‘We were not emotional enough’: Cultural liberalism and social contract imaginaries in the Colombian peace process

The 2016 Peace Accord signed between the Colombian government and the FARC-EP guerrilla was narrowly rejected by the public in a polarising referendum. This article focusses on government officials in the Office of the High Commissioner for Peace, the government institution in charge of peace negotiations, and explaining the peace process to society in an innovative strategy called ‘peace pedagogy’. These officials resorted to rational communication about peace and repudiated the accord’s opponents, whom they perceived as right-wing populists. They then self-critically analysed the referendum loss as due to their strategy being ‘too rational’ and ‘not emotional enough’. Drawing on the anthropology of liberalism, this article characterises these officials as ‘culturally liberal’: liberal ideology was enmeshed in their cultural worldviews, including a perceived binary between rationality and emotions, and a contractarian imaginary of state-society relations as above politics. This both contributed to the loss of the referendum, and confounded their attempts to analyse the result. The normative model of the social contract, enmeshed in real-world interpretations of state-society relations, thus creates inexorably political effects.

acronyms. Throughout the negotiations they adopted increasingly creative strategies, including an online course, informative booklets, and workshops with civil society organisations and media outletsall with this dissemination logic.
Their emphasis on techno-rational explaining was partly shaped by the peace negotiations' confidentiality principle. All official communication about the peace process was tightly controlled to shield the talks from national politics, because anything the government said publicly could affect the negotiations. However, as public interest in the peace process grew, the OACP received requests from organisations, universities, and state institutions across Colombia to attend events and present official information about the peace process, and this became known as 'peace pedagogy'. In these spaces, audience members often called on officials to respond to the Democratic Centre's messages about peace. Officials referred to these as 'myths', which they believed had to be repudiated with 'realities'. OACP official Hugo said, 'Our narrative was about deconstructing myths with truthful information'. Thus, both the confidentiality principle, and the fact that peace was increasingly polemicised as a clash between Santos and Uribe, led officials to emphasise technical accuracy, to avoid seeming political.
I will argue here, however, that this logic of rational explaining was also cultural. Officials construed the Democratic Centre as targeting emotions with dishonest politicking, and their own views as rational truths. After four years of negotiations, Santos announced there would be a referendum, presenting this as an opportunity for Colombians to consent to the Peace Accorda contractarian imaginary predicated on the normative ideal of rational consensus in state-society relations. The Constitutional Court ruled that the government had a duty to inform, but was forbidden from campaigning (Statutory Law 1806 of 2016), effectively restricting the government's communication in the month before the vote to peace pedagogy. 2 The institutional liberalism of the court thus reinforced the existing separation between pedagogy and politics, dovetailing with the OACP officials' culturally constituted perceptions of the political landscape. Their postreferendum self-critical analysis, 'we were not emotional enough', hinged on their perception that the 'No' side had won because it connected with people's emotions. This analysis relied on a false dichotomy between rationality and emotions common among some variants of liberalism, particularly among liberal responses to so-called 'post-truth' populism (Mazzarella, 2019). These officials often said things like 'I'm not santista' or 'our work is not political'. They were short-term contractors hired for this ad hoc work, not classic technocrats, 3 although commentators often described them as such, but they were not partisan, in line with the institutionalised separation of technocracy and politics in many states. Yet this separation is also produced culturally, through the officials' conceptual conflation of 'politics' with 'emotions'. To distinguish this phenomenon from explicit adscriptions to political liberalism, I term it cultural liberalism, meaning the way that liberal ideology intertwines with cultural repertoires (including, in the Colombian case, race, class, region of origin, and education), leading to liberal values being expressed and reproduced in ways that mask their ideological nature.
In this article, based on 13 months of embedded fieldwork within the OACP, I analyse how cultural liberalism played out among government officials charged with doing peace pedagogy in Colombia. I argue that a key dimension of their cultural liberalism was what this special issue calls 'contractarian thinking' (Burnyeat and Sheild Johansson, 2022, this volume); namely, an imaginary conjuring expectations and assumptions about statesociety relations derived from the normative model of a social contract. I argue that this contractarian imaginary was evident in government officials' ideal of a state above politics, and that this both contributed to the loss of the referendum and dominated their attempts to analyse their loss. The OACP sought to enact a liberal social contract by responding to 'myths' with 'realities', but this obscured the political nature of communicating about peace as the government. This focus on the OACP officials' contractarian imaginary contributes to anthropological literature examining liberalism in situ, and its effects on political processes.

The social contract and the anthropology of liberalism
Contractarianism is central in liberal ideals of state-society relations. Classic social contract theory, developed in the Enlightenment, proposes a philosophical model in which individuals form organised society via a common agreement that regulates their coexistence, and creates and legitimates an authority whose laws they consent to because of their rational understanding that cooperation with each other and with authorities is in their best interests (Boucher and Kelly, 1994). While variations on contractarianism have come in and out of fashion, and critiques have emphasised the model's exclusionary nature and its implications in histories of inequalities, the ideal of political life as a fair agreement between free, equal, and rational individuals remains a powerful normative ideal, particularly in societies where the absence of such agreement is evident (Boucher and Kelly, 1994: 28).
Contractarianism, like other liberal ideals, has travelled widely. Western political theories underpin many forms of political organisation. As this special issue argues (Burnyeat and Sheild Johansson, 2022, this volume), loosely shared assumptions from social contract theory, varyingly interpreted and appropriated in different cultural contexts, imbue state infrastructures across time and space. The anthropological approach to the social contract outlined in our introduction proposes interrogating ethnographically the real-world effects of contractarian thinking, examining how people conceive of, appropriate and reinvent this widely circulating model of political life. This article argues that contractarian thinking among government officials in Colombia is enmeshed in the wider cultural and ideological framework of liberalism. This is important for understanding social contracts anthropologically, because ideas 'travel' not in isolation, but within webs of associated assumptions.
'Liberalism', a political practice originating in Europe and the United States in the Enlightenment, developed different emphases in different contexts, and was introduced in societies like Colombia through colonialism. It consolidated in the post-war era, when liberal democracy was imagined to triumph over two anti-liberal others, fascism and communism (Fawcett, 2014). While liberal governance is associated with tenets such as the rule of law, individual rights and freedoms, the market economy and the separation of church and state, anthropologists have also analysed liberalism as a world-making project which travels and mutates across time and space (Ansell, 2021;Dzenovska, 2018;Fedirko et al., 2021;Schiller, 2013). Despite variations across societies, or 'liberalisms' (Mouffe, 2005: 10), and debate and pluralism in liberal canons, liberalism retains a common outlook, including the centrality of the rational individual, and the idea that states can improve society using objective science and knowledge (Gray, 2003). Anthropology of liberalism examines 'actually existing' (rather than 'ideal') liberalisms, exploring how people appropriate, mobilise and reinvent liberal values in different cultural contexts, and how these values intertwine with everyday practices and institutions, shaping both formal politics and intimate relationships and subjectivities (Fedirko et al., 2021).
I use the term 'cultural liberalism' to articulate a difference from 'political liberalism', which I see as the explicit adscription to a (variant of) liberal ideology. Dzenovska (2018: x) distinguishes between the normative 'system of thought and action' of liberalism-onpaper, and the 'contested projects of remaking people and institutions in the name of political liberalism' that is liberalism-on-the-ground. Yet while President Santos was politically liberalhis family was influential in the Colombian Liberal Party (more on which later) and throughout his career he adopted variants of liberal political thought as governance strategies (including Tony Blair's Third Way)many OACP officials saw themselves as 'not political'. Nevertheless, I see them as culturally liberal because liberal ideology imbued their worldview, including their belief that rational explaining could overcome what they saw as emotive populism, and their subsequent self-critical analysis that they had been 'too rational'. Liberalism conveys itself as a kind of common sense, and cultural liberals often do not see their views as ideological, yet it is an ideology, if we understand ideologies as culturally shaped political subjectivities (Haugbolle, 2018). Variations of liberal ideology exist in relationship to a historical philosophical canon, but liberal ideas are also actualised and contextualised through the cultural everyday.
I take cultural liberalism as the cultural embedding and expression of what Mazzarella (2015) calls the 'liberal imagination'; a contemporary global framework (with local variations) which idealises rational public debate, and rejects the irrational, emotional crowd, spellbound by charismatic politicianshow the Santos government saw Uribe's 'No' campaignas primitive and uncivilised. When people vote for politicians from whose policies they stand to lose, the liberal imagination construes this as civic ignorance, often proposing education as the solution. In Brazil, Ansell (2021) underscores the liberalism of government officials who travel the country explaining that clientelism is bad and encouraging voters not to engage in transactions with politicians; in post-socialist Latvia, Dzenovska (2018) analyses how various actors sought to inculcate tolerance and other liberal values to become more 'European'. If people do not agree, they have not understood it yet, and we need to explain it betterthis is liberalism's logic. The Santos government's belief that they could 'teach' society about peace, using rational, objective truths, and educate people to vote in their own best interests, was deeply liberal, as was their subsequent lament that people had been moved to vote by their emotions, rather than by their correct understanding of the peace agreement.
I suggest that these officials' assessments were influenced by a contractarian imaginary a particular strand of the liberal imaginationof neutral, apolitical state-society relations. The interpretation by actors, including state officials, of their actions as technical, not political, is a familiar theme in anthropology (Candea, 2011). Curtis and Spencer (2012) critique the tendency to highlight the political nature of such beliefs for employing a normative conception of 'politics' instead of an ethnographic one. They propose to 'treat statements about the limits of the political with the same respect we would accord any other ethnographic statement' and 'recover the wider cultural logic within which it is located' (2012: 8-9). The OACP officials associated 'politics' with emotions, manipulation, having hidden interests, and seeking to win power at any cost, rather than acting for the public goodcharacteristics they ascribed to the 'No' campaigners, whom they saw as populists. Through this association of politics with populismwhich I later show was connected to the cultural ideals among Bogotá elites of Euro-centric modernity and civilisationthey seemed to value politics negatively, and accentuate the contractarian ideal of rational, non-political state-society relations. Their view echoes the traits of liberalism which Mouffe (2005) critiques for failing to acknowledge the inevitability of conflict in human coexistence, and believing in the possibility of rational universal consensus. This blindness to the confrontationality of politics, Mouffe claims, contributes to exacerbating conflicts, by not making space for them to be processed calmly: the political effects of liberal fantasies of rationality and consensus.
The culturally liberal OACP officials sought to transcend politics, seeing their role as 'just explaining' the Peace Accord. Yet this was paradoxical, first, because they were speaking on behalf of the government, which is difficult for any audience to construe as not political; and, second, because it suggested that peace was not political, yet the Peace Accord was inherently tied to social and economic relations, and deeply rooted political antagonisms.

The 2016 Peace Accord: A new social contract?
Political liberalism was introduced into Colombia by creole elites early in the life of the republic. The Colombian Liberal Party was founded in 1848, influenced by Vice-President Francisco de Paula Santander, the father of Colombian Liberalismalthough he also inspired the Conservative Party, founded in 1849 (Delpar, 1981). Early Colombian Liberalism was influenced by the French Revolution, yet differed from its European philosophical roots as party members advocated reform of colonial practices and institutions. The Liberal and Conservative parties monopolised Colombian politics until the 1991 Constitution allowed other parties to emerge. Across Latin America, conflicts between Liberals and Conservatives were common post-independence, over issues like decentralisation and the role of the church in public life, yet nowhere was this conflict so violent, prolonged and entrenched as in Colombia; and bipartisan wars in the 19th century fostered further violence in the 20th and 21st centuries (Deas, 2015). After a Liberal-Conservative war in the 1940s and 1950s, the parties signed a power-sharing agreement to end the violence. Yet the sectarian legacy fostered other conflicts, particularly over land distribution and social inequalities, and a new conflict erupted in the 1960s, known as the internal armed conflict, between the Colombian state and the FARC and other leftist guerrilla insurgencies, into which paramilitarism and drug-trafficking merged in the 1980s.
The 2016 Peace Accord was the latest in a series of attempts to end Colombia's conflict. Three earlier attempts to negotiate with the FARC failed, and there have been (semi-) successful negotiations with other guerrilla and paramilitary organisations (Pizarro, 2017). The accord had five substantive points and one procedural: (1) Comprehensive Agrarian Development Policy; (2) Political participation; (3) Ending the conflict; (4) Solving the problem of illicit drugs; (5) Victims' rights; (6) Implementation, verification and public endorsement. It went beyond disarming a guerrilla group, promising structural reforms seen as key for preventing resurgence of violence, including land redistribution, broadening democratic participation, and redressing the rights of over 8 million conflict victims to truth, justice, reparations and guarantees of non-recurrence. Many saw it as an attempt to build a new social contract (Rettberg, 2020).
The contractarian imaginary underpinning recent efforts to end the conflict is not new; it informed many of Colombia's previous peace processes. In 1991, a new constitution was drafted following negotiations with four guerrilla groups, with participation in the constitutional assembly a key incentive for their disarmament. The 1991 Constitution expanded rights for ethnic populations, broadened political participation, and made 'peace' a constitutional right; but it failed to deliver, and violence escalated. In 2001, a pact called the 'Ralito Agreement' was signed secretly between paramilitary leaders and politicians, which sought to formalise a new social order and preserve the property rights of large-scale landowners who feared losing the land and power they had accumulated. The pact claimed to 're-found the nation' and 'sign a new social contract' (Semana, 2007), and eventually led to the paramilitaries' disarmament. As Morris (2021) argues, this justification of a new social order was rooted in social contract mythology, though many self-proclaimed liberals would have distanced themselves from the pact and its implications. Thus, the idea that the re-ordering of land, property and society would lead to peace in Colombia is reinstated in different political conjunctures, inflected differently by those who use it.
The 2016 Peace Accord's central concept was 'territorial peace' (paz territorial), which acknowledged regional differences and promised bottom-up, participatory peacebuilding processes and increased decentralisation. Peace accords are produced through a process of contestation between enemies; the Santos government's view of the accord was only one side. The FARC saw 'territorial peace' as an emancipatory project to redress the structural inequalities that were the raison d´être of their armed struggle, but for the government, 'territorial peace' was about 'closing the gap between the rural and the urban worlds', as officials often said.
In one pedagogy session I witnessed in a small town in a rural area, OACP official Pilar presented the six points of the accord to an audience of some three hundred local farmers, schoolchildren, local state officials, and members of victims' collectives. She told them: The negotiations in Havana sought to end the conflict. In Havana, they didn't sign peace. It is with you, the territories, that real peace is built, from below; in the accord this is called territorial peace. The accord establishes a 'what'some plans, some reformsbut the 'how' is defined in the communities, depending on local requirements. This address to 'you, the territories' invokes the audience as the subjects of peace, and incorporates a pre-emptive response to the common criticism that 'peace is not signed, it is built' (a phrase oft-repeated among social movements complaining that the negotiations were overly top-down). She explained point one (rural reform), saying, 'the objectives of point one are to make structural transformations in the territories, generate dignified living conditions for people in the countryside, reduce poverty levels, promote equal development, and close the gap between the rural and urban worlds.' She listed two measures agreed in point one to achieve those objectives: the creation of a land bank to distribute land to campesinos, and development plans designed with participation of local communities.
This notion of 'closing the gap between the rural and the urban' reflects a common perception of two Colombias, divided geographically between central (civilised) regions of the country, under state control, and marginal (barbaric) areas, known as 'the territory' (el territorio), inhabited by guerrillas, paramilitaries, drug-traffickers, and coca-growers (Ramírez, 2015;Serje, 2011). Serje (2011) argues that these 'wild areas' were discursively produced as the 'reverse of the nation', depicted in layered reiterations throughout history, since colonisation. As the 'peripheries' are produced in the national imaginary, so is the centrethe nation is defined in contrast to its other, an evolutionist paradigm of civilisation and modernity. Both diagnoses of problems in these regions and proposed solutions are cast in terms of finishing a perennially truncated state-building project, civilising the territories and integrating them into the nationvia the economic and political interests of the centre. 'Territorial peace' is the latest instantiation of this perception. It interprets the country's conflict as due to underdevelopment of peripheral areas, and proposes 'closing the gap' as a civilising project that will pave the way to peace. This conveyed progressive elites' aspirations for the peace process to unleash structural change within liberal parameters. As OACP official Simón told me, the peace process was 'an opportunity to do many things we haven't been able to do because of the conflict and the state's weakness … the peace process is to get the FARC to demobilise, but it's also leverage for other things', like land redistribution and political reform.
The government's view of peace belonged to a liberal peace paradigm. 'Liberal peace' refers to an international peacebuilding framework stemming from liberalism, upholding democratisation, the rule of law, human rights, free and globalised markets, and neoliberal development; this paradigm is often criticised for conceiving of states as neutral, implementing policies rationally and apolitically (Richmond, 2006). It has continuities with the cultural liberalism of Colombian elites, particularly their goals of state-building and 'civilising' the countryside, the idea of peacebuilding as forging a new social contract, and the ideal of a state that is independent from politics (Ramírez, 2015). Santos strategically courted international involvement in the negotiations, and the transnational community of peace experts, with their recognisable discourses on 'comparative expertise' and 'lessons learned' from other peace processes, and their methodological toolkits and funding frameworks (Autesserre, 2014), shaped the peace process. This shared liberal paradigm relied on the relationship between Colombian governing elites and the liberal international community (Burnyeat, 2022: ch. 7). Yet there was also a distinctive cultural framework that made officials especially responsive to this international vision of peace, which the term cultural liberalism invokes.
The rationale of peace pedagogy was that the peace process involved a body of knowledge, based on objective facts, about which the OACP was the authority; and that the government should explain this knowledge to society. Theoretically, for social contracts to function, citizens must know their terms, which allows them to give consent (Rawls, 1971: 15), for example, via referendums. Government peace pedagogy was an attempt at establishing state presence across through Colombia via teaching citizens about the peace agreement and obtaining their consent; a kind of pedagogical statecraft. It is itself an example of the contractarian vision promised in the Peace Accord.

Colombian cultural liberalism: Peace as a public good
Liberalism has impacted Colombia culturally, that is, beyond capital-L Liberalism (the Liberal Party's programmes), and the creation of liberal democratic institutions, which might together be termed political liberalism. Dzenovska (2018: x) notes that the elements of actually-existing-liberalism in Latvia were both transnationally recognisable and locally specific; likewise with liberalism in Colombia, which itself is multiple and constantly evolving. Cultural elements intertwine with political elements; the division is an analytic one, as political structure and culture are mutually co-constitutive. Yet commonalities among officials in the Santos government show how liberalism has a cultural side, relevant to understanding their analysis, 'we were not emotional enough', and its rationality/emotions binary.
Take, for example, Valentina, who became director of the peace pedagogy team after the referendum, but had worked in the OACP for two years previously. Hardworking, driven by the opportunity she believed the peace process represented for Colombia to end the conflict and build a better future, Valentina was fairly representative of the officials I got to know during my fieldwork. She was in her late thirties, with the pale skin and dark hair common among upper-middle classes in the capital Bogotá and other Andean regions. 'Whiteness' in Colombia was fluidly construed after independence to encompass selectively the mestizo population, and the process of mestizaje (mixing) as a civilising process of 'whitening' the Indian population, meaning to make them more culturally Spanish, especially regarding Spanish liberal notions of morality and progress (López Rodríguez, 2019: 33-4). Colombian elites considered the Andes the best place for consolidating a 'white' nation (López Rodríguez, 2019: 14), in contrast to the parallel construction of exotic, barbaric tropics, where Indians and Afro-Colombians lived (Serje, 2011). Race and region interlinked in imaginative geographies, wherein the cooler climate of higher altitudes was imagined to foster European-like civility and rationality, and the 'tropicality' of the lowlands was seen as a hotbed of pestilence, bestiality, immorality and libidinousness (Koopman, 2021).
Valentina dressed unpretentiously, in jeans and flat boots. Both elites and grassroots communities found her approachable, generous, and passionately committed to peace. Her uncle had been kidnapped by the guerrillas, and this motivated her to study Peace Studies at the Sorbonne. She spent ten years working for the German Development Agency (GIZ) in Sri Lanka, Germany and Colombia, and was struck by Germany's 'never again' narrative about the atrocities of the Second World War, which made her hopeful Colombia could also overcome its violent past.
Working at GIZ, Valentina witnessed the historic distrust between communities and the state, which anthropologists have documented in ethnographies of rural or war-torn regions, evident in widespread narratives about state 'absence' or 'abandonment', with many communities simultaneously desiring the service-providing state and rejecting the repressive state, which has brutalised civilians in the war (Ramírez, 2015;Serje, 2011;Tate, 2015). The ideal of what the state ought to bethe 'aspirational state' (Tate, 2015) is held up as an impossible yardstick against everyday realities. This imagined state is benevolent, impartial, independent from politics (Ramírez, 2015)a contractarian imaginary. Valentina decided the only way the state would improve was if good people went to work for it. 'That's why I put on the shoes of the state', she said, evoking a personal responsibility towards the common good. 'My heart has always been in serving others, and we have to do that from lo público'a concept linking the public sector with the public good. She frequently expressed awareness of the state's problems, saying that to change it, one had to strive to 'do things differently as the state'. She explicitly contrasted 'lo público' with 'lo político', the political, which she saw as electoral, masking hidden interests.
Like Valentina, and like Santos and Jaramillo who had both studied and lived abroad, most OACP officials were from Bogotá, educated at Colombia's top private universities, some with foreign postgraduate degrees. They shared a personal commitment to peace, often working evenings and weekends without overtime. Like other officials analysed in ethnographies of state institutions tasked with aspects of the Santos government's peace policy (e.g. Dávila, 2017), they were young, metropolitan, educated, and wellintentioned, aspiring to enact a 'virtuous state' (Dávila, 2017), different from the one they perceived as responsible for the harms their institutions were intended to redress. For many, it was their first time working for the state; some, like Valentina, had come from international agencies or local non-governmental organisations, motivated by the peace process. Many commentators celebrated that they and Jaramillo were 'technocrats' who were 'not political' (e.g. Silva, 2017), creating a binary between technocracy and politics that was viewed positively, echoing Valentina's defining her 'public' work in opposition to 'politics'. I foreground here an example of 'politics' as an ethnographic category, and of 'boundary work' between the political and the non-political (Candea, 2011;Curtis and Spencer, 2012). I do not mean that Valentina and her colleagues were misguided in labelling peace pedagogy 'non-political'. Their policing of this boundary was crucial to enable them to get the messages about peace they believed were important out to their audiences within a restricted context. As Valentina told her team before a series of pedagogy sessions, 'We have to be completely technical, to avoid political issues', meaning that if they did not stick to their script, their space could be closed down, or 'politics' could harm peace. However, I also note, following Eliasoph (1998), that people change their ideas of what counts as 'political' in different contexts, and that anthropologists can build dialogues between ethnographic and analytic categories. After one particularly fraught pedagogy workshop in which sceptical audience members accused the OACP of doing 'government propaganda', the team met to debrief. Valentina said, 'this is not about the government'; Myriam said, 'this is political but not politicalelectoral', and Valentina replied, 'this is politics but not politiquería [politicking]'. This suggests they viewed peace pedagogy as 'political' in some wider sense, in contrast to the electoral 'politicking' they eschewed. Their care in (re-)drawing this political/non-political boundary was thus important to their sense of peace as a public good, and their liberal valuing of rationality.
Political elites in Latin America frequently model their expectations about how states should behave on idealised and partial visions of European liberalism (Abelin, 2012). The OACP officials were influenced by Euro-centric liberalism in multiple ways: Enlightenment-era political liberalism forged the Colombian Republic and the historical evolution of Colombian governing elites; Santos subscribed to political Liberalism; the OACP officials were shaped by their international relationships through their studies and professional experiences abroad, and by their contact with international elites who encouraged the adoption of a liberal peace paradigm. The contractarian imaginary was embedded as an implicit normative ideal in many of these influences.

Populism: Liberalism's internal other
The shared conception among OACP officials of peace as a public good existed dialectically with their perception of Uribe as a populist threat, and the barbaric past this threat evoked. Uribe was popular throughout his 2002-10 presidency, despite multiple human rights, corruption and drug-trafficking scandals. Underpinning his appeal were the Colombian ideals of white masculinity and the paisa identity, as people from the Antioquia department are known, associated with savvy entrepreneurship, fervent religiosity, an ethic of hard work, adventurousness, and earthy realism (Viveros, 2013). Regional identities in Colombia intersect with race and class in social hierarchies; bogotano elites consider themselves exemplary of order and modernity, and other regions as wild and backward.
With the increase of drug-trafficking in the 1980s, Colombian elite identities, regionally divided since colonialism, further divided along two lines: 'the more cultivated, established affluent class, and the ostentatious nuevos ricos [new rich]' (Bunce, 2019: 178-80). Bogotano upper classes, such as the Santos family, employ symbolic boundaries to differentiate themselves from 'the chauvinistic or religious conservatives, regional political classes, and families who have suspicious sources of wealth' (Bunce, 2019: 178-80). They reject those they see as parochial, overly religious, lacking in refinement and an international outlook. Their internationalist cultural liberalism exists in opposition to the far-right's ultra-conservative nationalism.
Uribe's critics jeered at his provincial celebration of regional identity: he wore peasant clothing and used rural metaphors, depicting himself as one of the people, despite being part of a land-owning elite. He projected a different kind of state-society relationship than that of the liberal contractarian imaginary, similar to what Forbess and Michelutti (2013) call 'fictive kinship'; where a 'charismatic relational bond' is produced between popular leaders and citizens, often in ritual spaces, like Uribe's weekly visits to different municipalities to engage with local communities live on television.
The Bogotá elites' dislike of regional elites was exemplified by Mauricio Rodríguez, a journalist, ex-ambassador to the UK, and brother-in-law and adviser to President Santos, who told me he thought Santos was 'too rational and technical'. He should have tried to 'sell [peace] emotionally', said Mauricio; 'I sometimes write bullet points for his speeches, and he always takes out any emotional phrases.' He contrasted this with Uribe's success in the referendum, inciting the 'negative emotions' of 'hatred and fear'. Santos' speeches were well-received internationally: 'They applaud him in Oxford, not in a Colombian town', because Colombia is 'not a rational country', it is 'emotional, visceral, impulsive'. Santos was 'a statesman', a 'man of institutions', who tried to 'bring Colombia into modernity' by seeking peace, when the country was 'still living in the eighteenth century'. He 'lacks Uribe's common touch'; whereas Uribe was 'a populist, a caudillo [autocratic leader]', who 'uses the popular diminutive, la casita [little house], mijita [mate], asks after everyone's families'; he was 'the traditional paternal figure that resonates in all weak societies. Someone who protects, orders, controls.' Culturally liberal bogotanos saw Santos as the rational statesman, and Uribe as the populist, premodern othera view which arguably says more about bogotano elites than about Uribe. Sánchez (2016) critiques the liberal global stereotyping of Latin American politicians, from Simón Bolívar to Hugo Chávez, as backward populists, which construes populism as a threat to rationality. While many bogotano elites supported Uribe during his presidency, and Santos initially was deferential to his predecessor, as Uribe's opposition to peace grew, they increasingly saw him as this stereotype populist 'other'. Recent political theory (Arditi, 2007) critiques this liberal view of populism as an aberrant pathology, and sees it rather as inherent to liberalism, an internal, haunting other, with a reciprocally influential relationship to mainstream liberalism. Yet the interpretation of populism as deviant is widespread among cultural liberals worldwide, including the OACP officials, due to the enduring imprint of the liberal dichotomy between civilised reason and primitive emotions. As Mazzarella (2019) argues, liberals see the use of emotions in politics as populist, and therefore undesirable. Confronted with Uribe's success in delegitimating the peace process in public opinion, liberal officials in the Santos government reaffirmed their belief in the false rationality/emotions binary: populism was their internal other; the foil against which they construed themselves.

Conclusion
Both the peace process itselfwith its central concept of 'territorial peace' that reiterated a historic dichotomy between modern urbanity and uncivilised rurality and proposed to build peace by 'closing the gap' between themand the idea of explaining the peace agreement to Colombian society via peace pedagogy, for citizens to endorse it rationally via the referendum, reflected liberal values regarding state-building and a contractarian imaginary of apolitical state-society relations. Peace pedagogy developed both from the OACP officials' culturally liberal tendency to value rationality, and from their response to and repudiation of Uribe's emotive politics. Uribe's campaign and the OACP's peace pedagogy were not equivalentsone included spreading lies intentionally; the other disseminated the contents of a text. However, the very idea of a gospel-like text referent as objective truth, about which the government educates society, reveals the culturally liberal fantasy of an impartial state and a rational society, the former legitimated by the latter's consent. The OACP sought to enact a liberal social contract by responding to 'myths' with 'realities', but this obscured the political nature of peace pedagogy, by virtue of the OACP speaking as the government, and the political nature of the peace agreement itself.
The idea of distinguishing peace pedagogy from politics, underpinned by officials' cultural repudiation of what they saw as populism, clashed with Santos' desire to silence Uribe's opposition to peace electorally, in the referendum. When the Peace Agreement was signed on 24 August 2016, Santos announced that its 297 pages would be disseminated widely, 'for all Colombians to know it … so that when you vote in the referendum, you have all the information.… Nobody will be able to say they did not have the chance to know the Agreement' (Notiamerica, 2016). This cast the referenduma political-electoral exerciseas rational deliberation. Yet as Spencer (2007: 177) argues, politicians who claim to be above politics use this claim politically, suggesting they are superior to their opponents. As Michelle Obama said of Donald Trump's political bullying in 2016: 'When they go low, we go high.' In the contractarian imaginary, referendums allow citizens to choose or legitimate policies. But this ideal of public rationality obscures the confrontational nature of elections. Following Mouffe's (2005) critique of liberalism, we might argue that Santos' discourse of the referendum as an exercise in deliberation for the public good failed to make space for conflict, polarising the country more deeply.
Using an anthropological approach to social contracts to examine the OACP's peace pedagogy foregrounds how elites' diagnoses of Colombia's problems, and their proposals for solutions, are underpinned by contractarian thinking. Officials in the Santos government evaluated the state-society relationship as a dysfunctional social contract, and saw the peace process as a chance to build a new one. This liberal contractarian imaginary sees society as bound together by common agreement, a public sphere of rational deliberation. Peace pedagogy sought to model the contractarian ideal of a state-society relationship above politics. But it was founded on a paradox. Cultural liberalism construes binaries between rationality and emotions, pedagogy and politics, but the state-society relationship is inescapably political, first, because people's perceptions of the state fluctuate with their interpretations of national politics and, second, because the statesociety relationship is itself one of actual and potential contestation, sometimes even violence; not harmonious non-conflictual consensus. State and society co-produce each other, but cannot share equal responsibility towards a public good such as peace, because their relationship is never not political.
The referendum lost for many reasons, a full analysis of which is beyond the scope of this article. 4 The 'No' only won by some 50,000 votes. In Bogotá, and in many peripheral, conflict-affected departmentsthe so-called 'territorios'the 'Yes' side won. The wealthier, more populated Andean departments, which had not experienced the conflict so directly, voted 'No'. Commentators interpreted this as 'the victims voting "Yes"' because they understood the costs of war (Semana, 2016c). Deeper analyses complicate this narrative, highlighting correlations between higher education and socioeconomic levels with voting 'Yes', and right-wing voting histories with voting 'No' (Rodríguez, 2020); others posit a racial divide between regions with more Indigenous and Black populations voting 'Yes', and an implicit racism in the 'No' campaign (Courtheyn, 2016). Many abstained because they perceived the referendum as politicking, and distrusted all politicians. 'I don't like either Santos or Uribe', an airport security official told me, days before the referendum, 'so I won't vote.' On referendum day, Hurricane Matthew hit the typically pro-peace Caribbean coast, preventing people getting to voting stations. Multiple factors thus created a 'perfect storm'.
The Santos government's culturally liberal belief that rational explanations could overcome what they saw as populist disinformation may have been an additional factor in the outcome, but one could equally argue that peace pedagogy was simply insufficientit was a small team, never having more than twelve people, to cover a country of 50 million. Nevertheless, the project of explaining in a polarised context, as if their pedagogy script were objective truth, failed to recognise that political perceptions are intertwined not only with emotions but also with cultural frameworks and power relations, and that speaking on behalf of the government is always a priori political, because the government is one actor in a contested public space. And peace itself is inescapably political: it involves a settlement between warring parties and negotiations between the government and the establishment; peace accords reform political systems and depend on political will for implementation; and, crucially, public support for peace processes is contingent on citizens' perceptions of politics.
Nevertheless, post-referendum, multiple pro-peace sectors shared the OACP's belief that Colombians should 'learn the Peace Agreement' to defend it against attacks by future governments. The agreement was renegotiated and began to be implemented in December 2016, but the Democratic Centre's candidate, Iván Duque, won the 2018 presidential elections on an anti-accord platform. Nevertheless, multiple groups in civil society and politics continued to fight for the transformations they believed were possible and necessary for Colombia to achieve peace, often risking their lives in a context of increasing violence, by demanding implementation of the agreement. Seen in this light, peace pedagogy succeeded in arming people with a political tool with which to lobby the government.
As Appel (2019: 281-2) argues, on the page we can imagine and advocate for the end of liberalism and its historical violences. Off the page such imaginings meet limits. In Equatorial Guinea, while critiquing liberalism, Appel acknowledges it would be absurd not to recognise that a liberal government would be better than the current dictatorship. She argues that scholars can commit warily to pragmatic liberal reforms, while nurturing more radical, anti-liberal projects. Liberalism is not over. A peace process is an example of the kind of opportunity that requires us to engage pragmatically with liberalismeven while imagining more radical ends. After all, it is difficult for political alternatives to emerge if those who champion them keep being murdered, as they often are in Colombia. The social contract, like many liberal ideas, is simultaneously limiting and even harmful in its individualist, anti-political assumptions, and productive in its power to channel imaginaries of alternative futures, such as the promises of the Colombian peace agreement.