A necessary condition for democratic consolidation in post-authoritarian Latin America is the institutionalisation of civilian control over the armed forces. In turn, institutionalising civilian control under democracy requires reducing militaries’ power, or their autonomy and influence vis-à-vis civilian policy making.Footnote 1 Because military coups, juntas and repression are largely ghosts of Latin America's past, the crucial task of reducing the armed forces’ political power may seem straightforward.Footnote 2 The problem is that regional militaries’ power is assuming new forms. Civilian governments increasingly employ militaries to fight crime,Footnote 3 populists frequently use militaries to legitimise their rule,Footnote 4 and elites regularly lobby militaries to challenge left-wing populists.Footnote 5 These trends could produce tutelary regimes, wherein the armed forces moderate civilian policy preferences and political selections. They also could increase the risk of coups as regional democracy erodes under popularly elected but tenuous civilian leadership.Footnote 6 Building knowledge about the sources and expressions of these new forms of military power is therefore urgent for understanding the state of democracy in Latin America. One such form, to be examined here, is contestation of civilian leadership and policing missions via social media – an emerging, understudied trend that reflects and reinforces militaries’ political leverage by providing them with unfiltered access to public audiences.
These dynamics and trends are especially acute in Brazil, which boasts Latin America's largest democracy and military. The military ruled Brazil repressively from 1964 to 1985 and has retained broad political power since democratisation.Footnote 7 Nonetheless, scholars have generally considered Brazil successful in institutionalising civilian control.Footnote 8 Recently, however, amidst the multi-pronged crises of ‘Brazil's Illiberal Backlash’,Footnote 9 the Brazilian military has been seemingly executing a highly visible and ‘shocking … return to politics’Footnote 10 that challenges this scholarly consensus.
The Brazilian military's increasing power has been most evident in the areas of internal security and national leadership selection. In February 2018, the Army assumed unprecedented command and control over all law enforcement agencies in the state of Rio de Janeiro in order to combat organised crime. In April 2018, the Army commander seemingly threatened Brazil's Supremo Tribunal Federal (Supreme Federal Court, STF) via Twitter, warning it against releasing ex-president Luiz Inácio Lula da Silva from prison and thus against permitting him to seek a third term as president. In January 2019, right-wing populist Jair Bolsonaro, a former Army captain, became president with as many former or current military officials in cabinet positions as there had been during the height of the military dictatorship.Footnote 11 As the armed forces’ governing role increased, some generals attempted to constrain Bolsonaro's more extreme impulses,Footnote 12 engendering persistent civil–military contention.Footnote 13
To help us understand these developments, in this paper we examine a key source of the Brazilian military's power: ‘military contestation’,Footnote 14 operationalised as challenges to civilian control through ‘military discourse’Footnote 15 or public communication by military leaders. Specifically, we seek to explain why military contestation occurs in Brazil around some issues but not others. Most research emphasises that military contestation occurs because there are few external constraints upon military power. In contrast, we argue that discursive contestation in Brazil also stems from an internal factor: the potential for conflict over military identity and, particularly, institutional objectives to undermine the armed forces’ cohesion. This conflict occurs between two factions in the military that we refer to as interventionist and legal-institutionalist.
The interventionist faction prioritises accruing latent power through ‘military prerogatives’Footnote 16 and activating that power through intervention in civilian politics to fulfil the armed forces’ objectives of ensuring Brazil's economic development, national security and political stability during crises. In contrast, the legal-institutionalist faction prioritises maintaining latent power but perceives that intervening in civilian politics risks undermining the military's long-term interests, especially its internal cohesion.Footnote 17 Within a given issue area, some issues fuel identity conflict by introducing a trade-off between activating military power and preserving military cohesion. Conflict over which objective to pursue decentralises authority and threatens institutional cohesion, motivating military leaders operating under democratic governments to use discursive contestation to convey their prerogative-related preferences to civilian politicians and demonstrate their concern for cohesion to the rest of the armed forces.
We build this argument as follows. We first review scholarship on Brazilian military prerogatives, contestation and power. We then present our argument that intra-organisational identity conflict between legal-institutionalists and interventionists can lead to discursive contestation. Next, we conduct a plausibility probe of this argument. We specifically utilise Twitter postings from Army Commander Gen. Eduardo Villas Bôas in order to analyse contestation in Brazilian military discourse, supplemented as part of a generalisability check with key informant interviews and with articles from the ‘OMNIDEF’ news bulletin published by Brazil's Escola Superior de Guerra (National War College, ESG). We focus this analysis on military discourse between 2016 and 2018 regarding the selection of national political leaders and the utilisation of soldiers for internal security missions. Following a discussion of the implications and opportunities for future studies that stem from our argument and analysis, we conclude with a brief commentary on the lack of military contestation under Bolsonaro and posit that expanding the empirical scope of our research to the consolidation period of Brazil and other Latin American democracies would illuminate further the political roles of military self-perception.
Prerogatives, Contestation and Military Power in Brazil
The Brazilian military's power to shape national policies varies across two dimensions: latent power, embodied in military prerogatives; and active power, embodied in policy influence.Footnote 18 Alfred Stepan conceptualises the military's prerogatives as ‘those areas where … the military as an institution assumes … an acquired right or privilege, formal or informal, to exercise effective control over its internal governance, to play a role within extramilitary areas within the state apparatus, or even to structure relationships between the state and political or civil society’.Footnote 19 Prerogatives cover issue areas ranging from participation in government and independence from political oversight to roles in the nation's internal and external security.
Prerogatives are a type of latent structural power that the military retains regardless of whether it is in open conflict with civilian authorities.Footnote 20 Prerogatives are latent because, although the power exists, the military may refrain from exercising it. Prerogatives become the levers to influence policy when their presence leads civilians to appeal for military intervention in the political system, when the military uses this power to regulate political conflict or remove contentious issues from the political agenda, and when, through articulated military contestation, the military disregards or challenges the civilian government's authority in an effort to alter policy decisions.Footnote 21 Broader prerogatives signify less civilian control over the armed forces and, thus, less consolidated democracy because more powerful militaries can come to dominate and even overthrow civilian authorities.Footnote 22
David Pion-Berlin and Rafael Martínez complement Stepan's conceptualisation of military prerogatives by suggesting that military power under democratisation in Latin America encompasses both autonomy from, and influence over, civilian political elites. Military power includes authority over civilian leadership and policies, control over public defence enterprises and human rights legislation, and leadership over internal and external security. As the armed forces’ power increases within each issue area, civilian control over the military decreases, civil–military relations become less democratic, and democratic consolidation declines.Footnote 23 Synthesising these definitions, we define prerogatives as latent power which the military must activate to influence policy and which, to institutionalise control, civilian politicians must prevent the military from activating.
Stepan suggests that a key mechanism for transforming prerogatives into policy influence is ‘articulated military contestation’, or the challenge to civilian authority that is ‘clearly and persistently conveyed so as to be intelligible to relevant military and political actors, whether publicly stated or not’.Footnote 24 Contestation includes but is not limited to military discourse, or ‘the content of national security doctrine’Footnote 25 within institutional texts and statements. The military uses contestation to defend prerogatives in a way that constrains implementation of civilian political initiatives. In turn, the degree of contestation depends on how extensively civilian and military authorities clash over prerogatives.Footnote 26 Three sources of civil–military conflict over prerogatives engender military contestation: civilian efforts to hold the armed forces accountable for human rights violations; challenges to military autonomy by altering the armed forces’ missions, structures and hierarchies; and attempts to reduce the armed forces’ fiscal authority. When civilian authorities engage in such efforts, the military responds with extensive articulated contestation that weakens civilian control and threatens democracy.Footnote 27 In sum, civilian politicians’ threats to the Brazilian military's prerogatives can spur contestation that transforms the latent power embodied in prerogatives into active power over national politics, at the expense of civilian control and democratic consolidation.
While Stepan suggests that civilian threats to military prerogatives engender military contestation, other scholars argue that the armed forces’ identities spur contestation and, thus, activate power. João Martins Filho and Daniel Zirker posit that the post-authoritarian, post-Cold War identity crisis over the Brazilian armed forces’ changing political role led military leaders to use defensive, nationalist discourse to challenge politicians’ measures aimed at asserting civilian control under President Itamar Franco (1992–4).Footnote 28 Discursive contestation then dissipated under President Cardoso (1995–2003) as politicians grew increasingly apathetic about civilian control and as military leaders’ internal attempts to orient the armed forces’ role and identity away from domestic security intensified.Footnote 29
Despite conceptualising military contestation differently, scholars of civil–military relations throughout Latin America agree that both identities and prerogatives play an important role in determining when the armed forces are likely to engage in contestation. Ruth Stanley argues that threats to prerogatives are insufficient because articulated contestation is ‘fundamentally about non-material, ideological resources [like] values, the self-image of the military, the interpretation of the past, and the implications this has for the present and future order of the democratic polity’.Footnote 30 Maiah Jaskoski similarly asserts that, in Ecuador and Peru, the armed forces shirk counternarcotics missions not only because of threats to prerogatives but also their belief that the military should avoid unpredictable civilian-assigned missions that do not protect soldiers from prosecution for human rights violations.Footnote 31 David Kuehn and Harold Trinkunas find that contestation via officers’ resignations or threatened resignations, militaries’ resistance via non-cooperation with civilian authorities, and military coups stem from a combination of two factors: left-wing populists’ threats to prerogatives and elites’ appeals for the military to intervene because of populists’ perceived illicit enrichment and/or disregard for institutions.Footnote 32 We presume that the latter finding concerns identities insofar as, before they engender contestation, elites’ appeals to the military are filtered through the military's perception of its role as the country's protector.
Taken together, civil–military relations scholarship in Latin America therefore suggests that it is not challenges to prerogatives or identities in isolation that shape contestation, but rather a combination of the two. Understanding variation in Brazilian military contestation consequently requires examining how identity and prerogatives relate to the contestation that shapes the Brazilian military's power. We turn now to theorising this relationship.
Military Contestation, Discourse and Identity in Brazil
Why does military contestation occur in Brazil around some issues but not others? In this section, we develop the following argument: the more likely it is that the policy issue at hand risks engendering intra-military conflict, the more likely it is that the military will challenge civilians discursively. We first specify our conceptualisation and operationalisation of military contestation and discourse, and relate these concepts to identity scholarship. We then theorise the identity-based rationale for military contestation, revisit scholarship on identity-based divisions within the Brazilian armed forces under the 1964–85 military dictatorship, and propose an update to this scholarship for the democratic period. We conclude by underscoring internal and external factors that engender conflict between these identity-based divisions, and predict variation in military contestation based on these factors.
Military contestation encompasses mechanisms ranging from refusing to execute civilian leaders’ orders to public or private challenges to civilian authorities’ decisions or authority. We focus on articulated contestation in terms of publicly observable military statements and texts that challenge the government's authority or demand policy changes, by studying the role of discourse vis-à-vis identities and institutions.
Vivien Schmidt, in her theory of ‘Discursive Institutionalism’, argues that discourse entails both what policy actors say and where they say it.Footnote 33 Discourse thus serves different purposes depending on the context in which it occurs. A key contextual factor is institutional authority, or the level of autonomy and influence that elected officials and bureaucratic agents enjoy within the policy-making and implementation processes. Institutional authority varies between being concentrated in a few such policy actors and diffused across multiple policy actors. When institutional authority is concentrated, authoritative policy actors engage in more communication with the public to gain support for their position while minimising communication with other, less authoritative policy actors. They utilise this strategy because the public sanctions actors who fail to communicate about the issue and communication with the public can help increase support for their position. Conversely, when institutional authority is deconcentrated, a given policy actor emphasises coordinative communication with other policy actors while minimising discourse with the public because communicating with the public can undermine private negotiations.Footnote 34
The military can use public discourse to gain external support for its position, but it also uses discourse for internal purposes. Rawi Abdelal et al. claim that ‘much of identity discourse is the working out of the meaning of a particular collective identity through the contestation of its members’ because ‘[i]ndividuals are continuously proposing and shaping the meanings of the groups to which they belong’.Footnote 35 In their framework, contestation encompasses intra-group (e.g. intra-military) disagreement over the meaning of a collective identity. The meaning of a collective identity, in turn, stems from the shared rules, goals and worldviews that group members must follow, as well as the distinctions that members draw between their group and other identity groups. This framework implies that discourse results from not only external factors, but also from internal conflict over groups’ identities. A corollary framework suggests that political actors’ discourse contains ‘[n]aturally occurring messages that one might assume to constitute identity messages’.Footnote 36
Applying this framework to the Brazilian armed forces, we claim that military leaders use discursive contestation for two reasons. First, contestation allows military leaders to communicate their preferences for more prerogatives to civilian politicians while increasing public support for their position, thus pressuring politicians to acquiesce to military demands.Footnote 37 Second, discursive contestation allows military leaders to demonstrate concern with cohesion to the broader armed forces. Yet, although the armed forces disagree with civilian authorities on many issues, frequently they avoid discursive contestation. We argue that a key factor explaining discursive contestation is whether the policy issue at hand risks engendering intra-military conflict.
Despite being a hierarchical institution, the Brazilian military has internal divisions over how to utilise power. Stepan argues that, during Brazil's military dictatorship, these divisions split the armed forces into two factions: the military-as-government and the military-as-institution.Footnote 38 The military-as-government encompassed those who exercised or sought to exercise short- and medium-term power to ensure national development, security and stability in the face of Cold War crises and threats. The military-as-institution included others who operated and maintained the armed forces on a day-to-day basis and prioritised the armed forces’ long-term interests in terms of seeking rents through military budgets and responsibilities over strategic sectors of the national economy, in terms of sustaining societal prestige and, most importantly, in terms of preserving internal cohesion. In other words, the military-as-government focused outwardly on how to use the armed forces to determine and pursue Brazil's national policy and political objectives while, prioritising the status quo to a greater degree, the military-as-institution concentrated inwardly more on sustaining the armed forces as an organisation over risking reputational damage through public engagement.
For Stepan, a military official's position in government was the primary determinant of his/her faction during the military regime. Those within the military-as-government were typically senior officials by virtue of their proximity to power while those in the military-as-institution were largely non-senior officials. Yet seniority alone did not determine a given official's faction. Rather, preferences over power are what mattered most. The two factions agreed on the benefits of gaining latent power through increased prerogatives but generally disagreed about activating this power. The military-as-government preferred to activate power through political intervention to confront crises and threats whereas the military-as-institution preferred to avoid activating power if intervention appeared likely to undermine the armed forces’ long-term interests, especially its cohesion.
Senior military officials certainly played an increasingly prominent role within the executive cabinet and bureaucracy under presidents Michel Temer (2016–18) and Jair Bolsonaro (2019–22), but governments continued to be elected democratically.Footnote 39 As the military's ability to control the government's everyday working is more limited under democracy, describing the armed forces’ internal divisions based on officials’ positions in government would require stretching Stepan's concepts of military-as-government and military-as-institution beyond the bounds of authoritarian rule. Given that the main difference between the two factions is their goals rather than their position in government, we use a broader term to classify the divisions within the armed forces that applies across a range of regime types: legal-institutionalist, for military officials who prioritise protecting the armed forces; and interventionist, for those who prioritise governing the country. Consequently, we assume the following. Legal-institutionalists and interventionists are identity groups that conflict over objectives for the military. Senior officials’ authority within the Brazilian armed forces becomes deconcentrated when legal-institutionalists and interventionists conflict; otherwise, it remains concentrated. Militaries being hierarchical institutions, senior officials prefer concentrated to more diffuse authority.
Cohesion and hierarchy erode under the threat of intra-military conflict, making senior leaders consider private discourse within the armed forces insufficient for compelling other members to close ranks and obey the chain of command. When the armed forces face internal divisions, messages designed to prompt obedience may lose their intended meanings while trickling down the ranks. Military leaders therefore consider discourse beyond the armed forces necessary to reinforce cohesion and hierarchy. Concurrently, military leaders consider public discourse necessary to convey their displeasure to politicians regarding attempts at curtailing military prerogatives and, thus, to negotiate these prerogatives’ preservation or reinstatement.
Policy issues that engender such conflict are those wherein the interventionists’ activation of the latent power embodied within prerogatives to address crises and threats risks undermining the armed forces’ cohesion, raising legal-institutionalists’ concerns about the armed forces’ long-term interests. Facing such policy issues, military leaders must balance activating the armed forces’ latent power vis-à-vis politicians to address crises and threats, on the one hand, and fostering the armed forces’ cohesion to preserve long-term institutional interests, on the other. Senior military officials’ discursive contestation of civilian control is a means of achieving this balance insofar as it both communicates to other military officials the need for cohesion and pressures politicians to acquiesce to military leaders’ preferences regarding the extent of prerogatives to address crises and threats.
Although discursive contestation might not resolve crises and threats, it positions the armed forces as protectors of the nation amidst such conditions while preserving the armed forces’ cohesion and, thus, their long-term interests. When policy issues do not engender conflict between legal-institutionalists and interventionists, senior officials may contest civilian control around these issues, but such contestation is not public because there is limited need to foster internal cohesion discursively. Consequently, although the policy issues of political leadership selection and internal security missions subsequently analysed in this paper might seem to be similar, discursive contestation will occur in some instances but not others.
Plausibility Probe through Discourse Analysis (2016–18)
The armed forces are ‘semi-closed institutions’Footnote 40 into which researchers have limited insight and in which militaries seek to project cohesion through public discourse,Footnote 41 while much discourse and contestation occurs behind closed doors.Footnote 42 To probe our theory's plausibility, we therefore must deduce ‘observable implications’,Footnote 43 or phenomena that we would expect to see in the real world if our arguments were plausible.Footnote 44 The main implication is that, where discursive contestation occurs over an issue, said issue should have potential to engender conflict between military factions. We should therefore observe discursive contestation (conversely, no contestation) when pursuit of military prerogatives appears to risk (conversely, does not appear to risk) undermining military cohesion.
To determine whether these observable implications hold, we used discourse analysis or ‘the qualitative contextualization of statements and practices in order to describe social meanings’Footnote 45 based on the ‘assumed identity messages’Footnote 46 contained within statements, as well as the political processes surrounding these modalities. Central to understanding contestation was classifying which military faction made the statement when contestation occurred. We considered a statement to come from the legal-institutionalist identity group when it advocated for new policies designed to reduce civilian power over military policies and members or opposed policies that would increase civilian involvement in internal military affairs. In contrast, statements by the interventionist identity group would emphasise using soldiers to address the country's emerging social and political issues. Although the military can exercise articulated contestation within various power areas, we examined contestation related to internal security and selection of political leaders – the two issue areas in which contestation was most acute during the period of interest. Prioritising these issues allowed us to hold constant, as far as possible, the political stakes over which the military challenged civilian control.Footnote 47
We first examined statements from the Brazilian armed forces to identify discourse regarding internal security and selection of political leaders and determine – based on their content – whether these statements contained military contestation. We considered a statement indicative of military contestation if it included a message that demanded a government policy change or criticised civilian policies. If a statement discussed internal security or leadership selection but explicitly rejected military intervention over the issue, or simply described the issue without challenging government policy toward it, we did not consider it military contestation. We then consulted secondary sources to identify instances where discursive contestation did not occur around internal security and leadership selection despite apparent civilian challenges to military prerogatives. In addition, we examined contemporaneous media accounts of political dynamics around the Brazilian military's statements in order to assess whether identity conflict between interventionists and legal-institutionalists regarding the military's pursuit of prerogatives was likely and whether such likelihood was correlated with contestation in one instance but not another within a given issue area. The more likely we found it that internal identity conflict was associated with military contestation, the more confidence we would have in our theory's plausibility. Below, we further detail our data before presenting the analysis.
Our primary statements for discourse analysis were media interviews, public speeches, and 817 Twitter posts (‘tweets’) produced by the Brazilian Army Commander, Gen. Eduardo Villas Bôas (@Gen_VillasBoas) between 2016 and 2018. Although his staff wrote the Twitter posts, he ‘always [defined] the posts’ themes and spirit’.Footnote 48 We examined this period because, whereas scholarship on Brazil's democratic transition generally holds the optimistic view that military power decreased after authoritarian rule,Footnote 49 the years 2016 to 2018 encompassed an unexpected increase in military power with respect to political leadership selection and internal security missions.Footnote 50 The military and, specifically, Villas Bôas assumed a central role in national politics during this period partially because, amidst Brazil's crisis (stemming from economic recession, intensifying polarisation and distrust in political parties, the ‘Lava Jato’ corruption scandal and spiralling public insecurity),Footnote 51 the armed forces were Brazilians’ most trusted institution. A June 2017 survey showed 40 per cent of Brazilians had high trust in the armed forces, 18 percentage points higher than the second most trusted institution (the press).Footnote 52
We analysed Villas Bôas’ discourse because it was indicative of civil–military and intra-military conflict.Footnote 53 Amidst mounting civil–military tensions in April 2018, an Army spokesperson confirmed that Villas Bôas as ‘the Army Commander [was] the authority responsible for expressing the institutional position of the Force’.Footnote 54 After retiring from the Army and becoming an advisor to the Gabinete de Segurança Institucional (Institutional Security Cabinet, ISG, Brazil's national security council), Villas Bôas said, during a Federal Senate ceremony in his honour in August 2019, ‘everything that was done in my period of command was because I had certainty and confidence regarding … the attitude of the entire Army, flanked by the Navy and the Air Force. I had absolute confidence in the cohesion of the Army and of its purposes, values, and attitudes.’Footnote 55 These statements illustrate how, given his institutional position and personal perspectives, Villas Bôas symbolically represented and discursively sought to represent the armed forces as cohesive, suggesting that his discourse was probably sensitive to intra-military conflict and, thus, is a crucial source for probing our theory's plausibility. That the Army is Brazil's largest armed service, having dominated national politics historicallyFootnote 56 and increased its power in recent years,Footnote 57 further illustrates the importance of Villas Bôas’ discourse. Additionally highlighting his importance in the country's politics is that, between 15 and 18 June 2018, four presidential candidates (Geraldo Alckmin, Ciro Gomes, Fernando Haddad and Marina Silva) visited Army headquarters to discuss military strategy with him.
Although indicative of civil–military and intra-military conflict, Villas Bôas’ discourse is an unlikely example of our theory because his apparent sensitivity to cohesion suggests that he leant toward the legal-institutionalist faction. This identity group is less likely than the interventionist group to use discursive contestation in an attempt to convert latent power (i.e. prerogatives) into active power (i.e. policy influence) because legal-institutionalists consider activating power risky for cohesion. Therefore, we will have considerable confidence in our theory's plausibility if its implications are evident within Villas Bôas’ discourse.
Despite being a bellwether for military contestation and cohesion, Villas Bôas’ discourse is an imperfect sample and indicator of such phenomena because explaining institutional behaviour based on individual behaviour risks ‘ecological fallacy’.Footnote 58 Indeed, ‘[m]any [Brazilian] military leaders and unit commanders have social media accounts, a strict [departure] from the past's strictly hierarchical and vertical discipline where the highest military leader was the only one permitted to speak for the institution’.Footnote 59 Villas Bôas was not the only military actor whose discourse, on social media or otherwise, reflected and reinforced the Brazilian armed forces’ political power. Moreover, Brazil's ‘enormous bottleneck in advancement to the rank of general’ and officers’ ‘subsequent reduction to reserve status at an early age (around 50 years old)’Footnote 60 suggests that Villas Bôas was a demographic outlier among the armed forces as a general in his mid-60s. Building our probe around his statements and tweets alone is insufficient; it is crucial to triangulate our findings with multiple sources of military discourse.Footnote 61
We therefore complemented our analysis of Villas Bôas’ discourse in two ways. First, the lead author conducted 18 semi-structured interviews about military power with researchers and military and police officials in Rio de Janeiro during August 2019.Footnote 62 Interviews lasted between 45 and 90 minutes and occurred in locations ranging from government offices to private residences. The author analysed transcripts to identify how interviewees perceived the relationship between the armed forces’ pursuit of prerogatives, on the one hand, and cohesion and contestation within a given issue area, on the other.
Second, we examined the OMNIDEF weekly news bulletin published by the ESG. According to its strapline, OMNIDEF ‘aggregates the main news stories published each week by national and international magazines, newspapers, think tanks and academic journals’ concerning national defence and international security.Footnote 63 We reviewed 743 OMNIDEF stories published between January 2016 and December 2018 in the bulletin's ‘Public Policies of Defence’ and ‘Public Security and Intervention in Rio de Janeiro’ sections, which include Brazil-focused stories that most immediately concern the Brazilian armed forces and were most relevant to probing our argument's plausibility. This section of our article reflects sources of military contestation insofar as news stories that the ESG decides to highlight, and the excerpts that it selects for summarising these news stories, reveal how senior leaders of the ESG – considered by scholars to be constitutive of Brazilian military identity and thoughtFootnote 64 – perceive civil–military and intra-military dynamics. Regarding observable implications, we should find that OMNIDEF's ‘Public Policies of Defence’ section provides implicitly contentious content mostly regarding issues that seemingly risk institutional cohesion and issues around which Villas Bôas too employed contentious discourse. The more that we find such evidence, the more confident we will be that Villas Bôas’ discourse reflected broader military perspectives in terms of sensitivity to cohesion and that our theory about military contestation as a function of internal cohesion concerns is plausible.
Table 1 previews our discourse analysis findings. In what follows we first present the leadership selection and internal security findings based on our analysis of Villas Bôas’ discourse and author interviews and then discuss our ESG/OMNIDEF analysis, before outlining conclusions and implications.
Summary of Discourse Analysis Findings
Leadership Selection
In the leadership selection realm, there are differing conceptualisations of military involvement with democratic politics. Stepan considers active-duty military officials’ presence in the presidential cabinet a prerogative that constitutes latent military power.Footnote 65 Pion-Berlin and Martínez more broadly conceptualise this realm as democracies’ ‘autonomy to elect their own policymaking leaders and make their own policies, free from military oversight, threats, vetoes, or any form of undue pressure’. They posit that military power over leadership selection encompasses three dimensions: the extent of the armed forces’ influence over whom the people delegate as their political representatives; the number of military officials occupying cabinet positions; and the extent of the armed forces’ electoral interference.Footnote 66 Whether conceptualised narrowly or broadly, military power increases with the armed forces’ authority over political leadership selection within and beyond the cabinet. Discursive contestation that enhances military power vis-à-vis leadership selection thus represents a serious challenge to civilian control.
Since democratisation in the 1980s, Brazil has generally progressed in reducing the military's power over leadership selection. Pion-Berlin and Martínez write, ‘[t]he military exerts no autonomy over the choice of political office holders and no longer has a presence within the cabinet’. Nonetheless, the military ‘retains decision-making latitude, as it exerts effective pressures [including through discursive contestation] on policy decisions outside of external defense that have occasionally resulted in policy reversals and personnel changes’.Footnote 67 One of our interviewees suggested that this latitude manifested itself especially around budgets, where he saw ‘evidence of the military-as-institution asserting itself in government informally’ without attempting to rule.
Potential for military interference in leadership selection was evident during two events between 2016 and 2018, with differing outcomes. In March 2016, Brazil's National Congress brought impeachment charges against leftist President Dilma Rousseff for ‘fiscal backpedalling’:Footnote 68 delaying transfer of funds to state-owned banks that oversee payment of welfare programmes to avoid exceeding budgetary limits. During the impeachment trial, the military remained above the political fray, resisting both pro- and anti-impeachment pressures. Rather than remaining silent, it actively opposed involvement with the proceedings and criticised those who urged military intervention.
In a speech delivered on 19 April 2016, Gen. Villas Bôas rejected intervention, saying, ‘the armed forces do not exist to oversee government or to overthrow governments. We have to contribute to legality, making conditions for the institutions to continue working and to find ways to overcome the current situation.’Footnote 69 In a speech dated 21 April 2017 recalling the impeachment trial, he said that the military had rejected leftist politicians’ requests to declare a national defence decree, which would have enabled deploying soldiers to prevent pro-impeachment protests.Footnote 70 Even Rousseff would acknowledge that the military had avoided interfering in the process. During an interview on 31 July 2019, she said, ‘I do not think that, during the parliamentary coup [against me], the Army acted in support of the coup.’Footnote 71
Following Rousseff's impeachment and Vice President Michel Temer's ascent to the presidency in August 2016 amidst deepening national crises, there was growing public support for military intervention to restore stability, ensure development, and provide honest governance. Villas Bôas again opposed intervention, calling it ‘very sad that people see military intervention as an alternative. People do not even realise that this is absolutely anachronistic, having seen what happened in Turkey’, referring to the failed 2016 coup attempt against President Recep Tayyip Erdoğan.Footnote 72 Even more harshly, Villas Bôas said, ‘[t]hese freaks, these crazy people come looking for us here and ask: “How long will the armed forces leave the country sinking? Where is the responsibility of the armed forces?” And what do we answer? I answer with Article 142 of the Constitution. It is all there.’Footnote 73 In addition to detailing military missions, Article 142 broadly circumscribes the rules governing the armed forces’ involvement in politics.
The military's non-involvement in the Rousseff impeachment process and its subsequent opposition to intervention under right-wing President Temer, on the one hand, contrast starkly with its discursive contestation of leftist ex-President Lula's habeas corpus appeal, on the other. Luiz Inácio Lula da Silva, co-founder of the Partido dos Trabalhadores (Workers’ Party, PT), had been Brazil's president from 2003 to 2010. In July 2017, a lower court condemned Lula to nine years and six months in prison for corruption and money laundering.Footnote 74 Lula's lawyers filed a habeas corpus appeal with the STF, Brazil's highest court, to overturn the sentence. The result was open military involvement in leadership selection.
On 3 April 2018, the day before the STF ruling, Villas Bôas tweeted, ‘[i]n this situation that Brazil is currently living through, it remains to be asked of the institutions and the people, who is really thinking about the good of the Country and future generations and who is concerned only with personal interests?’ He subsequently tweeted, ‘I assure the Nation that the Brazilian Army shares the anxiety of all good citizens to repudiate impunity and give respect to the Constitution, social peace and Democracy, while it remains attentive to its institutional duties.’ Several Army leaders responded positively. Gen. Gerson Freitas tweeted, ‘[o]nce again the Commander of the Army expresses the concerns and anxieties of Brazilian citizens in uniform. We are together, Commander @Gen_VillasBoas!’ Gen. Pinto Sampaio tweeted, ‘[a]s the distinguished historian Gustavo Barroso said: “We all pass away. Brazil stays. We will all disappear. Brazil remains. Brazil is eternal. And the Army must be the vigilant guardian of Brazil's eternity.” *Always ready Commander!!*’ Gen. Antônio Miotto tweeted, ‘Commander!!!!! We are in the same trench!!! We think alike!!! Brazil above all!!! Strength!!!’
An interviewee posited two explanations for this military contestation:
Villas Bôas’ tweets were the only moment when he crossed the line [vis-à-vis civilian control] … We have to ask to whom it was directed. To the STF, saying that the military was going to overthrow it if it let Lula leave prison? To the internal public of the armed forces, where [Villas Bôas] has prestige and could have acted so that the troops did not?
Villas Bôas himself later noted that he had not tweeted in isolation. Rather, he had held discussions with the military's High Command, consisting of 15 four-star generals, before releasing the tweet.Footnote 75
A potential explanation for why Villas Bôas engaged in discursive contestation around Lula's 2018 habeas corpus appeal but not Rousseff's 2016 impeachment is that, within the context of the still ongoing Operação Lava Jato, Villas Bôas and, by extension, the armed forces considered impunity for Lula's corruption more of a threat to Brazil's political stability than they did Rousseff's impeachment. This external concern could have encouraged military contestation. Yet, this was not Brazil's or even the PT's first corruption scandal. The federal authorities had implicated nearly two dozen members of the PT and allied parties in the 2004 ‘Mensalão’ corruption scandal, which involved paying legislators to support Lula's congressional agenda. The armed forces had been concerned with the scandal and some military officials had even discussed openly whether Lula would win re-election, but they had avoided entering the political fray with contestation.Footnote 76 External political concerns like corruption thus appear insufficient for explaining the military's discursive contestation over leadership selection.
Another explanation lies in the armed forces’ rightist ideological leanings, a particularly strong antipathy to the PT that cuts across factions. While the armed forces’ ideological leanings plausibly influenced their reaction to Lula's habeas corpus proceedings, we would expect them to have taken a more interventionist approach to the ‘Mensalão’ corruption scandal, as well as to Rousseff's impeachment, if ideology was the central factor driving contestation. In both cases, however, they largely stayed out of the political fray. They also avoided intervening under Temer, whose Movimento Democrático Brasileiro (Brazilian Democratic Movement, MDB) had helped ensure the PT's presidential victories under both Lula and Rousseff.
A stronger explanation concerns conflict between interventionists and legal-institutionalists over the content of military identity vis-à-vis objectives for the armed forces. Rousseff's impeachment probably aligned with both groups’ preferences, insofar as interventionists considered it a means of stabilising Brazil and reigniting national development amidst corruption scandals, while legal-institutionalists worried that the military would lose cohesion by opposing the process. Hence, neither group had incentives to contest impeachment. Even for interventionists who considered Rousseff's ouster beneficial for the armed forces’ power, discursive contestation was unnecessary because the legislators increasingly supported impeachment. If they discursively contested Rousseff, the interventionists would expend their political capital needlessly because impeachment was largely a foregone conclusion. Therefore, partially due to alignment between interventionists and legal-institutionalists for impeachment, little contestation occurred. An interviewee suggested as much:
Villas Bôas spoke for the Army [in his tweets]. However, informally, there was potential instability. The Army still has symbolic elements that encourage being reactivated in crises and consider themselves responsible for institutional stability … [These elements surfaced when, in 2016,] Villas Bôas received more politicians in his office than Temer. Politicians knock on the barracks doors, too. Politicisation is due also to civilians.
The interviewee thus acknowledged (1) intra-institutional divisions that echo the interventionists versus legal-institutionalists distinction; (2) these divisions’ catalytic effect on the armed forces’ political behaviour, notwithstanding the absence of military contestation; and (3) Villas Bôas’ representativeness and responsiveness as an Army Commander concerned with cohesion. The interviewee implied that, had intra-military conflict been more acute in 2016, Villas Bôas would have used discursive contestation then in an attempt to preserve cohesion. Indeed, Lula's habeas corpus appeal probably exacerbated identity conflict over the armed forces’ objectives between interventionists and legal-institutionalists.
As national crises deepened, there were growing clamours, especially among retired military officials, for the military to exercise more power vis-à-vis national leadership. Pro-coup sentiment was deepening among interventionists. In September 2017, Army Gen. Hamilton Mourão, who became Bolsonaro's vice president in 2019, declared, ‘[t]here will arrive a time when we will have to impose a solution and this imposition will not be easy, it will bring problems’.Footnote 77 A week after Mourão's statement, Villas Bôas tweeted an image of a closed-door meeting of senior Army officials writing, ‘I was with active-duty and retired generals in RJ [Rio de Janeiro]. The objective was to personally advise the members of the @exercitooficial. #Coesão [@Official Army. #Cohesion].’Footnote 78
For legal-institutionalists, Lula's re-election risked undermining the military's cohesion because his campaign and victory would increase anti-corruption fervour and pro-coup sentiment. Despite his corruption conviction and the ongoing crises that some Brazilians attributed to his presidency, ex-president Lula retained widespread popular support. In a March 2018 poll, he received 33.4 per cent of vote intentions versus only 16.8 per cent for then-Congressman Bolsonaro. When Lula was not a possible candidate, however, Bolsonaro jumped to the top of the list against assumed PT replacement candidate Fernando Haddad, who received only 2.4 per cent of vote intentions.Footnote 79 If the STF had approved the habeas corpus appeal and Lula regained his political rights, he would probably have won re-election in 2018. Consequently, legal-institutionalists had limited choices. Given widespread support among the public and retired officials for military intervention to mitigate the crises, legal-institutionalists probably considered preventing Lula from seeking re-election the sole path toward maintaining cohesion. In an interview on 11 November 2018, Villas Bôas noted:
I acknowledge that there was an episode where we were actually at the limit. That was [my] tweet the day before the [STF] vote on the issue of Lula. There, we were consciously striving knowing that we were at the limit. However, we felt that things could get out of our control if I did not speak up because other people, reserve military officers and civilians who identified with us, were speaking more emphatically.Footnote 80
This statement suggests that Villas Bôas’ discursive contestation through the 3 April 2018 tweet was a means of reducing identity conflict between interventionists and legal-institutionalists by signalling that he understood their concerns and, therefore, was taking a unifying position by objecting to permitting Lula's candidacy. As one interviewee stated, ‘there were rumours … about the possibility that [Villas Bôas’ contestation] had been an attempt to control the troops … In Villas Bôas’ perspective, the troops were going to speak out and the crisis was going to deepen if he did not speak out himself.’ Concurrently, Villas Bôas was asserting the military's prerogative over leadership selection by discouraging the STF from granting Lula's habeas corpus appeal. If the courts had released Lula at that time, there would have been no legal recourse to prevent him from running for and probably winning the 2018 presidential election. This outcome could have engendered even greater military involvement in national politics, risking more identity conflict, as had occurred during the dictatorship of 1964–85. Villas Bôas himself noted this fear, stating, ‘we were concerned with stability because the worsening of the situation would then fall into our lap. Prevention is better than cure.’Footnote 81 Therefore, preventing Lula from seeking re-election through contestation was, paradoxically, how legal-institutionalists sought to protect cohesion and prestige and preclude the armed forces’ broader military intervention.
Internal Security
Scholarship on the prerogatives that constitute latent military power and on military power itself emphasises the armed forces’ constitutionally sanctioned role in ensuring internal security and their institutional authority over police agencies responsible for internal security.Footnote 82 Military power increases and civilian control decreases as these roles and authorities expand. These roles and authorities, moreover, risk jeopardising military cohesion. The shift to internally focused roles aimed at combating threats from societal actors has historically contributed to the Brazilian military's politicisation.Footnote 83 These roles ‘normally lie outside the professional domain of the [military], and as such, outside the realm of unquestioning obedience or established military doctrine’.Footnote 84 Law enforcement-focused internal security roles also risk undermining cohesion by exposing the armed forces' ineffectiveness against criminal organisations, making soldiers vulnerable to corruption and prosecution for human rights violations, and eroding the armed forces’ legitimacy and prestige.Footnote 85 Such unintended consequences can undermine the discipline, hierarchy, political neutrality and inter-personal and institutional trust necessary for military cohesion.
The Brazilian military's power vis-à-vis internal security roles and policing authorities is embodied within Garantia da Lei e da Ordem (Guarantee of Law and Order, GLO) operations.Footnote 86 Under GLOs, state governors request and the president authorises military deployments to Brazilian states during internal security crises. Governors can also choose to subordinate their states’ preventive law enforcement agency, the Military Police, to the Army as a reserve component.Footnote 87 While GLOs reflect civilian control in these ways, civilian attempts to curb the military's internal security roles are a main source of contestation.Footnote 88
Between 2000 and 2018, presidents deployed the armed forces 106 times under the GLO, with 19 of these missions responding to ‘urban violence’.Footnote 89 In 2017, Gen. Villas Bôas engaged in discursive contestation over GLOs several times in ways associated with identity conflict over the armed forces’ objectives between interventionists and legal-institutionalists. On 22 June 2017, he criticised the increasing frequency of GLOs before the Senate Commission on Foreign Relations and National Defence, stating that this trend was causing a stir among the armed forces: ‘We do not like this type of work. We do not like it.’ Seeing soldiers point heavy weapons at women and children in the narrow alleyways of Rio de Janeiro's Favela da Maré (a high-crime, low-income neighbourhood) during Operação São Francisco (April 2014–June 2015) had shown him that ‘we are an ailing society’. He concluded, ‘we need to rethink this model of [military] utilisation because it is wasteful, dangerous and ineffective’ as, once the military withdrew from the Favela da Maré, ‘everything returned to how it was before’.Footnote 90
After President Temer decreed Operação Rio de Janeiro on 27 July 2017 for fighting crime, Villas Bôas again engaged in discursive contestation to enhance military prerogatives in terms of ‘judicial security’, which would give soldiers protection from civilian oversight for human rights violations during GLOs. On 7 August 2017, he tweeted that Operação Rio de Janeiro ‘requires judicial security for the soldiers involved. As commander I have the duty of protecting them. The legislation needs to be revised.’ On 10 August, he tweeted, ‘I reinforce my position: the law needs to be modified.’ Observers perceived these tweets as pressuring Congress to pass Chamber of Deputies Legislative Project 44, a 2003 proposal placing homicides by federal soldiers and state military police under the jurisdiction of the military justice system.Footnote 91 During an interview on 2 October 2017,Footnote 92 Villas Bôas further pressured Congress to protect soldiers from civilian justice during GLOs, noting that failure to do so could make soldiers conducting Operação Rio de Janeiro more hesitant and less effective in fighting crime and, by having civilians discipline soldiers, could undermine military hierarchy. The pressure worked and Congress took up the issue. On 10 October 2017, he tweeted, ‘I am grateful for the passage of [Project] 44, which will guarantee judicial security to my men while on GLOs.’ President Temer ratified the law on 13 October 2017 following Federal Senate approval.
Villas Bôas engaged in discursive contestation for a third time on 30 December 2017, criticising civilians’ increasing reliance on GLOs. Following another GLO to provide security during a Military Police strike in Natal, Rio Grande do Norte, he tweeted that ‘the armed forces have been used three times, in 18 months’ in the state. He added, ‘I am worried by the constant use of the Brazilian Army in “interventions” [GLOs] …’ In an interview in January 2018, he also expressed concern that GLOs could enable criminals to corrupt soldiers.Footnote 93
These three instances suggest that the Brazilian military's discursive contestation over GLOs occurred because expansions of this internal security role exacerbated identity conflict between interventionists and legal-institutionalists over the armed forces’ objectives. On the one hand, interventionists considered GLOs a means of activating the power embodied in prerogatives to ensure internal security amidst crises induced by criminal organisations and strikes by the Military Police. The intensification of GLOs had also coincided with increased military power through a growing budget.Footnote 94 On the other hand, legal-institutionalists considered GLOs detrimental to the armed forces’ cohesion. The less ‘judicial security’ that soldiers had, the more that discipline for alleged human rights violations and corruption during GLOs would fall to civilian courts. Under these conditions, the armed forces also had less power to sanction soldiers and, thus, to foster cohesion. Moreover, irrespective of judicial security, potential human rights violations and corruption during GLOs risked undermining cohesion by making soldiers appear undisciplined, insubordinate and untrustworthy, not to mention ineffective and illegitimate. Given legal-institutionalists’ concerns with cohesion, the resulting identity conflict probably motivated Villas Bôas’ discursive contestation to communicate his concerns to other military officials to the effect that GLOs threatened military cohesion, while also applying increased pressure on civil authorities to expand military prerogatives in terms of ‘judicial security’. To summarise, Villas Bôas’ discursive contestation clearly seems associated with intra-military identity conflict in these three instances.
The military's contestation of civilian authority surrounding GLOs contrasts with its behaviour during the 2018 Federal Intervention in Public Security in Rio de Janeiro State (hereafter ‘Federal Intervention’), a mission concurrent with the aforementioned GLO, Operação Rio de Janeiro. Following Governor Luiz Fernando Pezão's request for assistance after a crime wave disrupted Rio de Janeiro's annual Carnival celebration, President Temer declared the Federal Intervention on 16 February (to continue through to 31 December 2018). The Federal Intervention focused on ending ‘the serious compromise of public order [by] planning, coordinating and executing actions that effectively seek the incremental recuperation of the operational capacity’Footnote 95 of Rio de Janeiro's state security agencies. The military assumed a more strategic internal security role than during GLOs because the Federal Intervention granted the Army unprecedented authority to command, control and reform law enforcement agencies.Footnote 96 Temer could have granted this authority to civilians but designated Army Gen. Walter Braga Netto as ‘Federal Intervener’ and Army Division Gen. Richard Nunes as Rio de Janeiro's Secretary for Public Security, a role usually reserved for local politicians or police leaders.Footnote 97
Exposing inter-faction conflict between military leaders who supported or opposed Temer's decision,Footnote 98 the Federal Intervention seemingly aligned with interventionists’ objectives more than GLOs had because it gave the armed forces a more strategic internal security role. The Federal Intervention also aligned more with legal-institutionalists’ objectives because this strategic role jeopardised cohesion to a lesser extent. Soldiers would continue to conduct Operação Rio de Janeiro, another GLO risking corruption, ineffectiveness, human rights violations and loss of legitimacy and prestige that could undermine military discipline, hierarchy, political neutrality and trust. The Federal Intervention, however, gave the armed forces more authority over GLO Operação Rio de Janeiro. Senior officials therefore could impose discipline and hierarchy over soldiers on the streets with more autonomy to foster cohesion. It also allowed the military to be selective regarding soldiers’ internal security roles. When asked about policing in Rio de Janeiro's high-crime, low-income communities, Braga Netto indicated that there were no planned occupations of favelas. Rather, state police forces would enter favelas to conduct specific operations while Army units would remain outside, reducing soldiers’ risk of being both victims and perpetrators of violence.Footnote 99 The military lost four soldiers and killed 46 people during the Federal Intervention.Footnote 100 In contrast, Rio de Janeiro state police forces lost around 25 members and killed nearly 1400 civilians,Footnote 101 suggesting that police officers, not soldiers, bore the brunt of armed confrontations with criminal groups. Such alignment between identity groups probably dampened intra-military conflict and, thus, military contestation, as evidenced by how Villas Bôas challenged civilian control through means other than publicly discernible discursive contestation.
It was not by addressing microphones or tape recorders but, rather, behind closed doors that Villas Bôas reportedly challenged civilian control over the Federal Intervention. During a national security council meeting on 20 February 2018, he expressed concerns with the potential establishment of a civilian commission to hold the military accountable for human rights violations during the intervention.Footnote 102 Other meeting attendees told journalists that Villas Bôas was ‘fearful of the intervention’ because ‘he knows that various crimes will fall on the shoulders of [the military] … [a]nd he does not want those soldiers to suffer’ judicially.Footnote 103 He nonetheless refrained from critiquing the intervention publicly. A possible exception is a speech he gave on 24 August 2018 wherein he said, ‘[d]espite the intense work of those responsible, the approval of the people, and statistics that demonstrate the decrease in levels of criminality, the military component is, apparently, the only one engaged in’ the Federal Intervention. ‘No … sector of the local governments has applied itself, based on socioeconomic measures, to modify the low indices of human development, which keeps the environment prime for the proliferation of violence’, he added.Footnote 104 Because this comment concerned civilian agencies’ performance more than it concerned the military's internal security role, we do not consider it evidence of discursive contestation over the latter. That senior military officials involved more directly in the Federal Intervention engaged in clearer discursive contestationFootnote 105 poses a greater challenge to our argument's generalisability, as discussed in the subsequent section.
An alternative explanation is that Villas Bôas challenged civilian control over the GLOs of 2017 but not the 2018 Federal Intervention because, as with domestic military operations elsewhere in Latin America,Footnote 106 GLOs are unpredictable, such that contestation is simply a means of achieving more military prerogatives vis-à-vis human rights. Compared to the strategic Federal Intervention, tactical GLOs involve more contradictions between fighting armed actors (for which soldiers prepare extensively) and protecting civilians (for which soldiers do not prepare extensively) and, therefore, entail greater risk of soldiers committing and being prosecuted for human rights violations.Footnote 107 That Villas Bôas engaged in discursive contestation in order to persuade legislators to shift human rights accountability during GLOs from civilian to military courts in October 2017 but engaged in less military contestation afterwards could support this logic. However, he engaged in discursive contestation over the GLO operation in Rio Grande do Norte in December 2017 and January 2018, after politicians had already granted the armed forces greater judicial security during GLOs. Furthermore, the Federal Intervention was itself also unpredictable. The ‘Federal Intervener’, Gen. Braga Netto, wrote that the mission ‘constitutes an unprecedented, extraordinary activity and, consequently, there are neither references of best practices nor any historical models’.Footnote 108 Indeed, while Article 34 of Brazil's 1988 Federal Constitution granted the national government the prerogative of assuming control over Brazilian states’ authorities in order to ‘put an end to the grave compromise of public order’,Footnote 109 among other objectives, the Federal Intervention in Rio de Janeiro was the first since Brazil's democratisation.Footnote 110 These considerations illustrate how prerogatives are insufficient to explain discursive contestation. To understand contestation, we must also consider identity.
Generalisability Check through Media Analysis (2016–18)
While our findings concerning leadership selection and internal security based on Villas Bôas’ discourse suggest that identity conflict contributes to military contestation, our analysis of ESG/OMNIDEF articles offers several implications regarding these findings’ generalisability.Footnote 111 More specifically, this generalisability analysis suggests that our theory of identity conflict helping to explain military contestation is plausible even though it may apply more to internal security than to other issue areas and even though Villas Bôas’ statements and tweets are imperfect indicators of military discourse. We discuss the analysis and implications below in further detail.
Among the 28 OMNIDEF articles from January 2016 to December 2018 that were interpretable as military contestation, 13 (46 per cent) concerned issues around which military identity conflict seemed likely, while six (21 per cent) concerned issues around which Villas Bôas exercised discursive contestation the same year. All 13 articles that were both interpretable as military contestation and related to issues of likely military identity conflict concerned internal security. Seven articles addressed GLO operations before the Federal Intervention. Between August 2016 and August 2017, four of these detailed the armed forces’ concerns that civilian politicians were using GLOs to task the armed forces not only with policing but, also, humanitarian activities – both missions well beyond the military's conventional purview of national defence.
During the Federal Intervention, nine articles from February to November 2018 addressed the mission's political dynamics in a way that was interpretable as either criticising or, via advocacy, attempting to pressure civilian politicians. These articles addressed four issues: (1) how President Temer's decision to entrust Gen. Braga Netto with the Federal Intervention was based on political risk management and how Temer delayed funding the Federal Intervention; (2) how Braga Netto advocated for increasing security spending in Rio de Janeiro and for concluding the Intervention in December 2018 independent of apparent civilian preferences; (3) how retired Gen. Augusto Heleno advocated for loosening soldiers’ rules of engagement against criminal organisations despite human rights concerns; and, generally (4) how the armed forces worried that their intensifying anti-crime role was ineffective and would lead soldiers to commit more human rights violations.
The analysis of ESG/OMNIDEF articles suggests three important implications. First, with half the articles interpretable as military contestation also indicating likely military identity conflict, the theory that contestation stems partially from anxieties around institutional cohesion is indeed plausible. If cohesion and contestation were uncorrelated, we would not expect half the instances of apparent military contestation to concern an issue that could accentuate divisions within the armed forces. External factors like civilian control certainly help us to understand military contestation, yet any explanation that overlooks internal factors like identity conflict illuminates this phenomenon incompletely. Second, with only 10 per cent overlap among ESG/OMNIDEF articles, Villas Bôas’ discourse imperfectly generalises to Brazilian military identity and contestation writ large. Our argument may apply to senior military officials with Army-wide purviews like Villas Bôas, rather than elite military institutions like ESG or senior military officials involved more in policing missions. Third, our argument may generalise most clearly regarding internal security because all articles that were interpretable as military contestation and likely identity conflict concerned this issue area. Villas Bôas’ contestation over Lula's habeas corpus appeal may have been an outlier and, as an issue area, leadership selection may entail intensive conflict over military identity only on rare occasions. The sensitivity of direct military intervention in politics may also lead only the most senior military leaders to engage in the practice, with other less senior leaders fearing sanctions for such behaviour. More research regarding elite institutions and the distinction between Army-wide versus mission-specific discourse, as well as issue areas beyond internal security, would help assess how the argument holds beyond our empirical scope.
Implications and Opportunities
We have argued that identity conflict between interventionists and legal-institutionalists engenders discursive contestation to convey both preferences over specific prerogatives to civilians and concern over cohesion to the rank-and-file military. Discourse analysis of Brazilian Army Commander Gen. Villas Bôas’ 2016–18 public statements and interview transcripts regarding leadership selection and internal security support our claim. Analysis of ESG/OMNIDEF-cited articles lends further support while suggesting that our theory applies most strongly to senior military officials’ behaviour and internal security. In the realm of leadership selection, Villas Bôas did not exercise discursive contestation around the 2016 Rousseff impeachment but did do so during the 2018 Lula habeas corpus appeal. Identity conflict over the armed forces’ objectives was limited in the former case, while there were strong internal cohesion concerns in the latter. Regarding internal security, Villas Bôas exercised discursive contestation around the 2017 Rio de Janeiro GLO operations but not around the 2018 Federal Intervention in Rio de Janeiro because, given concerns over cohesion, identity conflict vis-à-vis the armed forces’ objectives was probably extensive in the former but limited in the latter.
Our argument and findings suggest that the Brazilian military's recently increased power stems in no small part from its discourse and identity. This may help explain why, during Bolsonaro's presidency, the military's power in terms of senior officials’ ability to alter the right-wing populist's policies seemingly diminished. Legal-institutionalists within Bolsonaro's government may not have used discursive contestation extensively to express dissatisfaction and activate the power embodied in prerogatives because, given strong support for Bolsonaro among junior officers that comprise the legal-institutionalist faction and senior interventionist officials’ political co-optation, overall identity conflict remained limited.Footnote 112 In contrast, President Lula's recent re-election (2022) may lead to renewed contestation, given his low level of support among junior officers and the removal of retired military officials from cabinet positions that have traditionally been reserved for civilians.
Another finding is that the Brazilian military has used Twitter to contest civilian authority in diverse ways. That Gen. Villas Bôas contested leadership selection in 2018 (Lula re-election) but not 2016 (Rousseff impeachment) while contesting internal security missions in 2017 (GLO operations) but not 2018 (Federal Intervention) suggests that, as Brazil's political crisis intensified, the military became more strategic in using Twitter to challenge civilians. Armed forces’ contestation via social media may respond not only to distinct dimensions of civil–military relations but also to evolving political dynamics at the national level.
Beyond those related to generalisability, we see two main research opportunities stemming from this paper. One involves re-examining Brazil's democratic transition to allow an understanding of contemporary civil–military relations and democratic politics, including how the armed forces’ identity and discourse have changed since 1988. The military's recently increased autonomy and influence should lead us to ask whether senior officials believe that they truly left power and handed it back to civilians in the first place and, if so, how their self-perception has evolved.Footnote 113 A further opportunity involves probing our argument in other Latin American countries where military power has recently expanded amid democratic crises. As Nicole Jenne and Rafael Martínez underscore, such states’ increasing domestic usage of the armed forces has reflected and reinforced militaries’ power while undermining public confidence in the ability of democratic governance and civilian authorities to resolve political problems.Footnote 114 Compared to Brazil, these dynamics may be less intense in Argentina, Chile and Uruguay, similarly acute in Colombia, Ecuador, Mexico and Peru, and even more pronounced in Bolivia, El Salvador and Nicaragua, given relative levels of democratic consolidation and domestic military deployment. If our argument holds outside Brazil, we expect the degree of military contestation over leadership selection and internal security missions via social media to reflect these distinctions and to become more evident as democracies’ economic, ethical, political and security crises deepen.
Acknowledgements
The authors thank Rikhil Bhavnani, Yoshiko Herrera, Caileigh Glenn, Kate Carter, participants in the 2019 American Political Science Association panel and participants in the 2019 Latin American Colloquium and Comparative Politics Working Group at the University of Wisconsin–Madison for their invaluable comments on this paper.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0022216X23000664
