Trust, distrust, and mass atrocity prevention: The Central African Republic

Abstract

The article calls on academics and policymakers who focus on mass atrocity prevention to engage with Trust Studies. This is needed because trust and distrust are commonly identified as a significant factor in destruction processes, yet there remains no substantive engagement with these concepts. The article combines Trust Studies, interdisciplinary research on the Central African Republic (Anthropology, Sociology, African Studies, and Political Science), and primary sources to analyse social and political trust dynamics through an exploration of (a) leadership, (b) outsourcing, (c) identity politics, and (d) witchcraft. It makes a twofold contribution. First, it provides a more informed understanding of the mass violence that took place in the Central African Republic through a historical analysis of trust dynamics. Second, it considers the implications for mass atrocity prevention, as it argues that the mainstream commitment to ‘rebuilding trust’ is built on misguided assumptions. The case study holds broader implications for both Trust Studies and mass atrocity prevention. Ultimately, it calls for interdisciplinary research to aid our collective understanding of the multifaceted roles that trust and distrust play in mass violence.

Introduction

This article calls for a new interdisciplinary research agenda that brings together (a) studies on mass atrocities such as Genocide Studies and the Responsibility to Protect, and (b) Trust Studies with its focus on concepts such as trust and distrust as well as their application to the real world. This is needed because trust and distrust are cited as important factors when explaining mass atrocities (genocide, crimes against humanity, ethnic cleansing, and war crimes) and feature prominently in mass atrocity prevention strategies, but these studies do not engage with research on trust and distrust.

Consider Sémelin’s seminal study, in which he argues that one of the most fundamental causes of mass violence is distrust, as collective memories are forged and manipulated into thinking that ‘it is from THEM that all our suffering arises. We cannot trust them. Those people are not like US’.¹ We see this time and time again in literature. During the ‘Mytilenean Debate’ (427 BC), it is argued that their betrayal of the Athenians represented a ‘serious breach of trust’ and that only ‘violent retaliation’ could address this.² The ‘religious cleansing’ of Spain in the early 17th century is said to have taken place because ‘a court faction argued successfully that new Jewish and Moorish converts could not be trusted’.³ In Rwanda, Hutu extremists saw what was taking place in Burundi and concluded ‘you can’t trust the “Inyenzi” (cockroaches)’, which paved the way for the 1994 genocide.⁴ In Iraq, Al-Qaeda’s number two Ayman al-Zawahiri wrote to Aby Musab al-Zarqawi warning that ‘Shi’a Muslims could never be trusted’.⁵ In each case, the underlying logic was that the group in question cannot be trusted and must be destroyed before ‘they’ harm ‘us’. Such thinking is also embodied in mass atrocity prevention strategies. For instance, the United Nations Framework of Analysis for Atrocity Crimes views mass atrocities as caused by ‘[w]idespread mistrust in State institutions or among different groups as a result of impunity’,⁶ and ‘mistrust between opposing parties based on past or present breaches of commitments or agreements’.⁷ From this perspective, mistrust and distrust play an important role in mass atrocities, as people are killed en masse because perpetrators do not perceive they can be trusted. Yet other than these fleeting references, we do not see any substantive engagement with the concepts in question.

To illustrate this further, let us turn to a more contemporary case, which remains a grossly under-researched case of mass violence, the Central African Republic (CAR). Between 2013 and 2015, an estimated 3,000–6,000 people were killed, 825,000 internally displaced, and 423,000 forced to flee CAR, although the United Nations acknowledges these numbers are a ‘radical under-estimate’.⁸ A striking statistic is that the Muslim population in the capital Bangui may have been reduced by up to 99 per cent in just a couple of months.⁹ Explaining why these mass atrocities took place, academics have cited distrust as a key factor.¹⁰ Indeed, Gatfoaui went as far as to proclaim ‘It is all about Trust!’, though this was in a short blog piece that provided little by the way of in-depth analysis.¹¹ Significantly, the importance of trust was identified by policymakers as critical for preventing further mass atrocities. Jan Egeland (the former UN Undersecretary-General for Humanitarian Affairs) identified ‘mistrust’ as a key theme within a ‘deeply divided nation’ and called for ‘maximum pressure placed on the transitional government’ … to ‘rebuild trust between the divided communities’.¹² More recently, in the wake of post-electoral violence in 2021, Mankeur Ndiaye (Head of the United Nations Multidimensional Integrated Stabilization Mission in the Central African Republic [MINUSCA], and the UN Secretary-General’s Special Representative for CAR) called for a ‘genuine and deep reconciliation process to rebuild trust among communities’.¹³ What we see, therefore, is that on the one hand, trust-related issues are cited as a fundamental facilitator of mass atrocities, whilst, on the other hand, policymakers stress the need to rebuild trust to prevent future mass atrocities. Although one can understand why policymakers focus on rebuilding trust, a deeper engagement with Trust Studies begins to illustrate that such thinking is built on misguided assumptions.

The article makes two contributions. Primarily, it sheds new light on understanding the mass atrocities in CAR by providing the first in-depth study of trust dynamics through a historical analysis of social and political trust. It argues that high levels of political and social distrust help explain why there was a coup, why it was successful, and why the discourse of threat gained traction and helped radicalize people to perpetrate atrocity crimes. Breaking this down, political distrust towards elites helps explain both the coup itself and its ultimate success. President Bozizé was viewed as increasingly untrustworthy, as he reneged on various commitments, and, as a result, he lost both internal and external support as his ability, integrity, and practices were rejected. Following the coup, social distrust helps explain why the discourse of threat took hold. When one juxtaposes both social and political distrust, the complexity of trust dynamics begins to show the magnitude of the task at hand for anyone seeking to address trust-related issues in CAR.

The second contribution is to mass atrocity prevention, as it calls on academics and policymakers to reassess the ‘rebuilding trust’ approach for four reasons. First, it assumes trust did exist, broke down, and can, in turn, be rebuilt; however, in the example of CAR, trust never existed in the manner implied. This is important because if academics and policymakers misunderstand how trust and distrust operate, the normative recommendations put forth may be at best flawed or, at worst, counterproductive.¹⁴ Second, it fails to understand the multidimensional nature of distrust. The studies identified above all speak to the destructive nature of distrust, but they fail to consider the normative value of distrust.¹⁵ Distrust can act as a form of protection, which seems particularly relevant when analysing what Gerlach refers to as ‘extremely violent societies’¹⁶ such as CAR.¹⁷ Accordingly, it would be an error to think that all types of distrust could or should be eradicated. Third, such thinking embodies an overly simplistic view of the relationship between trust, distrust, and cooperation, as it proceeds on the basis that trust is a priori for cooperation, whereas CAR shows that cooperation can occur where trust is absent. Fourth, mainstream approaches fail to consider that external actors (the United Nations, great powers, and regional states) are viewed with distrust because they have a track record of exploitation and harm in CAR.¹⁸ Accordingly, there are tensions to navigate for any external actors seeking to address trust-related issues in CAR, as trust and distrust play multifaceted roles. Distrust may shape day-to-day behaviours and act as a source of protection; however, distrust in political elites can incentivize coups, which can create conditions that enable mass atrocities, especially in an environment where social distrust is entrenched and can be manipulated.

This article is structured in four parts. First, it draws on Trust Studies to lay the theoretical foundations for the study. In so doing, it highlights the contested nature of trust and distrust, the complex relationship between them, and the normative value of distrust. Shifting its focus to the case study, the second and third sections draw on Stoneman’s differentiation between ‘political trust’ (in the government) and ‘social trust’ (between individuals and groups).¹⁹ Section two analyses political trust through a focus on leadership and outsourcing, while section three looks at social trust by analysing identity politics and witchcraft. Finally, the conclusion reflects on the key findings whilst calling for further interdisciplinary research.

Trust and distrust

Studies on trust can be divided into two strands. First, academics focus on trust at the domestic level. Historically, political theorists such as Hobbes, Descartes, Mill, Kant, and Rawls invoked the concept, even if briefly. Over time, interdisciplinary studies arose which saw trust studied in relation to organizations,²⁰ democracies,²¹ post-communist societies,²² and civil society.²³ These reflect the multiple contexts in which trust is deemed to be doing something of notable importance. Second, studies on trust at the international level tend to be rooted in the discipline of International Relations but draw on interdisciplinary research, as they analyse how trust shapes international affairs.²⁴ At first glance, one may be forgiven for assuming that CAR fits neatly into the former, but so-called fragile states are more complex. What goes on in CAR is shaped extensively by the external involvement of the UN, the French, the Russians, and regional governments, which exposes global–regional–national–local linkages.

Defining trust can be a challenging proposition, as the term remains contested. Interdisciplinary studies have provided myriad interpretations, with trust being variously described as a ‘mental or psychological state’,²⁵ ‘a judgment’,²⁶ an ‘emotional attitude’,²⁷ or a ‘a belief’,²⁸ to name just a few. For the purposes of this article, trust is defined as ‘the expectation of no harm in contexts where betrayal is always a possibility’.²⁹ This definition is helpful because it begins to illustrate the relationship between trust and related concepts such as vulnerability, risk, and uncertainty.³⁰ The actor[s] involved choose to trust on the basis ‘that potential trustees will do what is right’,³¹ whilst at the same time aware that their expectations may be disappointed.

This brings us to the next key consideration, identifying the factors that people look for when deciding whether to trust others. Mayer, Davis, and Schoorman’s seminal article on trust in organizations put forward three characteristics: ‘ability’, ‘benevolence’, and ‘integrity’, the so-called ABI model (which is utilized below to discuss political trust in CAR).³² They review the existing literature and acknowledge the many different dimensions of trust that are raised but argue that these can be condensed into the characteristics listed above. Ability refers to the competency of the trustee, in that a trustor would only trust actor X to do task Y if the former thought the latter had the necessary skills and expertise to complete the job at hand. Benevolence refers to the perception that the trustee views the trustor as someone that will do good by the trustee, at least in relation to this specific issue. Here, we see the threat of betrayal loom large, in that there is an expectation that the trustee will not betray the trustor. The final characteristic is integrity, as the trustor will only trust the trustee if they believe that they will abide by a set of principles that they themselves align with.³³ The ABI model dominates the discourse, even if academics alter it as they study different aspects. For instance, Wheeler’s study of interpersonal relationships sees him alter the ABI model to look at ‘benevolence’, ‘integrity’, and ‘capacity’ instead of ‘ability’.³⁴ Stoneman’s study on trust in government focuses on ‘perceived intentions’, ‘perceived ability’, and ‘perceived motive’.³⁵ What we see, therefore, is that whether academics are studying interpersonal relationships at the international level (Wheeler) or trust in society and government at the domestic level (Stoneman), the underpinnings of the ABI model persist; they are discussed below within the context of political distrust, which brings us onto the issue of types of trust.

Stoneman differentiates between ‘contractual’ and ‘paternalistic’ trust and sees ‘social trust’ as an example of the former and ‘political trust’ as an example of the latter.³⁶ Everyday social interaction is very much dependent upon social trust being formed, but this goes beyond simple day-to-day activities, as social trust can play a critical role in civil war³⁷ or conflict management.³⁸ Political trust is different as this requires individuals to place their trust in elites and public institutions. The nature of these two types of trust is different. As Stoneman explains, we can have ‘thick’ trust, which refers to relations between humans, and ‘thin’ trust, which refers to trust in a government.³⁹ For example, the bonds of trust that underpin the vows made on a wedding day are thicker than the trust placed in a political party when casting a vote at the ballot box. Critically, however, as Stoneman’s study shows ‘trust in government has social origins’.⁴⁰ In other words, we should not think of social and political trust as operating in vacuums detached from one another, because the level of trust in society at large can have implications for the level of trust we have for political elites.

A further dimension, performance, must also be factored into this analysis. As Mishler and Rose explain, ‘trust in political and social institutions is contingent on economic and political performance’.⁴¹ This is important for so-called fragile states such as CAR, as they are often portrayed as the benchmark for underperforming governments. Therefore, it is necessary to say a few more words on the relationship between weak institutions and trust. Hutchinson and Johnson’s study of capacity, trust, and legitimacy views trust as critical for regime legitimacy.⁴² Their research on 16 African countries from 2000 to 2005 led them to conclude that ‘high institutional capacity is associated with increased levels of individual trust in government across African countries’.⁴³ In so doing, they uphold Bratton, Mattes, and Gyimah-Boadi’s view that individuals within African states trust governments based on performance rather than a commitment to abstract ideas.⁴⁴ The study has significance precisely because CAR is a yardstick for underperforming governments, which is underpinned by outsourcing, concessionary politics, and hollowing out the state.

When it comes to the relationship between trust and distrust, Lewicki, McAllister, and Bies highlight that the two concepts were traditionally viewed at opposite ends of a continuum.⁴⁵ From this perspective, trust and distrust do not coexist. Trust is viewed as ‘good’ and distrust as ‘bad’, with the former viewed as a prerequisite for ‘social order’.⁴⁶ We see this thinking again and again as trust is viewed as ‘a precondition for cooperation’.⁴⁷ Fukuyama goes as far as arguing ‘a nation’s well-being, as well as its ability to compete, is conditioned by a single pervasive cultural characteristic: the level of trust inherent in the society’.⁴⁸ Rooted in the idea that trust has an economic value, Fukuyama views high levels of trust as having ‘major [positive] consequences’, whereas a ‘decline in trust’⁴⁹ or ‘widespread distrust’ results in ‘transaction costs’ which create barriers to cooperation and prosperity.⁵⁰ Over time, revisionists have challenged this approach on two fronts. First, trust and distrust can coexist. For Lewicki, McAllister, and Bies, the two concepts are ‘linked’, as they define trust as ‘confident positive expectations regarding another’s conduct’ and distrust as ‘confident negative expectations regarding another’s conduct’.⁵¹ Going further, Cook, Hardin, and Levi argue that societies can function perfectly well in the absence of trust.⁵² Many mechanisms that underpin a healthy society do not actually depend on or embody trust relations. Second, the normative value of distrust comes to the fore.⁵³ This has two dimensions, in that distrust can stimulate ‘the development of improved institutions’, which in turn may ‘facilitate cooperation’.⁵⁴ Therefore, rather than being seen as a fundamental barrier to cooperation, distrust can facilitate it.

The normative value of distrust is captured in Benjamin Franklin’s famous quote ‘distrust and caution are the parents of security’. The statement begins to highlight that a high level of trust does not necessarily aid peace and harmony. As Baier explains, ‘exploitation and conspiracy, as much as justice and fellowship, thrive better in an atmosphere of trust’.⁵⁵ In ‘a moral test for trust’, Baier challenges the assumption that to trust is positive: ‘[w]hen the trust relationship itself is corrupt and perpetuates brutality, tyranny, or injustice, trusting may be silly self-exposure, and disappointing and betraying trust, including encouraged trust, may be not merely morally permissible but morally praiseworthy’.⁵⁶ As Baier goes onto explain, within the context of a ‘corrupt system’ it may be unwise for individuals to trust anyone with anything.⁵⁷ This sentiment has been reiterated in several different contexts since. In Lewicki, McAllister, and Beis’s study on the coexistence of trust and distrust, they argue that ‘unconditional trust appears to be an extremely dangerous strategy for managing social relations’.⁵⁸ Trust, therefore, can have ‘two faces’, in that blind trust can be harmful whilst ‘sceptical trust’ is of value.⁵⁹ Distrust, therefore, can act as a form of protection, and this is precisely what we see in CAR.

The Central African Republic

Between 2013 and 2015, CAR experienced mass atrocities unprecedented in the country’s post-colonial history, yet it remains a grossly under-researched case study. For example, the Oxford Handbook of the Responsibility to Protect was published in 2016 and includes 12 case study chapters, none of which are on CAR, which only appears four times in this landmark text.⁶⁰ To give another example, the journal dedicated to the study of the Responsibility to Protect, Global Responsibility to Protect, has published just one article on CAR, and this was 10 years on from the Séléka uprising.⁶¹ The majority of the studies on CAR have been done by anthropologists, sociologists, and African Studies experts. When explaining why the mass atrocities took place, we are told that they are a product of colonial legacies,⁶² a historical culture of violence,⁶³ elite struggles over power,⁶⁴ regional dynamics,⁶⁵ inequality and fear,⁶⁶ the failures of international interventions⁶⁷ and peripheral neglect, outsourcing, and the uniqueness of the coup.⁶⁸ From a mass atrocity prevention perspective, what we see is a mosaic of factors put forward, and the purpose of this section is to shed new light on trust dynamics, a hitherto-neglected factor in analyses of CAR.

To explain why a focus on trust and distrust is necessary, let us reflect on the conclusions drawn in Carayannis and Lombard’s seminal study on CAR:

while the violence was expressed using a religious idiom, it had less to do with doctrinal differences or hatred, and more to do with uncertainty, mistrust, and manipulation whose unfortunate long roots in CAR are beginning to bear fruit.⁶⁹

The statement is important for three reasons. First, the authors rightly reject the mainstream portrayal that the atrocities formed part of a religious cycle of violence between the Séléka (mainly Muslim) and anti-Balaka (mainly Christian). Second, offering an alternative explanation, they ask us to consider the social conditions which facilitated the mass violence, in particular, the role that mistrust and manipulation played in creating an environment in which cracks became chasms; but again, the authors do not engage with the concepts. Third, these social conditions did not arise overnight and have long historical roots. Accordingly, this article does not focus on 2013–15 but instead analyses how political and social distrust has been constructed through pre-colonial, colonial, and post-colonial practices, with a predominant focus on the post–Cold War era.

Political distrust

CAR went from its first democratic elections (1993) to unprecedented mass atrocities (2013–15) in just two decades. To help explain this, the section focuses on individual leaders in the post–Cold War era, prior to analysing the historical practice of outsourcing and implications that stemmed from it. These provide a more informed understanding of why there was a successful coup in 2013 but also highlight the multifaceted nature of distrust, as it can act as a source of protection but also foster a rebellion.

Political leadership

Stoneman explains that leaders and leadership act as a critical component of building trust in government.⁷⁰ Here, the characteristics of the ABI model, and variations of it, are important, as citizens judge aspects such as ability, benevolence, integrity, motives, and intent.

President Ange-Felix Patassé (born in Paoua, north-west CAR) became the first democratically elected leader of CAR, yet his record of human rights violations and corruption undermined his plea for the people of CAR to trust him. In his first term (1993–9), he had to overcome several attempted coups, as the ‘southerners’, who generally supported his predecessor President André Kolingba, rose up against the ‘northerners’, who backed Patassé. In a bid to hold onto power, Patassé made a radio appeal: ‘Central Africans … you elected me to lead you. Trust me to re-establish peace and national unity, trust me to rebuild our country … Together, we will triumph.’⁷¹ To return to the characteristics associated with the ABI model, the appeal is grounded on a plea: if you give me a chance, I will prove my ability to unite this country, and I will not betray you. This formed part of a broader promise that he would govern with, rather than against, his critics.⁷²

Although Patassé won a second election in 1999, his role in human rights violations and major corruption betrayed his promise of peaceful intent and, in doing so, fuelled distrust. Regarding the former, following a failed coup in 2001, there were widespread extrajudicial killings, torture, and ill treatment of civilians and soldiers suspected of supporting the coup.⁷³ Patassé is also said to have ordered the punishment of civilians, which led to Congolese rebels (supporting Patassé) conducting ‘systematic’ rape ‘against all women’ of ‘all ages’.⁷⁴ In other words, grave human rights violations were used in an attempt to hold onto power, but this ultimately failed. Regarding the latter, in 2005 he was tried in absentia, sentenced to 20 years, and fined five billion francs for embezzling seventy billion francs.⁷⁵ Notably, the money was embezzled whilst he failed to pay state employees, with salary arrears mounting up to the point where they became a key factor in the 2003 mutiny.⁷⁶ Reflecting on the 2005 ruling, a businessman claimed the sentence should have been more severe because Patassé had ‘betrayed the people of the CAR’.⁷⁷ When one considers that Patassé was the first democratically elected leader, his actions fuelled widespread distrust and undermined the hope that democracy would bring an end to elite-led violence.

If we return to the ABI model, it is evident that the ability of Patassé to govern, at least in terms of his power, was propped up by external forces. Over time, his track record of corruption and human rights violations undermined any perception of benevolence or integrity, which helped pave the way for political actors to side with Bozizé. As Mehler explains, by this point ‘getting rid of Patassé at any price was apparently the order of the day’.⁷⁸ Bozizé seized power in 2003 and stated his goal of ‘presidential elections … and parliamentary elections if necessary’.⁷⁹ Accordingly, it would appear Bozizé was trying to address the trust deficit which typically follows a coup. Within three days, the Republic of Congo’s foreign minister Rodolphe Adada proclaimed: ‘[h]e spoke of openness, of reconciliation … we believe that the Central African Republic can trust a man who says the kind of thing that we have heard’.⁸⁰ To return to the ABI model, the statement begins to illustrate that external actors expressed their view that they did trust Bozizé’s motives and intentions, and therefore he should be given an opportunity. He was not just someone who had seized power, but someone who could be trusted to do the right thing.

Unfortunately for the people of CAR, Bozizé accelerated corruption and outsourcing to the point that International Crisis Group famously labelled CAR a ‘phantom state’.⁸¹ The government was involved in ‘rampant corruption’ and ‘consolidated power in the hands of the President’s immediate family and associates’, which in turn is said to have led to ‘legitimate grievances’ against Bozizé.⁸² This further undermined the government’s ability to perform, which, as discussed above, only fostered further mistrust, as citizens distrust underperforming regimes. If this was not bad enough, widespread human rights violations made things worse. In 2007, Human Rights Watch released a report entitled ‘State of Anarchy’ documenting widespread human rights violations in CAR since mid-2005.⁸³ The report drew attention to the fact that ‘the vast majority of summary executions and unlawful killings, and almost all village burnings, have been carried out by government forces, often in reprisal for rebel attacks’ and also that the ‘level of civilian fear in northern CAR is palpable’.⁸⁴ To return to the normative value of distrust, trusting the government in this environment would be ‘silly self-exposure’, to use Baier’s terminology.⁸⁵

In what could be seen as an attempt to rebuild trust, in 2008 Bozizé requested that CAR be added to the agenda of the UN Peacebuilding Commission, which then created a fund to address ‘persistent political instability, the presences of multiple armed groups, and the limited State authority and services outside the capital, Bangui’.⁸⁶ At the time, the Commission focused on security sector reform and disarmament, demobilization, and reintegration; however, Bozizé ‘proved reluctant to implement many of the reforms to which he had committed in the dialogue’.⁸⁷ For Orlin, a critical failure arose, as the United Nations Integrated Peacebuilding Support Office in the Central African Republic (BINUCA) could not get President Bozizé to ‘uphold his end of the bargain’, which, it is claimed, was the very same problem that the Bangui Accords faced when the United Nations Mission for the Central African Republic (MINURCA) failed to persuade President Patassé to implement the agreement.⁸⁸ Time and time again, therefore, presidents in CAR made commitments (along with other actors such as political parties, trade unions, civil society) to ‘lend full support to all initiatives to promote dialogue and national reconciliation’ but subsequently failed to see through these processes.⁸⁹ This further embedded and fostered a culture of distrust in political elites. If we define distrust as ‘confident negative expectations regarding another’s conduct’,⁹⁰ both Patassé and Bozizé created a negative expectation that they would not fulfil promises made and should not, therefore, be trusted.

The downfall of Bozizé owed a lot, evidently, to his lack of integrity, because his ability to govern, like other leaders in CAR, was propped up by regional actors. If they had continued to support Bozizé, he would have remained in power. By betraying the regional powers, the last dimension of the ABI model, ability, fell apart, as he lost external support. To better understand this, let us turn to Marchal’s focus on the regional dynamics and the regional distrust that emerged. For instance, the region contributed 5 billion CFA francs to a Disarmament, Demobilization and Reintegration programme (DDR) in CAR in 2008, though ‘not even two billion of CFA francs were actually used for it’.⁹¹ The integrity of Bozizé was thus a significant problem, as he was viewed as untrustworthy. Against this backdrop, individual relationships began to deteriorate. The most important was between President Bozizé and President Deby of Chad, whose power and influence within CAR led him to be seen as the ‘kingmaker’.⁹² Between 2006 and 2012, their relationship got worse, but, despite this, in 2012 Deby tried to prevent further turmoil in CAR by calling for a ‘national dialogue, if not some kind of power sharing’, only for Bozizé to accept then ‘renege on his commitment soon after’.⁹³ The betrayal further undermined any sense of integrity and reinforced the view that Bozizé was never committed to change. To give another example, Marchal explains that Congolese leader, Denis Sassou Nguesso, ‘like many heads of state in the region, felt that Francois Bozizé never tried to convert to head of state’.⁹⁴ The broader issue was that this came at a time when regional elites were questioning not just Bozizé’s benevolence and integrity but also his ability. The president of Gabon, Ali Bongo, ‘had given up hope on Bozizé … long before the Séléka emerged’ and saw him as ‘dead wood’.⁹⁵

The Séléka rebellion was unprecedented. Whilst CAR had experienced many coups, this was the first time that a rebellion in the periphery had managed to take hold of the capital Bangui.⁹⁶ Distrust in the government played a key role in facilitating the coup, as the Séléka was made up of multiple rebel groups who cited ‘the government’s unmet obligations under various peace accords as well as the failure of the DDR process’.⁹⁷ From their perspective, if the president of CAR continued to renege on his commitment to peace, they had little incentive to uphold the status quo. Of course, a commitment to rebel in and of itself does not guarantee a successful coup, and here the distrust that Bozizé had created at the regional level helped facilitate his downfall, as President Deby supported the Séléka’s advance. Even then, though, regional heads of state tried to prevent the fall of Bangui by deploying MICOPAX; meanwhile, Bozizé managed to secure the support of South Africa, which deployed 400 soldiers to bolster his defence.⁹⁸ Again, we see Bozizé’s position being propped up by other actors. In January 2013, the Libreville Agreement was signed, which, in hindsight, was Bozizé’s last chance. Speaking at the time, Margaret Vogt (then head of the United Nations Integrated Peacebuilding Office in the Central African Republic) told the UN Security Council that the Séléka had advanced with little resistance from the national army because of the ‘depth of decay within the armed forces’ and that the soldiers had ‘simply dropped their weapons and melted into the bushes’.⁹⁹ Against this backdrop, one may be forgiven for thinking that Bozizé would seek to implement the agreement, but, instead, ‘he refused to implement the agreements provisions’, and ‘MICOPAX [Mission for the consilidation of peace in the Central African Republic] tactically withdrew its support for Bozizé, making no effort to implement the Libreville Agreement’.¹⁰⁰ The fact that the Séléka could take over the capital in three months highlights that the last dimension of trust – ability – had fallen.

Following the post-2012 mass atrocities, the final report of the Commission of Inquiry into CAR claimed:

The only goal of successive corrupt governments was personal enrichment of the political leaders and members of their families through embezzlement of public funds, looting of public corporations, and illegal exploitation of precious minerals and other natural resources while a very large majority of the people lived in abject poverty.¹⁰¹

The statement is important because it underlines the historical track record of corruption in CAR, yet it also raises the problematic issue of how accurate it is to say that trust is to be rebuilt, in a society in which political elites have always been regarded as untrustworthy. When one considers the track record of lies and embezzlement, it is difficult to imagine that the citizens would ever trust political elites. Rather than proceed on the assumption that trust broke down and needs to be rebuilt, we need to better understand the deep roots of political distrust itself. The focus now shifts to the historical practice of outsourcing.

Outsourcing

Although very little is known about CAR’s pre-colonial era, it seems that high levels of distrust were common. Drawing on Christian Prioul’s landmark study, Smith argues that pre-colonial CAR was ‘politically, socially and economically, organised in response to the devastating impact of sustained slave raiding’ and that within this era there was a ‘pervasive atmosphere of distrust’.¹⁰² Whilst one may question what such historical events have to do with the present day, it is important to note that historical sources of distrust can have a direct bearing on contemporary developments. In Nunn and Wantchekon’s study, Slave Trade and the Origins of Mistrust in Africa, they combine individual-level survey data with historical data on slave shipments and find that ‘individuals whose ancestors were heavily raided during the slave trade are less trusting today’.¹⁰³ The generational impact is put down to two things. First, stories told from one generation to the next shape societal values, beliefs, and norms. Second, the weakening of legal and political institutions fuelled mistrust in the government.¹⁰⁴ It is this latter point that is the focus here, which brings us back to the aforementioned issue that trust in governments can be based on performance rather than a commitment to specific ideals.¹⁰⁵ In CAR, outsourcing has fuelled long-term distrust in the government precisely because it has undermined its ability to perform the tasks that citizens view as the primary functions of the state.

The roots of outsourcing in CAR can be traced back to French colonialism, which was predicated on ‘neglect punctuated by outburst of arbitrary brutality’.¹⁰⁶ Looking at what was going on in Congo Free State, Paris decided to subcontract ‘1,233,000 sq km to forty private companies that were granted a thirty-year monopoly’ and, in so doing, effectively outsourced its sovereignty.¹⁰⁷ Whilst the companies and arrangements changed over time, the common thread was that there was very little real investment and the actors involved committed widespread human rights violations. As Lombard notes, the French government were ‘repeatedly made aware’ of ‘widespread murders and torture’ – which in today’s terms could constitute crimes against humanity – and would pass protection legalization but were unwilling to invest the money needed to create institutions that could oversee and enforce these laws.¹⁰⁸ If we accept the depiction of pre-colonial CAR as one made up of a ‘pervasive atmosphere of distrust’,¹⁰⁹ it seems that the French embedded and accentuated distrust through their own brutality, as well as turning a blind eye to both human rights violations perpetrated by companies and the need for institutions that could foster trust.

Post-colonial practices have further entrenched the colonial practice of outsourcing. As Smith explains ‘a pattern of outsourced governance rooted in colonial practices – “concessionary politics” – has become engrained’.¹¹⁰ Whether a leader comes to power through a democratic vote or via a coup, the pattern of behaviour is a familiar one. Drawing a similar conclusion to the aforementioned Commission of Inquiry, in 2015 the Advisory Group went further as they linked neglect to the 2013–15 atrocities:

The successive ruling elites and their entourage never demonstrated any sense of responsibility or accountability towards the populations they were meant to administer. Poor leadership and governance and the neglect of regions must therefore be seen as the principal causes of the current conflict.¹¹¹

The statement captures the sentiment expressed by the majority of academics working on CAR, as the post-colonial era embodies a business-as-usual logic. This is not unique to CAR, and many elites have maintained their position in society by stripping the state of its resources to pay those who help them hold onto power.¹¹²

To illustrate the implications of outsourcing for distrust, let us consider two examples, the judiciary and security, for these are widely considered to be central pillars of any state. Regarding the former, the final Commission of Inquiry report explains that the problem was being reported in 2009: ‘[t]he justice system is plagued by a lack of resources, severely limiting its capacity to address impunity’.¹¹³ To underline what a ‘lack of resources’ means, consider that by 2014 (during the civil war), there were only three functioning prisons in the whole of CAR.¹¹⁴ When successive leaders fail to invest in creating a system that can enforce the law, it fuels a climate of violence, as people know they can commit even the most heinous crimes with impunity.¹¹⁵ It also fosters distrust in the government, as people perceive that the leaders are so corrupt that they would rather spend foreign aid on themselves than protect citizens. This, of course, is accentuated by the fact that political elites are often the perpetrators of gross human rights violations themselves. This creates an environment in which grievances juxtaposed with impunity acts to incentivize coups and human rights violations.

Regarding security, the latest chapter in how outsourcing fuels distrust is unfolding in the ‘security’ provided by the Russian ‘semi-state force’ the Wagner Group.¹¹⁶ In 2017, Touadéra’s government accepted Russia’s offer to broker what became known as the 2019 Khartoum Agreement, which was a peace agreement between the government and 14 armed groups. Since 2021, however, the government has waged a brutal counteroffensive with the support of Wagner and Rwandan troops to prevent another coup; yet some analysts claim that Wagner are beyond the control of the government.¹¹⁷ Although Wagner have helped reduce the threat posed by armed groups and brought stability to certain parts of the country,¹¹⁸ their violent methods have drawn international condemnation. In 2021, a UN special Commission of Inquiry found that ‘systematic and grave human rights violations’ are being carried out by all parties to the conflict, including the Wagner Group.¹¹⁹ Echoing this finding, the Armed Conflict Location and Event Data Project found that civilian targeting accounts for 52 per cent of Wagner’s violence in CAR. The fact that they often operate independently in CAR (as opposed to in Mali, where they work alongside state forces) has seemingly increased their level of impunity.¹²⁰ At the local level, there are many citizens who welcome the improved security situation that has arisen due to Wagner pushing armed groups out of certain towns.¹²¹ However, there are many who tell a different story: ‘[t]he first time they came, I was very happy we all were – finally our suffering from armed groups will end because they’re here to help the government and save us … but eventually we realised [what] they were doing and we ran for our lives’.¹²² When one combines such eyewitness accounts with reports that Russian disinformation campaigns are fanning ‘distrust and instability’,¹²³ the implications of outsourcing security to the Russians does not bode well for the future, even if it ensures regime survival in the short term.

To bring the focus on political distrust to a close, the leadership style of the political elites in the post–Cold War period, combined with long-term practices such as outsourcing, created an environment in which many citizens understandably distrusted the government. Both Patassé and Bozizé proclaimed that they could be trusted, yet both oversaw grave human rights violations and reneged on commitments to the point that they were viewed as untrustworthy. Within CAR, this creates an environment in which coups become more likely, as the actors involved see little incentive in continuing to engage with elites that do not uphold agreements. Furthermore, the actors involved may know that they have a reasonable chance of success, which was the case in 2013, even though the Séléka arose in the periphery rather than the capital. If we accept that ‘trust is the expectation of no harm in contexts where betrayal is always a possibility’, the expectation in CAR at the time was that the government would both betray and harm. Here, we see the multifaceted role that distrust plays, as on a day-to-day basis it can act as a form of protection but can also catalyse rebellions against elites that are viewed as untrustworthy. The role of distrust is also pivotal in explaining why the coup was successful, as regional players withdrew their support for Bozizé as he failed to uphold commitments such as the Libreville Agreement. At this point, however, it is important to recall that a successful coup is one thing and mass atrocities are another; to appreciate why a discourse of threat took hold in the aftermath of the coup, it is important to understand the long-term social trust dynamics at play.

Social distrust

When it comes to the relationship between social distrust and mass violence, it is important to bear in mind that mass atrocities are not extraordinary societal episodes that represent a rupture in reality but are, instead, part of longer-term processes. Lombard and Batianga-Kinzi argue that the mass violence in CAR is rooted in everyday violent practices such as ‘popular punishment’, in which people are killed for being witches.¹²⁴ To link this back to the concept of trust, the point here is twofold. First, these everyday violent practices stem, in part, from distrust and therefore need to be factored into our understanding of what drives mass atrocities. Second, these practices in and of themselves also fuel distrust, which acts as a source of protection and, as a result, cannot necessarily be fixed as suggested in mainstream accounts. This section explains the perceived threats that exist in both the visible and invisible world through an exploration of identity politics and witchcraft.

Identity politics

Within CAR there is no consensus over ‘who is Central African’. Individuals may be legally Central African, but they are still viewed as outsiders, for example, if they come from certain groups and are therefore viewed as a threat. The failure to forge a consensus over who is Central African prevents the bonds of trust being established.

To provide some overarching context, let us turn to Geschiere’s study, The Perils of Belonging: Autochthony, Citizenship, and Exclusion in Africa and Europe.¹²⁵ The study analyses the role of autochthony, ‘to be born from the soil’, in the post–Cold War era. Whereas many hoped that globalization would bring about a new liberal era of universalism, he argues ‘the return of the local’ is evident in very different contexts around the world, with questions such as ‘who belongs? and how one can prove belonging?’ coming to the fore of international relations.¹²⁶ The concept of autochthony is evidenced to have ancestral roots which underpin constructions of identity, as individuals and groups appeal to X and Y, which may be real or not, when creating a sense of belonging. Autochthony has been ‘especially associated with Francophone Africa’¹²⁷ (which of course has relevance for CAR) and is extremely powerful, as it has a perceived authenticity which derives from people feeling that they are naturally born from the soil.¹²⁸ The implication of this is that individuals and groups may be viewed as not belonging and, in turn, as legitimate targets of oppression and violence, with political leaders such as Gbago of the Ivory Coast using similar thinking to inspire violence.¹²⁹

Applying the concept of autochthony to the case of CAR, Collins and Vlavonou highlight that relations between Muslim minorities and non-Muslims are not determined by religion but by the question of ‘who is a Central African?’.¹³⁰ Through a combination of archival research and interviews, the study reveals how far reaching hostility towards Muslim groups has been. Analysing newspapers between 2003 and 2018, the authors found that negative stories were linked to religion but only if the person in question happened to be Muslim. Collins and Vlavonou argue that the media routinely portray Muslims as ‘social ills’ and ‘essentialised’ to the point that ‘a typical Muslim, is always untrustworthy’.¹³¹ This forms part of a broader government-led discourse of discrimination, which allows the government to regulate ‘Muslim access to the state and associated benefits, notably regarding blunt considerations of “proper” citizenship’.¹³² Muslims face a double-edged sword, in that the discourse frames them as untrustworthy foreigners whilst government practices prevent them from securing official documents, the outcome of which is that Muslims are targeted because they are seen as foreigners. The authors detail how specific events, such as ‘two adolescents … found dead in a Chadian Muslim’s car’ in 2011, have had societal repercussions; the family of the person who owned the car was ‘attacked and accused of witchcraft’ and ‘several mosques were destroyed, and Muslims killed’.¹³³ Simply speaking, it seems the common discourse presents Muslims as illegitimate citizens and, as a result, legitimate targets of attack. As Kilembe explains, ‘[i]n CAR, Muslims are automatically considered foreigners’ and having an ‘Arab name can lead to harassment, swindling, exclusion, and degrading treatment’.¹³⁴

To better understand why this is the case, it is important to consider how French colonialism has shaped identity politics in CAR, which is evident in the language of Sango. Although few people spoke Sango when the French arrived in CAR, France promoted it to serve their own interests of aiding communication to facilitate their influence and control.¹³⁵ The number of people speaking Sango grew significantly as it was easy to learn, but it still had colonial linkages, and it was not until the post-colonial era that it become more popular. As Lombard describes, legally one can be Central African, but if a person cannot speak Sango, they are viewed with suspicion and their documents perceived as fake.¹³⁶ Here we see the role of distrust at the social and cultural level working on a day-to-day basis. Sango may have been constructed, but it appears to be a critical fault line that divides ‘insiders’ and ‘outsiders’ as it shapes perceptions of ‘foreignness’. For example, in interviews at local sites in 2019, Glawion, Van Der Lijn, and de Zwaan found that locals were eager to see displaced people return to the area but do not extend this welcome to ‘foreigners’, whom locals, including Muslims, defined as including non-Sango speakers.¹³⁷ Sango is so embedded, therefore, that those who cannot speak it are viewed as untrustworthy. Placing this within the broader context of the state, Lombard argues that ‘Sango seems more enduring a firmament of nationhood and belonging – of safety amid danger’.¹³⁸ If anything, it would seem that Sango speakers share a common bond of distrust against those who do not speak Sango.

To return to Trust Studies and debates over the relationship between trust and distrust,¹³⁹ in CAR it is evident that trust and distrust exist at the same time. For instance, Bierschenk and Sardan conducted anthropological research at five sites in CAR and found that these sites were ‘far from harmonious’ but that ‘[o]pen and public confrontation between warring parties is usually avoided because the individuals involved know that they may need the cooperation of the other individual in the future’.¹⁴⁰ It would seem that the individuals involved make rational calculations that they will not confront one another because they may need each other’s help at some point in the future. They engage in cooperation, therefore, and in so doing trust that person B will fulfil Y (and vice versa), but as authors go on to explain, this does not mean the conflicts are resolved. The conflicts are ‘supressed’ and ‘shift in disguised form to other areas of their complex relationships’ which then operate ‘against a latent background of conflict’.¹⁴¹ The picture painted therefore, is one of day-to-day activities that operate under a permanent state of suspicion. In turn, it raises questions over whether high levels of trust exist in the manner implied in attempts to rebuild trust.

It is evident that prior to the Séléka coup, the question of who belongs, and the implications of this for society as a whole, provided fertile ground for the violence that was to follow. The fact that the Séléka included a lot of ‘outsiders’ (including mercenaries from Chad and Sudan) only heightened distrust. As Carayannis and Lombard explain, ‘the failure of some among them [Séléka] to speak Sango fed the perception among people living in these places that their lives, lands, and livelihoods were being despoiled by rapacious foreigners’.¹⁴² From a trust perspective, the motives and integrity of the individuals involved were questioned from the outset. Not just because of what they were doing (staging a coup) but also because of perceptions regarding who they were (non-Central Africans), even though the group did include Central Africans, some with legitimate grievances. This brings us full circle, as the concept of autochthony was then evident in the anti-Balaka movement, which led to allegations of genocide and ethnic cleansing. As Collins and Vlavonou explain, ‘former President Bozizé’s co-ethnic Gbaya, who were heavily implicated in the anti-Balaka, propagated a discourse of autochthony, making claims to be “true” Central Africans’.¹⁴³ This invoked a historical narrative in which non-Central Africans had to be destroyed because they could not be trusted.

Distrusting the invisible

To understand CAR, one must factor in the role of invisible forces, recognized as ‘witchcraft’ and ‘sorcery’ in English. These are not unique to CAR and have been studied by anthropologists throughout Africa.¹⁴⁴ The studies address eurocentrism, economics, power relations, symbolism, morality, and, of relevance here, trust. Regarding the latter, Geschiere’s study Witchcraft, Intimacy and Trust documents his changing position towards witchcraft, which centres on intimacy and trust. Geschiere explains that in 1971, when conducting fieldwork among the Maka, his first encounter with these practices filled him with ‘excitement’ but over 40 years later, the ‘ugly effects’ of such practices led him to change his mind as these invisible forces are a source of ‘great unrest and fear’, which fuel violence and tear families and communities apart.¹⁴⁵ Whilst he questions whether a solution can be found, he believes that academic study can help and holds on to the fact that ‘people do manage to establish trust, even though one’s intimate relations are heavily charged with danger’.¹⁴⁶ Such thinking reinforces the view that trust and distrust can operate at the same time, whilst beginning to illustrate the magnitude of the task at hand, as there is no easy solution.

From a trust perspective, the belief that invisible forces exist and that they have their own agency is problematic, because how can you trust the motives and intentions of other actors if you do not think they determine their own behaviour? To better understand this, it is first important to consider that ‘the majority of Central Africans are part of the ethnic-linguistic cluster known as the “Ubanguian” peoples’ for whom ‘witchcraft is a substance housed in the human body. This substance is like a small animal.’¹⁴⁷ From this perspective, all human beings have something within them that they are not able to fully control, there is ‘hidden agency’.¹⁴⁸ This can lead to both positive and negative interpretations. Regarding the former, some believe that great leaders must have ‘cultivated and expanded their witchcraft substance’.¹⁴⁹ In other words, the individuals in question managed to gain control of this aspect of themselves, something which remains unachievable for everyday Central Africans. Regarding the latter, there is a common belief that the substance will grow, and, because the individual involved cannot control it, it poses a threat to both them and society at large.¹⁵⁰

The cultural view of invisible forces poses a fundamental challenge to liberal attempts to rebuild trust, as it is evident that this is not a crisis in trust that can be fixed. Furthermore, any actor, whether local, the state, or external, faces a pressing dilemma. To illustrate this, let us take a real-world example. In 1991, Geschiere explained that in the village Ntdoua (Cameroon), an elderly man was ‘dragged’ to the authorities for allegedly killing ‘several villagers with his witchcraft’.¹⁵¹ The prosecutor investigated the allegations but dismissed them due to a lack of evidence and released the old man, who was then killed by the villagers a few months later. This led to 17 young men being arrested. In a discussion with Geschiere, the new prosecutor explained the ‘dilemma’ he faced, as on one hand, he had an obligation to investigate and charge the young men if found guilty, because they cannot take the law into their own hands, but, on the other hand, ‘condemning them would also be dangerous since this would confirm the general idea that the state was inclined to protect witches’.¹⁵² The example raises the question which, as of yet, has not been answered: how can the problem of witchcraft and sorcery be resolved? In large part, the failure to resolve this lies in the fact that there is no shared understanding of the problem. Within certain cultures, the problem is the existence of witches and sorcery; within other cultures, the problem is the belief in, and reaction to, the existence of witches and sorcery. Accordingly, it once again underlines the magnitude of the task at hand and requires actors to gain a more informed understanding of the complexities that surround trust at the local level.

To bring the analysis of social distrust to a close, identity politics and underlying narratives of distrust surrounding witchcraft provided fertile ground for hatred, fear, and violence to flourish in the aftermath of the 2013 coup. Everyday discourses in CAR embody narratives that reject the idea that certain groups such as Muslims and non-Sango speakers are legitimate citizens. At the same time, the belief that invisible forces shape people’s behaviour, although not unique to CAR or Africa, sees suspicion and distrust fuel violence daily. At this point, it is worth noting that ‘every single day someone is killed in CAR for having perpetrated witchcraft’.¹⁵³ The culmination of these factors was that CAR was susceptible to mass atrocities taking place. Whilst liberal attempts to rebuild trust may be well intentioned, a more informed understanding of the trust dynamics involved means facing up to the fact that some trust-related issues – for example, around practices such as witchcraft – will not be solved anytime soon.

Conclusion

Collective knowledge of mass atrocities and mass atrocity prevention has advanced significantly in the 21st century; however, it is difficult to identify a factor that is cited as often as [dis]trust yet remains so under-researched. The article paves the way for a new research agenda that brings together Trust Studies and mass atrocity prevention, primarily by highlighting the conceptual complexities and debates surrounding trust and distrust while also analysing how these concepts relate to the mass violence witnessed in CAR.

The article makes two contributions. First, it provides a more informed understanding of the mass violence that took place in CAR. The study reveals that there is widespread distrust in both political relations (in the government) and social relations (between individuals and groups). These have historical roots, as pre-colonial, colonial, and post-colonial practices have fuelled both. The new era of democracy did nothing to address these, as political elites further entrenched and manipulated distrust in the post–Cold War era. High levels of political and social distrust acted as key factors in enabling the mass atrocities to take place between 2013 and 2015. The former helped catalyse the coup and its success, whilst the latter was manipulated to radicalize people into perpetrating mass atrocities. Second, it argues that the mainstream commitment to ‘rebuilding trust’ is built on misguided assumptions, as trust did not exist and break down as assumed in this approach. Indeed, it would be naive to think that distrust can be eradicated, and nor should it be because (a) cooperation can occur in the absence of trust and (b) distrust has a normative value. Whilst actors should seek to address the negative side of trust dynamics, they need to hold a more informed understanding of the complexities involved, whilst also being aware that they themselves may be perceived as untrustworthy. If it is the case that CAR is dominated by people who understandably distrust the government, external actors, and ‘others’ in society, and furthermore, this distrust acts as a source of protection, it begins to show the magnitude of the task at hand for anyone seeking to address trust-related issues in CAR.

Whilst a greater focus on recent developments is beyond the scope of this article, there are widespread concerns of mass atrocities in the near future. The Early Warning Project categorizes CAR as an ongoing case of mass killing and the 16th highest-risk country for new mass killing in 2023.¹⁵⁴ Whilst the government of CAR’s relationship with the Wagner Group has gained international attention, it is also important to note that the government’s increasing authoritarianism is of serious concern.¹⁵⁵ From a trust perspective, this may exacerbate distrust in elites, thus facilitating further anti-government offensives, which often create an environment in which atrocities are perpetrated. Also, we may see government and non-government actors manipulate distrust to radicalize people into further mass violence. Ten years on from the Séléka coup, the risk of mass atrocities in CAR remains ever present.

Although the article focused on CAR, it raises observations, concerns, and questions that feed into broader debates over the causes of mass violence and mass atrocity prevention strategies. Going forward, academics and policymakers need to develop a more informed understanding of trust dynamics within the context of mass atrocities. This is no easy task and will require many different methodological and disciplinary approaches. We need to better understand the multifaceted roles of trust and distrust, and their relationship with related aspects such as cooperation and, finally, consider not just how to, but who can, facilitate trust creation, if indeed, this is what is required on a case-by-case basis. Only then can actors begin to build more effective mass atrocity prevention strategies, which are desperately needed as mass atrocities continue to cause immeasurable harm around the world.