‘The future of methane gas is electricity’: electrification as infrastructures of repair

The Rwandan government set out a strategic masterplan for development in 2000 that explicitly identified electrification, infrastructure and energy as key components of social and political repair. Rwanda Vision 2020, revised in 2012, pointed to ‘low infrastructure development’ (Government of Rwanda 2012: 5) as one of the ‘major challenges facing Rwanda today’ (2012: 4), with a need for ‘a significant step up in energy investment’ (2012: 5). The government set a goal that ‘by 2020 at least 75% of the population will be connected to electricity (up from 2% in 2000 and 11% in 2010)’ (2012: 14). This plan took ‘the devastation that marked the nation in the immediate aftermath of the 1994 genocide against the Tutsis’ (Government of Rwanda 2000: 3) as its point of departure, and was framed as a ‘reconstruction of the nation of Rwanda and its social capital’ (2000: 10).

That is, expanding energy production and electrification was central to the government's vision for post-genocide repair and creating prosperous futures. I suggest that we can think of the projects literally powering the new Rwanda, Rwanda Rushya, as infrastructures of repair that, consistent with a rich anthropology of infrastructure, both reflect the techno-politics of governance and render visible quotidian practices of citizenship, as well as recursively reshape lifeworlds and landscapes (Anand et al. 2018; Braun and Whatmore 2010; Chalfin 2017; Harvey, Jensen and Morita 2016; Larkin 2013; Mitchell 2002; Von Schnitzler 2017). I consider infrastructural repair not at face value as fixing deteriorating pipes, nor as ‘healing’. Indeed, to the contrary, in Rwanda the past twenty-five years of post-genocide social ‘repair’ has been authoritarian, often violent, and preoccupied with coercive unity (Burnet 2012; Chakravarty 2016; Doughty 2016; Eramian 2017; Ingelaere 2016; Jessee 2017; McLean Hilker 2012; Purdekova 2015; Sommers 2012; Straus and Waldorf 2011; Thomson 2013). Rather, I think of infrastructural repair in terms of how it aims to ensure ‘that spectacular failure is not continually engulfing the systems around us’ (Jackson 2015, emphasis in the original). The new barges, pipelines and powerlines, I suggest, are part of an architecture of repair intended (ostensibly) to suture together the post-genocide social fabric, ensuring that the ‘spectacular failure’ of genocide does not ‘engulf’ the region again. What do energy infrastructures, in particular, illuminate about the process of repair itself, about efforts to control violence, people and nature?

Methane has been central to electrification as repair. As one Rwandan engineer explained in 2016, ‘The future of methane gas is electricity’ (interview, December 2016). KivuWatt's operations in 2016 increased the capacity of the national grid by 25 per cent, from 100 megawatts to 125 megawatts. By late 2017, the national grid capacity had grown to 208 megawatts, with 60 per cent from hydropower, the rest from methane, diesel engines and, to a much smaller extent, peat and solar (interview, Director of Generation at National Electricity Control Centre, October 2017). The government was actively expanding power generating capacity in order keep up with 10 to 12 per cent annual growth, in order to stabilize the national grid, attract manufacturing and export power (interview, October 2017). In late 2018, plans were in place to expand methane by another 125 megawatts, as well as to add 80 megawatts from peat, 30 megawatts from solar, and 30 megawatts of hydropower, and transmission lines to Uganda and DRC were completed, with a contract signed for a line to Burundi (interview, Energy Development Corporation manager, August 2018). By 2019, electrification through the central grid remained paramount, even as the government had begun to use micro-grids to more quickly expand rural access (interview, August 2018) – perhaps unsurprising given the highly centralized Rwandan state, given that an electrical grid, Boyer wrote, is ‘subtly inclined to encourage demand, to expand itself, to solicit further dependency on its powers … Grid must be understood as the organization of enabling power that allows any invention of statecraft to occur in the first place’ (Boyer 2015: 533).

A wide body of work shows that power plants, oil rigs and dams can be central sites of reproducing state violence, and demonstrates that people generate new modes of being and forms of sociality through improvising entanglements with power infrastructures, whether through work focused on the energy source itself – such as fossil fuels (Appel et al. 2015; Barnes 2005; Behrends et al. 2011; Campos-Serrano 2013; Chalfin 2015; Leonard 2016; Watts 2004), hydropower (Folch 2019; Lord 2016) or alternative ‘clean energies’ such as wind or solar (Cross 2013, 2016; Howe 2019; Howe and Boyer 2016) – or on the powerlines, meters and grids that create illumination (Degani 2013, 2017; Mains 2012; Winther 2008). The materiality of energy operations provides one way explicitly to see and analyse the world-making dimensions of energopower (Boyer 2014) and the forms of fugitivity that operate with and against it.

Existing ethnographic work on extraction infrastructures in Africa and elsewhere emphasizes how these operations risk leading to ‘natural disasters’ (Bond 2013; Cepek 2012; Kirsch 2001; Weszkalnys 2014), and how infrastructures such as pipelines invite violence as embodiments of the unequal control over resources (Adunbi 2015; Kabamba 2013; Leonard 2016). In the case of methane extraction, as I explore below, discourses of ‘risk’ of natural disaster have been motivated to justify development of extraction infrastructure rather than to prevent it, and the absence of conflict over infrastructure is the more compelling question, suggesting the nature of carceral repair in Rwanda.

‘Freshwater timebomb’: producing and capturing risk

Rwandans living along the lake have long identified the methane as risky and dangerous, without trying to capture it: local histories reveal stories of how methane pulls swimmers down, causes their fish nets to sink, or traps fish deep in the lake so they cannot be caught. Two women who live on Kivu's northern shore told me, ‘In Kibuye where you will go, they will show you where you must not swim at any time of day; where you can go to swim and be attacked by gas’ (Hari ahantu gazi iba hano hafi hariya ku Kibuye i Karongi aho muzajya bazahita babereka ahantu utagomba koga kuri buri saha, ushobora kujyamo ako kanya gazi igahita igufata) (interview, October 2016). A businesswoman in Gisenyi told me in October 2017, ‘There are some places which contain gas which can magnetize you to the extent that you may vanish but there are some areas which are suitable to the extent that you can swim’ (Hari uduce tubamo gazi ku buryo ishobora kugukurura ukaba wazika hakaba n'ahandi hameze neza ushobora kwoga) (interview, October 2017). A fisherman explained, ‘When gas is too much it attracts them (fish) and we miss them … The gas sometimes pulls our fishnets and we miss them’ (Ni uko ahubwo hari igihe iyo gazi ari nyinshi irazikurura tukazibura … Hari igihe nk'iyo gazi idukurura nk'imiraga tukayibura rimwe na rimwe) (interview, September 2016). A Congolese teacher in Gisenyi said, ‘The fishermen know that they can't drop their nets where there is methane because the methane pulls the nets and burns the nets’ (Les pécheurs, ils savent là où il faut ne pas déposer les filets, où le gaz attire les filets, le gaz brulent les filets) (interview, October 2017). Here I explore how ‘capturing risk’ has been mobilized to justify extraction, and suggest, by extension, that producing and capturing risk is central to repair.

The methane gas is dissolved in the deepest of Kivu's stratified water layers, trapped by the density of upper layers. Scientific, state (colonial and independent) and corporate interest in Kivu's methane as threat and as resource began with the first scientific paper published on it in 1937 (Damas 1937). From 1952 to 1962, the Belgian Chemical Union conducted research on Lake Kivu's gas content (LKMP 2011: 5; Schmitz and Kufferath 1955), leading to a new Belgian-owned brewery on Kivu's shore that, beginning in 1963, pioneered using small-scale methane extraction to provide fuel for beer fermentation boilers (LKMP 2011: 5). Concern about the potential threat posed by Kivu's dissolved methane increased after a 1986 release of carbon dioxide in a similar lake in Cameroon suffocated approximately 1,700 people. By comparison, Kivu has one thousand times more gas, and two million people live in the densely populated lake basin.

A 2002 eruption of Nyiragongo volcano, on the northern shore of Kivu, prompted new studies into the lake's methane, amidst concerns that volcanic activity could destabilize the lake and prompt a ‘devastating degassing’ (Schmid et al. 2002), though this research found that a ‘gas outburst in Lake Kivu is not to be expected from future eruptions’ (Lorke et al. 2004: 778). In 2005, the scientific risk took on more urgency when a highly regarded paper argued that methane levels in the deep sub-layers were increasing more quickly than before, and that the gas therefore risked oversaturation within the century, possibly triggered by one of the area's frequent earthquakes, landslides or volcanic eruptions (Schmid et al. 2005). That is, in an area that in recent decades had been noteworthy for its political violence – civil war, genocide, military insurgencies – the locus of danger suddenly shifted to this ‘freshwater time bomb’ (Anonymous 2009), as one news source called it.

At this point, the Rwandan government and private sector partners began pursuing projects to extract the methane (see also Doevenspeck 2007). The government built a pilot plant (KP1) in 2007 that generated 3.5 megawatts of power. KP1 was intended to serve as a prototype to convince other firms to invest in extraction, and it worked. Shortly thereafter, Contour Global, an American company, signed a governmental concession for 100 megawatts of Kivu's methane, and began building a plant through their local operation, KivuWatt. Symbion, later purchased by Shema Power Lake Kivu Limited through investor Irvine Laidlaw, secured a concession to upgrade KP1 (50 megawatts), and to build a new extraction and processing facility on the northern shore (56 megawatts). The fact that the Rwandan government granted companies concessions based on volume of megawatts provided, rather than methane extracted, suggests that the future of methane was always already electricity, despite other viable proposals for alternative uses of methane – including to provide methane for individual cooking stoves or to power city buses (interviews, May 2017). By 2019, a third concession was granted defined still in megawatts, though this supplier proposed to bottle methane for cooking.

The Rwandan government and corporate partners have consistently asserted that the methane extraction project has two simultaneous aims: (1) to reduce dangerous levels of unstable gasses dissolved in the lake and thus prevent it from exploding; and (2) to provide much-needed power to meet increasing demand in energy-strapped countries (Rwanda and the DRC). As KivuWatt tweeted in March 2016, ‘Our #KivuWatt increases electricity access while solving critical risk’. This frame of capturing risk was crucial to justifying and mobilizing public-private partnerships, international scientific expertise, international funding and expropriation of land. It made the project a fait accompli. As KivuWatt's Economic and Social Impact Assessment in 2009 indicated: ‘The project represents a major positive benefit through the reduction of the risk associated with the catastrophic event and the “do nothing” scenario is not an option’ (SinclairKnightMerz 2009, emphasis mine).

That is, the threat reduction frame suggested that not doing extraction was not even a choice – a sleight of hand that silenced opposition and distracted from motivations such as profit. Contour Global's efforts were often described in nearly humanitarian terms, as a ‘noble undertaking’ (Baker 2016), with little reference to their enviable profit margins. The emphasis on threat reduction in a ‘risk society’ (Beck 2009) is familiar in post-genocide Rwanda, as the government uses a particular interpretation of history to frame certain populations as risky or dangerous through policies that consolidate power in the hands of a few while disenfranchising vulnerable populations who live in fear (Burnet 2012; Meierhenrich 2008; Vidal 2001).

Interestingly, while Rwanda Vision 2020 initially referenced the ‘large deposits of renewable methane gas in Lake Kivu’ (Government of Rwanda 2000: 19), it did not include the language of threat, which suggests that initiatives to extract methane were underway prior to the risk framework emerging. By 2017, the science that motivated the risk framework came into question. At an international scientific workshop convened by the Lake Kivu Monitoring Program (LKMP), the government agency monitoring the safety of extraction operations, a Belgian hydrologist consulting for KivuWatt presented research based on new gas measurements, arguing that not all available realistic data had been used for the comparison of gas profiles presented in the initial paper, and that there is ‘no evidence for any increase of CH4 in deep water in recent years’, and ‘No strong evidence for any increase of CO2 in deep waters from the 1970s. We are very close to a steady state’ (author's fieldnotes, May 2017). LKMP commissioned a new gas study in May 2018, with teams of scientists from Switzerland, Germany, France and KivuWatt, and indicated intentions to present findings in October 2018 (interview, LKMP staff, August 2018). By early 2020, in a white paper buried on a government website, these experts concluded: ‘The 2018 measurements do not confirm the previous hypothesis that the CH4 concentrations were increasing during the last decades in Lake Kivu. They rather indicate approximately constant concentrations since the first observations in the 1950s within the uncertainty range of the present and previous measurements’ (Schmid et al. 2019: 1, emphasis mine).

That is, the ostensible scientific fact of the need to ‘capture risk’ – the whole premise that justified the extraction project – was revealed to be tenuous, even as the infrastructure had already transformed the lake and shorelines. Meanwhile, people living along the lake expressed ongoing concerns about gas, suggesting the risk frame was simply one among a deeper history of interpretations of methane, and hinting at distrust in government projects of repair. As a businessman in Gisenyi explained in 2017: ‘Many of us know that the more they extract it the more it can be spread everywhere in the lake. The more they extract it the more it can be spread in the lake and that we may face serious effects’ (Abantu benshi rero twe tuzi ko uko bayicukura niko irushaho gukwira mu Kivu hose uko bayicukura niko irushaho gukwira mu Kivu ko dushobora kuzagira ingaruka zikomeye cyane) (interview, October 2017). Or as a woman involved in selling fish said to me, ‘I hear that there will be a change such that all people could be killed, as far as Ruhengeri, and that all this area could be on fire. Is that true or not? Is that people's rumours? Due to the methane gas. What can happen afterwards if this is true?’ (Nonese ko numva ngo nyuma hazaba impinduka ngo abantu bashobora gupfa bagashyira kugera za Ruhengeri hose ngo hano hose hakaba umuriro. Nibyo cyangwa ntabwo aribyo? Ni impuha z'abantu? Kubera ikibazo cya gazi metani. Ese biramutse bibayeho buriya byagenda gute nyuma yaho?) (interview, October 2017).

‘It has to be implemented by the marine army’: protecting the pipeline

When KivuWatt's extraction barge launched in 2015, it became the first permanent installation in an otherwise pristine lake, surrounded by undulating hills. Anchored in the lake, the barge's pumps pull the water up from 450 metres deep, where the atmospheric pressure change causes the methane to bubble out. While the degassed water is immediately reinjected into the lake, the methane is pumped through a pipeline, submerged 20 metres below the surface, 13.5 kilometres to shore. Buoys strung across the surface of the lake trace the pipeline's gentle arc to where it emerges for only a few metres just south of the main part of the town of Karongi. There, KivuWatt's new power plant converts the gas into electricity, which is then dispersed through the national grid via a conduit of newly constructed high voltage powerlines that thread across the hillsides to the National Electricity Control Centre in Kigali.

The pipeline quickly provoked controversy among fishermen. The Economic and Social Impact Assessment indicated that in order to secure the pipeline from sambaza fishermen's nets, which sink to 150 metres below their dugout canoes, civilians on the water would need to stay 250 metres away, though boats would need to be able to cross ‘so as not to hinder access to fishing grounds’ (Sinclair Knight Merz 2009: 16). Yet, when KivuWatt began full-scale extraction in January 2016, fishermen were told to maintain a one-kilometre distance from any KivuWatt project infrastructure, and were prohibited from crossing. Marine security confiscated nets, impounded boats and detained fishermen who transgressed.

Staff at LKMP mediated between fishermen and KivuWatt in mid 2016. KivuWatt insisted that it was not the company that mandated the farther distance, but rather Rwanda's marine security (soldiers) and border patrol (interview, October 2016). Though government officials told me in 2016 that they reduced the perimeter distance and resolved the issue quickly, as late as 2018 fishermen still reported a wide variation of perceptions of the distance they had to maintain, whether or not they were allowed to cross the pipeline, and who was enforcing the rules, suggesting resolution as perhaps more evasive. As one sambaza fisherman explained:

The place where we used to fish was reduced because of the pipelines they put in the water, so we no longer fish there. However, since they installed the pipelines, the rule was established saying that fishermen must fish 1,000 metres or one kilometre away from the pipelines. Marine army emphasizes on it but also the state itself because it was explained at the district level. They announced that it would be implemented by marine army. (Aho twaroberaga kubera imipira bashyize mu mazi siho tukirobera. Noneho aho bashyiriyemo imipira habayeho itegeko kuva ku mupira ujya kuroba ugomba kuroba kuri metero igihumbi kuri kilometero imwe. Marine ya gisirikare ibishyiramo ingufu ariko na Leta ubwayo nayo kubera ko byavuye ku rwego rw'akarere. Babishyira mu nshingano kubera yuko bagomba kubishyira muri marine armee.) (Interview, November 2016)

While fishermen (and they are almost uniquely men) on the whole expressed nominal support for the extraction project, consistent with the politics of unity and ‘rehearsed consensus’ in Rwanda (Ingelaere 2010), their concerns punctuated our conversations. They voiced resentment about how the pipeline reduced access to their fishing grounds and livelihoods, and uncertainty about their futures, as several companies planned to build more extraction infrastructure on the water. They consistently explained that while few had any direct interactions with the methane extractors, in the words of a seasoned fish seller, ‘The conflicts come with marine police and marine security’ (interview, September 2016). Fishermen constantly described being subject to heavy winds (umuyaga mwinshi) that could blow them into the restricted zone or tangle others’ illegal nets with their own legal ones, suggestive of their broader vulnerability to arbitrary enactments of state power and propensity to be buffeted by forces outside their control.

Many watermen further explained that their strategies to cope with these changes subjected them to further capture. Many now crossed into the DRC to fish, to circumvent both the pipeline and the new two-month periods during which the government closed the lake to fishing. In one typical story, a young man described to me in October 2018 the second time in fourteen months that his father's fishing team had been held in jail on Idjwi island:

This happened when they were fishing in the night and the illegal fishnet of Congolese fell into in our sambaza fishing net due to the unfavourable weather. Such an illegal fishnet of Congolese was torn since the wind and storm were heavy and were moving our team of fishing quickly … The Congolese in the process of finding the solution arrested our team of fishing. (Personal communication, October 2018)

Overall, lakeside residents described feeling ‘stuck’ (Jefferson et al. 2019) with(in) the increased presence of military security patrolling the pipeline.

This pipeline, which was purportedly built to minimize the risk of exposing the surrounding population to methane, is now framed as itself at risk, and in need of protection by private and public entities from precisely that population. This securitization is a common effect, especially in connection to cultures of risk and system vulnerability around the world (Beck 2009; Lakoff and Collier 2010). The formal justification to displace fishermen, circumscribe their movement, prevent them from fishing, and even confiscate their boats and nets, is (ostensibly) politically neutral: it is about protecting the pipeline so the pipeline can protect the people.

The people now perceived not as vulnerable but as threatening are these small teams of low-income fishermen in wooden canoes, fishing in the dark with hand-knotted nets and petrol lamps to attract sambaza. Many were accused or convicted of genocide crimes. They passed through prison stays as well as many forms of official state National Unity and Reconciliation programmes, such as ingando solidarity re-education camps (Thomson 2011). Meanwhile, most of the people living along the shore remained unable to access electricity, or the service economy that the new electrification purports to bring. As one fisherman said, ‘But for us, there is nothing. We did not get any benefit as we dwell here. Have we got that electricity? We do not have electricity!’ (Ariko nkatwe ntacyo. Nta nyungu twabonyemo nkaho twatuye ahangaha, uwo muriro se twari twawubona, nta mashanyarazi tugira!) (interview, November 2016).

In contrast to how Cross has described the electricity grid in rural India that at once ‘registers expectations of modernity and government and citizenship even as it maps experiences of abandonment and invisibility’ (Cross 2016), I suggest that in western Rwanda fishermen are neither abandoned nor invisible through their lack of access to the growing grid, displacement from proximate fishing grounds, and increased surveillance by military security linked to KivuWatt. Instead, the infrastructures that seek to capture risk also entangle and entrap marginal people and places through layered processes of depoliticization, control, exploitation, submission, surveillance and exclusion. That is, the new energy projects created new opportunities for surveillance, detention and capture by state military security, on the water and off, what I call ‘carceral repair’.

In invoking repair as carceral, I join those who underscore how, taking inspiration from Foucault (1975), the carceral state operates not only through prisons and law enforcement but through a wide range of institutions including mental health programmes, social welfare projects, and credit and debt programmes (Hinton 2016; O'Neill and Dua 2017; Wang 2018). Carceral repair invites us to see how carceral logics can operate through energy projects, and how they can be mobilized in the name of repair.

‘Rwanda's Guantanamo?’: capturing land and people on Kivu

KivuWatt's pipeline was inserted into an environment saturated with soldiers and state power, and methane extraction extended this carceral reach. This was particularly true on the north shores, where soldiers routinely traversed town, moving between the border with Goma, DRC and the heavily patrolled military security zone that occupied several miles of shoreline between Rubavu and Karongi. The offices for Symbion (later Shema Power) were located in this military zone, prohibited to civilians, only accessible by passing with permission through a checkpoint manned by two armed soldiers. By 2019 this road had been opened, but along the water the length of the lake, marine security from Rwanda as well as the DRC patrolled the international border. Highlighting capture in analysing methane extraction brings these soldiers and their strategies of containment to the centre.

Throughout 2016 and 2017, for example, soldiers oversaw expropriation of land and construction as a new road was carved by day labourers into the hillside outside Rubavu to where the Shema Power plant would be built in Busoro. Residents and community mediators (abunzi) described to me people's unresolved concerns, including receiving insufficient or delayed payments for their homes and land, or having their houses destroyed as collateral damage in landslides after the initial bulldozing for the road. In the same period, Human Rights Watch released a report detailing grave concerns over land cases in precisely the areas of Rubavu and Karongi in which I was working (Human Rights Watch 2017b), quickly followed by others alleging extrajudicial executions in Western Rwanda and torture and unlawful detention across the country (Human Rights Watch 2017a, c). The intertwining between methane extraction and these reports – which prompted aggressive responses by the Government of Rwanda on the radio, newspapers and Twitter, including ad hominin attacks on Human Rights Watch's director – suggest how new energy landscapes are being reshaped by violent state carceral logics.

Meanwhile, the same government military who patrol the pipeline also guard the island of Iwawa, near the Congolese border, a few kilometres from the extraction barge. The formerly uninhabited island has since 2010 housed the Iwawa Rehabilitation and Vocational Development Centre, created by presidential decree and managed by the Ministry of Youth. Supporters say Iwawa is a place for ‘idlers, thieves or bandits who are particularly troubled’ (interview, June 2018), mostly men between the ages of eighteen and thirty-five who are homeless, unemployed and/or drug-addicted, to get treatment and job training. From its inception, Iwawa has been ‘controversial’ (Rosen 2012), as critics suggest that rather than a rehab centre, the site serves to explicitly target political dissenters, and to exile, intimidate and contain mostly men (Human Rights Watch 2015), and that Iwawa should be more aptly called a ‘reform prison’ (Aglietti 2014), ‘training camp’ (Brouck et al. 2014) or even ‘Rwanda's Guantanamo’ (Bikorimana 2014). The few anthropologists who have done research in Iwawa suggest that the spectre of death is central to how it functions (Lovgren and Turner 2019). Iwawa's proximity to the pipeline amplifies fishermen's anxieties of detention, and reminds them that, while they fish at the geographical margins of state power, the carceral state is ever present. Meanwhile, KivuWatt's country manager explained to me in 2018 that they were designing a corporate social responsibility project that involved ‘supporting kids who have been abandoned, who are in a reformation space’ (interview, June 2018). While he hesitated to give further information, and nothing has yet been publicized, this suggests that the government is soliciting, or accepting, KivuWatt support in projects of social repair at the level of resocializing vulnerable youth – suggesting a complicity or entanglement of the capture logics of the government and the energy company.

I suggest that Iwawa might productively be conceptualized as another kind of energy project – a ‘waste-to-power’ plant of sorts, where men deemed otherwise disposable are captured, and produce energy in the form of manual labour. This is similar to how, in Rwanda's post-genocide justice process called gacaca in the mid 2000s, community-service work given as part of a reduced prison sentence for genocide crimes (Travaux des Intérêts Généraux) converted genocide perpetrators (waste) into physical laborers (power) who rebuilt the country manually through construction and digging roads. Thinking energy alongside capture brings these ethnographic entanglements into the same frame and invites further questions about how carceral logics are complicit in producing geographies of disposability that themselves produce power. That is, the waste that is generated by power-intensive ways of life under late capitalism – whether landfills or unemployed people – comes to itself be framed as something that can produce further power. Waste-to-power projects (Ahmann 2019) and waste generated from power projects (Hecht 2018) are ethnographically rich sites that reveal dynamics of struggle living within political-economic inequalities generated by new energy frontiers.

Overall, methane infrastructures bring into relief the captures, I suggest, that are part of the broader process of repair more widely, as we see in land seizures and forcible detentions in Iwawa – what I am calling ‘carceral repair’. Along with new power plants, new pipelines, powerlines and roads reach across Rwanda's hillsides, connecting, relocating, bypassing and ensnaring people, in ways that reinscribe historically deep patterns even while producing new entrapments. Methane extraction both reflects and reproduces how formerly dangerous people and places are captured through electrification projects fuelling fuller incorporation of Rwandan elites into global domains of power. Similar dynamics of dispossession and entanglement occur through Kigali's urban infrastructure and design (Shearer 2017) and through the materiality of rural villagization projects (Wendel 2014). Carceral repair underscores how energy extraction and electrification are methods of state violence predicated on capture and containment.

Carceral repair and anti-carceral energies: repair as a politics of habitability?

The Kigali Convention Centre and the rest of Kigali's newly illuminated skyline, where cosmopolitan Rwandans watch international sports and African leaders plan the continent's future, is predicated on increased vulnerability to capture of men and women on the lake at the nation's farthest edge. This, I suggest, is not an accidental externality but integral to the infrastructure of (carceral) repair. Defining the methane as dangerous, then capturing it to ostensibly reduce risk, then using the resulting infrastructure to contain the people, is part of the logic of social repair.

My analysis here is consistent with an emerging body of cultural anthropology that has emphasized how conditions of contemporary life are shaped by containment (Doughty 2019), inviting us to see politics not as ‘abandonment’ but rather as a ‘matter of captivity’, a ‘politics constituted by the tracking and capturing of humans and animals’ (O'Neill and Dua 2017: 5). The methane extraction projects on Lake Kivu suggest that energy infrastructures and energy politics should be analysed ethnographically as carceral spaces alongside clear sites of caging ranging from prisons to detention camps, forcible rehabilitation centres or infectious disease quarantine centres (Gomez-Temesio 2018; Jefferson et al. 2019; O'Neill 2018), and analyses of surveillance directly linked to border control or justice systems (Besteman et al. 2018; Cooper 2018; Jusionyte 2018). Considering energy projects in relation to captivity is nothing new; capture has always been central to the design of, and lived experience in, energy frontiers – as resources, landscapes and people are harnessed to produce power. Global capitalist expansion was fuelled by capture of energy sources from Africa, from kinetic energy of enslaved bodies, to gas and oil today. These intertwined legacies endure, manifest for example in the Niger Delta, where many people hold the idea that victims of the slave trade were transformed from black bodies to black crude oil, and thus oil is ‘ancestral promise’ (Adunbi 2015: 6–7).

In characterizing Rwanda's post-genocide rebuilding initiatives as carceral repair, I am perhaps merely glossing what many have already described as the authoritarian and often violent nature of the Rwandan state. Many have noted that the Rwandan government is preoccupied with capturing people and ideas deemed risky to neutralize threats. Through ethnographic attention to the methane project on Lake Kivu, I aim to bring together analyses of post-genocide unity programming with more standard development projects, and illustrate how the carceral logics commonly identified in so-called reconciliation programming – manifest in ingando ‘solidarity’ re-education camps and mandatory participation in grassroots gacaca courts, for example – pervade other seemingly neutral initiatives such as roads or powerlines. Carceral repair as an analytic emerges from the continuities between people's lived experiences with expanding energy infrastructures and their engagements with other state projects of post-genocide rebuilding, many of which are also experienced as confining, constraining and punitive. For many Rwandese, specifically for many of the most vulnerable, new power projects and increased electrification nationwide mean increased social control through surveillance, dispossession and even capture. Carceral repair sharpens attention to how carceral logics infuse the everyday, in large part produced through energy politics and energy infrastructures.

Identifying energy projects in Rwanda as carceral repair complicates official characterization of these projects as laudable governmental efforts to protect people from environmental risk, and opens up a new set of questions about what anti-carceral energy projects would look like. How might energy and electrification infrastructures allow us to imagine repair in ways that undo rather than replicate logics of capture? What possibilities do energy infrastructures and practices provide for producing alternative socialities and disruptions? Can energy projects generate what Stacy Langwick, writing about ‘gardens of remedy’ in Tanzania, calls a revolutionary ‘politics of habitability’ as an antidote to toxicity generated by the past (Langwick 2018)? Langwick explores how local practices of cultivating ‘gardens of remedy’ are challenging techniques of governance implemented in the name of health:

They do not lend themselves easily to dystopic or utopic visions … They are experiments in creating spaces for a new politics of habitability, one that continually asks: What kinds of lushness can be cultivated in twenty-first century Tanzania? What relations enable bodies and landscapes to grow ampler, denser, more productive and more potent? What ongoingness and what forms of continuance do our gardens support? (Langwick 2018: 436)

What would it look like for energy projects to promote new politics of habitability, and how might it work against logics of extraction, carcerality and disposability? Ethnographic work on Lake Kivu suggests that such projects would need to combine the creative and cutting-edge use of local resources that characterizes methane extraction in Lake Kivu with decarcerating logics to promote more capacious, resilient and sustainable forms of social rebuilding, power and care.

Meanwhile, the women, men and children living on the farthest edge of Rwanda, fetching water from the lake, swimming in it, fishing in it, selling the fish it gives, still know that the methane remains a dangerous beast beneath the water, whose unpredictable threat may well outstrip efforts by the state and corporate partners to harness it.