Police guns and private security cars. Ordering the state through socio-material policing assemblages in Nairobi

Abstract

In this article, I show how the work of heterogeneous security and policing assemblages in Nairobi hinges upon and reproduces physical urban borders, and consequentially enacts social orders. While these assemblages enrol a diverse collection of people and objects, I liken their work to that of the state: some urban residents are considered as belonging to safe spaces and in need of extra protection, while others are considered dangerous and targets of policing activities. I draw on one year of ethnographic fieldwork with private security companies and police patrols in middle- and upper-class Nairobi. In Nairobi, armed police personnel are commonly seen in vehicles that are marked with the logos and colours of security companies or private vehicles. These arrangements are not only based on agreements between companies’ managers, urban residents and police, but rely on what specific infrastructures (such as road or radio networks) and various objects (such as guns and cars) afford. These material elements are not insignificant details. Rather they become central to the unfolding of the patrols. They contribute to the work of security and policing assemblages of categorizing Nairobi’s residents as either dangerous and potentially criminal subjects or as residents in need of extra protection.

Introduction

In this article, based on one year of ethnographic fieldwork in Nairobi, I look at heterogeneous policing and security assemblages, mostly in middle- and upper-class neighbourhoods, specifically in the form of joint patrols. In these joint patrols, armed police officers ride in the vehicles of private security companies, and together with private officers they patrol some of Nairobi’s neighbourhoods. I show how these particular arrangements unfold in relation to other ongoing social processes, material conditions and technologies: the socio-economic statuses of the people involved, mobility and communication infrastructures (such as roads and VHF radio networks), guns and cars. Policing and security assemblages enact the state as a process of groups making: they contribute to delineate groups of people in terms of safety and danger and, consequentially, in terms of belonging (or not) in the community of citizens who deserve (or not) protection.

In Nairobi’s congested traffic, the sight of armed police personnel in vehicles that are clearly marked with the logos and colours of private security companies or other private vehicles is quite common. While police are usually armed with assault rifles, private security companies in Kenya are not allowed to carry firearms. ¹ These arrangements unfold in Nairobi where, on the one hand, policing work is often reduced to and identified with the use of firearms. A gun-wielding police officer becomes a very specific and relevant element of Nairobi’s socio-material policing assemblages. On the other hand, the logistic and financial challenges of the Kenyan National Police Service usually translate into minimal availability of cars for patrolling, of other assets and of officers to carry out regular police work. The vehicles made available by security companies or residents become necessary tools for policing the city. Vehicles themselves, however, come with their own particular conditions of possibility, such as a functioning road network. This article shows how a diverse constellation of objects and technologies contribute to the unfolding of security and policing practices ² in Nairobi.

The consequences of opening up to such a multitude of elements and complexities are not to overlook power relations between, for example, the police and the security guards, or between the wealthy residents and the police. Instead this analytical avenue highlights how seemingly mundane objects afford specific power relations and the categorizations of groups of people in relation to security and policing concerns. It does not dismiss the symbolic meaning of objects and material infrastructures (such as, for instance, the authority symbolized by gun-wielding policemen), rather it goes beyond this to show how such objects become active and relevant in producing social orders.

The policing and security assemblages of Nairobi work towards the knowledge, classification and realignment of different urban residents into categories that are concerned with danger and security. As such, they fit into what Trouillot (2001) defines ‘state effects’, which often take place beyond national institutions and government sites. Trouillot (2001) suggests to ‘track down these practices, processes and effects whether or not they coalesce around the central sites of national governments’ (131), as long as they are concerned with the production of individualized subjectivities, the realignment of these subjects into collectivities and the creation of tools and knowledge towards their classification and regulation (126). These are the types of practices and effects visible in the security practices I analyse in this article.

Adopting an assemblage approach brings to light three critical and interwoven elements that require emphasis. First, groups (of Nairobi residents, in this case) are not pre-existing starting points but they are made and unmade (cf. Latour, 2005) through the unfolding of policing practices. Second, the making and unmaking of such groups as deserving more or less protection happens without a necessary and a-priory dominance of social dimensions over material elements. The social and the material are necessarily assembled together and enact the state together, symmetrically. Finally, such an approach shows how the state is not a fixed entity under which policing takes place. Rather the state comes into view as an effect (Trouillot, 2001) of policing practices.

In the coming sections I first provide an overview – though partial – of Nairobi (in)security issues, giving particular relevance to the narratives and the various policing activities that took place in the neighbourhoods where I conducted my ethnography. Next, I explore the academic debates on the role of infrastructures and objects, especially in relation to security and policing. I then suggest that we think in terms of assemblages, as heterogeneous networks in which human and non-human elements are symmetrically enrolled. This, I propose, helps to shed light on specific political consequences of policing and security practices. I show that objects and infrastructures, such as private security cars, police guns or road networks, are not just details of policing activities but actively contribute to modes of ordering Nairobi’s residents as dangerous and criminals or as subjects in need of protection.

(In)security and policing in Nairobi

Ever since Nairobi was founded at the end of the 19th century, (in)security has constantly been on the agenda for administrators and politicians, police, companies and urban residents (Anderson, 2002, 2011; Colona and Jaffe, 2016; Musoi et al., 2013; Ruteere and Pommerolle, 2003). Currently, private security companies, police and residents’ associations are the actors who extensively deal with the insecurity problems of the city. Particularly, the presence of private security companies in urban areas is ever increasing (Colona and Diphoorn, 2017), both through personnel and through technologies and artefacts such as alarms, metal detectors, cameras, barbed wire, etc. Following structural adjustments in the 1980s and 1990s, the funding for police and the public sector in general plummeted, as some of my interlocutors recalled. Today the number of private security officers dwarfs the amount of police personnel by five to one.

The empirical material I discuss in this article comes from participant observation during night and day patrols in private security cars, with and without police. Additionally, I attended various security fairs where the abundance of technological devices on sale was mirrored by an abundance of conference panels articulating who and what the danger seemed to be, and how the technological wonders on display could manage this danger. I carried out several interviews with residents and various stakeholders. I lived in three different middle-class neighbourhoods. In one of these, a private security company hosted police personnel in its car that was usually stationed next to my apartment block. Long and casual chats with the officers there were priceless learning moments.

Throughout my fieldwork I learned how danger is dealt with in these neighbourhoods of Nairobi. My interlocutors often cited their own experiences to explain what they believed were the most relevant dangers: armed home invasions and robberies, car-jackings, cons and frauds, and violent thefts. Among other things I was struck by how some specific security threats were given emphasis in newspapers, commercials and everyday talk. For instance, public venues could become the target of a terror attack, such as the al-Shabaab strike on Westgate in September 2013 or on the Westland DusitD2 Hotel in January 2019. Domestic or commercial workers could become involved in ‘inside jobs’: complicit in passing information about residential or commercial premises to criminals, or stealing from their bosses themselves. And lastly, a diffuse concern about ‘thugs’ hiding behind any corner and ready to commit robberies.

The neighbourhoods where I conducted my fieldwork were part of the northwest constituencies of the city. The wealth and class ³ geography of the city can be interpreted, with a few exceptions, on an east–west line. In connection with the former colonial management of the city, ethno-racial categories ⁴ intersect with class (cf. Murunga, 2005). Low-income African residents mainly populate the eastern neighbourhoods, the city centre is the business district, while moving westwards there are middle- and higher-income neighbourhoods. Upper-class neighbourhoods are either pockets surrounded by middle-class areas, or are altogether further removed, taking the place of former white farmlands towards the south-western or north-western fringes of the city. While the overwhelming majority of the residents of the low-income areas in the east of the city are of African origin, the middle-class areas house both African residents, Asian residents that moved to Nairobi during and after the colonial times, and an ever-growing group of European and American ‘expats’ who live and work in Nairobi. Based on my own observations, in the upper-class areas, wealth became the main factor influencing residential access, rather than ethno-racial markers.

The joint patrols I explore in this article took place primarily in one of the upper-class neighbourhoods, and were generally initiated or directly organized by the residents. Thus, the analytical contribution of this article is necessarily affected by the modes and places of production of its empirical material. Most of the residents of these neighbourhoods expect the police to be the institution in charge of policing, while their resident organizations and the private security companies seem to be held in an ancillary position (see Colona and Diphoorn, 2017). In the following vignette however, the work of companies and associations seems to shift. Rather than just facilitating the police to be in charge of policing and security provision their role seems to be a necessary component for the joint-patrol arrangements to work.

At the time of my research, the residents association of an upscale neighbourhood, Greenwoods, sponsored a security partnership between the police and a private security company, Maximum Security. ⁵ The night patrol team of the company regularly hosted two police officers with their guns, and the association paid them 250 Kenyan Shillings per shift, on top of their regular salary. ⁶ At the beginning of the night shifts, the Maximum Security team drove to the local police station to pick up the officers. These routines did not necessarily play out smoothly. Conflict often arose between private security officers and police regarding the comfort of the car, or the absence of a VHF radio on the police frequencies, or the inadequate extra pay, among other issues. One night, however, everything went as well as anyone could hope for.

James, the commander of the Maximum Security’s patrol team, drove us to the police station. A sleepy policeman opened the gate, which was freshly painted in the colours of the Kenya Police flag (a stark contrast to the generally run-down aesthetic of the place). The gate’s renovation was also sponsored by Greenwoods residents’ association. The police officers who were assigned to our team quickly boarded the car, and I moved from the front seat to the back to make space for the senior police officer in charge. He directed the driving with movements of his chin, showing James which turns to make. Unexpectedly for me, we drove towards Bonde. Bonde was a poor urban settlement a stone’s throw away from the upscale Greenwoods, but the team was not authorized to patrol there. As we approached the settlement, one of the private security officers warned me to make sure I knew that this place was where all the ‘thugs are, live, or come from’.

The physical border between the two neighbourhoods is a major highway that we had to cross to enter Bonde. As we left the highway the roads immediately turned rough. After about a hundred metres of a slow bumpy climb over a small hill, we turned left. The street was now smaller and darker. As we kept driving it became even narrower and busier with the lights and loud music of local bars and restaurants that catered to the local residents. This was a busy – but peripheral – road of the neighbourhood, and people had to stop and step aside to make space so that our car could slowly make some progress. During this time the police officers held their rifles high and visible through the car windows. With severe facial expressions they scrutinized many of those who happened to be on the street, in what seemed a very well-rehearsed performance.

As the road ended we drove back to green and lush Greenwoods. In Bonde there were no more roads accessible by car, and James seemed in a hurry to get back to his official area of operations. The large tarmacked roads in Greenwoods, many of which were lit by street lights and seldom used by pedestrians, allowed for a more dynamic and faster driving. We were able to patrol every single road of the neighbourhood. None of them was left unchecked, and we drove up into dead-end alleys where James put on the high beams to make sure that the gates of the wealthy residences had not been breached. The streets were quiet and large enough that we could park the patrol car in reverse against a tall fencing wall on a major road. Here the police officers relaxed and seemed unconcerned about the surroundings. One of them was sitting in the car, resting his cheekbone on the muzzle of his rifle. The other, strolling up and down a stretch of the road, killed some time by kicking grass at the edge of the curb, which did not conform to the straight lines of the neatly manicured flowerbeds. In contrast to the patrol we had just carried out in Bonde, the body language of police officers communicated safety and tranquillity.

This vignette raises a few points I would like to address. Certainly, there are important issues concerning the socio-economic power differentials between different groups of residents and between them and the police or the private security officers, as I and Diphoorn (2017) have suggested elsewhere. In this article, however, and in relation to the concerns raised by this volume, I set to unpack how some policing assemblages, in which the police are only one component, come to rely on various objects and technologies, such as vehicles and firearms, or on the physical and infrastructural (planned or unplanned) makeup of the city, such as streets, highways and hills – among others. These, I show, are not mere details of these patrol arrangements, but rather critical aspects of their implementation. Through the diffused work of these assemblages, the residents of poor urban settlements become the object of policing, while in the middle-upper class areas the policing is for the residents.

Borders, policing and assemblages

Objects, technologies and infrastructures such as walls, roads, gates and barbed wire (among others) are entangled with policing and security practices (see Bremner, 2004; Caldeira, 2000; Smith, 2015) as tools helping to border spaces. Bordering practices have a longstanding role in affecting classifications of spaces as public or private, state or non-state, actively contributing to govern and manage populations within such spaces. People inhabiting these spaces are consequently categorized as ‘us’ versus ‘them’, ⁷ ‘dangerous’ versus ‘non-dangerous’, or ‘criminal’ versus ‘in-need-of-protection’.

Infrastructures have the capacity to contribute to such distinctions. Graham and Marvin (2001) point to their double-sidedness and their potential to splinter urban spaces, even in the event of their own failure (Graham, 2010). While infrastructures facilitate the lives of some people, they simultaneously make access to spaces and services more difficult or impossible for others, highlighting what Rodgers and O’Neill (2012) call ‘infrastructural violence’. Considering that Graham and Marvin (2001) extend their analysis beyond classic infrastructure such as roads, water and electricity to include, for instance, cyber-technologies, it becomes clear why Graham (2011) would later talk about ‘ubiquitous borders’, which come to be embodied in different materialities and practices. Various technological artefacts and objects such as gates, urban CCTV systems and metal detecting scanners become obligatory passage points and security infrastructures.

In urban areas very tangible physical barriers, such as perimeter walls (Caldeira, 2000), gates (Smith, 2015), highways (Rodgers, 2012) and houses (Bremner, 2004), contribute to bordering spaces within and between neighbourhoods, and are often enrolled in security or policing practices. Anthropological analyses often focused on the symbolic relevance rather than the materiality of objects and technologies (Hull, 2012), or the influence these might have on ‘the social’ (Latour, 2005). For instance, in Hansen and Stepputat’s (2001) work, ‘buildings, monuments, letterheads, uniforms, road signs [and] fences’ become ‘symbolic languages of authority’ as ‘the materialization of the state in series of permanent signs and rituals’ (8). Caldeira’s (2000) ethnography of crime and segregation in San Paulo (Brazil) talks about a ‘city of walls’. She shows how fears reacting to the ‘talk of crime’ push middle- and upper-class Paulistanos to retreat into fortified enclaves. Walls for Caldeira (2000) become a forms of class distinction that ‘confer status’ (258) within a cultural-symbolic understanding of Brazilian society.

Some scholars have attended more directly to the materiality of these technologies and infrastructures. Rodgers (2012) emphasizes how the design of the Managua highway purposefully promoted the mobility of wealthier city dwellers at the expense of poorer communities, whose residents were often considered criminal. In the Kenyan context, Smith (2015: 142) talks of ‘fortress Nairobi’ to show how tall perimeter walls, gates and privately operated public roads become ‘material and architectural manifestations of security in Nairobi, and the way in which these shape modes of living in the city’. In the analysis of bounded spaces in so-called post-apartheid Johannesburg, Bremner (2004: 465) highlights how ‘the house’ becomes a ‘site of double domestication’: a place where ‘rituals of humiliation’ are mobilized to identify who can come in or stay out, or as a place where a black person can be either admitted as a security personnel or be considered a potential criminal. In these analyses, however, objects and technologies remain (somewhat passive) tools in the willing hands of humans, or are portrayed as manifestations of social processes, effectively excluding them as part of the same social world they manifest.

In this article I build upon these sensitivities and emphasize the active work done by objects and technologies to shape security practices. To do this, I turn to an assemblage approach inspired by Actor-Network-Theory and Material Semiotics. Generally, approaching security practices in terms of assemblages has proved to be fruitful: Abrahamsen and Williams (2011), for instance, talk of global security assemblages to show how the privatization of security is entangled with state and international governance, which often plays out at the urban level. With a focus on the African continent, the term ‘assemblage’ helped in tracing the dynamic and fluid connections, often volatile, (Higate and Utas, 2017) between various protagonists of security, such as private providers (Diphoorn, 2017), state actors (Reno, 2017), secret societies (Albrecht, 2017), ex-militias (Mynster Christensen, 2017) and groups that in Kenya are often described as gangs (Rasmussen, 2017). In Jerusalem, Volinz (2018) suggests the term ‘modular security toolbox’ to show how through the enlistment of specific ‘security modules’ (e.g. private security companies, security infrastructures, local authorities, etc.) state security actors effectively assemble the state. Similar to the way that objects and technologies appear as passive tools in human hands, the ways in which security assemblages are approached in these works tend to foreground a reality that is ‘out there’ (Law, 2004) rather than as a dynamic process of assembling (Neyland, 2008), whose components contribute to engender important political consequences.

With this article, I contribute to these debates on security assemblages, by advancing a more symmetric analysis of objects and materiality in policing and security practices, and show how they contribute to urban b/ordering. I consider assemblages as socio-material networks of ‘human’ and ‘non-human’ elements that, symmetrically and heterogeneously, become relevant and are put to work – for the case of this article – within security and policing practices (see Colona, 2019). Heterogeneously and symmetrically are here to indicate that the human and non-human elements together mediate social processes without a pre-established hierarchy (Latour, 2005: 38–39). As Latour (2005: 76) explains, it is not about establishing ‘absurd symmetry’ between humans and non-humans. Rather, it is about foregrounding a theoretical and methodological sensitivity (Law, 2004), which ‘simply means not to impose a priori some spurious asymmetry among human intentional action and a material world of causal relations’ (Latour, 2005: 76).

So, returning to the vignette above, in the unfolding of the patrolling activities, the vehicle, the guns and the diverse road network are symmetrically relevant as much as the residents, the private officers and the police. The human and the non-human elements of this assemblage together mediate very different ways of policing two neighbourhoods in Nairobi. When I suggest that we think of these objects and technologies as mediating social processes, it is not about some sort of liberal agency or intentionality. It rather means thinking of a form of ‘relational materialism’ where objects and people ‘never act alone’ (Abrahamsson et al., 2015: 15, emphasis in the original). Thus, neither objects nor people liberally act out of the intentional will of some ‘isolated individual actors’. This approach emphasizes the profound relationality of policing and bordering practices (Kyed and Albrecht, 2015) that comes into existence through socio-material assemblages. These lenses allow for an analysis of the social and political consequences that security and policing assemblages engender.

Security, subjects and the state

Policing and security are exceptional themes by which to explore relations between ‘state’ and ‘citizens’ (Kyed, 2009), or to explore how security becomes a central concern of state apparatuses (Glück, 2017). Specifically, security practices offer a special lens to learn how state and political subjects are simultaneously and mutually enacted (Colona, 2019). Deploying ‘joint patrols’ in some neighbourhoods and not others, having police guns as an important element of this activities, or using private security vehicles as the technology of mobility for patrolling, contribute to enact different groups of urban residents as particular categories of subjects. The belonging of different residents to different categories emerges as one of the state effects that Trouillot mentions, and it is through this process that state and political subjects are enacted simultaneously.

Talking about different categories of subjects must be understood as ways to be and ways to be perceived in the world, and as such they are possible only when certain distinctions are socially and historically made. For example, Hacking (2002: 99–100) shows how labelling certain people as ‘perverts’ was only possible after specific sets of behaviour were identified as perversions. Rather than using groups and categories of people (i.e. dangerous, criminals, in need of protection, etc.) as incontrovertible starting points, I emphasize how groups are ‘being made or unmade’ (Latour, 2005: 27). To this end, thinking with assemblages offers a privileged window to unpack the various social and material elements that contribute to these categorization processes.

In relation to policing and security practices, looking at how groups are being made is particularly relevant for two reasons. First, policing and security practices, such as the joint patrols, aim to distinguish between people that are dangerous or safe, and, on this basis, to grant or deny them access to malls, offices, flights and much else. Constructions of safety and danger thus inform who belongs to a given space and who does not, who deserves protection and who must be excluded and removed, easily leading into categorization of ‘us’ versus ‘them’. Thus, security and policing come with and expose specific modes of ordering ⁸ people. In turn, these modes of ordering become instances of state formation, not as an institution or a locus of authority and/or sovereignty, but as a process of performing and enacting a political community. ⁹ These are examples of those ‘state effects’ that Trouillot (2001) described and that do not necessarily coalesce around state institutions.

Second, these modes of ordering do not begin and end within the social realm. They are not effects as if they were dependent variables influenced by pre-dating social processes (e.g. social and economic disparities or ethno-racial tensions in some of Nairobi’s neighbourhoods). Certain material conditions of possibility, such as the road network in the vignette above, contribute to shaping specific policing and security practices. It is in this way that the work of objects, technologies and infrastructures contributes to the categorization of certain groups of residents as (not) fully belonging to a space or a community.

The sensitivities brought by this approach resonate with and complement many contributions in political anthropology and anthropology of the state. In these disciplines, the state, rather than as a natural, reified and stable entity, has come to be approached through its fluid processes of formation (e.g. Buur, 2001; Ferguson and Gupta, 2002; Gupta, 2012; Hansen and Stepputat, 2001) and through those localized relations and activities that bring it into existence. Anthropological and ethnographic research on the everyday provision of security not only attested the importance of non-state collective actors in policing practices, worldwide and especially in the African continent, but also how these were constantly entangled with state apparatuses (Albrecht, 2012; Buur and Jensen, 2004; Diphoorn, 2016; Meagher, 2007). Buur and Jensen (2004), for instance, highlight the complex ambiguity of this type of policing that remains somewhat oppositional to the state apparatuses when residents take security and justice into their own hands.

Neoliberal explanations are often brought into the analysis to make sense of the outsourcing of security to individuals, companies or groups of citizens (e.g. Sundar, 2010). Yet, the role of the residents (and their associations) that emerged from my fieldwork in Nairobi fits this box uncomfortably. These residents’ initiatives, in fact, derive from the expectation that it is exactly the state police who should take care of security. What follows is that in Nairobi, contrary to now old arguments (e.g. Helman and Ratner, 1992; Zartman, 1995), ‘the state’ has tremendous importance in the everyday lives of urban residents, especially through its security organs such as the police. Depending on the neighbourhood, police presence is either sought after and paid for, as I show in this article, or avoided at all costs as van Stapele details in this volume, since they are seen to bring insecurity (see also Price et al., 2016; van Stapele, 2016). However, they are never irrelevant.

The heterogeneous assemblage of policing in Nairobi

The Kenya Police Force was created during the British colonial rule and was initially divided in two main branches: the regular police and the administration police. The latter, commonly known as the ‘tribal police’, was often considered to be a political tool in the hands of local administrators and chiefs. Mainly recognizable from combat fatigues, it now consists of three main units: the Rapid Deployment Unit, the Rural Border Patrol Unit and another unit that takes care of the security of government buildings. The regular police, usually wearing dark blue uniforms, are considered to be the main police body, and are deployed for more traditional tasks of crime control, investigation and law enforcement. The Regular Police consists of several units: the General Service Unit, the Traffic Police, the Diplomatic Police Unit (DPU) and some more, each with a different mandate. Following the implementation of the Constitution of Kenya in 2010, the new police service (not a force anymore) is being restructured. Among these changes, the Administration Police and the Regular Police were brought under the single command of the Inspector General of the police. Generally, however, the police were considered to be ineffective, inactive, incapable of dealing with crime, and often in collusion with criminals (cf. Musoi et al., 2013; Omenya and Lubaale, 2012), corrupt, ill-trained and lacking basic equipment. These are some of the reasons that were brought up by my interlocutors to contextualize and justify the involvement of private security companies and residents in the joint patrols.

In Nairobi there are different configurations through which different types of co-operations between police and other groups take place, and it is by no means a recent practice (Ruteere and Pommerolle, 2003). The DPU, for example, had a formal Memorandum of Understanding with some private security companies. The DPU catered to the security needs of the specific population of diplomats and UN personnel, who often reside in some of the areas where I conducted my fieldwork. Representatives of these groups of residents also took part in regular meetings to share crime intelligence, and adapted their strategies accordingly. Another less regulated co-operation is within the Cash in Transit operations of private security companies (cf. Diphoorn, 2015, 2019). In these arrangements Administration Police personnel regularly travel in the ‘follow-car’, a private security car that drives behind an armoured vehicle transporting valuables through the city. The police officers in the follow-cars are always armed. ¹⁰ Another instance is what one of my interlocutors described as ‘gentlemen’s agreement’. Through personal acquaintances in the ranks of the different police stations, security companies’ managers call for the help of the police for securing events or venues which they have been contracted for. These personal ties are also called upon to guarantee the armed presence of the police in their mobile response teams during their patrols in different neighbourhoods. The necessary material arrangements required for these practices to unfold, such as the armoured vehicle, the police with firearms and other vehicles, show how mundane material conditions allow for specific policing practices to take place.

At times, although not as often, armed police officers can be seen in unmarked private vehicles as part of initiatives organized and led by some resident groups. In Forestgrounds, a middle-class neighbourhood whose residents are predominantly of Asian descent, a community policing group does exactly that. ¹¹ This group was born at the time of the first multi-party elections in 1992, when seven presidential candidates for the first time in the history of the Kenyan Republic ran for office. Since these elections generated fear of violence and insecurity, a small group of Forestgrounds’ residents with a police background organized themselves in a community policing organization, which over time grew in size, capacity and legitimacy in the neighbourhood. Under the umbrella of this organization, four volunteers pick up four police officers from the local police station every day. The eight-strong party is usually split in two private vehicles, each hosting two civilians and two police officers. They patrol together every day from six in the afternoon until midnight. When security intelligence suggests higher risk, the patrols continue until six in the morning. The community policing organization also operates a control room and a VHF radio network that provides a bridge between the residents and the police. Residents of these neighbourhoods are so deeply used to this ‘symbiotic arrangement’ (Colona and Jaffe, 2016: 6) that while there is a conceptual distinction between the community policing organization member and the police officer, residents have come to understand them as a co-presence.

From these examples it is already possible to understand the relevance of the police and various technologies and objects of Nairobi security assemblages. In Nairobi, as I and Diphoorn (2017) argued, the police act as a ‘coagulating agent’ (8) of security provision activities. They are the ones who make these arrangements possible and meaningful for all the different actors involved. At the same time such arrangements could not be performed and imagined without the vehicles provided by private security companies or residents’ associations, the VHF radio network the community policing group provides, or the guns carried by the police officers. It is the working of all these heterogeneous elements of the assemblages that contributes to bordering the city’s spaces and, consequently, their inhabitants, in relation to their access to or being targeted from security provision and policing practices. That is, the spatial bordering of the city necessarily enables a mode of ordering urban residents.

Similar to the episode described in the vignette, some neighbourhoods such as Greenwoods become spaces where police officers are put on the ground during day or night shifts. It suggests how these areas are enacted as worthy and in need of extra security provision, possibly with a gun. The people living in these neighbourhood become the beneficiaries of these patrols: residents to be protected. Other neighbourhoods are either excluded from such activities or targeted as the hotbeds of dangerous criminals to be policed: police presence is granted there only to show that the police are watching and menacingly flaunting guns.

Residents’ associations often distribute specific tasks to some of the members who are elected to the managerial board. Once I had a conversation with the resident who was delegated to manage security for Greenwoods residents’ association. I had previously heard rumours that NW Security, a competitor of Maximum Security in the neighbourhood, was planning to implement a co-operation arrangement with the police independent of the residents’ association, possibly as a marketing strategy. The feelings of the security delegate regarding this revelation were clear: if NW Security does it, for whatever reason, and guarantees an armed police presence in the neighbourhood, that would comply with the principle behind the original partnership arrangement between the association, Maximum Security and the police. ‘Having police around in my neighbourhood’, he said, was the reason why the association promoted and financed the arrangement in the first place. The normative goal he strived after was clear: it is the armed police the desired actor for the policing of their territory. At the same time, it implicitly assumes that there are specific and necessary financial and material conditions for this to happen, such as the availability of a vehicle to host armed police officers.

Cars and guns

Joint policing patrols rely on a particular set of objects and material infrastructures beyond the guards and the companies, the police and its officers, and the residents and their associations. Cars, guns, money and other technological artefacts are similarly relevant. Specific types of streets or mobility infrastructures and spaces influence how these policing practices unfold. These objects, technologies and infrastructures are not meaningless details or a neutral background of the agreements between the police, the companies and the residents’ associations, but elements that enable and mediate policing and security practices.

A manager of the community policing organization that was active in Forestgrounds explained their operation as the ‘eyes, ears and wheels’ ¹² of the police. Community policing organizations and residents’ initiatives are often considered the ‘eyes and ears’ of the police. In such terms, civilians are expected to provide to and share information with the police, while they are not an active part of their operations. In practice, residents are expected to co-operate by relaying information. The specific addition of ‘wheels’, however, shifts the balance. The manager reinforced the necessity for the police to be mobile by saying that, when emergencies happen, ‘we take them [the police] there’. The stress he put on this aspect of the arrangements shows, first, that the police are in a position in which they need to be taken somewhere to do their expected job. Second, it implies the necessity of a car as the appropriate technology of mobility for such an arrangement to hold.

Residents and private security companies alike initiate these arrangements and go out of their ways to have armed police officers join them. All these efforts suggest that the residents expect – and clearly want – the police to do the policing. Or at least a sort of policing that requires the presence and the possibility to use firearms. The police officers therefore are sought after and paid extra in order to sit with their guns in the private security cars that are patrolling these neighbourhoods. Their armed presence in wealthier neighbourhoods is considered a good to be financed and materially enabled. Also, according to the police, guns are not marginal appendages of police officers. Officer and gun become a very specific heterogeneous assemblage of policing that points to a symmetry of these two elements: police officers and guns.

Once I was visiting my friend Ibrahim, a private security officer who lived in a poor urban settlement where he was celebrating his wedding. While he and some other people were showing me around, two men came through a door with rifles slung over their shoulders. Ibrahim elbowed me, aware of my interest in security and police, and whispered to me: ‘look, police’. I asked how did he know who they were, and he simply replied: ‘haven’t you seen the guns?’. Another time, during a late-night patrol in Forestgrounds, the mobile response team and I were patrolling in the neighbourhood. At a major road junction in the area an armed police officer was walking alone. The reaction of the team leader was very interesting: ‘What is a gun doing alone at this time of the night?’. After I asked some clarification over the meaning of his question he explained to me that police guns should always be at least two in number, so they can provide safety for each other. I find this synecdoche to be illustrative of how guns and police officers become such a blurry heterogeneous assemblage whose elements are difficult to distinguish.

Guns and policemen, and vehicles and private security officers are symmetric and necessary elements for these policing arrangements to unfold. Vehicles allow mobility in an efficient and relatively quick way, bringing the gun–officer assemblage where it is needed or wanted. During a patrol night in a wealthy neighbourhood of Nairobi, one police officer affectionately tapped his short-barrel AK47 hanging over his unmarked blue rain jacket and said: ‘I give them backup, they give me mobility’, referring to the private security company officers and their cars. He described himself and his colleagues as if they were the muscle helping out the powerless security guards. During the same night he proposed a thought experiment, as an educational device for myself, while the security guards were checking a residence after a false alarm. If the false alarm we were attending to had been an actual emergency, he suggested, and if he had not been there with the private security car, he would have probably been idle at the local police station. In that case, the residents would have to call the station and arrange transportation for him and his colleagues. It would have taken a long time, and the ‘thieves’ would have already left by the time he or one of his colleagues reached the scene. Here the car is not a detail of the arrangement. It is as central as the gun and the police officer himself, and suggesting an a priori hierarchy would not do justice to the complexities of the relations that these arrangements rest upon.

Furthermore, in these policing practices cars and guns are put in relation to even more technologies and people. An alarm needs to be installed, activated and connected to the control room of a security company, somewhere else in Nairobi, so that a patrol car can be dispatched. There must be a controller in that room who sends the local response team and their vehicle to the exact address, so that the gun–police officer assemblage can be present and offer backup in case of an emergency. Additionally, a response vehicle has to negotiate roads that are not always in good condition, and at certain times of the day traffic becomes an impediment to a quick response. Thus, operations managers of security companies need to always reconsider strategic locations where they can locate response vehicles in relation to recent traffic patterns, road works, construction sites, customer geographical distribution and time of day.

Private security companies have mobility thanks to their cars, but lack the firepower provided by the police. Vehicles are not just patrolling tools of private security companies, but become a condition of possibility that enables the police to be present over a large territory. The relations between all these elements in such socio-material assemblage have social and political consequences for the resident of Nairobi. First, they suggest that the kind of policing that is expected is more a repressive kind of policing (Francis, 2012) rather than a proactive one (Clarke, 2006), which is more frequently associated with community policing activities. Second, the specific work afforded by socio-material security and policing assemblages suggests that residents of wealthy neighbourhoods become the subject of protection, rather than a subject of policing: guns and cars are there to ensure that their gates and perimeter walls are intact and effectively bordering the lives of the residents from whatever could be happening outside their private spaces. Through the same work afforded by these assemblages, people in Bonde (the poor urban settlement we drove through in the vignette) become subjects of suspicion and policing rather than protection. They are considered and enacted as criminal and dangerous.

Physical borders, social orders

The joint patrols I analysed in this article rely on various technologies and infrastructures: a gun, a vehicle, an alarm able to send an activation signal, a control room to receive it and a radio network to dispatch a patrol vehicle. Among the most relevant and necessary conditions of possibility are those mobility infrastructures that spatially make possible for these arrangements to unfold. That is, roads that are large enough for a vehicle to drive through are required to make sure that the response team and the police can reach the location they are dispatched to. In the example from the vignette, the type of road network profoundly affected the type of policing. The crowded narrow street in Bonde afforded only slow progress and high proximity to the residents, who were considered and enacted by the armed police officers as dangerous and threatening. The large and paved roads in Greenwoods, on the other hand, allowed for fast driving and afforded a quick presence of the police if an emergency occurred.

The way in which different road networks affect the type of policing that is possible was made clear to me by Chakur, a regular interlocutor of mine who was born and had lived for more than 40 years in Forestgrounds. Before we met he had been an active member of the community policing organization active in his neighbourhood, and as such he volunteered and patrolled with police officers in private cars. During one of our long chats in his living room he described to me the routines in the organization. He recalled when he and his buddies chased ‘thugs’ who – allegedly (my addition) – had just committed a crime in or around Chakur’s neighbourhood. He told me that they often escaped by entering Small River, a poor urban settlement at the fringe of Forestgrounds. When that happened, the chase was over. Chakur and his fellow volunteers could not follow the fugitives further into Small River, as they believed it was too dangerous for them, especially if the chase had to be continued on foot in the small alleys of the settlement. As many others frequently reminded me (both from the community policing organizations and from the private industry), when thugs enter a ‘slum’ the chase is over. In a poor urban settlement, thugs are at home, Chakur explained, and they were far away from the reach of the community policing volunteers and their patrol cars. Chakur’s explanation foregrounds the ‘slum’ as a homogeneous dangerous space where, independently of who you are, if you are a thug, you are welcome and at home. By contrast, this becomes a space where police and community policing volunteer are either unwelcome or in danger.

The enrolment of police, guns or vehicles in socio-material security and policing assemblages does not happen in a vacuum, it is entangled with multiple material borders that actively mediate the policing. These borders are not only the backdrop where policing happens and where people are either protected or targeted by these practices. These physical borders and infrastructures become part of the policing practices themselves and have consequences for the kind of patrolling that is possible. These consequences are not just technical corollaries deriving from the possibility to drive (or not) to a place or over a road. As Chakur’s story suggested, the impossibility of driving and chasing suspected criminals into poor urban settlements reinforces social imaginaries of these spaces as criminogenic. In turn their inhabitants are essentialized to (potentially) criminal subjects.

Concluding remarks

Objects and devices are not simply an extension of someone’s will, they mediate social processes. In this article I showed that socio-material security and policing assemblages in Nairobi enrol vehicles, guns, alarm systems, police, private security companies and their officers, and the residents of middle- and upper-class neighbourhoods. I highlighted the relevance of some of these elements, such as cars, guns and road networks, and suggested that they do more than just symbolic work. Both humans and non-humans in the assemblage never act independently from each other (Abrahamsson et al., 2015), but together contribute to bordering city spaces and ordering urban residents simultaneously. This ordering is one of the most relevant consequences of policing and security provision practices. I suggested – following Trouillot (2001) – that we should consider this ordering as a state effect.

The various elements of socio-material assemblages that are mobilized in security and policing practices hinge upon and reproduce physical borders within the city. For instance, some spaces are accessible by patrol cars while others are not. These physical borderings of urban space contributes to how urban residents are enacted; they do not only reflect spatial or socio-economic rationales. Security and policing practices shift spatial and socio-economic realms into that of criminality and danger. Thus, some people are considered to be in need of extra armed protection, while others – often living in poor urban settlements – are seen as dangerous and criminal subjects.

Like the state effects described by Trouillot (2001: 126), the security and policing practices explored in this article are concerned with the production of individualized subjectivities, their realignment into collectivities and the creation of tools and knowledge towards their classification, regulation and management. The bordering (re)produced through security and policing practices affords specific modes of ordering residents as (not) belonging to the neighbourhood, the city, a community or a state. It is in this respect that security and policing practices are processes whose effects reverberate beyond the realm of security provision, into that of order-making (Kyed and Albrecht, 2015). They allow for certain categories of people to emerge (Hacking, 2002) and to be essentialized as criminal or dangerous, and thus ‘strangers’, ‘outsiders’ or ‘alien’ (Isin, 2002) to a specific (political) community. Others instead emerge as belonging citizens who are worth of protection.

In this article I bring together theorizations about the state in political anthropology and sensitivities promoted by thinking in terms of assemblage. I suggest that we approach the processes of categorizing residents through the work of security and policing assemblages as ‘state effects’ (Trouillot, 2001). Thinking in terms of assemblage, with its characteristic sensitivity to the relational enactment of reality, opens up the state to a new analysis. It allows to recognize the state in its process of becoming rather than as a predefined entity and as the obvious source of power and authority, which would then trickle down to the police. Here, the state is enacted as the effect of categorizing people as dangerous or in need of protection. Thus, thinking in terms of assemblage shows how state power and authority are made and unmade simultaneously with policing and security practices. In turn, the latter make and unmake groups of residents as categories of people who deserve security or not, and as categories of people who belong (or not) to the state.