Sheltering from domestic violence: Women’s experiences of punitive safety and unfreedom in Cambodian safe shelters

Abstract This paper explores the experiences of Cambodian domestic violence survivors who have fled their abusive partners to live in NGO-run safe shelters. Through in-depth interview research undertaken in 2016, we explore the stories of seven women whose experiences speak to tensions between having safety from violence and freedom to live as they choose. The pervasive impunity of the legal system means that Cambodian society operates as a safe space for perpetrators of domestic violence and spatially excludes survivors from it to guarantee their safety from injury and even murder. Just as violence against women has been described as a major area of ‘unfreedom’, we contend also that safe shelter provision in Cambodia, albeit essential, does not necessarily afford freedom from violence, but rather a punitive safety from it which can curtail women’s bodily integrity. Survivors are too often being excluded from decision-making processes in the shelter and treated as passive recipients of physical safety. Making the argument that safety and freedom are not coterminous, we contribute to recent feminist scholarship in geography and aligned disciplines focused on the significance and workings of safe space for marginalised groups. As such, the paper complicates singular viewpoints of safe spaces as enabling environments which can challenge oppressive forces both inside and outside of them.


Introduction
I feel safe here because of the security. This place is secret and he cannot come here. We have a security guard. The fences are high and nobody can look inside … I didn't have this feeling of safety in my home before. But now I am safe, I do not have freedom. We do not talk with people outside and we cannot go out alone. This is not a normal place. This is not how you live on the outside. I feel worried I am forgetting how to do that.
Puthea arrived in the safe shelter with her two children having fled her violent partner. She had reported the physical, emotional, and economic violence for the past two years of their nine-year relationship to the police, but they took no action. Not held accountable, her husband went on to damage her hand so severely that she could no longer work as a construction labourer to support the family's financial needs. Puthea took the difficult step of contacting Banteay Srei, a Khmer NGO that works with women survivors of violence but does not have a long-term stay safe shelter. She was then referred to a Cambodian Women's Crisis Center (CWCC) shelter so that she could pursue a court-sanctioned divorce and legal division of land ownership without risk of retaliation. While she feels a sense of physical protection from the security of guards and fences around her in the shelter, she no longer has freedom (seripheap). Not only this, but eight months into her stay and uncertain when the legalities will conclude, she voices fear and a lack of confidence about going out alone and functioning in normal (tomodah) society in the future. Puthea's concerns over her diminishing confidence and social skills are not unusual. Indeed, in this paper we explore the stories of seven women who are living in one of three safe shelters hidden from their perpetrator(s) and whose experiences speak to tensions between having temporary safety from violence and freedom to live as they choose.
Our contribution to this theme section on 'Safe Spaces of Refuge, Shelter, and Contact' draws on survivors' narratives to demonstrate how safety and freedom are not necessarily coterminous. Just as violence against women has been described as a major area of 'unfreedom' (Agarwal and Panda 2007;Nussbaum 2005), we make the argument that safe shelter provision in Cambodia does not necessarily afford freedom from violence, but rather a punitive safety from it which curtails women's freedom of movement and autonomy further.
To explore women's experiences of safe shelters, the paper begins by looking at existing geographical and aligned literatures concerned with 'safe spaces' and notions of freedom. After this we provide more information about the project's research methods, a brief detailing of its ethical guidelines, and key contextual information on domestic violence and gender norms in Cambodia. Shedding light on how women individually and collectively navigate, cope with, and make sense of their lives whilst living in the safe shelter, the empirical analysis is divided into two interrelated parts: the first, concentrating on women's curtailed freedom of movement; and the second on their rule-based existence in the shelter. Both show how there needs to be greater attention paid in shelter interventions to supporting women's bodily integrity and emotional needs.

Safe spaces and unfreedom
While feminist geography has a long lineage of scholarship on women's experiences of fear in unsafe spaces (e.g. see early seminal work by Valentine 1989), the paper that follows shifts this focus to their encounters with 'safe spaces'. In fact, in the last five years, geography has seen a burgeoning of scholarship on a diversity of 'safe spaces' for different groups (see The Roestone Collective 2014 for a review), from child contact centres (Morrison and Wasoff 2012), universities (Arao and Clements 2013), gay neighbourhoods (Hanhardt 2013), to theatres and other sites of night-time leisure (Held 2015;Richardson 2015). First then, our paper expands the spatial remit of this work to the safe shelter and their use by domestic violence survivors. By doing so, it goes beyond the Anglo-American focus of interdisciplinary academic research which currently exists on safe shelters concentrated on the United Kingdom, Canada, and the United States (e.g. Bowstead 2015;Lyon et al. 2008;Warrington 2001;Woodhall-Melnik et al. 2017).
Second, our paper pushes forward scholarship that singularly frames safe spaces as enabling environments 'to challenge the forces that oppress' (Bowpitt et al. 2014(Bowpitt et al. : 1255. Lewis et al's (2015: np) writing on women-only feminist gatherings in the UK, for example, argues that 'once women are safe from harassment, abuse and misogyny, they feel safe to be cognitively, intellectually and emotionally expressive' in this environment (emphasis in original). Such separatist safe spaces engineered by civil rights, feminist and queer activists are understood to offer individuals the opportunity to 'speak and act freely, form collective strength and generate strategies for resistance' (Kenney 2001:24). The data in this paper complicates this one-sided analysis by showing how the safe shelter does not provision a space to speak and act freely. In a similar way that women on college campuses were restricted to their dorms to keep them safe from potential assault in the mid-twentieth century (Stengel 2010), women accessing safe shelters in Cambodia are restricted to the separatist space of the safe shelter. The 'degree to which safety is an appropriate or reasonable expectation for any honest dialogue about social justice' (Arao and Clements 2013: 139) is therefore a mute question.
On this point, third, we argue that the case study shelters' almost exclusive focus on physical safety is to the putative neglect, and further detriment, of women's feelings of wellbeing, independence, personhood, self-esteem, and control. As Hiscock et al (2001: 50) set out, 'people need more than just adequate sustenance and shelter to live happy and fulfilled lives'. They assert furthermore, that 'autonomy is a mixture of freedom to do what one wants and to express oneself as well as freedom from any need to have one's actions approved by others and from any need to conform with others' expectation of oneself' (2001: 53 emphasis in original). Encompassing these ideas, the paper that follows develops the argument that just as domestic violence is an unfreedom hidden within women's families, there are also unfreedoms which are re-inscribed afresh in safe shelters which demand greater recognition and action.
Developing this line of inquiry, fourth, we make the point that freedom and unfreedom are important concepts for making sense of women's experiences in safe shelters. While there is a weight of social, political and poetic discourse on 'freedom' that cannot be discussed here (see Agarwal and Panda 2007), Sen's (1999) usage of the term, as its most basic, suggests that freedom arises from the capabilities to choose what you have reason to value. Capabilities ask, '"What is she actually able to do and to be? … as a dignified free being who shapes his or her own life, rather than being passively shaped or pushed around by the world' (Nussbaum 2002: 124). 'Bodily integrity' is an overlooked capability in Sen's vision of freedom and includes 'being able to move freely from place to place; to be secure against violent assault, including sexual assault and domestic violence' (Nussbaum 2003: 41;also 2005). The safe shelter both helps and hinders this view of bodily integrity. While it secures women, quite literally, from domestic violence, it does not permit freedom of movement. The functioning of the shelterits confining tendencies and rigid rulesare interlocking factors stymying women's freedom.
The safe space of the shelter thus provokes 'the possibility of reinscribing relations of domination' (Stengel 2010: 536). As DeWard and Moe (2010: 122) qualify in their work on women's narratives of homeless shelters in the United States, ' … while shelters are distinct from institutions, such as prisons and some mental health hospitals, wherein people are confined against their will and are not free to leave, there is an element of coercion within them. although women were free to leave, this 'freedom' was mitigated by the consequences of living homeless on the streets or otherwise without secure access to shelter, food and clothing'. In our case study, this unfreedom is imposedcoerced evenas a necessary means of ensuring women's safety from their perpetrators. In other words, safety from domestic violence does not mean freedom from it; rather women's safety in the shelter and their adherence to its rules is a continued necessity. In sum then, the paper raises some of the challenges that women face in being capable of choosing the lives they want to lead. It also signals the importance of organisations being more considerate of, and where possible adaptive to, women's expressed needs so that survivors can feel greater agency over day-to-day existence in the shelters they reside in for months and sometimes even years.

Researching survivor experiences of safe shelters in Cambodia
In this paper we draw on the narratives that women shared with Naomi in her research with seven women living across three shelters run by the CWCC. Their shelters in Phnom Penh, Banteay Meanchey, and Siem Reap can variously accommodate between 40-100 clients and have the stated objective to provide women and children survivors of violence with 'access to immediate protective services, such as the drop-in crisis center, emergency medical care, basic counseling and relevant information' (CWCC 2018: np). Access was officially agreed with the Executive Director of CWCC and introductions facilitated with shelter managers to take the research forward. The CWCC preferred and often insisted that Naomi remained in the meeting room to undertake the interviews with little opportunity to see other common or private areas of the shelter except for using their toilet. Painted in a magnolia colour reminiscent of Cambodian schools with discoloured women's rights posters on the walls, and iron bars on many of the windows, the safe shelters each had a set of swings with some seating and flower beds which women tended to as their children played.
After a substantive ethical review and approval process followed with our university, the interviews with survivors were subject to their informed consent. The information sheet and consent form were read to women given low levels of literacy and as per Cambodian custom a finger print was provided to confirm consent. Survivors were reassured that any identifiable information would be removed from their interview transcripts and their experience sharing would not impact upon the care or services they were receiving in the shelter. The interviews were undertaken in Khmer language with the support of an experienced female interpreter disassociated from the CWCC and the shelter.
In this paper we focus on the interviews of all women who were domestic violence survivors residing in the CWCC shelters at the time of research (2016), namely Puthea, Lina, May, Channou, Socheata, Sopheap and Nipha. To be clear, these seven women were the only domestic violence survivors residing in the CWCC shelters in the 6-month period of the research which Naomi conducted. Ranging in age from 29 to 36 years-old, the women were all Khmer and had all been married to their violent partners, be this a traditional or officially registered marriage, for between 9-12 years (see Brickell 2014; Brickell and Platt 2015 for information on the distinction between marriage types in Cambodia). All had children, most of whom stayed in the shelter and others who were living with grandparents or (ex-) husbands. The women's stays ranged, at point of interview, from 1 month to one year and all (except Channou) had experienced physical violence in combination with emotional and economic violence. Their duration of stays in the shelter cross-cuts the demarcation of shelter accommodation as 'emergency' (a few days up to a few months) and 'transitional housing' (six months to one year or more) in UNWOMEN (2013: 9) guidelines.
The absence of research in both academic and non-academic realms on safe shelters in Cambodia was identified in Katherine's study (2012)(2013)(2014)(2015) on the implementation and enforcement of Cambodia's first ever domestic violence law ratified in 2005 . Addressing this identified knowledge gap, Naomi undertook research in every shelter in Cambodia which were accepting domestic violence survivors in 2016. While in this paper we focus only on the CWCC-run shelters, the larger study also undertook 10 interviews with women who accessed an emergency 'transit shelter' run by the local NGO Banteay Srei. Key stakeholder interviews with shelter staff, domestic violence advocates, and policy makers also formed part of this suite of research across the two organisations. This limited number of case study shelters reflects the lack of access that most women in Cambodia face (MOWA 2015: 3).

Domestic violence in Cambodia
Both the authors' projects were designed and undertaken in the context of insidious and widely experienced domestic violence, its prevention and elimination having shown 'limited progress' according to United Nations Committee on the Elimination of Discrimination against Women (2013). Nationally representative data from a Partners for Prevention (2013) study shows that 1 in 4 women (25.3%) in Cambodia have experienced at least one act of physical or sexual violence or both perpetrated by an intimate partner in their lifetime. There is a clear hiatus between these daily realities of violence suffered by women and girls, and the rights espoused in (inter)national legal frameworks including the Universal Declaration of Human Rights, the Cambodian Constitution, and the country's 2005 'Law on the Prevention of Domestic Violence and Protection of the Victims' which classes severe bodily injury as a criminal offence and recommends that the Penal Code to be used to pursue a criminal conviction (Article 17). In a country with 'vitally absent' rule of law and pervasive impunity (United Nations Human Rights Council 2018), shelters offer a 'last resort' for women escaping domestic violence as criminal convictions are rare and it is not uncommon for perpetrators to pay their way out of jail (Brickell 2015;Burns and Daly 2014). Given the low salaries of public officials, poor training and patron-style leadership, bribes are systematically used to leverage the dropping of criminal cases across the country (Broadhurst and Bouhours 2009).
The imperative placed on harmony, consensus and the collective in Cambodia is also so pervasive that many victims are still pressured into local reconciliation (samroh samruol) with the same husband, often on a repeated basis (Brickell 2015). Meaning to smooth over and seek harmony, this practice is precluded under DV law in cases which fall under the Penal Code, yet it is habitually used. The official glossary to DV Law is clear that this 'communication process between quarrelling parties … aims at maintaining family life' (Royal Government of Cambodia 2007: 11). Such is the preponderance of local reconciliation, that CEDAW (2013: 4) has made overt reference to the state's propensity to 'dispose of cases of violence against women through mediation' (our emphasis).
In this context, Cambodia's National Survey on Women's Health and Life Experiences (UNWOMEN 2015) found that although 90 percent of women beaten by their partners needed medical treatment, nearly half did not try and receive any help at all. Cambodian domestic violence victims trying to divorce, or enter a safe shelter, are de facto stepping outside the social and moral order which Cambodian tradition and government rhetoric reinforces (Brickell 2015;Lilja and Baaz 2016). The idea that women should accept and endure domestic violence is related, in part, to 19th century normative Cambodian poems such as the Chbab Srei (Rules for Women) (see Brickell 2011 for fuller exploration of these texts). The women's code deals predominantly with obeying and respecting spouses by keeping 'fire in the house'. This Cambodian Buddhist expression embodies the idea that in order to maintain a harmonious household, women are responsible for suppressing three fires of potential conflict connected with their relationships to parents, husbands and 'others'. Women should not bring fire from outside into the house, not take fire inside the house outside, and should take care not to spread or overheat fires. It is not only commonplace that women in Cambodia are blamed for the violence they encounter, but they can also feel judged for defying norms of familial harmony and togetherness. These are all factors which women living in the country's shelters contend with.

Unfreedom of movement outside the shelter
Here is a place for the victims. They must make it safe. There has to be rules but I want to live like normal people. I want to see the outside. When I am living in my own place I talk with other people who live close by, my friends. I miss that … we do not go out to visit other people. We stay in here with the other women and children and staff. It's nothing like my life outside. It is not like a normal place so I do not feel like I am normal. I am different because I stay here. It's like being cut off from the outside world. (Puthea) All of the women in this paper lived in the same province as the shelter and this curtailed their ability to move freely outside given the threat of meeting or being identified by their partners who would then know their whereabouts. Many of the women had not seen their family or friends since arrival at the safe shelter. None had internet or social media access. However, all the women had old-style Nokia mobile telephones on pay-as-you-go that enabled them to maintain contact with their family and friends via phone calls (some of the women were illiterate and could not text message). In the CWCC safe shelters, the women's mobile telephones brought new life to their restricted and otherwise controlled space, blurring the line between the realms of private and public. While women described their mobile phones as one of their most important possessions, the devices could not fully mitigate the unfreedom that women express in the interviews. As Puthea (above) notes, the shelter is not a 'normal' place, it is 'cut off from the outside world' precluding contact with anyone other than survivors and staff. The safe shelter is viewed as a place where women find their lives defined by their status as a survivor of violence. As Lina and Channou elaborate, the designing in of physical safety is lived and felt almost as a form of captivity, Both Lina and Channou explain how there is a strict regulation of their movement with limited opportunity to leave the closed space of the shelter during their stays. While safety remains their utmost priority, the interviews show how their spatial confinement, isolation, and lack of agency over this echoes the coercion and control experienced in their abusive relationships they are trying to break free from. Channou desires more 'freedom' and in her interview she expressed her frustration of not having more control over her visits outside the shelter. These are what Haaken and Yragui (2003: 50) describe as 'border tensions' in shelter provision which includes 'the problematic question of how to draw physical and social boundaries between safe houses and the wider community'. Cambodian safe shelters define these boundaries without any consultative route for women to discuss their individual circumstances and differential risks in the vicinity of the shelter.
The CWCC felt it was too unsafe for survivors to leave the safe shelter alone and the women were chaperoned on the rare occasion when they did leave. Being seen in public with a support worker reinforced feelings of stigma and shame for many of the survivors who felt this dependency marked them out. As Lina went on to explain, We can go to the market together with one [member of] staff. When we go there I am sure some people think that we are bad wives, or bad women, because we faced the violence … I would like to change what they think, but I am not allowed to do anything outside of here … they do not talk to us because they do not think we are normal like them. They can recognise us … we are together in one group and with the staff. Lina longed to be treated with equality and perceived as 'normal' by people in the local market and community. As Padgett's (2007) research with homeless persons living with mental illness in the US found, engagement in everyday activities such as grocery shopping offered them greater parity of experience with their non-mentally ill peers. Lina's interview however spoke to the challenges of this therapeutic practice as a domestic violence survivor living in a country where victim-blaming is prevalent. The infamous Cambodian proverb 'men are gold, women are white cloth' implies, for example, that if gold is dropped in dirt, it can be polished; whereas cloth remains stained. Women's internalisation of this sentiment was strongly evident across the interviews. As Socheata explained further, They know about us when we go to the market. All the people around here know about this place and they know why we stay here. They [local people] know we are not allowed to meet with people outside so they are afraid to talk with us.
When questioned about whether the women were permitted to talk with people outside of the safe shelter, Socheata responded with uncertainty, It does not really happen because we always have the caretaker with us. We are in a small group together so we just stay in the group. If someone tried to speak with us I think the organisation staff would handle it. Of course people say hello, but we never stop for long enough to talk. I can see people are not sure about talking with us.
Everyday interactions are vital to an individual's health and feelings of normality and interaction with other people can build self-esteem and enhance decision-making capacity. That these visits were chaperoned and their purchases limited to essential items reduced the capacity of their market visits to instil a sense of freedom and personhood. As Sopheap elaborated, I feel happy when I am made up. I really enjoy make up. But I cannot do it here. It is sad for me … before I came here I would feel happy when I went to the market to buy some clothes, and make up to use at home. I felt like I was a princess when I went shopping. I cannot do that here because I must obey the rules of the shelter. We cannot go out alone and if we go with someone then we can only buy food for cooking. When I see clothes and make up in the market I want to buy them. But I have no money … inside here we do not get money to buy something for ourselves. (Sopheap) Like other women, 31-year-old Sopheap left her home suddenly and had to abandon the majority her possessions. Some of the women had time to pack official documents (e.g. their birth certificate and/or identity card) and clothes. One woman came to the safe shelter without any shoes. Without the means or license from the shelter to buy clothes and make up Sopheap felt her life had been reduced to the bare essentials without freedom of selfexpression. This was felt keenly given that Sopheap previously earned a monthly wage in a garment factory. That taking pride in one's appearance and feeling special (for Sopheap like a princess) is part of the rebuilding process has a been identified in research with domestic violence survivors living in US shelters (Neuman Allen and Wozniak 2010). But living in the shelter, becoming financial dependent and unable to take everyday mundane decisions, had brought a diminishing sense of confidence and bodily integrity over how women look and feel. In the next section, we reflect on the unfreedoms that Sopheap and other women encountered during their stays in the shelter.

Everyday unfreedoms in the safe shelter
We have a schedule for the cleaning, cooking and skills training. We must keep the shelter clean and take care of our rooms … some rules are simple like at home, for example we should not drink or take drugs, we must cook and clean. The hard rule is about the freedom. We cannot go along and we are not allowed visitors here. I do not like this rule. (Thida) I am so far from my own place. We live inside here and I feel only. I want to talk with more people, to have more freedom … to do stuff I choose. We have rules here I have to respect. Sometimes I do not want to respect the rules and do something I choose. But I can't. If I do not respect the rules I cannot stay here. (Maly) A recurrent feature of the interviews with survivors living in the shelters centred on the everyday unfreedoms that they faced and connected to this, the myriad rules which govern it. The time-space of the shelter is dictated to women by its staff without any consultative functions or flexibility. The women were always required to complete their daily tasks and follow the time schedule, irrespective of their physical or mental state that day. As Maly and other participants in their interviews noted, the punitive affect and logic of the shelter rules means that if they are not adhered then women must leave. Many interviewees felt they were again engaged in a trade-off between control over their own time and access to services and support. Instead of being granted time to focus on their emotional wellbeing and healing processes, women were consumed by their daily duties. The proportion of their time spent on cooking, cleaning, and childcare also reinforced gender norms and responsibilities tying women to the social reproductive work of the family.
These unfreedoms also extended to women's lack of choice over sharing or having their own family bedrooms. All but one of the women staying in the CWCC safe shelters did not have any access to private bedrooms. Instead, they shared their bedrooms, bathrooms, and the kitchen. There were up to six people (a mix of women, their own children and unaccompanied children) staying in one bedroom. When the women entered the safe shelter, they were assigned a bedroom based on where there was space. None of the women could choose their own room or who they stayed with. All six women who had children with them wanted a bedroom of their own to allow greater privacy.
Maly was the only participant who was fortunate enough to have a bedroom of her own. This privacy was key to Maly's feelings of control and autonomy. Maly had her own bedroom because the room was only big enough for her and her two children, not because of any preferential treatment. The 33 year old, had been staying in the safe shelter with her children for just over a year. She was eager to exit the safe shelter but was unable to leave because her case had not yet been processed by the court. Albeit frustrated about the delays relating to her case, her private room made the waiting less problematic: Fortunately, I have my own room and it is a small room. If it is big, I have to stay with the other people … l can feel happy because I have my own private room. Nobody can enter into my room. My case is taking too long, but I have my own room so I can live here longer.
Maly's previous marital home had been characterised for eight years by physical, psychological and economic violence. These experiences of violence had a direct impact on the characteristics she now sought from her accommodation: Home is safety and security. We did not have that before -me and my children. I cannot call here my home because it is temporary, but it is like what I want a home to be like. It is safe. We can be alone and we are safe … but if I did not have my own room, I cannot stay here.
Research into temporary accommodation has repeatedly advocated for the benefits of residents having access to both private (bedroom, bathroom, kitchen) and communal spaces (outside areas) (Datta 2005;Kellett and Moore 2003). Having access to spaces that are typically private in a normative home can be key to creating feelings of independence and control. While these findings are based on Western case study research, in Cambodia what is classed as the 'normative home' deviates from these ideas however with it being common for extended families to be living in one house and sharing sleeping spaces. Withstanding this, when the all-female shelter staff were asked about the women sharing bedrooms, they consistently acknowledged that it was not ideal. However, the shelter manager from Siem Reap stated there was no possible alternative due to budget and space constraints. Reflecting this, several women stated they felt unable to ask for a private space because they were already indebted to the organisation and needed to show respect. Channou stated: Sometimes it is hard because I do not have a private place to be with my son. But I stay here for free so I cannot expect to have everything how I want it. That is the organisation's decision … I sleep with my child and share everything with the other women and children. We stay together, eat together and share everything. One the one hand, this teaches us all to share. But on the other hand, I miss having my own space. Either I cannot be alone with my son, or I cannot take time away from him. I must always stay with with him and take care of him. If I were at home, I could ask my relatives to help me … if I had my choice I would stay alone. Confinement to the shelter and the lack of private space within it, meant that Channou and her child either could never be alone together, or she did not have any relief from parental duties. The need for women to do everything together and be ever-conscious of sharing also posed an additional burden that women spoke of. As Heath and Scicluna (2017: 53) write, 'living in close everyday contact with non-family members (non-kin) in shared living arrangements necessitates an often extraordinary degree of physical and emotional intimacy, whether actively sought or not'. Sharing of responsibilities, resources, and also space were forms of emotional labour undertaken in the shelter. For many women this meant putting collective needs before their own for fear of friction and disagreement otherwise.
Discussing the sharing of bedrooms, Lina highlighted, for example, that the placement of objects, photos and furnishings would involve negotiating a space for each woman and could lead to unwanted conflict and arguments. As a result, she minimized her own desire to engage in decorative practices so as to avoid any chances of conflict: We share our rooms. So what if I want to decorate but another woman does not want to? Or I choose to post pictures the other people in my room do not like. We do not want to fight and we do our best to avoid problems. We want a peaceful place. So it is easier if we do not decorate because then there is nothing to disagree about.
In research into the role of hostels and temporary accommodation, Busch-Geertsema and Sahlin (2007) argue that sharing spaces such as the bathroom and kitchen increases the risks of in-house conflicts in comparison to homes that are self-contained or where residents have agreed to share space based upon friendship or family relations. Negotiating space and carrying out typically private tasks in a collective sphere was particularly challenging for the women in the CWCC shelters. Nipha, for example, had a photograph of her family which she locked away rather than display for similar reasons: Nipha's uncertainty about displaying her photograph illustrates the lack of control she had over her own decision-making processes. She wanted to display her photograph but felt unable to do so because she was living under the ruling of the safe shelter. Nipha expressed how her photograph gave her feelings of normality and helped her remember an identity that was not characterised by violence. It projected a future self because it allowed her to imagine a happy, secure, and reliable family without violence. Research conducted by Pable (2013) in Florida, America argues that the inclusion of possessions in temporary accommodation can assist with psychological recovery and reengagement with society. As a result, the author argues that shelter designs should embrace the importance of residents' possessions. In Nipha's case, there could be psychological benefits to having her photograph in constant view, yet this was restricted by an uncertainty of whether she could do so and the potential jealousies and tensions that might arise if she did.
In contrast to most women interviewed, Maly stated that she had no desire to decorate her bedroom, despite it being her own private space and her stay having exceeded one year. She stated, I do not know when I have to leave here so I do not want to decorate my room. This is just a temporary place. Moreover, if I do the decoration my children will start to think this is their home.
Lina's feelings about the walls reflecting the temporary nature of the safe shelter is in keeping with findings from research by Parrott (2005) which explored how residents relate to their accommodation through the decoration of their personal space in a medium secure psychiatric unit. In some cases, patients refused to decorate their bedrooms (although they were free to do so) because of a perception that fixing objects to the walls symbolised being fixed in the institution. For some women in the safe shelter the affirmation of temporality was marked through a lack of decorative practices and asserted hope that their freedom from the safe shelter would be immanent. Across the interviews then, possessions held an ambivalent position in women's lives in the shelter, as sources of tension yet comfort, as present reminders of home life, or absent presences of a hoped for freedom in the after-life of the shelter.

Conclusion
Shelters in Cambodia are commonly viewed as the 'last resort' by survivors given their separation of women, literally and metaphorically, outside of 'normal' society. Despite a strengthening of women's legal rights in the country, Cambodian society persists as a safe space for perpetrators and it is domestic violence survivors who are spatially excluded from it to ensure their safety from injury and even death. Just as Sen (1992: 125) affirms the general existence of 'systematic disparities in the freedoms that men and women enjoy', these injustices are manifest in women's shelter stays. To pursue their legal claims against violent partners, it is women who are effectively excluded from public life and live behind walls and bars that their perpetrator will likely never see.
Survivors describe being 'cut off' from the outside world and judged when they do enter it, thus showing the lack of entitlement which women have to bodily integrity in mainstream society given the immobility and perceived surveillance they face. Women's interviews discussed in our paper also demonstrate how their bodily integrity is compromised by a non-negotiable onus on rules in the shelter itself. On this front, survivors are too often being excluded from decision-making processes in the shelter and treated as passive recipients of physical safety. So while women were safe from domestic violence, survivors did not feel freedom to express themselves, their needs, and identities in the shelter. Best practice guidelines to enable this were also not followed given resource and space constraints. In their guidelines, UNWOMEN (2013) note, for example, that where possible, 'facilities should have space and infrastructure which allows for individual privacy, reflection, expression as well as collective activities'. This includes, inter alia, a bedroom for each woman with her children, a bathroom shared by no more than two rooms/women; and adequate storage for women's belongings such as on-site storage units.
In conclusion, the design and running of safe shelter provision in Cambodia does not necessarily afford freedom from violence, but rather a punitive safety from it which curtails women's freedom of movement and autonomy further. While 'safe spaces are traditionally assumed to provide temporary relief from potential threats to ones' wellbeing within a secure and private physical structure' (Hassan Nur et al. 2018: 97), this paper has explored women's accounts of what this safety feels like, and amounts to, beyond the protection offered from violent partner. Making the argument that safety and freedom are not the same, we have contributed to recent feminist scholarship in geography and aligned disciplines focused on the significance and workings of safe space for marginalised groups. As such, the paper has problematized overly celebratory viewpoints of safe spaces as enabling environments which can challenge oppressive forces both inside and outside of their walls.