A predisposed view: State violence, human rights organisations and the invisibility of the poor in Nairobi

The article examines the challenges faced by human rights organisations in documenting cases of torture, focusing on the particular example of Kenya. The analysis is situated in the context of both widespread human rights violations and a vibrant human rights community in Kenya. There is considerable evidence that the urban poor are particularly vulnerable to torture and ill-treatment. However, the article suggests that human rights organisations often fail systematically to document the experiences of survivors living in conditions of poverty. The empirical material for this article was produced during three stages of data collection between May 2014 and September 2016, including a household survey examining AFRICAN HUMAN RIGHTS LAW JOURNAL

exposure to torture and ill-treatment in an informal settlement and indepth interviews with human rights practitioners, survivors and members of the community. The authors focus on the particular case of torture and ill-treatment rather than wider forms of state violence, such as extrajudicial killings, although the two often overlap. The aim is not to provide a legal analysis of mechanisms around the prevention of torture and illtreatment. Rather, it is to provide a sociological and anthropological analysis of the obstacles to the effective human rights documentation of violations experienced by the urban poor. The article argues that three

Introduction
The body of Willie Kimani, a human rights lawyer representing a person who had laid a complaint against the police, was found in a river one week after he had disappeared along with his taxi driver and his client. Kimani and his client were on their way back from a court appointment. Kimani, his client and their driver were tortured before being killed. Many human rights organisations put all else aside and organised a demonstration demanding that the authorities stop extrajudicial killings. The demonstration featured the dramatic use of redstained white T-shirts and three white coffins. This occurred in July 2016, and arguably was the first time that human rights organisations had gone to the streets on such a scale since the promulgation of the progressive Constitution in 2010. According to some activists, it was time for Kenyans to 'take back our streets'. The Law Society of Kenya told BBC News that Kimani's death was a 'dark day for the rule of law in Kenya'.
Three years previously Dennis was the victim of a vicious attack by the police. In 2013, around the time of the national elections, he was arrested with three of his friends. At the time of the arrest, Dennis was 29 years old and lived in South Eastleigh, working as a casual labourer. The young men were taken to different police stations. Dennis was tortured 'to say what he didn't know', as he put it. He was also injected with some substance. Dennis surmised that the purpose of the arrest and the ill-treatment was to intimidate the young men into not participating in any public demonstration. After a few days, Dennis was taken to hospital, treated for his injuries at no cost (a social worker helped him waive the bill) and sent home in his bloody clothes and with KSH 200. At the time he was interviewed as part of this research project in September 2015, he still was not able to walk. Dennis's case did not become a nationally-recognised human rights case, even though his case was later picked up by local paralegals who made contact with a prominent human rights organisation. However, Dennis was fortunate. The three friends with whom he had been picked up, who also were young and underemployed members of the urban poor, all died in police custody.
There are many differences between the two cases above. However, what interests us in this article are the differences in terms of how human rights organisations dealt -and did not deal -with them, and what that might tell us about the potential and limits of human rights work in conditions of entrenched poverty and violence. We want to explore why human rights organisations 'missed' Dennis's case. We suggest that this was not simply a case of 'bad luck', but that it occurred in the context of a wider predisposition which means that human rights organisations systematically fail to document the forms of torture and ill-treatment experienced by the poor. By predispositions we mean the existence of a number of structural and conceptual factors working against cases such as that of Dennis being followed up by human rights organisations. This is an issue that goes well beyond Kenya. As the UN Special Rapporteur on Extreme Poverty and Human Rights has stated: 'When the situation of people living in poverty is addressed in development or human rights frameworks, their civil and political rights are often completely ignored' and, as a result, human rights work can fail to 'address the distinctive ways in which people living in poverty are affected by police brutality'. 1 While human rights organisations are aware that their documentation of state violence only represents the 'tip of the iceberg' and, more particularly, leaves out the experiences of many people living in poverty, they can find it hard to address the gap. 2 This article attempts to identify some possible reasons as to why that is the case. We ask what the obstacles are that prevent the poor from being 'seen' by human rights organisations, and how the inability to see them impacts on human rights work in urban Nairobi. This analysis is situated in a context of both widespread human rights violations and a vibrant human rights community in Kenya. 3 The violations by successive governments, since colonial times, through the Kenyatta and Moi regimes and to the present, as well as the history of the Kenyan human rights community, have been While this analysis often is insightful, it is also important to acknowledge the role that civil society organisations -often with large elements of local ownership -play in the process of strengthening democracy, the rule of law and social justice. 6 This broader discussion of Kenyan human rights NGOs informs our analysis, but we want to adjust the terms of engagement to focus explicitly on the forms of violence, the types of perpetrators and the types of victims that human rights documentation brings into the picture. Hence, we want to explore and understand human rights through their practices of 'seeing' and counting human rights violations. Our goal here is not to serve as an indictment of human rights organisations in Kenya. All the human rights organisations interviewed for the purposes of this article achieved a great deal with limited resources and under great pressure -and most often were brilliantly committed. What we suggest instead is that the social and spatial position of human rights organisations, their focus on particular forms of accountability, and their understanding of appropriate forms of victimhood, at a practical level, stood in the way of them being able to 'see' the poor. 7 This might be put differently: It was not simply the fact that an act of violence had occurred that helped it to become a human rights issue. Instead, for that to happen, a range of different features had to be in place. In particular, the victim had to be relatively easy to reach for human rights groups, and be seen as a viable case for criminal prosecution or other forms of legal accountability. Willie Kimani lived up to these criteria; Dennis did not. 8 At this point an important conceptual issue needs to be addressed and unpacked, namely, the issue of 'seeing'. Analytically, and in human rights terms, what does it mean to 'see' something? As  Kenya, 1992Kenya, -1997Kenya, (1999 indicated above, whether a particular case is seen as a human rights violation is not merely a matter of it being made known. Dennis's case was known to paralegals and they had informed a national human rights organisation. However, it still failed to become a case that human rights organisations were able to act upon. One way to approach this is through Cohen's seminal work on what he calls 'states of denial'. 9 Cohen's notion of denial involves what he calls a moral paradox of knowing and not knowing at the same time. For Cohen, denial is both an individual and a social practice of not wanting to see, or not being able to see and turning a blind eye. For Cohen, these forms of denying presume knowledge about the violation at some level. The human rights organisations that were interviewed for this research study know in some specific way, but cannot or will not (often for perfectly good reasons) respond in any great measure to the violations. In her insightful review, Moon suggests that human rights documentation and advocacy are closely interlinked because knowledge must always be tied to advocacy within human rights organisations. 10 One therefore needs to ask what 'seeing' implies and why certain violations can be seen and acknowledged while others are not substantially recognised by human rights actors. We focus on the particular case of torture and ill-treatment, rather than wider forms of state violence, such as extra-judicial killings, although the two can often overlap, and indeed they did in the research. Our analysis, however, not only concerns torture and illtreatment as defined in international legal frameworks. In real life, on the ground for victims and for local human rights organisations, it often is difficult to distinguish between acts of torture and killing, as the case of Dennis indicates. In this way, we use the case of torture and ill-treatment as a privileged access point for understanding larger issues around documenting and intervening in relation to human rights violations against the poor. What follows is not a legal analysis of mechanisms around the prevention of torture and ill-treatment. Rather, it provides a sociological and anthropological analysis of the obstacles to the effective human rights documentation of violations experienced by the urban poor, with a particular focus on the human rights categories of torture and ill-treatment.
The remarks above all assume that the poor are particularly vulnerable to torture and ill-treatment. We often associate torture and ill-treatment with particular forms of violence, involving state officials dragging political activists into dark dungeons and torturing them. While these forms of violence certainly occur, as Kimani's case illustrates, there are also many incidents that count as torture and ill- treatment under the UN Convention Against Torture (CAT) that are more 'everyday' and 'mundane' occurrences, linked, for example, to extortion and harassment of poor and marginal groups by police officers or other public officials. 11 According to a recent report by our research partner in Nairobi, the human rights organisation Independent Medical and Legal Unit (IMLU), close to one in every third Kenyan (30 per cent) has been subjected to ill-treatment during the past four years. 12 This is an increase of almost 30 per cent since the similar national survey conducted in 2011. 13 The police committed 78 per cent of these violations. Importantly, only 28 per cent of the incidents were reported to a public authority. The low levels of reporting state violence to human rights organisations were also established in a survey conducted in three informal settlements in Eastern Nairobi. 14 We return to the survey below. The empirical material for this article was produced during three stages of data collection between May 2014 and September 2016. The research study began with a mapping of the organisations engaged in documentation of torture and ill-treatment in Nairobi. On the basis of the mapping, we interviewed staff from nine Kenyan NGOs, three international non-governmental organisations (INGOs), one civilian-based organisation (CBO), one state organisation and one media organisation about the aims, results and difficulties in their documentation practices. The second stage involved a quantitative survey in one low-income area in Eastern Nairobi. This particular area was selected as IMLU, the partner in the research, had connections to a paralegal organisation. This connection proved essential to ensure access to the informal settlement areas. Designed as a householdbased victimisation survey, it covered socio-economic issues, social capital, exposure to torture and ill-treatment, perceptions of risk of torture and ill-treatment, and justice-seeking behaviour. A team of five enumerators assisted by five paralegals carried out 500 structured 11 While the Convention Against Torture distinguishes between torture and cruel, inhuman and degrading treatment, all violations meeting certain criteria fall within the Convention. In this article we refer to cruel, inhuman and degrading treatment under one as ill-treatment. Hence, in our analysis ill-treatment is also a legal category under CAT. In practice, this will always involve the participation, consent or acquiescence of a public official. interviews during March and April 2015. 15 The third stage consisted of interviews with identified victims among respondents such as Dennis. In this way, the qualitative interviews were embedded within the survey and not accidentally identified. The combination of these research techniques allows us to produce different perspectives on experiences of torture and ill-treatment, and triangulate against the information produced through human rights documentation. We organise our argument in this article in five parts. After this introductory part, we describe in more detail the context in which the research has been carried out with regard to violence and the human rights framework, in order to map out what we consider to be the under-perception by human rights actors of state violence against the poor. In parts 3 to 5 we explore in turn each of the three predispositions we have identified, namely, socio-spatial and institutional predispositions; a predisposition defined by privileging legal accountability; and a predisposition based on notions of who constitutes the good victim. In the final part we conclude our analysis by suggesting a need to consider in more detail how to overcome the predispositions and thereby deepen and consolidate human rights work in Kenya among the poor.

Human rights and violence in Kenya
The post-election violence of 2007 and 2008 constituted a watershed moment in Kenya. One of the effects of the violence was work on a new Constitution that was adopted in 2010 and arguably became one of the most progressive constitutions worldwide with a wide range of legal and institutional protective mechanisms. To a large extent, the Constitution also provided the perfect rallying point for many human rights organisations. Torture and ill-treatment were central issues dealt with during the discussions around the Constitution. Articles 25, 26 and 29 guarantee the right to life and the absolute prohibition of torture. Following the promulgation of the new Constitution, torture has been prohibited by the National Police Service Act 2011 (revised in 2014); the Kenya Defence Forces Act 2011; the National Intelligence Service Act 2012 (revised in 2014); the Chiefs Act 1998 (revised in 2012); and the Children's Act 2010 (revised in 2010).
However, despite the criminalisation of torture in these Acts, the ratification of the UN Convention Against Torture, and the wellcoordinated advocacy by human rights actors of an anti-torture Bill, the Prevention of Torture Act -which defines and criminalises torture and establishes a legal and institutional framework to support victims of torture -were only passed into law in April 2017. Furthermore, the government in 2013 passed a counter-terrorism Bill that gave the police wide-ranging powers to extract information. As the perceived and real threats from Al-Shabaab became more pronounced, more counter-terrorism legislation and state practices have been legitimised in the name of the war on terror. For instance, the police have made use of the counter-terrorism legislation in the security operation Usalama Watch where they arrested thousands of people in a Somalidominated suburb, and held more than 400 'suspects' in custody and charged approximately 70 persons with various offences. 16 The 2010 Constitution also led to the establishment of the Independent Policing Oversight Authority (IPOA) -a rare case of a civil oversight authority for the police in Africa. It was introduced with the new Constitution, with the mandate to carry out oversight of the entire police system. IPOA started operations in June 2012. It has the power to investigate; recommend prosecution upon investigation; monitor policing operations; review or audit investigations carried out by the Internal Affairs Unit of the NPS; conduct inspections of police premises, detention facilities; and also review patterns of police misconduct with a view to making policy and institutional changes. IPOA works closely with the National Police Service Commission (NPSC), another new institution with civilian influence on the recruitment, training, disciplining and promotion of police officers in Kenya. The Independent Policing Oversight Act (2011) stipulates that IPOA has the responsibility of investigating any death or serious injury suspected to have been caused by a member of the police. Following this mandate, it has published critiques of the police work and it can open closed police files to investigate the quality of evidence used in court cases. 17 In many ways, IPOA has been a success in terms of establishing oversight of the police. Despite the justified acclaim, IPOA is also vulnerable. It receives low government funding and relies heavily on donor support (bilateral aid agencies and the United Nations Office on Drugs and Crime (UNODC)) and employs a mere 25 investigators. 18 Police misconduct remains prevalent. In 2012 IPOA conducted a baseline survey with 5 082 households and 515 police officers across the country, which revealed that 33 per cent of the respondents had experienced police misconduct, including assault, falsification of evidence, bribery and the threat of imprisonment in a 12-month period. 19  police officers committing 26 per cent of the incidents. 20 Furthermore, and according to Kenyan human rights institutions and NGOs, in the past six years the number of extra-judicial killings involving Kenyan police and security forces is in the high hundreds. IMLU has recorded that the police killed 97 people in 2015. According to a social scientist working in Mathare, a large slum area in Nairobi, a conservative estimate is that at least one young man was killed every week between January 2013 and December 2015 in the area, and activists in Korogocho, another Nairobi slum notorious for crime, have estimated that police officers killed 25 young men in 2015 in that particular area. 21 These seemingly contradictory trends -intractable state violence and the development of a strong human rights culture -may be indicative of a Gramscian 'war of position' between conflicting senses of emergency; one that focuses on the need to curb state violence and one that focuses on the threats from crime and terrorism. Hence, while people in the cited surveys and reports are well aware of police violence and Kenyans generally view torture as 'bad', 57 per cent of the population approve of the use of torture in relation to national security. 22 Likewise, while many Kenyans would express horror at the extra-judicial killings of young men in the slums, the fear of crime often legitimises such slayings. These seemingly contradictory points of view were captured in interviews from our survey sites where women, mothers of endangered young men, expressed gratitude towards one particular police officer because of his rough style, including the execution of criminals, and because he issued warnings before shooting. 23 There is a vibrant human rights community in Kenya -particularly in Nairobi -and this community is well aware of the widespread human rights abuses perpetrated against the general Kenyan population, including the many living in urban slums. The question then arises as to why these same human rights organisations still have difficulty in 'seeing' violations against the poor such as Dennis. This is the subject of the following three parts.

Spatial, social and institutional predispositions
One way in which to start unpacking the predispositions against documenting torture against the poor is simply to map human rights organisations in relation to the location of the urban poor. Most national NGOs have head offices in Nairobi and operate through offices in the main towns (for instance, Mombasa, Nakuru and Kisumu) where they work with community-based organisations and local NGOs. Some organisations such as IMLU have a network of monitors to provide a link between victims and themselves. IMLU then refers victims to the nearest pro bono lawyer, doctor or psychologist. Within the human rights community, this network structure has the advantage that the monitors can operate in an informal manner so as not to attract too much attention. However, there are doubts as to the reach of this network. As a senior human rights activist puts it, 'it is not entirely clear when we say we have a network, what we mean by that. Who is in it and how strong are those relationships?' This uncertainty arguably is also a question of spatial make-up. The NGO offices are often located in middle-class areas, especially in the areas around Westlands and Kilimani. This is a considerable distance away from potential victims among the poor living in informal settlements. Many human rights organisations also rely on people coming to their offices to report violations. However, the cost of travel from Eastlands to Westlands and the time spent often constitute prohibitive barriers to reporting. As our survey indicated, very few people living in the slums would even know about human rights organisations, let alone where they are located. 24 For instance, when we carried out the survey people interviewed would not know of the existence of the paralegal organisation with which we worked but they would know the names of individual members, suggesting that people's relationships with structures often are personalised relations. Apart from the spatial outlay of Nairobi, space plays another role in that many human rights organisations literally are turned towards the UN in Geneva, the African Commission on Human and Peoples' Rights (African Commission) in Banjul and the 'international community', more broadly. As is the case with the geographical structure, this is hardly surprising. Most human rights organisations -and even state institutions such as IPOA -rely heavily on foreign funding. As part of the realisation of rights-based development, most foreign donors focus on the consistency between interventions and international conventions such as the UN Torture Convention, the Convention on the Rights of the Child (CRC), or the African Charter on Human and Peoples' Rights (African Charter). 25 Many national human rights organisations mirror the structure of international and regional human rights practices, and are highly specialised in specific areas of human Importantly, most human rights work does not take place with direct textual reference to international and regional human rights mechanisms. A major part of the documentation on torture and illtreatment on the ground is carried out by paralegals that are often, but not always, recruited from the communities within which abuses are taking place. However, some NGO staff stated that they faced a number of difficulties in relation to the paralegals. Although many human rights organisations might prefer to recruit from within affected communities, there are important challenges in doing so. For one, there may be few people with the desired skills living in the very poorest communities, and once they obtain the skills, they often move out. Furthermore, when paralegals live in the community they often face the risk of reprisal if they report abuse. A further key dilemma is the gap between the skills required to be a paralegal and the allowances available for paralegal work. NGOs that were interviewed preferred paralegals and other documenters with secondary education and a fluency in English reading and writing. Yet, the organisations can offer only low allowances. Besides the financial constraints, which are difficult to overcome as donors are reluctant to pay allowances at community level, some organisations admitted that there was little meaningful feed-back from the head offices to the 'people on the ground'. Both issues were confirmed by one of the paralegal units that we interviewed that referred cases to civil society organisations and received no feed-back or assistance on the cases, not even when it had been promised.
The issue of language points to one last structural issue relating to the inability to 'see', namely, class. In Nairobi language maps on to class in intricate ways with Kiswahili and English being the dominant languages of the elite. To say that class matters probably is not problematic, but what does it mean for the engagement of human rights organisations with the urban poor? We may use our own research process to illustrate some points. We decided to work in South Eastleigh as it was one of the places in Nairobi where IMLU had stronger relations through a paralegal organisation. We contracted an independent consultancy company to carry out the survey and agreed that the paralegal organisation would assign one member to each enumerator (of which there were five) to ensure the safety of the enumerators as well as to help establish contacts with residents who are justifiably wary of people with paper and pen. After conducting the survey, their experiences were collected in a survey process report. 26 The report describes language problems where English and Kiswahili were not sufficient and where enumerators needed translation from Kikuyu, Luo or Shen. It also describes the 'hostile residents'. For instance, one male enumerator reported: 27 Some respondents I approached in the field for interviews were unreceptive in that they did not give me a chance to explain to them what the research was all about but rather responded with sentiments that they are used to 'people who carry files around' [meaning researchers] and that they don't get any help after the surveys are conducted.
He, along with other enumerators, also discussed the harsh weather conditions, or probably the lack of air-conditioning, along with bad sanitation. However, the report states that '[t]he enumerators handled it [the sanitation] with grace'. 28 The enumerators carried out their work with relative competence and their account shows that they were conscious of their foreignness ('we stood out') and verbalised it in several instances as if they were in a foreign land, which of course they were in many ways. One of their most important reflections about this 'other' land has a direct bearing on our analysis. It is not only human rights organisations that fail to see violence against the poor. Often residents also fail to see that what they experience warrants attention or even the label of violence. It simply is what is to be expected. A female enumerator reflected as follows on this dilemma: 29 The respondents reflected a lack of knowledge about their rights as human beings, the laws, regulations and procedures related to the various acts of violence. Furthermore, they seemed to lack knowledge on what can be regarded as violence and the extent to which the various acts can be 26 Intermediaries in Development (IDC) 'Study on comparative analysis of the documentation of violence and ill-treatment in low-income countries -Nairobi-Kenya. Survey Analytical Report' (2015). 27 IDC (n 26) 12.  considered as violent. For example, most respondents considered being 'roughed up', slapped or punched around by police as 'normal' -these are things that the police do … Their lack of knowledge on what violence is, and the mandate of the police seemed to make them highly vulnerable to the acts of violence.
In the quote, she reflects on when and how residents in the survey area perceive something as violence and concludes that they see only events that lead to serious impairment as violent events. Apart from the obvious sense of talking about the social and spatialised 'other', her remarks also reveal what we may call a predisposition on the part of victims, namely, the normality of violence. Hence, we may conclude that it is not only the human rights organisations that fail to 'see' violations against the poor, but it is also the poor that 'fail' to see what happened as a human rights violation. As in the case of language, the ability to see events through a human rights lens maps onto what we could term spatialised class distinctions.
In summary, spatial, social and institutional predispositions exist against human rights organisations 'seeing' violations against the urban poor. They are geographically and socially far removed from the slum areas and institutionally their attention often is turned towards inter-organisational collaboration and relations to the international world of human rights, leaving access to poor areas to struggling paralegals or to chance. Similarly, poor urbanites do not understand their afflictions as human rights violations. This is not a moral condemnation or assigning guilt to either human rights organisations or to the urban poor, but a reflection of the structural dilemmas of human rights work. Nonetheless, these dilemmas have consequences.

A predisposition for legal accountability
Just before dawn, three young women from Eastleigh were on their way to work when they were shot by the police. The police were chasing two alleged robbers who disappeared into the slum. Having lost their initial target, the police fired their guns and called for everyone in the area to lie down. While lying down, two of the young women were wounded. The police called for a local mini-bus driver to take the young women to hospital. After having been treated in hospital the women were discharged after one day. The police officers accused the girls of hiding the suspects, but did not press any charges. 30 A neighbour reported the incident to a member of a local paralegal organisation who lived in the vicinity. One woman was suffering from particularly serious wounds on her leg. After carrying out the initial documentation, the paralegal arranged for the women to be taken to IMLU for treatment and legal counselling. IMLU assisted the women to access medical services (over a two-month period), carried out forensic documentation, psychological reports and rehabilitation plans.
As in the case of other human rights NGOs, IMLU requested the women to report the case to the police. The women had previously been too afraid to do so, but agreed to go to the police. When they entered the police station they saw the very police officer who had been implicated in the shooting and they became afraid. IMLU persuaded a lawyer to accompany them. The police were unwilling to register the case in the occurrence book and it took the pro bono lawyer another six months to get the officer in charge at the police station to accept the case for registration. After the women had obtained a police report and a P3 form (for the documentation of physical injuries), the police officers involved began threatening the girls that they would be charged with armed robbery if they pursued the case. The police officers also threatened the paralegals to a point where the latter had to move from the area for some months.
By the time the incident was reported to the police six months later, one of the young women had lost her job, her ability to rent a place, and she wanted 'treatment, justice, and compensation'. She had heard of people who had been compensated by the police and, in addition to the medical records at IMLU and the hospitals, two witnesses would support the case in court. In other words, the victim was initially confident that the evidence could lead to compensation for the injuries and lost income. However, police threats put an end to her hopes of compensation. Throughout the two-hour interview with members of the research team, she talked about the difficulty to secure a living and her fear of police officers, but little about justice in the form of holding individual officers responsible. The inability of human rights organisations to protect the victims (and the paralegals) from the police threats of reprisal removed her urge for legal accountability. Although the proof was available and the victim had hopes of compensation, the lack of protection was of greater concern to her than the prosecution of the perpetrators. 31 This case is indicative of a wider pattern of police practice, oscillating between assisting women who have been hurt in their pursuit of criminals and their own violent practices, threats and intimidation in protecting themselves against prosecution. However, the case also illustrates how legal accountability is a central priority in the actions of human rights organisations. Again, this is not a criticism of human rights practice. In an interview, one and half years after the incident, one of the women was grateful to the paralegals and IMLU for the treatment and legal assistance that they helped her to access. Nonetheless, legal accountability -or enabling work around legal accountability -was central to the way in which the organisation acted. We will elaborate in some detail on the logic of legal 31 Jensen (n 7). accountability before exploring the extent to which and how this might be part of a disposition against the urban poor.
The goals of criminal prosecution and reparations often dominate human rights documentation globally. 32 General Comment 4 of the African Commission, for example, sets out in some detail that '[s]tate parties are required to establish judicial, quasi-judicial, administrative, traditional and other processes to enable victims to access and obtain redress'. 33 The implicit assumption is not only that such processes are crucial in terms of individual cases, but they can also serve to reduce future acts of torture and other crimes through ending cultures of impunity. The UN Convention Against Torture similarly requires that 'all acts of torture are offences under … criminal law' and that states 'make these offences punishable by appropriate penalties' (article 4). Article 14 also requires states to ensure that a 'victim of an act of torture obtains redress and has an enforceable right to fair and adequate compensation'. This has been widely interpreted as implying legal forms of redress. 34 In this context, much time and effort are invested in training, networks and lobbying in order to further the aim of holding perpetrators to account and providing compensation for survivors. Hence, it is fair to say that legal accountability constitutes a central element in the way in which human rights organisations work.
There is nothing inherently wrong with focusing on legal accountability. In fact, it is part of the success of the anti-torture movement worldwide, and for good reasons. The model inherent in the fight for legal accountability and compensation may be summarised as follows: By ensuring that public officials are held to account we prevent impunity. This future public good is matched by an individual benefit for the victim in that proving legal accountability ensures the right to compensation.
However, the question arises as to how and to what extent this model predisposes against the needs of the poor. First, the model of legal accountability requires an investment of time and money, to the extent that local human rights organisations have limited resources to manage more than a few cases at any given time. Furthermore, it potentially comes with the price of not aligning with the wishes of the victims and residents in poor, urban neighbourhoods. To take the case of the three women set out above: While this case illustrates the power of the model, it is also fraught with uncertainties and possible misunderstandings. The three women were willing to engage in the process but as the case dragged on, the issues of protection (due to 32 Moon (n 10 police threats) and survival (a lack of livelihood and shelter) took on a more acute form. Their needs in some ways aligned with the narrative of human rights accountability up to a point, after which the discrepancy became ever greater. The situation has to be understood against a wider background of impunity and the widespread lack of trust that the Kenyan legal system can provide a measure of justice. Levels of intimidation against people reporting police abuse can be very high, and witness protection programmes are virtually non-existent. 35 Survivors and family members can, therefore, prioritise simply arriving home alive over legal forms of accountability. In our survey, only 30 per cent of respondents said that they would report incidents of torture and illtreatment to the police. 36 Of those that actually had experienced incidents of violence, only 13 per cent had reported the incidents to NGOs or paralegals. Only 25 per cent of the respondents reporting incidents of violence to the police felt that 'justice was served'. Perhaps most importantly in terms of the arguments presented here, only 4,6 per cent of respondents to the survey said that they would report incidents of torture and ill-treatment to NGOs.
This does not mean that there are no local strategies to engage with police violence well beyond the model of legal accountability. Paralegals, human rights monitors and other 'people on the ground' that we have talked to provide legal first aid, that is, the initial documentation of a case, informing the victim about the legal process, and taking actions such as accompanying the survivor to report to a police station. This clearly is in line with or at least not opposed to human rights strategies of legal accountability. At community level, however, there is no legal competence to carry out litigation. Instead, there is a deliberate attempt to solve the issue 'before it becomes a case'. As one paralegal said, 'most NGOs prefer to take cases to court, but we don't believe in the court process for family law because women do not go to the dock to give witness against their husbands'. Instead, the paralegals use mediation such as alternative dispute resolution (ADR), where they interact with each party, inform them about the options on how to resolve the issue at hand -court or mediation -whereafter most people agree to go for mediation. These are disputes between relatives, landlord and tenants, and neighbours (between citizens), not between citizens and the state. The many cases resolved through mediation remain undocumented and, therefore, do not figure in statistics on human rights violations.
Paralegals and human rights monitors rarely employ mediation in allegations of torture and ill-treatment. However, mediation may constitute a different and complementary approach to human rights violations or police violence that is more in line with the immediate needs of the poor in terms of their ability to live safely. If mediation were an option, perhaps more victims in poor neighbourhoods would come forward. This is not to suggest that the perpetrators of torture and ill-treatment should 'walk free', so to speak, but rather that the current focus on legal accountability only leads to few court cases. Such cases often involve well-known human rights defenders and not young men from poor neighbourhoods who have been severely beaten up by police officers. The fear of reprisal by police officers is an important obstacle to legal accountability, as we saw in the case of the three women. As an illustration, we took part in a one-day legal aid clinic in South Eastleigh in a community hall. Immediately prior to the meeting 'a killer police' was 'spotted around the area'. This 'sighting', according to those that did not dare face the danger, meant that only six people turned up.
In summary, while legal accountability is and has been central to address torture and ill-treatment globally it also privileges violations that can be legally documented with relative ease. This entails that a number of elements must be present, such as a victim who is willing to come forward and engage in a legal process, an organisation that has the capacity to engage with the matter (and no organisation can take on the number of cases that our research revealed in the three slum neighbourhoods in South Eastleigh alone) and much time and courage on the part of all parties, especially victims living in poor areas and local organisations such as the paralegals. These elements are seldom present. Even in cases such as this one where an organisation helps the victim to come forward, it is almost impossible to carry it through with legal accountability. Moreover, the victim's main concern may not be legal accountability but the ability to get on with his or her life.

The good victim
The third predisposition concerns the fact that the human rights movements tend to valorise torture victims as particularly heroic, often with good reason. Within this moral system, the political activist in a dictatorial regime assumes a privileged position. The claimed attributes of a 'good victim' are not only a reflection of suffering, but also the product of specific political contestations over moral deservedness. 37 A good example is the '56 ex-Kenya air force soldiers who were subjected to brutal torture following a botched coup attempt in 1982 in Kenya'. These soldiers approached IMLU 32 years after the event, and IMLU enrolled the survivors into its rehabilitation programme and wrote an article about the torture sequelae and the achievements of the IMLU group therapy. In this case, which resulted in compensation -the only compensation for torture ever to be paid 37 Jensen & Ronsbo (n 8). out in Kenya -the victims were portrayed by many as true heroes, defending the nation and revolting against a past autocratic ruler. Yet, 'good' victimhood is more than a set of moral criteria. It also requires holding attributes that can be recognised by appropriate structures, above all by human rights organisations and legal institutions. Some types of torture survivors are easier, both morally and practically, to recognise as such.
The case of Willie Kimani, quoted at the beginning of this article, is an excellent case in point. It involved people who were widely respected for their work on human rights in Kenya. For example, the daily newspaper, Standard Digital, wrote the following obituary: 38 Kimani was an investigator at International Justice Mission (IJM) for one and a half years before joining IPOA where he majored in investigations on police brutality and sexual violence against children ... His passion for fighting for the protection of victims of torture and extra-judicial killings drew him to the Independent Medico-Legal Unit before he joined IPOA.
He was only 32 years old but his passion for the protection of those who had suffered state violence had enabled him to achieve admirable results. The KNCHR uploaded a signed statement -as part of a human rights coalition -that 'Willie Kimani has dedicated his career to secure basic human rights and freedom for his fellow citizens' and that he was 'pouring his incredible passion into the fight for securing justice for the poor and transforming the criminal justice system'. 39 Another case, involving the slaying of Hassan Guyo, featured similar elements. 40 As in the case of Willie Kimani, the Kenyan National Commission for Human Rights uploaded a joint press release on the killing of Hassan Guyo that praised him for being 'a prominent human rights defender' and that 'Mr Guyo was a human rights defender and a pillar in the realisation of human rights at the vast marginalised Northern Kenya region'. 41 In 2013 Guyo was shot in the back by members of a military unit at a checkpoint. He had been arrested several times prior to this fatal encounter and had kept records of his torture and ill-treatment by the Kenyan security forces. The Kenyan state or Kenyan human rights organisations had not made any significant effort to provide him with protection. At first, the military and the police denied all knowledge of the incident. After intense media coverage, the military was forced to conduct an inquest. The inquest revealed that Guyo had been hit by a militarily-issued bullet. The Kenyan National Commission of Human Rights and a group of human rights NGOs provided evidence to the coroner. In some ways, Hassan Guyo was a 'perfect victim'. He had previously been tortured several times without his case being given a high profile by the human rights organisations. The police in Kenya shoot people every day. Yet, Guyo's death elicited a particular media, human rights and state response. Why might this be? First, on a sombre note, our material indicates that -in Kenya and beyondhuman rights cases can be easier to pursue once the victim is dead. Furthermore, survivors of torture can be easier to document once they are also murder victims. Living survivors are often too scared to seek redress. In contrast, the dead cannot run away and their bodies are (sometimes) available for documentation. There is no need to balance legal justice with a fear of reprisals. Institutionally, the recognition of Guyo (and Kimani) as victims was also made possible by the existence of legal and political procedures for recognition -forensic reports, a magistrate's court, a human rights community, interested journalistswhich, as we have argued above, are not equally distributed. Guyo's recognition was further enabled by his status as a human rights defender. This was what mobilised human rights interest and the press. The two young women who were accidentally shot by a police officer on their way to work were 'good victims' in the sense that they were innocent, young women making a living through hard work which included going to work at sunrise. The institutional recognition was enabled through especially the medico-legal reports and the victim and witness statements. Although the proof was available and the victim had hopes of compensation, the lack of protection was of greater concern to her than the prosecution of the perpetrator. In a meeting with the involved paralegals, one said that 'in most cases, people die from police shooting -the problem for the girls was that they survived'. 42 However, due to the duration of the case, the fear of the women due to the threats of the police and the hard work needed to be performed by cash-strapped NGOs, the women's case ended nowhere, neither in court nor with the satisfaction of redress and security for the girls.
The most common victims of torture and ill-treatment are young, indigent males such as Dennis whom we encountered in the opening pages of the article. Among the friends with which he was detained, he was the only one to return home and grateful to be alive, although he returned as a disabled person, being unable to walk, paralysed on one side of the upper body and suffering from lumps on his neck and back. It is illustrative that this case, involving one young man crippled 42 Interview with paralegals, 31 July 2014. by torture and three others dead, has not been picked up by any human rights group or become the subject of international alerts and media campaigns. These four young men -and numerous others killed by the police -in the public eye are 'bad' victims. They are stereotypically linked to violence and crime, often perpetrating violence, which is also borne out in our survey where they were the most likely perpetrators. Some are involved in criminal activities and sometimes they terrorise neighbours and kin. For instance, in the area where we conducted the survey, one woman had on several occasions been terrorised on her way to work. 43 In what became the final straw, she was held up early one morning. When she had no money and the young men did not want her old cellular phone, she had to promise to find them another phone. Since then, she dared not use the same road and, consequently, her livelihood was compromised.
Despite the violence that young men sometimes visit on their fellow community members, there is also the realisation, at least among some community members, that the police specifically and unfairly target young men. In the same slum area where the woman terrorised by young men lived, another woman explained how she had to provide sexual services to the police as well as pay money to get her children released from jail and the very real dangers of death or torture. This kind of sentiment found expression in informal settlements. Here we found a mural with the heading 'Our fallen soldiers', with at least 20 names written below. All these young men had been beaten and killed by the security forces. While only some of them allegedly had been involved in criminal groups, none of the cases have been taken up by human rights groups and none of the cases appeared to have become the subject of international alerts and media campaigns. As another study on police violence in a neighbouring slum in Mathare has revealed, grassroots activists in informal settlements perceived a 'gap' between their aspirations and working practices and those of larger NGOs. 44 While the allegations levelled against young men often are substantial, this clearly should not deprive them of protection. However, they do present human rights organisations with complicated dilemmas. As poverty can push people into moral compromises, the distinction between victim and perpetrator seldom is clear-cut, especially among poor young males. Their poverty alone exposes them to accusations of criminality and deviancy, which is further aggravated by their age and gender. In other words, the often compromised lives of poor young men are more complex than the documentation of the accidental shooting of young women on their way to work or the lives and deaths of largely middle-class human rights activists such as Hassan Guyo and Willie Kimani.

Concluding remarks
This article explored the way in which human rights organisations in Kenya are predisposed against documenting and engaging with the violence against the (urban) poor in Nairobi. Based on a victimisation survey on the prevalence and nature of violence in three Nairobi slum neighbourhoods, interviews with human rights organisations and documenters as well as in-depth interviews with several victims of torture and ill-treatment, we concluded that there are at least three predispositions working against the poor. First, human rights organisations and the urban poor do not share the same social or spatial world, even if they stay in the same city. The two are often separated by insurmountable distances, and the poor often see no point in directly accessing human rights remedies should these be available to them. Second, human rights organisations are concerned with legal accountability whereas the poor -while no enemies of accountability -can also think of justice in different ways incorporating security, livelihood and shelter. Finally, human rights organisations often find it easier to work with what we may term 'good victims'. This moral category often is off-limits to the poor who live complicated and morally-ambiguous lives. Human rights organisations may find it difficult to reach survivors of torture and ill-treatment who live in poor communities. This is not an indictment of human rights organisations in Kenya or elsewhere. We understand these predispositions as rooted in structures beyond their control. However, while this clearly is the case, there are also avenues for addressing the poor in different ways that do not render them invisible or difficult to assist. Access to justice, for example, not always only implies access to the courts. Legal accountability, in the shape of criminal prosecutions and redress, certainly is important, but this might not be the immediate priority for many survivors of torture. This realisation is making its way into, for instance, regional normative frameworks well ahead of global normative frameworks. Hence, the African Commission states the following in its General Comment 4 with regard to the victim-centred approach to redress after torture and ill-treatment: 45 A victim-centred approach to redress requires an analysis and full understanding of the harm suffered and of the victims' wishes. It needs to reflect their experiences and realities, so that the provided redress is responsive to their needs. States should ensure that victims have ownership of the redress process, and relevant actors providing redress are expected to work with the victims, and not on the victims.
Working more directly with the poor, and understanding the harm suffered and the victims' wishes, are of crucial importance. The predispositions outlined above affect the ability of human rights organisations to reach survivors of torture and ill-treatment in poor communities and suggests a need to rethink some approaches.
Through our analysis we see at least two possible implications for human rights work that must be taken into account by local, national and international human rights organisations and donors if they wish to design victim-centred interventions. First, legal accountability is not necessarily the place to start with human rights work. If one of the main reasons why survivors do not report their experiences is their fear of reprisals, it is necessary to consider protection even before the point of instituting a legal case. For people living in poor communities, protection, in the sense of feeling safe and secure, might also involve the ability to put a roof over their families' heads and to put food on their table. Now, for instance, protection usually is considered by human rights organisations a function of a legal process, but human rights organisations might also consider protection in relation to torture more broadly, also when there is no legal process. This is difficult work, and might involve thinking beyond the usual remit of human rights organisations, also in relation to funding that classically privileges legal accountability. However, a victim-centred approach is essential if human rights groups are to be enabled to bring about a measure of justice.
Second, and linked to the above, in order to respond to the forms of torture and ill-treatment experienced by people living in poverty, human rights groups could develop connections with the diverse grass-roots organisations that already work with poor populations, such as women's groups, savings groups, youth clubs, religious institutions and health organisations. Again, this may require redesigning funding schemes for human rights work. These organisations often do not think of themselves within narrow and legal 'human rights' terms but they may be interested in extending human rights protection given the right circumstances and incentives. They complement the work of paralegals as they not only often have a strong understanding of day-to-day life, but people living in poverty are more likely to trust them. Such organisations are well placed to identify victims and survivors and to provide the necessary support in the form of medical assistance, shelter and local knowledge.