Vulnerability, Space and Conflict-Related Sexual Violence: Building Spatial Resilience

This article is about the relationship between vulnerability and space. It examines how vulnerability moves and shifts in relation to space – and more specifically particular types of spaces – and how space both generates vulnerability and affects the experiential ‘living’ of vulnerability. With a specific focus on conflict-related sexual violence, and drawing on empirical data from Bosnia-Herzegovina, Colombia and Uganda, the article situates and discusses the relationship between vulnerability and space within a broader transitional justice framework. Underlining the need for transitional justice processes to give more attention to spaces of vulnerability, it foregrounds the notion of ‘spatial resilience’ as providing a conceptual frame for an ecological remodelling of transitional justice.


Introduction
Vulnerability is a concept that has variously been described as 'vague and nebulous' (Brown, 2011: 314), plagued with 'opacity and complexity' (Munro and Scoular, 2012: 195) and 'a buzz word lacking precision' (Cannon, 2008: 351). It is not the purpose of this interdisciplinary article either to add its own critique or to attempt new definitional 'clarity'. Taking as its basic starting point the idea that vulnerability is a 'dynamic phenomenon often in a continuous state of flux' (Adger, 2006: 274), the article's aim is to explore and elucidate some of the concept's spatial dynamics. Specifically, it examines how vulnerability moves and shifts in relation to space, and how space both generates vulnerability and affects the experiential 'living' of vulnerability (Murphy, 2011: 578). In this regard, 'space' is theorized not as a generic concept but, rather, as a latticework of different and intersecting spaces that create highly fluid and context-specific forms of vulnerability.
This research is not the first to discuss vulnerability and space (see, for example, Delor andHubert, 2000: 1563;Smith, 1996: 74-75;Watts and Bohle, 1993: 117). What is novel is its conceptual and empirical analysis of space and vulnerability in the particular context of conflict-related sexual violence. 1 The idea for this article developed out of the author's work in Bosnia-Herzegovina (BiH) and frustration with the frequent portrayal of women who suffered conflict-related sexual violence (the men are routinely overlooked) as 'some of the most vulnerable in the country' (Sesar, 2017). This research does not take issue with the concept of vulnerability per se, but with the essentialized notion that victims-/survivors 2 are vulnerable. Not only does this detract from contextual and spatial vulnerabilities, but it also neglects how individuals experience and manage those vulnerabilities. To cite Brown (2017: 425, emphasis in original), 'vulnerability narratives or discourses constitute something different from lived experiences of vulnerability, or what might be thought of as real vulnerability'.
Using qualitative interview data to explore lived experiences of vulnerability, the article links the concepts of space, vulnerability and resilience. In their daily lives, individuals interact with and negotiate different types of spaces (Plummer, 1987: 13). Viewed through a resilience lens, some of these spaces constitute potential resources that can be actively harnessed to help promote well-being. It was striking, for example, how frequently environmental spaces and natural resources -from land, water, charcoal and mountainsemerged from the interviews (and particularly those in BiH and Uganda) as a protective factor. Some spaces, however -such as a community or cultural space that assigns blame to victims-/survivors and stigmatizes them -can create and/or exacerbate vulnerability.
Looking at the bigger picture, it is precisely because vulnerability is not a given, or something fixed, that the concept of space is so important within a transitional justice context. This article argues that transitional justice, as a set of judicial and non-judicial mechanisms 3 for dealing with legacies of mass human rights abuses (United Nations, 2010) -and, more generally, international policy discourse surrounding conflict-related sexual violence -should give more attention to spatial aspects of vulnerability. Because these spaces form part of the social ecologies that victims-/survivors inhabit, they additionally provide important insights into the wider ecological legacies of conflict-related sexual violence. Spatial dimensions of vulnerability also point to the fact that transitional justice has a potentially important, yet unexplored, role to play in fostering 'spatial resilience' (Cumming, 2011), and thus in building networks within ecological spaces and systems that buffer the impact of shocks and disturbances (Nyström and Folke, 2001: 410). In this regard, the core originality of the article lies not in discussing vulnerability and resilience together, but in its spatial exploration of the two concepts as part of a proposed ecological reframing of transitional justice.

Methodology and Ethics
Doing research with human subjects can contribute to fostering, or increasing, potential 'pathogenic vulnerabilities' Lange, 2013: 2141); these include 'exacerbating feelings of embarrassment, shame, isolation and an inability to care for one's family' (Lange et al., 2013: 337). Such risks are especially acute when doing highly sensitive research with individuals who may already be facing social stigma and marginalization (Skjelsbaek, 2010: 50), and/or struggling to come to terms with their experiences. For Lange et al. (2013: 337), the crucial point is that researchers become part of the context of the vulnerable participant, 'and hence a potential source of new or increased vulnerability once research commences'.
In the context of conflict-related sexual violence, some scholars have accordingly expressed concerns about the ethics of seeking first-hand information from victims-/ survivors themselves (see, for example, Boesten and Henry, 2018;Campbell, 2018). These concerns must be taken seriously. The solution is not, however, to avoid such research or to allow 'key groups' like local non-governmental organizations to speak for men and women who have suffered sexual violence (Campbell, 2018: 480). This is potentially disempowering because it neglects the many reasons why some victims-/ survivors may in fact want to speak about their experiences (Campbell and Adams, 2009). It also promotes and entrenches the narrative of the 'vulnerable rape victim' (see Clark, 2018Clark, , 2019. This article embraces the view that 'even with the significant attention directed at conflict-related sexual violence, more research is needed' (Buss, 2014: 4). It is needed because many gaps still exist, including a critical 'resilience gap'. The following sections draw on first-hand research with victims-/survivors of conflict-related sexual violence in BiH, Colombia and Uganda, undertaken as part of an ongoing mixed-methods comparative study about resilience. The five-year study is exploring why some victims-/survivors of conflict-related sexual violence demonstrate high levels of resilience -ecologically defined as 'the capacity of both individuals and their environments to interact in ways that optimize developmental processes' (Ungar, 2013: 256) -while others do not. In short, the study is not only about victims-/survivors themselves, and their needs, but also about the wider social ecologies that contribute to fostering or impeding resilience -and the shared needs that exist across conflict-affected groups within these ecologies.
Extending the notion of ecology to transitional justice, the research will ultimately use the empirical data to develop a new ecological model of transitional justice that gives greater attention to the interactions between individuals and their environments -and the resources within those environments. In this regard, the selection of three very diverse case studies from different continents is important for analysing the concept of resilience within and across highly varied social ecologies -and for ecologically reconceptualizing transitional justice in ways that have cross-cultural relevance.
The Humanities and Social Sciences Ethical Review Committee at the University of Birmingham and the European Research Council (ERC), the research funder, granted full ethics approval for this research. The ERC additionally required the author (who is the principal investigator of the project) to secure ethics approvals from the relevant authorities in each of the three case study countries. 4 During the long and often protracted process of obtaining these multiple approvals, it was necessary to address a range of issues -from how to ensure that a research participant was actually giving informed consent and how to deal with incidental findings to issues pertaining to confidentiality, illiterate research participants, data storage and fair benefit sharing. The rigour of these ethical procedures greatly improved and strengthened the design of the project. As part of the informed consent procedure, for example, research participants were asked two questions: 'In your own words, can you tell me something about the research?' and 'Can you give me two examples of your rights as a research participant?' Asking these questions was important for establishing whether participants had in fact understood the information given to them and were therefore able to give informed consent.
During the first part of the research, completed between May and December 2018, a total of 449 victims-/survivors across the three countries completed a copy of the study questionnaire. The work of administering the questionnaires was divided between the in-country partner non-governmental organizations (NGOs) involved in the project and the three researchers (one based in each country). The purpose of the questionnaire was twofold. First, it was used as a tool to measure research participants' levels of resilience.
A key part of the questionnaire in this regard was the Adult Resilience Measure or ARM (Resilience Research Centre, 2016). Divided into three sub-scales, namely personal, relational and contextual, the ARM consists of 28 statements, including: 'I have people I can respect in my life' and 'My family stands by me during difficult times'. For each statement, a respondent is given a score of between one and five, resulting in a total ARM score. High ARM scores indicate that an individual has significant protective resources -such as spiritual resources, support from family and friends, and a sense of being part of a community -that offset the stressors in his/her life and thereby bolster resilience. Second, the questionnaire was designed to allow the researchers to explore potential relationships between ARM scores and other factors, including levels of trauma (measured through a Traumatic Events Checklist), the centrality of the sexual violence trauma in an individual's life (measured through the Centrality of Event Scale; see Berntsen and Rubin, 2006) and the number and types of current problems that respondents were facing.
This article specifically draws on the qualitative part of the study. ARM scores were used to divide the respondents from each country into four ARM quartiles, and each researcher chose five interviewees from the quartiles. The selection of interviewees was also shaped by the need to ensure demographic diversity (in particular gender, ethnic and age diversity) within the interview samples. In total, 63 interviews (21 in each country) 5 were completed between January and July 2019.
The three researchers collectively designed the interview guide to facilitate a narrative space for the telling of interviewees' multi-layered stories. This was achieved in two key ways. First, because the interview guide was informed by the study's ecological approach to resilience and by its overall aims in respect of transitional justice, interviewees were asked a wide range of questions about their lives (past and present) and broader social ecologies. Divided into seven sections (namely Life Today, War Experiences, Sexual Violence, Resources and Support, Resilience and Coping, Justice, and Closing Questions), some of the questions further expanded on particular parts of the questionnaire (for example, on sources of support). The crucial point is that by giving interviewees space to speak about different aspects of their lives beyond the sexual violence, the questions created 'interrupted spaces', a term coined by Bolzan and Gale (2011: 503) 'to describe an interruption in the usual life worlds of our research participants'. Second, Boesten and Henry (2018: 581) have posed the question: 'So do first-hand accounts add anything to research?' The answer to this arguably depends on the questions that we ask. Are they novel questions that have not been asked before? In asking them, what new information are we going to learn about sexual violence in conflict (Skjelsbaek, 2010: 3)? The interview guide deliberately asked questions that interviewees were unlikely to have been asked previously. These included the following: 'If you were to tell the story of your life, what title would you give it?', 'Are there parts of your war story that are important to you and which you are never asked about?' and 'Do you think that being a man/woman has influenced how you deal with challenges and adversity in your life?' The researchers, all of whom have extensive fieldwork experience in BiH, Colombia and Uganda respectively, conducted the interviews in the relevant local language/s (the author undertook all of the interviews in BiH). This further contributed to building a narrative space, by helping to put interviewees at ease. Highlighting this, the overwhelming majority of interviewees commented very positively on the interview experience (the interview guide included a question specifically about this). An interviewee in BiH, for example, told the author: How did I experience it? I love, I love that at least someone is fighting to carry this out . . . For people to know what happened, for the world to know, for everyone to hear so that, if at all possible, it does not happen to anyone ever again. (Author interview, BiH, 20 March 2019) An interviewee in Colombia explained to the in-country researcher that having no one to talk to meant that she had long had a lump in her throat. She continued: All of the interviews were recorded using fully encrypted voice recorders and NVivo software was used to code them. The codebook -developed by the author over a period of several months -continues to undergo revision and refinement as new ideas take shape. Using thematic analysis as the analytic method (Braun and Clarke, 2006), the emphasis has been on data-driven -as opposed to theory-driven -codes (DeCuir-Gunby et al., 2011: 137), and on identifying themes that help to make sense of a highly complex and unique dataset. While the development of data-driven codes is important for allowing the data to 'speak', it is not an atheoretical approach. Specifically, thematic analysis extends beyond the semantic level to include the latent level; and in latent thematic analysis, 'the development of the themes themselves involves interpretative work', meaning that the analysis is 'not just description, but is already theorized' (Braun and Clarke, 2006: 84).
While the research on which this article draws is not specifically about vulnerability or space, its focus on three diverse case sites (which are also internally diverse) -and on examining interactions between individuals and their wider environments and ecologies -has highlighted how vulnerability can move and shift in relation to space. In this regard, a useful analogy can be made with magnetic resonance imaging (MRI) technology. Fundamentally, if vulnerability is likened to protons (the nuclei of hydrogen molecules) in the body, different spaces exert different magnetic fields that force those protons to line up in different ways, spinning on their axes. These spins 'can be made to flip from being aligned against, to being aligned with the field, by applying a burst of a magnetic field that oscillates at the same frequency at which they are spinning' (Gowland, 2005: 176). In other words, vulnerability, like protons, responds and reacts to environments. Before exploring this empirically, it is important to look at how existing scholarship on vulnerability has theorized and discussed the concept.

Vulnerability, Situations and Spaces
According to Brown et al. (2017: 505), 'the ubiquity and elasticity of vulnerability generates a sense of familiarity and common-sense or assumed understandings which conceal diverse uses with enormously varied conceptual dimensions'. Notwithstanding these diverse uses, one common starting point is to approach vulnerability as an intrinsic part of being human. Fineman (2008: 1) maintains that 'vulnerability is -and should be understood to be -a universal and constant, inherent in the human condition'. To conceptualize vulnerability in this way foregrounds our corporeal vulnerability, and 'the inherent and continuous susceptibility of corporeal life to the unchosen and the unforeseen' (Harrison, 2008: 427). These innate vulnerabilities, moreover, become more enhanced in a globalized world where '[t]he vulnerable self depends upon others to live' (Ecclestone and Goodley, 2016: 180). While this interdependence exposes the 'porosity of our embodiment' (Munro and Scoular, 2012: 196), the increasing porousness of territorial boundaries contributes to the biopolitical blurring of biological and political bodies (Muller, 2004: 55).
It is precisely because human vulnerability exists within a broader global and political context that a blunt theorization of vulnerability that viewed all bodies as equally vulnerable would necessarily dilute the analytical and social utility of a concept that has obvious policy dimensions (Brown, 2017: 425). Scholars who emphasize a universal condition of vulnerability, therefore, also acknowledge more specific forms of vulnerability that exist alongside it. Fineman (2008: 10), for example, stresses that 'Undeniably universal, human vulnerability is also particular: it is experienced uniquely by each of us and this experience is greatly influenced by the quality and quantity of resources we possess or can command' (see also Lange et al., 2013: 333-334). The broader point is that there are different and multiple sources of vulnerability (Cannon, 2008: 355;Rogers et al., 2012: 12), which, in turn, draw attention to the concept's situational dimensions.
According to Rogers et al. (2012: 24), situational vulnerability refers to 'vulnerability that is context-specific, and that is caused or exacerbated by the personal, social, political, economic, or environmental situation of a person or social group'. Its duration can range from short term and intermittent to long term and enduring. The authors give the example of a family that is living in emergency accommodation after their home was damaged by flood waters. While their situation potentially makes them vulnerable, this vulnerability will be only temporary if, for example, they receive government assistance and have home insurance. However, 'if they live in a poor country, with no or limited support from government or non-government agencies, the effect of being displaced from their home may be catastrophic and enduring, rendering them more or less permanently vulnerable' (Rogers et al., 2012: 24). Situational vulnerability is thus a useful concept that has important ecological elements. It emphasizes the situations and environments that contribute to fostering vulnerability, and the relationship between vulnerability and resources. While the concepts of situational and spatial vulnerability overlap, it is argued that there are two key reasons for giving greater attention to the relationship between space and vulnerability (as opposed to simply situational vulnerability). One is more general and the second is specific to the issue of conflict-related sexual violence.
Turning to the first reason, Saatcioglu and Corus identify two core approaches to vulnerability within consumer research. One of these emphasizes underlying and systemic factors, meaning that 'certain populations (e.g. the homeless, ethnic minorities, poor elderly) that face various co-existing disadvantages (e.g. low-income, lack of job, lack of cultural acculturation, social stigmatisation, cultural and social exclusion) are multiplicatively vulnerable' (Saatcioglu and Corus, 2016: 232). The second approach conceptualizes vulnerability as a dynamic and shifting state (Saatcioglu and Corus, 2016: 232). These approaches, however, are arguably two sides of the same coin. Broadly, while the former accentuates situational vulnerabilities that may require complex and multi-dimensional solutions, it does not give the full picture and exists alongside more fluid and diversified spatial vulnerabilities that tell their own story.
Situational vulnerabilities, for example, may demand crucial external intervention. Lange et al. (2013: 336) argue that aid workers operating in areas that are prone to natural disasters, such as earthquakes or flooding, are situationally vulnerable. If these aid workers 'are regularly rotated out of that environment and into more secure areas, then their situational vulnerability will be intermittent', in contrast to the more enduring vulnerability of the villagers who are permanently living in these areas (Lange et al., 2013: 336). In other words, the aid workers are 'rescued' from a situation of vulnerability, whereas the villagers -without anyone to help them -are left situationally vulnerable. However, this narrative overlooks possible spaces within the overall situational vulnerability where the villagers' vulnerability may be highly fluid -and where they themselves are seeking to address and manage their vulnerability. Emmel (2017: 464) emphasizes that 'an explanation of vulnerability must include what people "do" to access, or to fail to access, resources and opportunities'. This 'doing' aspect of vulnerability can be overlooked when the focus is solely on broader situational vulnerabilities.
The second reason is that even if all of us are corporeally vulnerable, there is limited space for certain bodies to be acknowledged as vulnerable. Men, for example, can struggle with body image issues and develop eating disorders, yet anorexia and bulimia in men (and boys) remain under-diagnosed (Strother et al., 2012: 349). In a very different context, Ehrenreich describes how, after being diagnosed with breast cancer, she was denied the space to be and to feel vulnerable. She notes that 'in the seamless world of breast-cancer culture . . . cheerfulness is more or less mandatory, dissent a kind of treason' (Ehrenreich, 2001: 50). In contrast, the vulnerability of certain bodies is heavily emphasized, particularly bodies that have been sexually violated (Bergoffen, 2003: 121). In situations of war and armed conflict, victims-/survivors of sexual violence (and specifically women and girls) are often placed 'at the top of the victim hierarchy especially in relation to vulnerability' (Walklate, 2011: 188). This hierarchization of vulnerability, however, suggests that vulnerability is fixed and static; those at the top of the hierarchy remain there. Bringing a spatial dimension to discussions about vulnerability and victims-/survivors of sexual violence is important for demonstrating that vulnerability is not a closed space. Even though the term has its etymological roots in the Latin word vulnus or 'wound', it is significant that 'vulnus is also, even if a wound, an opening, and this opening can symbolize the opportunity that vulnerability can offer' (Davenport and Hall, 2011: 181).

Living Vulnerability
Notwithstanding the wealth of vulnerability scholarship, Brown et al. (2017: 506) note that: 'Much research focuses on theoretical debates and policy critiques and there are many fewer accounts that centre the empirical realities of vulnerability from the perspectives and experiences of various stakeholder groups, such as practitioners, service managers and service users/clients.' In the resultant 'gaps between theorisations and lived experiences of vulnerability' (Brown et al., 2017: 506), spatial dimensions of vulnerability are easily overlooked. This article, in contrast, foregrounds these dimensions through the novel use of a spatial lens. In their work on climate change, O'Brien et al. (2007: 75) refer to the concept of 'outcome vulnerability', which they define as 'a linear result of the projected impacts of climate change on a particular exposure unit (which can be either biophysical or social)'. Transposing this idea to conflict-related sexual violence, the crucial point is that if vulnerability is a 'linear result' of individual trauma, it thus has little room to move. The purpose of this section, therefore, is not to demonstrate that victims-/survivors of conflict-related sexual violence are vulnerable but, rather, to empirically explore how vulnerability is experienced in particular social, political and cultural spaces.
In his reflections on the ontology of space, Newman (1989: 211) identifies three categories: '(i) there are points of space; (ii) there are shapes, or more generally, spatial attributes, relations and so on; and finally (iii) there is space itself'. He terms this space 'the void' (Newman, 1989: 210). Regarding the relationship between categories (ii) and (iii), he submits that 'Shape concepts are independent of where they are applied, so they are certainly independent of the void' (Newman, 1989: 211). Similarly focusing on categories (ii) and (iii), this article's core argument is that vulnerability can be theorized as a shape that changes and shifts in relation to -rather than independent of -different spaces. Contrary to the position that 'vulnerability as a social concept addresses social rather than spatial dimensions' (Bohle, 2002: 3), this research submits that different spaces create or enhance particular vulnerabilities. To return to the earlier MRI analogy, just as protons communicate crucial information about the body, vulnerabilities communicate important information about the spatial environments in which they develop and with which they interact. In other words, vulnerability is an ecological concept that dialogues and interacts with different spaces within the systemic whole.
This section will use data from the interviews conducted in BiH, Colombia and Uganda to examine how vulnerability is intrinsically linked to particular spaces. In so doing, it demonstrates that vulnerability is a multi-dimensional concept 'encompassing physical, social, economic, environmental, and institutional features' (Roy and Blaschke, 2015: 24). It is these features, in turn, that constitute important spatial aspects of vulnerability, thus underscoring the need for fine-grained relational analyses of the concept. In Colombia, for example, some interviewees expressed feelings of physical vulnerability that necessarily had an important situational dimension, reflecting the longevity of the armed conflict and the fragility of the 2016 peace agreement between the Colombian government and the Revolutionary Armed Forces of Colombia (FARC) (Taylor, 2019). However, they also articulated more localized and context-specific physical-spatial vulnerabilities. Some were living in spaces where armed groups remain active. One interviewee revealed that 'where I live now is an area very . . . you have to know how to behave; you have to know how to talk because of the [armed] groups and organized gangs' (researcher interview, Colombia, 12 March 2019). Some interviewees were social leaders who, through their work and support for the peace agreement, were effectively challenging the control that armed groups exercise over certain spaces (Human Rights Watch, 2019). In this way, they had accrued additional spatial 'layers' of vulnerability (Luna, 2009: 128-129) within a more general situational vulnerability. Underscoring the reality of ongoing attacks on social leaders and human rights activists (Grattan, 2019), one interviewee explained: 'we are still threatened, we continue to be victimized. At any moment, well, we're in danger . . . Every day, people get killed and people are threatened' (researcher interview, Colombia, 4 February 2019). Because their work had made them vulnerable within public spaces, some of the Colombian research participants were registered in State protection programmes that necessarily impacted on their private lives. This is an illustration of 'the diminishing boundaries between public and private spaces' (Strauß and Nentwich, 2013: 725) -and of the ease with which vulnerability can move across such spaces.
Broad situational vulnerabilities also acquire a more acute spatial dimension when perpetrators continue living freely, leaving their victims fearing for their own safety. An interviewee in BiH reflected: In 1994, if they had locked them up, I would have had the strength then and I would endure all this much easier . . . Because, like this, you are afraid. If they were able to do it, then anyone can. Well, this is the hardest pain for a victim, for victims to be . . . I mean, for perpetrators to be left unpunished, and then the victim is afraid that [crying] they will do the same, or even worse. (Author interview, BiH, 2 June 2019) Within international discourse on conflict-related sexual violence, 'ending impunity' is a recurrent leitmotif (see, for example, United Nations Security Council, 2017: 5). In reality, it is impossible to prosecute all perpetrators, not least due to the enormous amount of time and resources that this would require and the reality that prosecutors necessarily have to make difficult decisions about which cases to prioritize and pursue (see, for example, Del Ponte, 2006). Hence, some impunity will always exist, and this can alter the use of public spaces and the feelings that they evoke (Hagberg, 2002: 233). In other words, alongside general situational vulnerabilities that exist in societies recovering from war/armed conflict and mass human rights abuses, including livelihood vulnerability (Maconachie et al., 2012), there are also more specific spatial vulnerabilities that critically shape the everyday living of vulnerability and the ways that individuals engage with and utilize their environments.
In Uganda, interviewees commonly expressed feelings of vulnerability in the context of community spaces. One interviewee described how her time in 'the bush' -referring to her period of captivity after being abducted by the Lord's Resistance Army (LRA)had made her vulnerable to verbal abuse in a cultural space where 'strong beliefs' exist that 'sex "in the bush" is inappropriate and carries negative cosmological consequences' (Porter, 2015: 87). She talked about 'cimo tok' (literally 'pointing at the back of the head') from members of the community to which she had returned. Her time in the bush, as both a physical and experiential space, had also made her young body vulnerable. She had become infected with the HIV virus, exposing her to new socio-economic vulnerabilities in a space where she no longer had the physical strength that she once did to dig her land (researcher interview, Uganda, 20 March 2019). This example illustrates how the use of a spatial lens can illuminate different and intersecting spatial vulnerabilities within a wider context of situational vulnerability that is partly a legacy of the protracted war in the north between Ugandan government forces and the LRA (Meinert and Reynolds Whyte, 2017).
In other words, the concepts of situational and spatial vulnerability tell different narratives operating at different levels. Focusing broadly on narrative text, Zoran (1984: 333) notes that 'space is not a neutral material just existing in the world; it has various functions relating to other planes of the text. Every element in space -actually every element in the text -has to be regarded.' Similarly, spatial forms of vulnerability have a function in the sense that they add contextual nuance and thickness to broader narratives of situational vulnerability. Just as vulnerabilities can travel across and between spaces, for example, different experiential spaces can also blur and merge in ways that create 'amplified causal looping' (Beck, 2011: 304) and new cross-temporal vulnerabilities. As one illustration, an interviewee in Colombia expressed a sense of situational vulnerability vis-a-vis her economic situation and general loss of trust in others. However, she also evinced a more specific spatial vulnerability in relation to her marriage and own body. Three years earlier, she had gone through an extremely traumatic labour, and different experiences of pain had become confused in her mind. Describing the labour, she reflected: It was like I was living . . . like I'd been taken years back to when . . . when the incident of sexual violence took place . . . My health was, well, it was to do with my private parts. I was reliving the pain I had experienced then, I mean, it brought back memories . . . It was 17 years ago [referring to the sexual violence] and well, the pain I had felt 17 years ago had been in my vagina and after the birth, it came back . . . I thought . . . in my mind . . . the wires were all crossed and I thought the pain was from that sexual violence and not from the labour. (Researcher interview, Colombia, 11 February 2019) This 'crossing of wires' had, in turn, affected her conjugal relationship. Within this very personal and intimate space where particular expectations existed, she felt physically and psychologically vulnerable and would not allow her husband to touch her. She would scream and push him away. More than a year passed before she was able to resume having a sexual relationship with him. In this example, thus, the interviewee expressed vulnerability linked to the blurring and merging in her own mind of different experiential and temporal spaces. In short, her vulnerability was located within a 'multidimensional space' that complexified the aetiology of vulnerability, drawing attention to 'a temporal ordering whereby an earlier event determined a later one' (Armstrong, 2012: 189).
Further illustrating a temporal-spatial dynamic, interviewees in BiH frequently demonstrated vulnerability in navigating the spaces of their past and present, highlighted by the large number of interviewees who talked about relying on psychotropic drugs such as benzodiazepines. Explaining how she copes on a daily basis, one interviewee disclosed that: 'I take a pill -thank God that you can take them without a prescriptionand I drink it, calm myself' (author interview, BiH, 6 March 2019). Another told the author: 'I take a Lexilium [Bromazepam] tablet . . . Almost every day' (author interview, BiH, 3 May 2019). This vulnerability qua dependency has broader spatial dimensions. BiH's experience of all-out ethnic conflict created a ready environment for what Pupavac (2004: 378) has termed 'international therapeutic governance'. A model that 'pathologizes war-affected populations as psychologically dysfunctional' (Pupavac, 2004: 378), it has encouraged medicalized forms of coping that have taken hold in socio-political spaces where many individuals -and not just victims-/survivors of sexual violence -have become deeply jaded and struggle to see any positive change more than 20 years after the Bosnian war ended (Arnautović, 2011). Interviewees who were still internally displaced within the country exhibited additional layers of vulnerability, complaining that community spaces did not always accept them. In the words of one interviewee, for example: We thought, when we arrived from the Serb territories [meaning Republika Srpska], that we would be welcomed with full hearts, but you saw immediately the rejection and . . . Even today, they say 'refugees'. 6 If something is stolen, it was the 'refugees'; if anything -the 'refugees'. (Author interview, BiH, 3 February 2019) All of the above empirical examples highlight vulnerability in relation to different types and levels of space. Highly significant in this regard, thus, is Brown et al.'s (2017: 498) observation that: 'How vulnerability is deployed in research is to some extent contingent on the historical, political and disciplinary context in which the concept is utilised.' These different macro-contextual dimensions, which are also spatial, help to shape the discursive spaces in which vulnerability is experienced and narrated. Interviewees in all three countries, for example, commonly expressed emotional vulnerabilities, from fear and distrust to feelings of sadness, emotional hurt and self-blame. However, there were also some important contextual differences. Particularly evocative in this regard was the strong contrast that emerged between Bosnian and Colombian interviewees when they were asked: 'If you were to tell the story of your life, what title would you give it?' Bosnian interviewees frequently responded in a way that reflected their emotional vulnerabilities. One interviewee gave her life story the title: 'I'm sorry that my life passed in sadness and suffering' (Žao mi je što je moj život prošao u tuzi i patnji). Another chose the title: 'From 1992, my life story is a sad story' (Od 1992, moja životna priča je tužna priča). After thinking about it for a few seconds, a third interviewee answered: 'Not to be alone' (Da nisam sama). Colombian interviewees often gave their life stories more positive titles. These included 'Woman warrior' (Mujer guerrera), 'My new dawn' (Mi nuevo amanecer) and 'Rebirth' (El renacimiento). The purpose of these examples is not to construct simplistic dichotomies between victims-/survivors in different parts of the world, but simply to emphasize the significance of broader social and discursive spaces in which vulnerability is lived and narrated.
Based on the author's years of experience of working in BiH, it is argued that very little discursive space exists for talking about victims-/survivors without emphasizing the common vulnerability themes of trauma, unmet needs and persistent everyday problems (Clark, 2019). The issue of conflict-related sexual violence remains highly politicized in BiH -particularly within the Federation -and is spatially situated within a broader identity politics, defined by a 'logic of competing victimhoods' (Helms, 2013: xi) and construction of rival ethnic narratives (Clark, 2014: 205). In this context, speaking about the challenges that victims-/survivors have overcome, how they deal with adversity or the goals that they might wish to achieve is politically far less expedient than reiterating how these women have suffered -and continue to suffer. As such, they serve as useful 'victimization markers' (Basic, 2015: 26).
In Colombia, different contextual factors -including the nature of the armed conflict and its longevity -have contributed to creating discursive spaces that encourage a culture of resistance, and more specifically women's resistance (Murdock, 2008). Several of the Colombian women who took part in the research were active in networks such as Ruta Pacifica de las Mujeres (Pacific Route of Women) and the Red de Mujeres Víctimas y Profesionales (Network of Women Victims and Professionals). The networks have helped some women -by giving them opportunities to gain a political education and to learn more about their rights -to take on new roles as social leaders, political activists and human rights defenders. According to one interviewee, for example, Thanks to the many organisations that have helped us . . . with that subject, I said: 'right [short pause], I have to take control', and so I've gained strength and that has helped me to go on and get some support for these other women too. So, I use my experience to help other women [who are also victims-/survivors of sexual violence]. (Researcher interview, Colombia, 4 February 2019) The fact that social leaders in Colombia continue to be targeted (Tamayo, 2018) means that leadership roles often entail new spatial vulnerabilities. These vulnerabilities, however, have an agentic dimension. They result not simply from what has been done to these women, but also from what they are actively doing to try and change the social and political spaces in which they live.
The above example highlights an important 'generative' dimension of vulnerability, linked to the creation of new realities (Ehlers, 2014: 116). This generative aspect is further evidenced in the various ways that some individuals actively engage with -or utilize -the spaces around them (and the resources within those spaces) as a way of managing or seeking to overcome their vulnerabilities. In BiH, a male interviewee conveyed a strong sense of emotional vulnerability, emphasizing his sense of shame and inability to talk to his family about the sexual abuses that he endured in a camp during the Bosnian war. However, he also stressed the importance of a nearby lake as 'something that keeps me going here'. When asked to elaborate, he explained: 'Well, I rest [here]. I rest mentally. Like this, I observe the ducks, fish. Pigeons come along. I feed the ducks. Like that, and I don't think about problems' (author interview, BiH, 10 April 2019). In Uganda, a male interviewee talked about a local river that no longer provided him with an income; the water had dried up. He was nevertheless trying to overcome his financial vulnerability by using the spatial resources around him, such as rock outcrops from which he took chunks of rock and processed them into coarse aggregate for sale (researcher interview, Uganda, 22 February 2019).
The concept of 'outcome vulnerability' (O'Brien et al., 2007) was previously introduced to problematize the notion that victims-/survivors of conflict-related sexual violence are vulnerable. This section has empirically explored the relationship between vulnerability and particular spaces, some of the spatial dynamics of the concept and the different ways in which vulnerability is experienced and managed. Taken together, the empirical examples illustrate how space plays a distinct role in the ecological legacies of conflict-related sexual violence. These are legacies that impact, inter alia, on families and communities, on systems of health, security, justice and education, and which affect 'the interrelationship of organisms and their environments' (Harvey, 1996: 5). Accordingly, spatial thinking about vulnerability has important implications for policy and, in particular, for the field of transitional justice. Spatial dimensions of vulnerability underscore the need for an ecological reframing of transitional justice theory and practice that gives more attention to vulnerable spaces. If spaces play a crucial role in generating and fostering vulnerability, the challenge is to transform these spaces as part of building spatial resilience. Nyström and Folke (2001: 407) define spatial resilience as 'the dynamic capacity to cope with disturbance and avoid thresholds at spatial scales larger than individual ecosystems'. Coral reef systems, for example, are vulnerable to diverse threats, and even individual reefs with high resilience become vulnerable when significant disturbances occur. In such situations, 'local ecological functions can only be maintained by resilience of a matrix of reefs in the seascape' (Nyström and Folke, 2001: 410). In other words, spatial resilience is about ensuring that small changes or disturbance events do not result in thresholds that fundamentally and irrevocably alter an ecosystem (Allen et al., 2016: 634); and about recognizing the 'dynamic interactions and interdependencies' between different systems (Nyström and Folke, 2001: 407), which together can reduce the impact of shocks and stressors.

Conclusion: Towards Spatial Resilience
When vulnerable spaces are ecologically conceptualized as forming part of broader systems 'with spatially located components, flows, interactions, and perturbations' (Cumming, 2011: 900), this foregrounds important intersections between vulnerability and resilience that remain critically unexplored in the context of conflict-related sexual violence. A key explanation is the marginalization of space and social ecologies. International policy discourse on conflict-related sexual violence places a strong emphasis on preventing and eliminating this scourge of war (see, for example, European Union, 2018; Hague, 2014;United Nations General Assembly, 2015;United Nations Security Council, 2018). Little attention has been given to spaces of vulnerability or to the transformation of these spaces. 'Transformation', in this regard, means ensuring that the possible 'cushioning' resources within these spaces -including support systems -are strengthened and fully utilized (Walsh, 2007). These resources are crucial for building social 'matrices' across systems that can help to reduce the destructive impact of conflict-related sexual violence.
From this perspective, it is not sufficient to stress the need for a 'survivor-centred approach' (see, for example, United Nations Office of the Special Representative of the Secretary-General for Sexual Violence in Conflict, 2018; United Nations Security Council, 2019). There is an imperative need for multi-level ecological approaches that view individual victims-/survivors in the context of their broader environments and the spaces that they inhabit, individually and collectively (Harvey, 2007). The example of stigma powerfully illustrates this point. Those who have suffered conflict-related sexual violence may lose their means to earn a living, become socially ostracized and/or feel unable to seek medical/ psycho-social help (Albutt et al., 2017;Christian et al., 2011). In BiH, 'More than two decades after the end of the conflict, survivors of wartime sexual violence continue to suffer the enduring effects of that crime, including socioeconomic marginalization and stigma' (United Nations Security Council, 2017: 22). In Uganda, some of the research participants had suffered additional stigma as a result of becoming infected with the HIV virus. A survivorcentred approach focused on individual needs will necessarily be deficient if it does not also address the wider social spaces that critically shape needs and vulnerabilities.
This research is accordingly advocating for a novel reframing of transitional justice that gives more attention to ecological legacies. As part of this reframing, the challenge is not only to ensure that victims-/survivors are and feel supported, but also to strengthen and enhance 'the ability of spaces to cope with diversity and change' (Kärrholm et al., 2014: 122) -and to thereby foster spatial resilience. In this regard, resilience should not be represented 'as merely an antidote to a prior problem [i.e. vulnerability]' (Furedi, 2008: 657). Rather, spatial resilience potentially offers a new heuristic framework for thinking about transitional justice and the realization of its key goals, such as peace, justice and reconciliation. Ultimately, responding to conflict-related sexual violence is not only about prosecuting the perpetrators, delivering 'justice' to victims-/survivors and addressing their needs. More systemically, it is also about dealing with the ecological legacies of sexual violence and giving attention to the spatial interactions between individuals and their wider environments. This is essential for addressing 'socio-spatial segregation' (Nail, 2018: 50) that is detrimental to the realization of core transitional justice goals.

Acknowledgements
I would like to thank three anonymous reviewers for their detailed comments and highly constructive feedback on this article.

Funding
The author disclosed receipt of the following financial support for the research, authorship and/or publication of this article: this work was supported by the European Research Council under grant number 724518.