienced differentially by Brazilian migrant women, depending on their intersectional social locations especially their race, immigration status and occupational status. The final section examines how Brazilian women access support from migrant infrastructure institutions and the violence they encounter in the process.

Conceptualising gendered infrastructural violence
The recent 'infrastructural turn' within urban theorising has explored how power relations are embedded, mediated and enacted through infrastructure (Larkin 2013;McFarlane and Silver 2017). Further links have been made with complex forms of endemic urban violence and emerging notions of 'infrastructural violence' (Rodgers and O'Neill 2012). Yet, only recently have these debates acknowledged the gendered dimensions of urban and infrastructural violence mainly in relation to poor communities in the global South (Datta and Ahmed 2020). We are interested in conceptualisations of infrastructural violence among overseas women migrants who originate from such cities but who reside in the global North situated within complex circuits of power, structural violence and global inequalities.
Interesting conceptual lessons for understanding migrant women's exclusion from support organisations can be learnt from research on infrastructural violence in the global South. Rodgers and O'Neill (2012, 401) define such violence as the 'processes of marginalization, abjection and disconnection [that] often become operational and sustainable in contemporary cities through infrastructure' . Infrastructural violence is rooted in ideas pioneered by Galtung (1969) around structural violence and incorporating a range of deprivations, including material, psychological and symbolic discrimination across space and time (Farmer 2004). While this work has recognised infrastructure as inherently relational, becoming 'real' as it links to organised practices (Larkin 2013;Star 1999), these debates have underplayed how women are disproportionately excluded from urban infrastructure in ways that create further gendered violence (although see Chaplin and Kalita 2017;Datta and Ahmed 2020). Much feminist research, especially on low-income urban neighbourhoods in the South, has shown how poor-quality infrastructure can generate direct forms of gendered violence against women which undermines their right to the city in terms of freedom and livelihood security (Beebeejaun 2017;Chant and McIlwaine 2016). Yet, there remains considerable scope to explore these among migrant survivors of gendered violence in other contexts.
A crucial distinction relevant to migrant women's experiences relates to passive and active forms of infrastructural violence. The former refers to harm caused by 'limitations and omissions' rather than direct consequences, while the latter purposefully develops infrastructure to control social norms and actively marginalise certain groups (Rodgers and O'Neill 2012, 406-407). The state is often the primary perpetrator of passive infrastructural violence in its failure to provide adequate facilities, which can disproportionately affect women (Chant and McIlwaine 2016;Datta 2016). In relation to India (Chaplin and Kalita 2017, 45) note that: 'This passive infrastructural violence, created by the failure of the state to provide adequate sanitation facilities, or effectively maintain those that do exist, exposes women and girls to gendered, caste-and class-based forms of both physical and emotional violence' (also Chaplin 2011). Yet it can also perpetrate active infrastructural violence through slum demolitions or violent militias acting on behalf of the state to restrict access to basic services to marginalised groups, especially women (Sawas et al. 2020).
The ways in which infrastructural violence allows impunity to flourish are especially relevant for migrant women survivors and relate to Nixon's (2011) concept of 'slow violence' referring to the gradual violence of environmental destruction. Conceptualisations of slow violence have been increasingly common among those exploring intersectional oppressions experienced by women of colour with relevance for research with migrant women. Indeed, migrants and women of colour living in poor neighbourhoods face routinized, everyday gendered infrastructural violence of discrimination and fear (Dominguez and Menjívar 2014). Exploitation of migrants has been conceived as a form of 'soft violence' perpetrated by employers against migrant domestic workers, or as 'slow violence' against asylum seekers, as well as 'legal violence' migrants with insecure immigration status (Menjívar and Abrego 2012;Parreñas, Kantachote, and Silvey 2021;Mayblin, Wake, and Kazemi 2020;also Cassidy 2019).
Such slow violence is often at the hands of the state which can act in racialised ways to exclude marginalised groups (Goldberg 2002), especially through police violence (Smith 2018;also Ferreira da Silva 2009;Segato 2016). This has been shown to be especially marked in the case of gendered and anti-Black state violence in the Americas where Smith (2016) develops the notion of 'sequelae' to refer to physical and emotional trauma affecting Black women resulting from state violence. Focusing specifically on the case of Brazil which is directly relevant to the empirical focus in the current paper, Smith notes that such trauma is less visible than the killing of young Black men at the hands of the police, but it generates a 'slow death' of pain, trauma, and depression especially among mothers in an 'anti-black genocide' rooted in racism and sexism (see also Caldwell 2007;Perry 2013). For Black women in Brazil, such 'invisible violence' also limits their rights to positive representations and privilege (Carneiro 2003, 122). In the current context, 'sequalae' and 'slow death' can be related to racialised and gendered infrastructural violence against Brazilian migrant women survivors of gender-based violence as they navigate state support systems (see below). These processes also feed into wider feminist geographical debates around the need to move beyond individual explanations of violence and discrimination against migrants to recognise the interconnections among different types of structural and direct gendered violences (Pain and Staeheli 2014;also Brickell 2020;Brickell and Maddrell 2016;Fluri and Piedalue 2017). This is especially relevant for understanding the experiences of migrant survivors of gendered violence.

Conceptualising migrant infrastructure
Gendered infrastructural violence can be productively explored in relation to how migrant women survivors of gendered violence negotiate support from statutory and non-statutory services that we suggest constitutes a form of 'migrant infrastructure' . Xiang and Lindquist (2014, S124) define this according to five interrelated dimensions: recruitment intermediaries, state apparatus and procedures for documentation, licensing, and training, communication and transport, NGOs and international organizations, and migrant networks. These activities and actors govern and facilitate migration. We have adapted the term to focus on 'statutory migrant infrastructure' comprising the police and judicial services, state social and health services, and 'non-statutory migrant infrastructure' referring to civil society organisations that support migrants and women. We also draw on the idea of 'women's equality infrastructure' (Brodie 2008, 160), denoting organisations and initiatives providing specialist support for women. However, recognising the complex interactions between state and civil society, we focus on non-governmental organisations emphasising how they support migrant women's needs when the state excludes or fails them (Turcatti 2021;Vacchelli, Kathrecha and Gyte 2015).
Research in a wide range of contexts has shown how state services for women survivors of gendered violence often fail them (Brickell 2015(Brickell , 2020. The barriers faced by migrant women when seeking help have been well-documented, especially in the US (Menjívar and Salcido 2002;Reina, Lohman, and Maldonado 2014). These include isolation, lack of information, insecure immigration status, language competencies, exclusion from criminal justice systems, fear and stigma (O'Neal and Beckman 2017;Vidales 2010). Despite a tendency to focus on barriers as individual level factors, more recently, they have been conceived as structural, reflecting multiple 'everyday bordering' , which are situated and intersectional processes (Cassidy 2019;Yuval-Davis, Wemyss, and Cassidy 2019). We suggest that migrant women's interactions with statutory migrant infrastructure can generate gendered infrastructural violence and create further direct gendered violence experienced in intersectional ways. However, non-statutory migrant infrastructure also provides women with much-needed support through providing avenues for safe reporting and access to justice (Gottardo and Cyment 2019).

Contextualising gendered violence against migrant women
The definition of gendered violence against women used here draws on the classic 1993 UN Declaration of the Elimination of Violence against Women defined as: 'Any act of gender-based violence that results in, or is likely to result in, physical, sexual or psychological harm or suffering to women, including threats of such acts, coercion or arbitrary deprivations of liberty, whether occurring in public or in private life' (UN 1993). Yet such definitions have been negotiated over time. Gender-based violence can include violence against men, sexual minorities or those with gender-nonconforming identities. Yet this term arguably de-politicises violence against women and girls who experience it disproportionately at the hands of men. It can also imply a binary, heterosexist interpretation at the individual level devoid of reference to structural and intersectional gendered inequalities (Hughes, Marrs, and Sweetman 2016). Gendered violence therefore needs to be viewed as a continuum where multiple types of physical, sexual and psychological violence against women and girls intersect and which are experienced intersectionally (Kelly 1998; see also McIlwaine and Evans 2018).
In terms of migrant women's experiences of gendered violence, it has been argued that they are more vulnerable than non-migrants. However, despite a tendency to 'hyper-fixate' on how such violence is more widespread within migrant communities (Fluri and Piedalue 2017, 541), the reality is much less clear-cut, hampered by lack of reliable data and prejudice (Menjívar and Salcido 2002). At the European level, a survey with 42,000 women across 28 Member States, showed that non-citizens in their current country of residence had higher rates of physical and/or sexual violence since the age of 15 by partners and ex-partners than those who had always lived in their country (27 percent compared with 22 percent) but lower rates of stalking (16 percent compared with 18 percent) (EU Agency for Fundamental Rights 2014, 189-190). These estimates also hide how gendered violence can be higher among certain types of migrants, such as refugees, migrant domestic workers, forced marriage migrants and those who have been trafficked. There are also dangers in over-stating levels of violence and culturally essentialising minoritized groups, especially in relation to so-called 'honour-based violence' (Montoya and Agustín 2013) (gendered violence where the main justification 'is the protection of a value system predicated on norms and traditions concerned with 'honour'' (Gill and Brah 2014, 73). Such approaches tend to pathologize migrant women and underplay the racism and xenophobia they face (Sokoloff 2008). There are also dangers in ignoring historical trajectories of colonialism together with global migration dynamics which are underpinned by racialised state power (Grosfoguel, Oso, and Christou 2015;Segato 2016) especially when migrants are Black women (Malheiros and Padilla 2015).

Contextualising Brazilian women's experiences of gendered violence in London
The focus of this study on gendered violence among Brazilians in London was driven by long-standing research interests by the authors as well as their activism with the wider Latin American community in the city. Following two prior research projects with the partner organisation, the Latin American Women's Rights Service (LAWRS) (see below), the need for more empirical data on gender-based violence among Latin Americans was identified as urgent (McIwaine 2010). The specific focus on Brazilians related to their position as the largest of the Latin American nationalities in the city, and because this was also part of a wider project with Brazilian partners in Rio de Janeiro (Krenzinger et al. 2018). In the context of a Latin American population of approximately 250,000 in the UK, with 145,000 in London, the most recent census of 2011 shows that there were 52,000 Brazilians in the UK with marked concentrations in London (60 percent) (McIlwaine and Bunge 2016, 8). Other research with 2,188 Brazilians in the UK in 2020 showed that they are youthful and well-educated with most migrating in search of economic and educational opportunities. However, many end up in elementary jobs such as cleaning mainly due to limited English language skills preventing them from working in jobs they may have previously held back home. The majority (63 percent) had a European passport, with 14 percent having no documents (Evans 2020, 15).
In terms of racial identities, three-quarters self-identified as 'white' (76 percent), and 19 percent as 'mixed' (Evans 2020: 9); this has also been noted ethnographically (Martins Junior 2020). The ethnic-racial classification in Brazil follows a system of self-declaration by 'colour or race' established by the Institute of Geography and Statistics (IBGE) (Petruccelli and Saboia 2013). IBGE's national census uses four traditional categories: white (branca), brown (parda), black (preta), indigenous (indígena) and yellow (amarela). This categorisation has been deeply controversial. Yet it remains the most commonly used set of racial/ethnic parameters used in research in Brazil and therefore why we followed it here. For terminological clarification, the use of the term 'black women' (mulheres negras) typically refers to the combination of IBGE's black and brown categories -a common fusion in Brazil especially used by Black movements incorporating pardos within the Afro-Brazilian population. Moreover, when further combined with the 'indigenous' and 'yellow' categories, the term 'non-white' tends to be used although it continues to problematically claim a central role for white groups as the norm (see Krenzinger et al. 2021). Although the research followed IBGE's classification, in this paper we adopt the term mixed race as opposed to 'brown' , considering the key critiques of 'colourism' and the liminal imperatives of pardos/brown groups within Brazil's degrading history of whitening (branqueamento) (Munanga 1999;Schwarcz 1993).
The challenges of racial identification among Brazilian migrants in London should thus not be under-estimated. While those who identify as white find it easier to migrate, once in the UK they become embroiled in constellations of representations of race, class, and gender as migrant 'others' located beyond the national imaginary (Martins Junior 2020). These intersectional identities, themselves rooted in colonial injustices, play out in London to create a diverse but fragmented community. What Martins Junior (2020) refers to as 'lighter-skinned' Brazilians from professional backgrounds often claim to avoid socialising with other Brazilians or visiting 'Brazilian places' as a form of distancing within a wider racist environment. Women were especially keen to distance themselves, reflecting pervasive processes of racialised exoticisation and hyper-sexualisation of Brazilian women in the UK and elsewhere (Martins Junior 2020; also Ribeiro 2016).
Turning to direct forms of gender-based violence in the UK in general, rates are high for all women with 44 percent experiencing physical and/or sexual violence by a partner or non-partner since the age of 15, which is higher than the EU average of 33 percent (EU Agency for Fundamental Rights 2014, 29). Women from migrant or minoritized backgrounds do not necessarily experience more gendered violence. For example, in 2020 those born in the UK experienced higher levels of domestic abuse than those not born in the UK (ONS 2021). However, minoritized women are more likely to access domestic violence support services than the general population (Bowstead 2015). Despite limited research, a recent study among Brazilians reported that 28 percent suffered violence, of which 65 percent were women (Evans 2020: , 33; see also McIlwaine and Evans 2020).
This must also be understood within a context of a 'hostile immigration environment' in the UK, whereby migrants have been increasingly subjected to punitive migration regimes that have systematically eroded migrant's rights. While this has been at the core of much of the British state's immigration policies for decades, since 2010 government discourses have aimed at actively discouraging overstayers and limiting rights to settle. The Immigration Acts of 2014 and 2016 have ensured ever more draconian enforcement of immigration controls by hospitals, banks, public sector organisations, private landlords and others (Yuval-Davis, Wemyss, and Cassidy 2019). Such policies have undermined migrants' rights to work and to access education, health care and housing as well as state social support (Hodkinson et al. 2021). This has been crucially important in influencing whether migrant women survivors report and seek help. A key dimension of this has been the No Recourse to Public Funds (NRPF) restriction attached to many temporary visas, including spousal visas, making women unable to apply for state benefits or assistance. Despite concessions when women are destitute and/or there are safeguarding issues around children, most women find it difficult to access support (Cassidy 2019). Although the NRPF restriction has been in place since the 1999 Immigration and Asylum Act, the hostile immigration environment has made it more difficult to access the concessions.

Methods
The paper draws on research conducted in 2016 and 2017 involving an online survey, interviews and focus groups with Brazilian women migrants and service providers in London. The small-scale survey was conducted in Portuguese with 175 Brazilian women and based on purposive sampling with the link distributed among various migrant organisations, Facebook groups and other networks of Brazilians in London. The survey was adapted from the EU Agency for Fundamental Rights (2014) questionnaire on gender-based violence. It addressed issues, such as whether gender-based violence had been experienced and if so, the types, frequency and spaces where it occurred as well as eliciting views on the perpetrators (in relation to Brazil and London). While there are constraints in conducting an online survey in terms of reaching the more educated with access to a computer, it had benefits of ensuring anonymity and confidentiality thus potentially making it easier to disclose sensitive information. Among those surveyed, most were young (74 percent under 50), well-educated (72 percent with university education), employed in professional and managerial jobs (53 percent), and self-identified as racially white (73 percent), with 21 percent identifying as mixed, four percent as Asian, two percent as Black and one percent as Indigenous. Almost half originated in São Paulo (42 percent), with 10 percent from Rio de Janeiro. The majority (70 percent) were married/in relationships, with just over half having children. Most women had entered on temporary visas (65 percent) although by the time of survey, 80 percent had regularised their status (mainly through obtaining EU passports leading them to become European rather than Brazilian migrants). While three percent stated that they were currently undocumented an additional 17 percent were unable to make any claims on the state for support because of their visa restrictions (see below). In addition, two-thirds had experienced insecure status at some point in their migration trajectory. Although the exact length of time it took for regularisation was not captured in the survey, it is a time-consuming and complex process which is currently based on residency for 10 continuous years.
The survey was complemented by in-depth interviews with 25 women among whom 20 were survivors of violence and who had accessed support from our partner organisation, LAWRS. This is a migrant feminist organisation with whom both authors have a long history of engagement through other research projects, trusteeships and advisory roles. Six focus group discussions were also conducted -five groups with women and one with men -with a total of 16 people, all conducted in Portuguese. Interviews were also undertaken with representatives from 12 service provider organisations supporting migrant women in London. These included four women's organisations, two generic Latin American migrant organisations, two Latin American women's organisations and two health organisations. Most of the interviews with women survivors and all the women focus groups were conducted at the migrant organisation with a trained counsellor on hand in case the interview process was distressing. They were all conducted in Portuguese and facilitated by a specifically trained interviewer approved by the organisation and by the authors. The primary criterion for interviewing at LAWRS was being Brazilian and having accessed support. This support (legal and welfare advice and psychological counselling) is provided free of charge and so resulted in those most excluded financially and in terms of English language competency and immigration status being more likely to be interviewed. For 20 of those interviewed, being a survivor of gender-based was also a pre-requisite. The interview sample of women were therefore more likely to have lower levels of education, work in elementary jobs, such as cleaning and have experience of insecure immigration status, with 15 identifying as white, seven as Black, and three as mixed, again following the IBGE classification which, as noted above, is highly contested.

Gendered violence against Brazilian women in London
In acknowledging that gendered violence is common among all women in London, this section briefly outlines how it is experienced differently by Brazilian migrant women depending on their intersectional social locations especially their race, immigration status and occupational status. Among those included in the survey, gender-based violence was pervasive in Brazil and in London with 82 percent experiencing at least one type over their lifetime. Among those interviewed where 22 out of 25 had experienced gendered violence, several women spoke of fleeing domestic violence and especially abusive partners to move to London. Juliana, 36, from Paraná who identified as white and who worked as a cleaner, migrated to London in 2007 to escape from her abusive husband. Her plan was to work and save to apply for Italian citizenship and then divorce him, all of which she accomplished.
Focusing on women's experiences of violence in London from the survey, psychological violence emerged as most prevalent (48 percent), followed by physical (38 percent) and sexual (14 percent) with almost half having experienced more than one type. The women interviewed also discussed experiencing a huge diversity of types of violence over their lifetime. For instance, Camila, 31 years old, who was mixed race and from Bahia, spoke of multiple forms of physical and emotional abuse by former partners and family members back home, only to be confronted by frequent sexual harassment by various clients in her job as a computer technician in London and by an immigration officer at Heathrow airport.
In terms of how racial identities affected women's experiences of violence, and recalling that the majority of the survey sample identified as racially white (73 percent), with 21 percent mixed, four percent Asian, two percent Black and one percent Indigenous, it emerged that mixed heritage women were more likely to suffer (66 percent) than white women (58 percent). In addition, two of the three Black women surveyed experienced violence. Among those interviewed, all those identifying as mixed or Black (10 of the 25) suffered gendered violence in London compared with 12 of 15 white women. According to the survey, experiences of gendered violence emerged as higher among those with insecure and temporary immigration status (60 percent) compared to those with permanent status (mainly an EU passport) (54 percent). Occupational status was also important with higher levels among those working in cleaning and caring jobs in London (61 percent) compared to those in professional and managerial activities (53 percent). This also relates to the higher incidence of gendered violence in the public sphere over the private where violence took numerous forms from sexual assaults to verbal abuse. Gabriela, 41, from Belo Horizonte and who was mixed race had been working in several cleaning jobs since she arrived in London in 2012. She spoke of how one of her clients began to sexually harass her in his home: After you get to know them, they think they've earned your trust … after a while I started to notice that when I arrived and said hello, he started to hug me, putting his hands round my waist … I make light of it, but the reality is that it shouldn't happen.
While all Brazilian migrants tend to be 'othered' in some way (Martins Junior 2020), experiences of gendered violence therefore varied according to social locations and identities. As a participant in a focus group who discussed her experience in a shop in London stated: 'the saleswoman kept looking at me. I am not sure if it is this context, but because we are Black, we suffer more … there is much more discrimination [against Black women] than against white Brazilian women.' This also played out in intimate relationships in London when couples are from different racial and nationality backgrounds. Those interviewed who were in abusive relationships with European partners, reported especially 'serious' forms of physical, psychological and sexual abuse (also Lopes-Heimer, forthcoming) (where 'serious' is defined as 'an incident that had the biggest impact on you, either physically or psychologically' -EU Agency for Fundamental Rights 2014, 14). Bianca, 71, from Minas Gerais, who identified as Black and ran a small business selling Brazilian snacks spoke of how she spent 16 years with her Portuguese ex-husband who was so physically violent that she has required medical attention. He regularly called her racist names and complained that she came from Brazil rather than Portugal. Gendered and intersectional power relations are rooted in structural and symbolic inequalities making it difficult for migrants to secure qualified jobs for in London partly due to language competencies, but crucially in relation to racialised and class discrimination. These are inscribed in the harms committed through the bodies of migrant women, and especially Black women, with these traumas reflecting wider geopolitical and colonial relations (Grosfoguel, Oso, and Christou 2015;Smith 2018).

Gendered infrastructural violence and migrant infrastructure among Brazilians in London
This section will explore how experiences of direct gendered violence against Brazilian women in London interrelate with indirect forms of infrastructural violence when they disclose, report or seek help from migrant infrastructure state institutions or civil society organisations (feminist and/or migrant supporting). Survivors of intimate partner, workplace and other forms of direct gendered violence are faced with further slow violence when they try to seek help. Actual and perceived, active and passive gendered infrastructural gendered and racialised violence serves to dissuade and deter women from seeking help as difficult as possible for them (Cassidy 2019; see also Segato 2016) with the hope that they will return home.

Disclosure, reporting and gendered infrastructural violence
Disclosure and reporting incidents of direct gendered violence in London were generally low. According to the survey, less than half (44 percent) disclosed an episode of gendered violence to friends or family or reported an incident to formal sources. Only a quarter reported to the police, with even fewer contacting a solicitor (seven percent), or specialist organisation (4 percent). Informal disclosure was more common than formal reporting suggesting that infrastructural violence played a part in dissuading women. These patterns must also be understood in relation to racial identities in that white women were more likely to report (63 percent) compared with mixed or Black (37 percent). In turn, more than half who had permanent immigration status reported, compared with less than half of those with insecure status (46 percent).
The reasons why women were reluctant to report suggests high levels of passive infrastructural violence that needs to be understood intersectionally. More than a quarter of women surveyed (29 percent) did not report because they thought that nothing would be done, with 15 percent not knowing how to report. Another 10 percent said they did not want to expose themselves to community opprobrium, while nine percent felt ashamed. Carolina, 53 from Paraná who identified as white and worked as a cleaner and selling Brazilian snacks noted: 'The shame of it! I felt so bad, so humiliated! … here I thought, "I'm nothing, I'm no one. They won't help me at all"' . Among those interviewed, Black and mixed women reported similar processes of feeling shame and community opprobrium (Rahmanipour, Kumar, and Simon-Kumar 2019). However, they were also more likely to mention fear than white women. Gabriela (see above) noted: 'I didn't have the courage, I was afraid that he would be arrested and then how would I explain that to my daughter?… I didn't have the courage to call the police.' Related to this, women only report as a last resort when they feel they can no longer cope with the abuse (Vidales 2010), as noted by service provider F: 'when women suffer violence, by the time that they talk about it … it is because it has already exceeded all the limits.' Women's relationships with intimate partners perpetrators were also crucial in their reporting. Indeed, eight percent of women surveyed stated that having feelings for the perpetrators was a reason not to report, while five percent were afraid of them or financially dependent. While on one hand, these reflected individual circumstances, on the other, they were linked with wider structural disadvantages of living in marginalised and isolated situations as migrants as Camila (see above) recalled: I was afraid of him turning against me, I was here on my own and I didn't want to go back to Brazil. I was ashamed to go back home and admit that nothing had worked out … So, I started to live with it … and I didn't report him so I wouldn't lose the opportunity of being here.
Migrant women's precarious situations of structural violence and exploitation provided fertile ground for abusive partners to behave with impunity as Márcia, 42, from Espírito Santo who was Black and worked as a health care assistant stated: Here you're on your own … he took advantage of that because he knew I had nowhere to go … I think abusive men take advantage of the fact that we're far away from family and friends.
This was closely related to immigration status. While most women in the survey and interviews had regular status, many had also experienced insecure status at the beginning of their migration trajectory (see McIlwaine, Granada, and Valenzuela-Oblitas 2019). It is also important to emphasise that most women with regular temporary visas had an NRPF restriction meaning women could not access public services. Manipulation of immigration status as a tool of control by perpetrators emerged as a core dimension of passive infrastructural violence dissuading women living in vulnerable situations from reporting and seeking help (Menjívar and Salcido 2002). This was most marked if women were undocumented as noted by service provider B: 'if the woman is in this country illegally [sic], and the husband is legal, he will do whatever he wants with her, because she is at his mercy.' This manipulation was also used by employers as in Sabrina's case. Sabrina who was, 45, from Fortaleza and identified as Black worked as a nanny, cleaner and courier for a family in London who confiscated her passport. When her boss began to sexually harass her, she felt powerless to report as she feared deportation: I'm sure he … did this knowing that I wouldn't go to the police, because I was an illegal [sic], as my visa had already expired. And he was right; I didn't have the courage to go to the police. This was further bolstered by Sabrina feeling she could not communicate well in English (also Reina, Lohman, and Maldonado 2014). These forms of passive infrastructural violence, which are experienced differently by Brazilian women also affect their engagements with service providers and judicial services.

Statutory migrant infrastructure and gendered infrastructural violence: criminal justice and health and social services
Somewhat paradoxically, given high levels of fear among migrant women to approach the police, they were the most commonly consulted form of migrant infrastructure sought by survivors (41 percent of women surveyed). Increasingly it has been acknowledged that migrant women report violence to the police, but that they often have negative experiences linked with institutional racism (McIlwaine, Granada, and Valenzuela-Oblitas 2019). This must also be situated within wider processes of racialised direct and indirect state and police violence against those identifying as Black migrants with insecure status (Perry 2013 on 'slow death'; Smith 2016 on 'sequelae' in Brazil; Mayblin, Wake, and Kazemi 2020 on 'slow violence' in the UK). Several service providers noted that women with insecure status were more reluctant to approach the police and often waited much longer and experienced more violence before contacting them. Service Provider C estimated that undocumented Latin American women are assaulted 60 times before their first call which compared to 35 assaults among women in general (see McIlwaine, Granada, and Valenzuela-Oblitas 2019, 19).
Language played an important role. For example, Maria, 32, from Rio de Janeiro, who was white, revealed in a focus group how while pregnant with twins, she was attacked with a knife by her former (Brazilian) partner who called the police. On arrival, the police took Maria into custody because her partner blamed her for instigating the incident as he could speak English and she could not. She spent three hours clarifying the situation with the help of a translator: 'I thought the world had ended, you know? Why was I being arrested, why was I inside a cell if I hadn't done anything wrong?' Maria was one of many Brazilian women who stated that the police had widespread discretionary powers (McIlwaine, Granada, and Valenzuela-Oblitas 2019;Erez, Adelman, and Gregory 2009). Service provider D discussed a Brazilian woman who went to the police to request a document so that her solicitor could apply for the renewal of a molestation order against her daughter's father; a policeman refused, claiming that the father had a right to see his child.
The police are the gateway to the judicial system which many Brazilian women and service providers reported as institutionally racist and sexist, trivialising women's concerns and favouring men's rights (as parents) over women's rights as survivors (Rose 2015). While the state may be complicit in such violence against women from all backgrounds, it is more disempowering for those in more precarious situations who may lack information, fear deportation if they have insecure status and face racial discrimination. Many women spoke of not being believed by the authorities because they were racialised and minoritized migrants (see also McIlwaine, Granada, and Valenzuela-Oblitas 2019). Gabriela, who we represented before, spoke of the failures of the judicial system when she took her husband to court for long-term domestic abuse of her and her daughter and the judge dismissed her charges: The justice system is very flawed … At no point did they look out for me, did they respect what I was going through or believe me … They didn't believe me, they didn't believe my pastor, they didn't believe my case worker, they didn't believe the psychologist.
These types of encounters, coupled with lack of information and resources, meant that women rarely prosecuted perpetrators meaning that men could act with impunity. According to service provider M, it was common for women to drop prosecutions because of fear, especially of being deported which is exacerbated by the demands for detailed evidence required for a case to proceed. The system becomes an 'active partner' and perpetrator of active infrastructural violence against migrant women (Burman and Chantler 2005;Cassidy 2019), as service provider F stated: What I hear most from women is "If [only] I'd known that it would be so difficult to [navigate] the system", because the abuse by the system often is worse … than the abuse they suffered at the hands of the perpetrator.
It is also important to acknowledge the inadequacy of the criminal justice system rather than blaming women for not reporting effectively or withdrawing cases (Brickell 2015;Hume and Wilding 2020). This is exacerbated by cutbacks in services for women, especially free legal advice as service provider F noted: cutting legal aid means that 'so many women wind up in court without a barrister, without knowing what they have to do there, without any preparation, without speaking the language' (also Vacchelli, Kathrecha and Gyte 2015). Some interactions with the police and criminal justice system were more positive, although they were a minority, as noted by a woman from a focus group discussion who identified as white: Brazilians tell you "Careful with the police, they'll arrest you!" so you get scared of the police. It is nothing like that! The police went to my place twice. … Yes, they were like angels that God sent me to help me.
Turning to accessing statutory health and social services, a quarter of women surveyed accessed this type of support. Some women spoke positively about accessing these services such as Ana Clara, 36, from São Paulo who was white and worked as a teacher who recalled: 'I became depressed, I went to the GP because of depression, but I didn't think it was to do with the relationship, I thought it was lack of vitamin D and sunlight. ' Yet, it is notable that those with more favourable viewpoints were more likely to be from more privileged backgrounds. In contrast, those subject to NRPF restrictions were particularly vulnerable, especially those requiring access to refuges. Cristina who worked as a cleaner and at the time had NRPF stated: 'If I'd had documents, Brent council would have removed him from the house and given me some support, but as I didn't … I felt like my hands and feet were tied, I had no way out. ' Institutional racism and constructions of migrants as 'undeserving' of state assistance played out through overt and covert infrastructural violence on the part of social services. For example, service provider B spoke of how a Brazilian woman survivor of intimate partner violence whom she had signposted to a social worker told her: 'You are strong, I don't see any marks on your body, why don't you go and do some work?' These processes amount to gendered and intersectional infrastructural violence committed by the state serving to disempower Brazilian women, albeit in different ways according to their social locations.

Challenging gendered infrastructural violence through non-statutory migrant infrastructure
Migrant infrastructure in the form of non-statutory migrant and/or women's organisations provide essential support for women in general and specifically for survivors in the face of infrastructural violence by statutory bodies (Turcatti 2021). 'Women's equality infrastructure' encompassing civil society was especially important in providing long-term tailored support (Brodie 2008). Indeed, accessing these organisations is increasingly viewed as forms of resistance (Piedalue 2019 on 'slow non-violence'). While the Brazilian consulate was the most visited organisation cited in the survey (58 percent), only a very small proportion consulted them in relation to gendered violence. Besides the police, migrant and/or women's organisations were the next most important source of support (for 39 per cent of women). Most found out about the organisations by 'word-of-mouth' and not surprisingly, those oriented towards Brazilians or Latin Americans and providing services in Portuguese were the most important. Miriam, 46, from Minas Gerais and mixed race who worked as a cleaner recalled: They gave me guidance and supported me from the beginning to the end, through two and a half years of court hearings … I won my case, thankfully, with the support of [Service Provider C] … as I don't speak English very well, I had the help of a lot of volunteer interpreters.
In addition to English language and legal assistance, psychological support for survivors was seen as essential by over one third of women surveyed. Carolina (see above) explained how therapy provided by a Latin American organisation had been pivotal in her life: 'the biggest impact was the therapy, because I would have jumped in front of a train or a bus, I would have destroyed myself.' This was invariably accompanied by childcare provision as well as wider non-judgemental support that reflected understanding of women's circumstances as noted by Service Provider F who said they became women's family and community, the only people who did not judge them (also Lopes-Heimer forthcoming).
Migrant infrastructure organisations allowed women survivors of all racial, occupational, class and immigration status backgrounds to navigate and mitigate state gendered infrastructural violence They provided a support community, specialist advice in their own language, and legal assistance. These were part of 'countering' or '(en)countering' (Sen, Kaur, and Zabiliūtė 2020), which entails challenging barriers to support in ways that range from the superficial to deeper structural changes (also Hume and Wilding 2020; Pain 2014b; Piedalue 2019).

Conclusions
This paper has explored how diverse Brazilian migrant women survivors navigate multiple forms of direct and indirect, passive and active gendered violence when they seek help. We have made a series of key contributions to understanding migrant women survivors' experiences through an infrastructural lens (McFarlane and Silver 2017). The first relates to the utility of acknowledging exclusion from migrant support services, themselves a form of infrastructure, as infrastructural violence that is underpinned by discrimination and xenophobia within a hostile immigration environment (Rodgers and O'Neill 2012). The second feeds into emerging work on infrastructural violence among women through the notion of 'gendered infrastructural violence' that we suggest must be acknowledged as deeply intersectional, especially among migrant women (Datta and Ahmed 2020;Sawas et al. 2020). This conceptualization has been informed by black feminist thinking around structural and symbolic violence, especially in relation to Smith's (2016) notion of 'sequelae' and Perry's (2013) 'slow death' (see also Carneiro 2003;Ribeiro 2016;Segato 2016). We argue that the intersectional oppressions of gender, race and class that permeate gendered infrastructural violence need to be foregrounded especially in the context of Brazilian migrant women.
This relates to the third contribution to feminist geographical debates on further acknowledging how gendered violence is multi-scalar, with direct intimate and interpersonal forms in the home and public sphere intersecting with state violence in passive and active ways (Brickell 2020;Jokela-Pansini 2020;Pain 2014aPain , 2014bPain and Staheli 2014). This is especially important in relation to migrant women survivors where analysing gendered infrastructural violence sheds light on racialized hostile immigration regimes and their inherently violent modus operandi (Cassidy 2019;Yuval-Davis, Wemyss, and Cassidy 2019). The final main contribution of the paper relates to the need to more fully acknowledge the role of migrant infrastructure and specifically migrant feminist organisations in ameliorating violence and fear experienced by survivors. While their role as caring communities has been noted more generally, it is essential to recognize them as providing a critical lifeline for migrant women survivors from Brazil and elsewhere.

Notes on contributors
Cathy McIlwaine is Professor of Geography and Vice Dean of Research for the Faculty of Social Science and Public Policy at King's College London. Her research focuses on gender, livelihoods, gendered violence against women and migration in several countries of the global South as well among Latin Americans in the UK. She also works with a range of artists and theatre groups, co-producing theatre, film and multi-media art installations. She is currently working on gendered violence and resistance among women in Maré in Rio de Janeiro, Brazil and continuing her work with Brazilian migrant women in London. Cathy has published eleven books including most recently Dores que Libertam: Falas de Mulheres das Favelas sa Mare, no Rio de Janeiro sobre Violencias (with Miriam Krenzinger, Eliana Sousa Silva, and Paul Heritage [eds] [2018] Appris Editora, Rio de Janeiro) and Cities, Slums and Gender in the Global South (with Sylvia Chant [2016], Routledge) and over 40 journal papers. She is a trustee at the charity, Latin Elephant and the Latin American Bureau and an advisor for the Latin American Women's Rights Service.
Dr Yara Evans is a Human Geographer with interests on international migration, political ecology and the bioeconomy. She worked at Queen Mary, University of London (2004)(2005)(2006)(2007)(2008) on research that examined the role of low-paid migrant workers in London's economy. She also carried out research on Brazilians in London/UK as a member of the Brazilians in the UK Study Group (2008Group ( -2015, documenting their socio-economic characteristics, cultural practices, financial behaviour, and connections with Brazil. She was also a tutor with the Centre for Development, Environment and Policy at the School of Oriental and African Studies, University of London (2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016)(2017)(2018). Between 2016-2019 she worked at King's College, University of London on research on Brazilian women's experience of violence in London. She has since worked at the Centre for Environmental Policy, Imperial College London carrying out sustainability assessment of socio-economic, health and policy impacts of bioenergy and bioproducts in EU Horizon 2020 projects.