The ‘pure apples’: Moral bordering within the Kenyan police

This article analyses various police reform initiatives in Kenya as a form of ‘moral bordering’. Drawing from ethnographic fieldwork conducted in Nairobi between 2017 and 2018, I analyse how police officers differentiate themselves from other police officers along (moral) ideas of reform and how this occurs in two divergent, yet interconnected, directions. The first is a process of bordering in: moral bordering occurs internally within the police and reform efforts aim to break down borders among police officers. The second is a process of bordering out: reform initiatives are designed in the urban centre and are aimed at spatially pushing the border externally, away from Nairobi. My approach to reform as moral bordering shows how borders can simultaneously take on disparate dimensions: with bordering in, borders are primarily social and symbolic, and with bordering out, borders take on a more spatial nature. This duality encapsulates the inherent friction that results from reform initiatives simultaneously moving in distinctive directions and differs from much of the (anthropological) work on the state police that analyses how the police themselves either enact borders or act as borders.


Introduction
I attended a workshop in February 2018 that was part of a pilot project aimed at fighting corruption in Nairobi, Kenya. Police officers, community members and media representatives were present to see how they could overcome their presumed differences and outline a strategy for further collaboration. During a coffee break, I was chatting with one of the community leaders and Wilfred, 1 a police officer working for the Internal Affairs Unit (IAU), an internal oversight unit within the National Police Service (NPS) that was set up in 2012.
At one point, the community member was defending the police, employing an often-used rhetoric that reforms should focus on improving police welfare and that the current (working) conditions of the police make bribery, and corruption in general, unavoidable. He then stated, rather boldly, that, 'all officers take bribes'. Wilfred's eyes opened widely and he snappishly interrupted him and stated that 'not all of us take bribes'. Thereafter he provided an example of a complainant who had come to the IAU and had offered a bribe, which the IAU investigator in question had refused. The community leader apologetically interjected him and said, 'No, I was referring to normal police officers; I know that the IAU is different'. Wilfred smirked, and then said, 'But we are also police officers'. There was an awkward silence and the conversation continued on about something different.
A little whilst later, after the community leader had left our conversation, Wilfred lowered his voice and whispered, You see, in Kenya, people talk about rogue officers and rotten apples, but here, almost all are rotten, that is true. So here, it is the other way around, here . . . we need to focus on the ripe apples, the pure ones, they . . . we . . . are the minority. *** This distinction made by Wilfred between the 'pure' and 'rotten' apples lies at the core of this article and is analysed as a process of moral bordering within the NPS in Kenya. The idea of 'rotten apples' is popularly used in various parts of the world to refer to police officers who operate outside the legal parameters of the state. These 'rotten apples' are habitually treated as the exception, as individuals who do not represent the general ethos and mind-set of the state police.
In Kenya, reference is also made to 'rogue officers' to clarify acts of police violence, corruption and other forms of (illegal) misconduct. Throughout the past two decades, police reform efforts in Kenya have revolved around eliminating the potentiality of officers going 'rogue', cultivating a policing culture that promotes 'pure' ones, and thereby enhancing accountability and legitimacy -pillars of democratic policing (see Manning, 2015). To accomplish this, the encompassing Kenyan police reform trajectory has included the creation of different command structures, oversight bodies, new training curriculums and much more. Yet, the idea of 'pure' police officers (and the entire reform project) has been overshadowed and disparaged by the vast amount of police misconduct and extra-judicial killings documented across the country, especially during the second half of 2017 (the time of the Presidential elections). Considering this, reform initiatives are either failing or operating in extreme isolation. Similar to what has been identified elsewhere (see Bayley, 2008;Hills, 2008), it seems that police reform efforts in Kenya have not fully materialised, especially in the peripheries of the state.
In this article, I analyse police reform initiatives as moral bordering by focusing on the perceptions of certain police officers, such as Wilfred, who are at the frontline of continuously enacting the moral borders between 'pure' and 'rotten' police officers. Drawing from (critical) border studies (see Brambilla, 2015;Newman, 2003Newman, , 2006Novak, 2017;Rumford, 2012;Yuval-Davis et al., 2019), this article sees bordering as a processual enactment of borders, which are points of differentiation by which individuals distinguish themselves from others. Moral bordering refers to the enactment of borders along moral questions of good and bad, and in this case, the (presumed) differentiation between the 'pure' and 'rotten' apples among the Kenyan police. I demonstrate how moral bordering occurs in two divergent, yet interconnected, directions. The first is a process of bordering in: moral bordering occurs internally within the police and is geared towards breaking down (moral) borders among police officers. The second is a process of bordering out: reform initiatives are designed in the urban centre and are largely aimed at spatially pushing the border externally, away from Nairobi -the 'headquarters of reform' -towards the peripheries -'the outskirts' (Simone, 2010: 39) and 'margins of the state' (Das and Poole, 2004). Police reform efforts thus also have a spatial dimension.
My approach to reform as moral bordering has analytical benefits for scholars working on both border(ing)s and policing across disciplines. The first is that it shows how borders can simultaneously take on disparate dimensions: with bordering in, borders are primarily social and symbolic, and with bordering out, borders take on a more spatial nature. Second, this approach encapsulates the inherent friction that results from reform initiatives simultaneously moving in distinctive directions. This friction not only emphasises the multidimensional nature of bordering, but also reveals the stark disparity between the idea of reform and the everyday reality of (state) policing. Third, by focusing on bordering within the police, my approach differs from much of the (anthropological) work on state policing that analyses how the police themselves either act as borders and/or how they enact borders. Moral bordering emphasises that bordering also occurs within the police as an organisation and that this transpires along morally contingent issues. Combined, this article aims to contribute to contemporary debates in (urban and political) geography on the spatiality of policing and geographies of morality in the urban realm.
In the first section of this article, I briefly outline the methodology that I used to research this topic. In the second section, I introduce my conceptualisation of moral bordering and thereafter, I highlight how this contributes to the existing scholarship on policing. In the fourth section, I sketch the Kenyan context and outline the key issues that have shaped the police reform landscape. The fifth and largest section portrays how police officers from two core units within the NPS, namely the IAU and the Directorate of Reforms, engage in moral bordering by differentiating themselves from other police officers. This article ends with some concluding remarks about how processes of transformation, such as police reform, act as fruitful trajectories to examine how police officers define and differentiate themselves, and how bordering can act as a lens to understand how internal differences shape the practices, norms and dynamics within institutions.

Researching moral bordering
This article draws from ethnographic fieldwork conducted between August 2017 and 2018 on a research project on police reform in Kenya. Largely ethnographic of character, the main aim of this research project is to understand the various mechanisms that monitor police behaviour and the actors involved in this process. I used various qualitative methods, such as participant observation, which entailed attending numerous events, workshops and meetings. Whilst attending, I interacted with the participants, observed interactions and recorded this through jotting down key words, which I extensively documented afterwards in more elaborate field notes. I also conducted approximately 180 interviews with an assortment of research participants, such as police officers, human rights activists, lawyers and members of civil society. These interviews ranged from semi-structured ones with guided topic lists to open-ended interviews that were predominantly steered by the interviewees. Sometimes interviews were recorded and transcribed, and other times I only made jotteddown notes that I extensively wrote up afterwards. All interviewees were informed of the purpose of my research and granted consent to their involvement.
It is important to mention that this project commenced during a rather volatile periodthe election violence of 2017 -and this shaped how and when I was able to access the NPS. Due to the high levels of police misconduct during this time, I initially focused on gaining access to civil society actors and understanding their perspectives. During this process, I was fortunate to also informally speak to approximately 15 police officers through a snowball method (very often meeting them by chance at events). Towards the end of my fieldwork, I formally requested permission to conduct interviews with police officers in Nairobi from the Inspector General's (IG) office. After formal consent was granted, I conducted a total of 75 formal interviews with police officers in Nairobi between June and August 2018. These 75 interviews were conducted at police headquarters and across six police divisions/subcounties and at 14 stations/posts. Although I tried to gain a representative sample in terms of rank, gender, ethnic background and police unit, the sample for officers was random and mainly selected by the officers in charge: he/she determined which officers could be interviewed. As a result, most of the officers were men of higher-rank (Inspector and above) from the Kenya Police Service. Similarly, these interviews ranged from semi-structured oneon-one interviews to informal group sessions with a few officers from various ranks.
From this particular group of police officers, 25 of them worked within the IAU and Reforms Directorate, two units explicitly engaged in reform efforts, as I will discuss later. As my research is specifically interested in reform, I did grant these officers more attention by conducting more in-depth interviews and (often) repeat interviews, collecting their life histories and engaging with them in informal discussions during public events, such as the workshop mentioned in the Introduction. Essentially, therefore, my interactions with those 25 officers were more comprehensive than the other officers who I interviewed rather sporadically (and often under supervision of their superiors) at police stations. However, despite being a 'minority', as Wilfred proclaimed, this group is telling of certain practices and opinions that constitute the police in Kenya and steer much of the (envisioned) moral bordering. In this article, and especially in the last section, I focus on their perceptions and portray how they experience being at the forefront of deciding who is 'pure' and 'rotten' among their peers.

Moral bordering
The study of borders is an increasingly multidisciplinary affair that analyses the numerous ways in which differences can be identified across physical, social and cultural terrains (Newman, 2006;Novak, 2017). One of the key contributions of more recent critical work on borders has been the debunking of the narrow Weberian-based perspective of borders as physical national frontiers and geographical demarcation points (Balibar, 2014;Newman, 2003Newman, , 2006Rumford, 2012). 2 Rather, borders can also be social, cultural and symbolic and are therefore not only spatially or physically constituted. Borders are increasingly seen as 'lines that provide socio-spatial criteria for defining a "here" and "there", (some of) "us" and "them", and what/who is and is not' (Novak, 2017: 850). With this broader and more critical perspective, borders are not only defined by states, do not rely on mutual recognition and do not need to be visible.
Another shift in our understanding of borders has been what Brambilla (2015) refers to as the 'processual turn', namely a revived focus on 'the bordering process, rather than the border per se' (Newman, 2006: 144, emphasis in original). Bordering refers to 'spatial and virtual processes -dynamic and shifting, multiscalar and multilevel -that construct, reproduce, and contest borders' (Yuval-Davis et al., 2019: 3). Bordering is a means by which we organise, assemble and manage our social and political landscapes. By focusing more on bordering, rather than the actual borders (physical or social), we gain several benefits. The first is a move away from a static notion of borders towards one wherein borders are mobile and relational: they are continuously constructed and deconstructed, confirmed and contested and perpetuated and impeded. As outlined by Yuval-Davis et al.: 'Bordering is continuously happening ' (2019: 4). The second benefit is a recognition of the involvement of a spectrum of people: bordering is not only performed by state entities, but largely occurs through 'the efforts of ordinary people' (Rumford, 2012: 897). The third is an understanding of bordering as a process of othering and social ordering (van Houtum and van Naerssen, 2002) that occurs along various pre-determined criteria, which can range from natural benchmarks, such as rivers, to physical entities, such as walls, but also social and cultural categories, such as age and gender (Newman, 2003: 17). Combined, as a conceptual lens, bordering allows us to understand the ways in which difference manifests itself through a variety of ways. Elsewhere, Parker and Adler-Nissen (2012: 780-785) discuss how bordering constitutes a process of 'inscribing' borders, whereby certain 'planes', such as topography, economic relations and linguistic commonalities, define when and how borders emerge. In this article, I am interested in understanding how borders are inscribed through moral 'planes', and more specifically, along (perceived) differences between 'good' versus 'bad' police officers. Understanding how certain moral ideas move across borders and spaces is relevant to further developing our spatial frameworks of (urban) security (see Glu¨ck, 2017) and geographies of morality more generally.
Within anthropology, there has been a revived analysis of morality, often framed as moral anthropology (Fassin, 2012) or as the anthropology of moralities (see Heinz, 2009;Zigon, 2008). Although there is not sufficient space here to delve into this ethical turn in anthropology (Mattingly and Throop, 2018), it is important to articulate how it shapes my formulation of moral bordering. I draw from Didier Fassin (2012: 5) and use moral as an adjective that 'designates what is viewed as good, or right, or just, or altruistic'. Moral bordering thus refers to the processes in which borders are enacted that, in both their production and contestation, touch upon issues in which there are (active) disputes over what is good or right, or just and unjust. Moral bordering is thus not about the bordering of moralities, i.e. the construction of borders around certain sets of values and norms that shape how people think or act. This perspective would present the idea of moral entities as distinct and 'separated from the other spheres of human activities' (Fassin, 2012: 4). Rather, as Fassin argues, 'moral questions are embedded in the substance of the social; it is not sufficient to analyse moral codes or ethical dilemmas as if they could be isolated from political, religious, economic, or social issues' (4). Taking this approach, I do not see police reform as a morality. Rather, police reform is a process geared towards transforming policing and this is a highly morally contested issue that is embedded within larger social, political and economic trajectories. As a form of moral bordering, reform efforts create moral categories and communities, such as the 'rotten' versus the 'pure' apples, and this occurs across various spaces, from the urban centre to the state's peripheries, and at various levels, from the individual (such as Wilfred) to the collective (such as the IAU).
In a sense, all human activities are laden with moral questions and all organisations experience a degree of moral bordering, i.e. people who distinguish themselves from their colleagues along moral issues. Moral bordering can thus be identified in and applied to various sectors. However, moral bordering is particularly pertinent within certain organisations, such as the police, where morality is centralised within their occupational cultures. The state police is an institution that is explicitly given the task to uphold certain moral values (Herbert, 1996). Morality defines everyday police work: it is for a reason that the police are habitually referred to as the 'Thin Blue Line' (Steinberg, 2009), as those standing at and embodying the border between anarchy and order, tasked with the heavy mandate of maintaining society's morality. Understanding how this occurs across spaces is relevant to scholars interested in the geographies of morality more broadly.

Police/policing border(ing)
The centrality of morality in police work is explicit in the way that police offices are engaged in the policing of borders. When employing a narrower perspective to borders, police officers (and other state institutions) are mandated to safeguard national borders and protect the citizens within them. When employing a broader approach to borders (as this article does), police work is also inherently about bordering, i.e. maintaining (social, physical and imagined) borders. Both the classic criminological studies on the state police (Bittner, 1970;Ericson, 1981;Herbert, 1996;Manning, 1977) as well as more recent ethnographies (Beek et al., 2017;Denyer Willis, 2015;Fassin, 2017;Hornberger, 2011;Jauregui, 2016;Martin, 2018) all demonstrate how police forces across the globe stand at the frontline of upholding particular moral (b)orders. 3 Policing is thus morally charged border work and it is also here where the city, as a site of insecurity and risk, is consistently the conceptual and empirical backdrop. Most anthropological studies on the police occur in areas that are referred to as 'no-go zones' and 'governance voids' (Kruijt and Koonings, 1999) and are regarded as the peripheries of the cities, those parts that are 'simultaneously included and excluded in the formation of a nation, municipality or region' (Simone, 2010: 4). In fact, most of the contributions to this Special Issue are based on police practices in so-called slums or informal settlements, where we see police misconduct embroidered in larger structures of marginalisation, social exclusion and stigmatisation.
Although there are exceptions (see Fassin, 2013;Glaeser, 1999), this (ethnographic) focus on state policing is predominantly outwards lookingbordering out -and provides little analytical scope to intra-police bordering, i.e. understanding what goes on within police institutions and among police officersbordering in. These dynamics have been granted more attention in the (primarily criminological) work on 'police culture', which highlights the differences among officers and policing bodies (Campeau, 2015;Loftus, 2009;O'Neill et al., 2007). More recent scholarship has emphasised the need to examine connections and linkages across police units, such as Chris Giacomantonio (2014) who focuses on organisational boundaries across police units and Chad Whelan (2017) who analyses security networks between police organisations.
In this literature, we can also identify several policing typologies, such as the distinction made by Reuss-Ianni and Ianni (1983) between the 'street cop culture' and the 'management cop culture', and Ericson's (1981) analysis of the differences between detectives and patrol workers. Other studies have highlighted other key markers of variance among police officers. For example, both Sinan C¸ankaya (2015) and Paul Mutsaers (2014) discuss how ethnic identity impacts how Dutch police officers experience their work and interact with citizens. And Theresa Ulicki (2012) and Marisa Silvestri (2017) highlight how gender differences, primarily the 'cult of masculinity', define police officers' opportunities. Combined, such studies emphasise the variety within police organisations and how this variety influences the work experiences of police officers and translates into a multiplicity of relationships with others. Simply stated: citizens have different experiences with officers from different age brackets, ethnic backgrounds, police units and so forth, and this variety must be taken into account when analysing everyday policing.
However, an analysis of how morality, or ideas of good or bad policing, creates differences (and thus borders) among police officers has not been granted analytical or empirical attention. This article aims to fill this gap by analysing how moral bordering occurs within the police itself. More specifically, it analyses how, in line with police reform efforts, categories of 'pure' and 'rotten' apples within the Kenyan police are constructed and narrated and how this bordering process occurs in two divergent, yet interconnected, directions. The first is a process of bordering in: reform efforts are geared towards breaking down moral borders within the police. The second is a process of bordering out: reform initiatives are designed in the urban centre and are largely aimed at spatially pushing the border externally, away from Nairobi, towards the peripheries of the state. This simultaneity of bordering in and out across moral landscapes draws from and contributes to contemporary debates in anthropology, criminology and urban and political geography.

Police reform in Kenya
In Kenya, the idea of police reform emerged under the Kibaki government in 2002, where various government programmes aimed to transform the security and justice sector (Ruteere, 2011) and eliminate the police's role as the instrument of control by and for the political elite (Akech, 2005;Auerbach, 2003;Hills, 2007). 4 Yet the need for transformation truly materialised after the post-election violence of 2007-2008. A Commission of Inquiry into the Post-Election Violence -also known as the Waki Commission -concluded that from the 1133 people that had died, over 400 were the result of police actions. 5 This report, and many others, reaffirmed the idea that substantial police reform was needed. In 2009, the National Task Force on Police Reforms was established, chaired by Philip Ransley (later known as the Ransley report), and in 2010, the Police Reforms Implementation Committee was set up to oversee the implementation of the recommendations from the Ransley report (Kivoi and Mbae, 2013;Skilling, 2016).

The NPS Act of 2011
Most of the recommendations were channelled through the establishment of the new constitution in 2010 and the NPS Act of 2011. 6 This Act primarily entailed transforming the Police Force into the Police Service and targeted the command structure of the police by instating an independent IG to command over the two Deputy IGs of the two existing police forces. In Kenya, British colonial rule had introduced state policing and had divided the corps into two sectors: The Administrative Police and the Kenya Police. The Administrative Police -popularly known as the tribal police -largely operated in the rural areas, dealt with matters pertaining to customary law and was regarded as a political tool to support provincial administrations and chiefs (Nyabola, 2016;Ruteere, 2011). The Kenya Policeregarded as the regular police -acted as the main state police force and was engaged with activities such as crime prevention and investigations. Historically, the two police units operated rather distinctively and the reform initiatives intended to unite them under one command -to break down borders. Yet despite these efforts, they were still seen as two separate police forces that operated independently from one another. At the time of writing (2019-2020), more dramatic efforts were being implemented to merge the two.

Police oversight
Another fundamental legislative change was the setting up of two oversight agencies to oversee police conduct. For external civilian-led oversight, the Independent Policing Oversight Act of 2011 was decreed and an oversight agency, the Independent Policing Oversight Authority (IPOA), was established. IPOA is an independent state institution that is required to investigate police misconduct, especially deaths and serious injuries caused by the police, review the functioning of internal disciplinary process, monitor and investigate policing operations and deployment and conduct inspections of police premises. On a global level, IPOA is regarded as highly progressive, having an extensive mandate that exceeds oversight authorities established elsewhere.
For internal oversight, the IAU was set up under Section 87 of the NPS Act. The IAU is responsible for handling police (mis)conduct internally and it receives and investigates complaints against police officers that come from members of public and police officers (Osse, 2016). Up till now, the IAU only has an office in Nairobi, but interviews conducted in March 2019 suggest that the first regional offices will be established in the near future. This decentralisation is an example of bordering out -moving away from the capital to the rest of the country.
In an audit report written in 2015 by the Kenya National Commission on Human Rights and Centre for Human Rights and Peace, the IAU was described as incapable of fulfilling its mandate. At the time of writing, many dimensions support this claim, such as its financial dependence on the NPS and the mere fact that it reports to the IG. Yet at the same, the unit comprises approximately 150 police officers, 7 such as Wilfred, all of whom voluntarily applied to work within the unit, and as I will discuss later, did so with the primary purpose of implementing change. Each of them, especially those that have been there from the start, have narrated the tremendous progress they have made. As a male officer who has been with the IAU since inception said to me: When we started, there was no culture of people being accountable, no culture of people dealing with complaints against them. Police officers had a Big Man syndrome, thinking we are the government and nobody tells me what to do. Our work was to do that and we had enemies at the start. And now, although many people cannot see the changes, steps have been made. Officers know that someone is watching them . . . they know that people can ask questions about what they do and this is already a very big step. 8 This sentiment -of the IAU having an impact -is both supported and contested: some people argue that the IAU is capable, albeit on a small level, whilst others claim that the IAU is just there for show and that it acts more as a way of 'protecting their own'.

'Reform champions'
In addition to some of the external and organisational dimensions of police reform, there have also been changes centred on operational and cultural change. The guiding mantra has been an emphasis on 'people-centred policing' and this has, for example, included a revived and increased focus on community policing (Brankamp, 2020;Diphoorn et al., 2018;Skilling, 2016). Furthermore, in early 2018, the new Service Standing Orders -the key Orders that dictate everyday police work -were launched and this was celebrated as a crucial step.
Many of these changes have been done through or with the assistance of the Reforms Directorate, which acts as the policy-making hub and advisory board of all police reform initiatives, ranging from training to fleet management. Much of this stems from the Reforms Handbook (National Police Service, 2017), titled the Roadmap to Transformation, which functions as a basic guide for police officers on how to implement police reforms, as stipulated under Section 132 of the NPS Act. The handbook acts as a collection of the various laws, policies and regulations that will 'provide the framework for police reforms' (xviii) and will 'stimulate commitment and foster open-mindedness to all police officers to implement the new and revised policing laws and regulations' (xv). According to one of the officers within this unit, 'it focuses specifically on ensuring that the police is professional; fosters a relationship with society; tackles corruption; enhances integrity, and respects human rights'. 9 A key project of the Directorate has been the identification, training and cultivation of 'reform champions', which refers to police officers that have been selected and trained to 'spread the word of reform out there' (emphasis added). Through a voluntary application and strict recruitment procedure, approximately 25 officers were selected and received 'reform training' in 2017. 10 These officers, also referred to as the 'pioneers' and 'champions of change', went out to police stations across six different counties and discussed, together with leading officers from the various regions, about how to implement the reform plans at the police stations. The idea is that these stations will be revisited throughout the next few years and that the 'champions' will stimulate further progress, which will in turn, 'cascade the right mentality' and create new 'champions'.
This project perfectly encapsulates how police reform is simultaneously a process of bordering in and bordering out. It is based on a conviction that reform should come from 'within', from police officers, and thus not from external donors or civilians. Police officers are thus internally responsible for breaking down borders and producing 'pure apples'. On the other hand, the reform champions have the primary aim of spreading the reform message 'out there', far beyond the capital and out into the margins, to places where reform efforts are not felt. Nairobi is considered the 'centre for reform', as the epicentre where change is aspired and where policies are concocted. During my interviews, a distinction was always made between officers operating 'out there', in contrast to those working in the city. In this regard, reform is a process of physically bordering out and ensuring that officers 'out in the bush' are trained and educated. As a result, reform efforts often skip over Nairobi, because, as one officer stated, 'the city is already reformed'. 11

'Reform is a joke!'
This perception of Nairobi as a city that is 'already reformed' is in stark contrast to the reality of everyday policing and the prominence of police misconduct in Nairobi. This was most palpable with the excessive police violence that marked Kenya between August and December 2017, now referred to as 'post-election violence of 2017'. Numerous reports from leading organisations documented the vast amount of police violence, each providing different figures on the casualties. A joint-report produced by Amnesty International and Human Rights Watch (2017) documents that at least 33 people were killed in Nairobi between August 9 and September 12. In another report written by the Kenya National Commission on Human Rights, at least 25 new cases of deaths were recorded between September 1 and October 25, in addition to over 100 injuries.
These incidents are not isolated to times of political change and uncertainty. Police violence, extortion, intimidation and extra-judicial killings are a daily reality for many inhabitants living in Nairobi's lower socio-economic neighbourhoods (Glu¨ck, 2017;Jones et al., 2017;Mathare Social Justice Centre, 2017;van Stapele, 2016van Stapele, , 2019. These studies, along with my own fieldwork observations, show that police interactions with the public are regularly hostile and far from the people-centred approach hallmarked by the NPS. Figures on police violence starkly contradict the idea of Nairobi as a city that is 'already reformed'. One is forced to conclude that reform efforts are either failing or operating in complete isolation, devised in police headquarters, but not actually shaping police practices, both in the margins of the urban and the state (see Diphoorn et al., 2018). Reform efforts are thus largely bordered in. As stated by a foreign policing consultant and expert: 'Change is only at Jogoo House, at the headquarters, yet at the ground level, there is no change, and this gets worse the further away from Nairobi you are'. 12 This perception, which was shared by others, highlights the spatial dimensions of reform and the role that cities such a Nairobi play in acting as both the potential and setback for change.
Police reform efforts are thus forms of moral bordering: a continuous enactment and negotiation of borders by which police officers are differentiated along moral understandings. By distinguishing between 'pure' and 'rotten' apples, police officers assemble and manage their daily lives. This is especially the case for those police officers, such as those working in the IAU and Reforms Directorate, who are specifically assigned to create such categories. It is safe to say that the majority of officers engaged in a certain degree of performative moral bordering during our interviews, wherein they would locate misconduct with their colleagues, rather than with themselves. Yet certain officers, such as Wilfred, have the explicit mandate to engage in moral bordering on a daily basis: it is their designated task to continuously seek out difference and identify those so-called 'rotten' officers.
In the following section, I focus on their perceptions and experiences of being at the frontline of reform, i.e. moral bordering, and analyse how they are engaged in constructing, changing and obliterating borders. I am therefore not analysing networks or connections across police units or departments, but rather, how certain police officers from various police units distinguish themselves from others along ideas of 'reform'. By doing so, this article places (certain) police officers (and their viewpoints) at the forefront of 'border management' (Newman, 2003: 18) and 'seeing like a border' (Rumford, 2012: 897), which implies 'taking into account perspectives from those at, on, or shaping the border'. As discussed, understandings of morality are not monolithic but vary greatly across officers (Herbert, 1996). Through analysing individual experiences and perspectives, we can uncover this diversify and acquire a more experiential understanding of the geographies of bordering.

'Reform means curing': Moral bordering in the police
Towards the end of 2017, members from civil society felt a need to 're-open' a dialogue with the police. Numerous meetings and trainings were organised and I attended one in December 2017 that outlined a revised police reform strategy. Several police officers were strategically invited: these were officers working for the IAU and the Reforms Directorate. When I inquired why, one male NGO employee explained to me: 'There is no point in inviting just any police officer. Not only will they not come, but they will not want to speak to us, especially now. So we invite those that are open and want to engage with us'. Later on, one of the police officers echoed this sentiment: Us, we want to change the police and to do that, we must engage with the communities and these organisations [NGOs], we need their help. But other police, they don't see that. They are different; they don't want change and feel these guys are attacking them. For change in the police, it is us who are responsible.

'Officers of integrity'
The police officers from the IAU and reforms directorate all defined themselves as 'different' from other police officers. As one of them said, 'Compared to other police, we are very different. People don't file complaints against us'. 13 This difference is also visually identifiable, particularly with IAU officers, who do not wear their police uniform, both in the offices and when outside, conducting their investigations. One female IAU officer who has been there since inception says that this was purposely decided, so that the officers would acquire and exude more of a 'civilian feel'. 14 Although this was also for themselves, it was primarily geared towards acting more as a civilian body for the public, as she explained: If you are a civilian and you want to complain about the police, and then you come to us, and I am in uniform, this can present problems. You might be scared. At that moment, you need to see that we are different, so that you will trust us. 15 Many see themselves as 'officers of integrity' and describe this as something intrinsic. As mentioned, these officers volunteered to work within these units and used their voluntary application as an example of this integrity. For most, the prime motivation to conduct this work was, as one female officer stated, 'to be a part of the process of changing the Kenyan police'. 16 Although they also shared other reasons for applying for these positions, such as the love for investigative work, the thrill of being part of something new and practical reasons to be stationed in the capital, all interviewees stressed how personal experiences of corruption and a sense of injustice motivated them. Many base this on depraved experiences in the past, such as one male IAU officer who discussed at length how he had personally been intimidated by his superiors on numerous occasions, but that he could not do anything about it: 'I knew that this was wrong, but there was nowhere to report it. Nobody to help. In the police, your senior decides how it is for you and nobody can change it . . . you just have to accept'. 17 For many, working here is an 'opportunity to change the police'. 18 Interestingly, for most of them, their motivations to do reform work were similar to their motivations for joining the police, namely about serving society. Many of them felt that by working for one of these units, they were serving the same goal. As one IAU officer stated, When I was an OCS [officer commanding police station], I worked for the community and interacted with the public to serve them. Now, I still work for the public, but indirectly; now I try to correct the police, to solve problems, so that they can serve the public better. 19 Several officers discussed how they want to act as role models for other officers, especially junior ones, because they missed this when they entered the police. One male officer told me: Even if you don't want to take bribes, you feel you have to. Around you, everyone does it, and you have to give to your superior. When I started, I was looking for someone to tell me: no, this is bad, but I didn't have it. I had to keep strong, to remind myself that I can be different. This was very difficult for me and many officers are experiencing this. Our role is to help them. 20 This statement also highlights the intrinsic nature of being an 'officer of integrity' and when I probed further about this, he explained: Yes, this is something inside, as a person. Either you want to do right or you want to do wrong. You see this at the training already, you can see which officers are good, have integrity, want to serve the public. But even with this, many also go wrong after some years in the police, because of their colleagues, their superiors. There is no one there to correct them and show them the correct way. This sentiment was supported by many officers: although being a 'good' officer was something natural, it was also shaped by one's colleagues and experiences and it is here where 'corrections' are needed. Moral borders are thus internal, within individuals, but can also be constructed and deconstructed through the practices and guidance of others. Reform efforts are aimed at doing precisely this, i.e. breaking down these invisible social borders.

'To correct the errant police officer'
This need to 'correct' was regarded as their core work. One officer mentioned that 'reform means curing and to be better' 21 and another male officer described the IAU as a 'correctional tool for the services'. 22 One female colleague described it as a 'place where things are right, where they are correcting things' 23 and another female and high-ranking officer stated that their job is to 'correct the errant police officer' and to transform them. When I asked her what the 'errant police officer' was, she defined it as a police officer who 'is not straight, doesn't care, doesn't follow the law, doesn't have and take responsibility, and doesn't serve citizens'.
For many, correcting was primarily geared towards two types of 'errant', so to speak: police inaction and corruption. Police inaction, which forms the majority of the complaints received by the IAU, refers to cases where police officers do not do something that they should. This can be as simple as not recording something into the occurrence book or not handing out a receipt of cash bail payment, but it also includes not showing up during a court session to provide a statement, failing to record malpractice by a colleague or refusing to investigate a complaint. Such instances are described as laziness and for IAU officers, these are often the simplest cases to investigate and solve, whereby a 'brief training session to educate them' 24 is often sufficient.
The second main 'errant' is corruption, and this is regarded as a much larger and systemic problem. According to the East Africa Bribery Index (Transparency International, 2017), the NPS is the most bribery prone institution in the region, with a 68% likelihood that one will encounter bribery with the police. One police officer within the reforms directorate described it as a 'wave that just flows. You cannot control it'. 25 Unsurprisingly, corruption was the most difficult topic to discuss during interviews: many became overtly uncomfortable when I mentioned it, or when I probed further after they themselves recalled an incident, they diverted the question, and sometimes literally stated that they 'wouldn't venture into that'. Fortunately, some were willing to discuss it and elaborated on how this lies at the root of most police problems. One officer narrated how: Every police officer is corrupt at least once, sometimes without knowing. This is also because corruption happens in many forms: most people think about bribes, the traffic cop asking for money, but corruption is also about abuse of office, abuse of authority and this happens a lot. Many times, officers do it and they don't even know. Our job is to make officers see this, and to understand the bad side of corruption, how it ruins our image with the people, and that it, eventually, it creates a bad police force. They need to understand that corruption doesn't help, in the end. 26 This perspective of corruption conveys a somewhat sympathetic view: the perception that sometimes 'officers do it and they don't even know' point towards ignorance, a lack of knowledge of what corruption is. In contrast, the 'officers of integrity' do know, as the same officer further describes: We, we have been selected, because we can see this and we know how to identify corruption. Of course, we have had struggles too and made mistakes, but when we made the mistake, we knew it was a mistake. You understand? So we need to help them see, to show them. But the problem is that some don't want to see and this is where our work becomes difficult.
He further discusses the various strategies he employs to get them to see how bad corruption is and although he prides himself on being 'successful' and 'changing others', he also narrates how some police officers cannot be reformed. These are officers that 'are in an occupation and not in a professional situation, aimed at professionalization. For them it's all about their earnings'. 27 Borders, in such instances, are experienced as static, fixed and indestructible. For them, bordering sometimes fails, and accepting this is, at times, very frustrating and emotional.
'But we are also police officers!' Interestingly enough, despite proclaiming themselves as different, all identified themselves, first and foremost, as police officers. This not only occurred during our interviews, but also during public events, where they introduced themselves as officers, similar to what Wilfred did at the workshop mentioned in the Introduction. For them, working for the IAU or Reforms Directorate entails working for a particular unit within the NPS. They do not distance themselves from the NPS as a whole, but rather, from a certain mentality held by their colleagues, which results in particular practices, or the lack thereof.
This process, of feeling like a police officer, but also distancing yourself from your colleagues, is not an easy task and all of them stressed at length the difficulties they face in 'investigating your own colleagues'. 28 As one of the reform champions mentioned to me: Doing our work is not easy, because you need to do things differently -you have to admit that the police has a problem and be willing to fight it head on. If you can't admit there's a problem, then you are in the wrong place. And if you can't encourage and convince others of this problem, you are in the wrong place. 29 Others highlighted how unpopular they are, and some claimed that their colleagues despise them. Sometimes they do not see them as real police officers or they see them as officers who are 'on the other side' and have 'betrayed them'. A male IAU officer described how his former colleagues at the police station where he had previously been based for a long time, perceived him: They [former colleagues] don't talk to me anymore. For many, I am seen as a betrayal. It is difficult. You grew up with them, in training, and now, they act like you are not police. But I am still an officer -I may not wear a uniform every day, but being a police is my heart. One day I hope they will see me like this. 30 Border work is thus far from easy, as has also been highlighted by other studies on internal oversight mechanisms (see Goldsmith, 1991;Savage, 2013). Yet for some, reaching out to one or two officers is sufficient. Although this is slow, it will eventually lead to change, and being a police officer is precisely what allows them to do this work, they contend. This is often contrasted with IPOA, the external oversight authority. As one IAU officer mentioned: To go to an OCS, to inspect his records, it is not easy. Our IPOA counterparts, they find it very hard, to get the police to get their records. IAU has not been finding it very hard, because even though they don't like us, we are police, and not civilians. They know that we are still one of them.
Similar to what has been documented in other research on police oversight systems (see Savage, 2013), individuals with a police background are often seen to be more effective as they understand the 'ins and outs' of police work. They 'speak the same language', 31 as an IAU employee stated. Another one affirmed this: 'We know the job that needs to be done, so we also know when it is not done well, but wrong. A civilian doesn't see this'. 32 Another more senior IAU officer confirmed this: 'We, as police, know how it works; we know how they collect bribes, how we are used by politicians, outsiders don't understand and they cannot see it'. 33 Therefore, despite the difficulties they face, they argue that reform must come from within, and not from outside. Although the literature on bordering highlights the involvement of numerous individuals (see Newman, 2006;Rumford, 2012), these sentiments emphasise that for many, moral bordering can only be executed by particular people with certain characteristics. For many of these officers, moral bordering is an internal affair.
These perceptions also underline the contradictive nature of bordering: on the one hand, these officers see themselves as different and they pride on this difference; it is what gives their work meaning. Yet on the other hand, their work is also defined by the breaking down of borders, about getting people from the 'other side' to join theirs and become 'officers of integrity'. What complicates matters further is the invisibility of these borders. Although all borders are socially constructed (Newman, 2003: 17), various markers of difference are visible, such as gender and age. A willingness to change is, in contrast, not something that can be physically identified. As one officer said to me: 'You often work on your instinct, because you cannot see from the outside!' 34 This invisibility makes moral bordering trickier; laden with secrecy and doubt, the border is defined by uncertainty and risk. This risk is amplified by the fact that down the line, officers working for these units will be transferred at one point, as all officers are. What will happen when they return to stations with colleagues who they may have investigated? And even more, what will happen if they themselves are investigated? These questions linger over many of their heads: it questions where bordering starts and ends and emphasises the continuous nature of bordering (Yuval-Davis et al., 2019).

Nairobi: 'The centre of reform'
Another major strain experienced by these 'officers of integrity' is the high levels of police violence, and it is here where the friction of simultaneous bordering in and bordering out really comes to the fore. As discussed, Nairobi is seen as the 'centre of reform': it is the place where reform ideas are devised; it is the 'hub of change and transformation', and implementing reform is about spreading the message outwards -bordering out. This supports much of the work on policing outside of Nairobi (see Brankamp, 2019;Mkutu et al., 2014;Mkutu and Wandera, 2013), particularly in Northern Kenya, where policing is largely militarised and executed as a form of 'emergency policing' (Brankamp, 2019: 70). Areas outsides of the capital are often characterised as 'operational areas' (Brankamp, 2019: 68).
In contrast, many interviewees described Nairobi as the place where people are educated and informed, particularly in comparison to rural areas, where people need to be 'sensitised' and educated. Several police officers felt reform was not necessary in Nairobi, because 'the city is already reformed'. This perception was affirmed when the IAU set out a new sensitisation campaign during April and May 2018, wherein they would visit several counties (ranging from Turkana to Kakamega) to inform the public about their work. Interestingly, Nairobi was not scheduled in. When I inquired why, I was told that Nairobi 'doesn't need it', because in Nairobi, citizens are already 'enlightened' and 'have more information and they know where to go'. 35 Furthermore, as the IAU's only offices are in Nairobi, the capital's residents have the advantage of proximity. Interestingly, some interlocutors did not regard certain urban areas within Nairobi, such as Mathare, as the periphery -'out there' always referred to the rural areas.
Yet on the other hand, Nairobi was also described as the 'most difficult' and 'complicated' place, where reform efforts have the most work to do. One senior police officer explained it to me as such: In Nairobi, you have so many different types of police stations, ranging from the high end Gigiri, to the low end, like in Kamukunji, which is the biggest police station in Kenya and is full of thugs. Reform is different in every station; this is different than in the rural areas, where things are the same, you can compare. To reform police in Nairobi is a bigger challenge. 36 The idea of Nairobi as a bigger challenge was supported by others. Nairobi is thus simultaneously described as the city that is 'already reformed' and as the 'biggest challenge to reform'. And furthermore, when talking about these challenges, they are referring to certain parts of the city -the peripheries -the areas that are simultaneously ignored and granted attention (Simone, 2010: 40). Therefore, although figures on police violence in Nairobi clearly indicate the urban nature of the problem, reform is seen as a physically outward looking project, with eyes gazing towards the rural. Bordering out thus indicates physically going outside of the city, implying that the border takes on a more spatial dimension. This reiterates the idea that borders can simultaneously be geographical and social, albeit in different ways. Furthermore, this contradiction of Nairobi unveils the friction and incongruity of bordering that unquestionably shapes the everyday work of these police officers.

Concluding remarks
In this article, I have approached reform as a form of moral bordering in order to understand some of the reform initiatives occurring within and by the Kenyan state police. I have specifically focused on the perspectives of certain police officers working for certain units who see themselves as 'officers of integrity' and differentiate themselves from their colleagues who are 'errant' and need 'correcting'. By doing so -seeing themselves as dissimilar and ascribing themselves the task of changing their fellow officers -they are engaged in moral bordering and this occurs both internally, within the police service (bordering in), and externally, towards the peripheries of the state (bordering out). These divergent yet complementary processes highlight the inherent friction that defines police work and this is particularly manifest in Nairobi, which is seen as both the centre and the failure of reform. This not only speaks to the disparate spatial imaginaries between Kenya's urban centre and its territorial peripheries, but also highlights that borders, although imagined, simultaneously take on a spatial and symbolic dimension. In our analysis of bordering, we should therefore not confine ourselves within a 'social-to-spatial analytical trajectory' (Novak, 2017: 849), but rather focus on the ways in which bordering occurs along social and spatial dimensions at the same time, albeit in divergent ways.
Another key dimension is the emotional experience of bordering. Although the interviewed police officers state that their aim is to change the police, they simultaneously need this difference to give their work meaning. If all officers were 'pure', their work would become obsolete. In addition, due to the simultaneous process of bordering in and out, these police officers habitually find themselves in contradicting situations. Several officers openly discussed how their work is often impossible and a selected few admitted that perhaps the police alone cannot change the police, but that they need assistance from others, such as civil society. With this in mind, moral bordering will (start to) look outwards again. As also highlighted by Pauschinger (this issue), an emphasis on the emotive dimension of bordering can further develop our understanding of the spatiality of policing and contribute to debates in the geographies of bordering.
Combined, this article has shown how police officers do not only maintain and act as borders vis a vis citizens and other actors but are engaged in internal border work all the time. Different ideas of what a 'good' police officer is and what is just or unjust in police work creates divisions among the police. These divisions and eventual divergent moral communities shape everyday policing practices and this needs to be further analysed, both empirically and conceptually. This is particularly so in light of contemporary dynamics, wherein police forces across the globe are increasingly subject to criticism for their excessive use of force. In line with Campeau (2015), times of change, such as police reform, are fruitful trajectories to understand moral bordering: it is during these phases where people are forced to question and situate themselves alongside moral borders and where sites of friction are most evident.
Furthermore, although this article has exemplified the centrality of moral bordering within the police, these processes can surely be identified in many institutions. More comparative analyses on moral bordering can provide more insight into the differences within institutions, but essentially also into the larger political, social and economic structures and conditions that define how we as societies co-create our moral landscape.
Tessa Diphoorn is an Assistant Professor at the Department of Cultural Anthropology at Utrecht University. She is currently working on a research project that ethnographically analyses police reform efforts in Kenya. Previously she conducted ethnographic research about private security in South Africa and her book, Twilight Policing. Private Security and Violence in Urban South Africa, has been published with the University of California Press (2016).