Exploring the conflict-readiness of parties: The dynamics of proclivity towards violence and/or conflict in Madagascar

Dr Velomahanina Tahinjanahary Razakamaharavo is the Visiting Postdoctoral Fellow in Global Transformations at the Hochschule für Politik München at the Technical University of Munich. She is also a Scientific Collaborator at Université catholique de Louvain, Louvain-la-Neuve, Belgium. She obtained her Ph.D. in International Conflict Analysis from the University of Kent’s Brussels School of International Studies (BSIS). She is a 2017 Women in International Security (WIIS) Next Generation Scholar.

Abstract

Since the beginnings of the anti-colonial struggle, Madagascar, a former French colony and an island in the Indian Ocean, has gone through nine episodes of conflict, ranging from political tension to high intensity conflicts. These changes of conflict intensity demonstrate that the proclivity towards conflict may take different forms in various episodes of  violence and conflicts in the country. This phenomenon may be explored by examining the causal configurations and the co-existence of positive and negative processes and mechanisms which are interacting and co-constructing each other. In order to untangle the intricacy behind the conflict-readiness of parties preparing for conflict at low, medium or high levels of violence, use is made of concepts and theories pertaining to peace, conflict, negotiation and mediation, conflict escalation and de-escalation to explore the roles played by the following factors:

1. local narratives and metanarratives.  
2. repertoires of action of the actors 
3. the actors’ framing of the conflicts
4. the actors’ polarising of public opinion
5. construction of the image of the self and the other 
6. conflict dimensions (socio-economic, cultural, political and global external) 
7. accommodation policies 

This paper argues firstly that proclivities toward violence/conflict in Madagascar are related to the coexistence of positive and negative elements, and secondly, that such proclivities are built partly upon the fact that liberal strategies for maintaining peace give rise to negative as well as positive effects on the dynamics of keeping that peace.

1. Introduction

Whenever an episode of conflict breaks out or recurs in their country, the Malagasy always say ‘tsy maintsy mipoaka ny sarom-bilany’ (‘the cover of the cooking pot will explode one day or another’). Regular episodes of conflict have occurred in Madagascar ever since the colonial period, despite the implementation of various initiatives aimed at addressing the incompatibilities between the conflicting parties. These episodes have been different in nature. Some involved socio-economic dimensions; but other dimensions from the remnants of colonisation also featured. These episodes seem to be linked to each other by common elements at play. For instance, all of them feature political dimensions, and the state is therefore always present as a party on one side of the conflict. The episodes have also differed in levels of intensity, ranging from mere political tension to high intensity conflict. From these variations it can be inferred that there are differences in the preparedness of parties for conflict at low, medium or high levels of violence. 

This article intends to contribute to the literature on peace and conflict processes in Madagascar. The majority of the existing scholarship tends to focus on specific episodes of conflict (e.g. Ralambomahay 2011; Ratsimbaharison 2017) or current events, and thus do not present a holistic picture of the varying proclivities towards peace and violence which emerge. Moreover, a great majority of the scholars working on Madagascar focus on the episode of conflict that occurred in 2009 and pay less attention to the earlier ones, such as those that broke out in 1971 and 1972 (e.g. Blum 2011). 

The present article uses Qualitative Comparative Analysis (QCA) methods and techniques to explain the principles behind causal configurations in recent history. 

First, it argues that the proclivity towards violence and conflict in Madagascar may be studied in conjunction with the coexistence of positive and negative elements where peace and conflict processes co-exist, interact, and co-construct each other. The ways in which these elements change and are changed in this process affect peace in a negative or positive manner and constitute the dynamics of the situation. Second, it suggests that those dynamics are built partly upon liberal peace prompting negative effects despite the existence of elements affecting the process in a constructive manner.

2. The episodes of conflict in Madagascar

The story of conflicts in Madagascar shows a highly unstable situation. There have been nine episodes which have occurred consecutively during the post-colonial period. The beginnings of the colonial struggle in 1947 saw more than 1000 deaths, including Malagasy from the Merina and Côtiers ethnic groups who were said to be pitted against each other by France. Later, in 1971, the farmers of the Southern part of the country rose up to protest against the state’s capitalism and its lack of interest in the people of the South in a rebellion that was led by a Maoist political party.¹Raharizatovo, Gilbert. Interview with the author in June 2014, Antananarivo (recording in possession of author). Recordings of all the following interviews are also in possession of the author. That event also led to more than 1000 deaths. In 1972, another episode started, this time led by young intellectuals issuing from the Maoist party and supported by young deprived people from the poor neighbourhoods of Antananarivo. They fought against the state and denounced the cooperation treaties which Madagascar had signed with France. French imperialism and the state’s education policy were at the centre of the issues at stake (Blum 2011:68). This resulted in 40 deaths and 100 wounded. Clashes with the political elite in 1975, involving 22 deaths, climaxed with the assassination of the new President, Colonel Ratsimandrava, who had just gained power and had held office for only six days. Ideological clashes involving socialist and communist ideas as well as identity issues, where the actors exploited the issue of ethnicity between the Merina and Côtiers communities for political advantage, were said to be among the problems during that episode.²Rajakoba, Daniel. Interview with the author in June 2014, Antananarivo. In 1985, the state intervened militarily to deflect a movement thought to be initiated by a group of people practising the martial art of Kung Fu. This group maintained that they ‘stood up to defend the population against the terrorising acts’ of a group of deprived young people called Tanora Tonga Saina (TTS, Unemployed Youth Aware of Their Responsibilities). This group was in fact ‘used by the state to crush any political attempt to counter its power such as the organisation of protests’.³ZOAM [Zatovo Orin’asa Malagasy or Young Unemployed Malagasy] member. Interview with the author in March 2016, Antananarivo. There were more than 100 deaths during that episode. 

The episode of 1991 saw a strong opposition demanding democracy; they confronted the powerful President Ratsiraka who had held office for around 15 years. Fifty people died and 300 were wounded during a large march organised by the opposition at the Presidential palace of Iavoloha. It is claimed that the President gave the order to shoot at the crowd which was marching peacefully. Institutional clashes in 1996, described by the actors as a crisis, led to the impeachment of the President.⁴Zafimahova, Serge. Interview with the author in March 2016, Antananarivo. There were no casualties during that episode. Later on, in 2002, another event occurred initiated by an electoral dispute.⁵Raharizatovo, Gilbert. Interview with the author in June 2014, Antananarivo. This episode, in which there were more than 60 deaths,  involved issues pertaining to identity such as belonging to ethnic groups (e.g.: Merina or Côtiers) or social classes (e.g. ‘Andriana’/noble or ‘Andevo’/slave). Lastly, the episode in 2009 was an event involving all the main political parties which had participated in the earlier episodes of conflict. President Ravalomanana was overthrown in a military coup led by Mayor Rajoelina. Since that time over 100 more deaths have occurred.

3. On levels of conflict intensity and stages of (de-)escalation 

First, this article groups these episodes of conflict according to their levels (steps in a process/developmental sequence) and intensity (magnitude of tension or violence reached in each of the episodes). Second, it identifies the stages of escalation. For example, a stage of escalation shows an escalation from a lower to a higher level (e.g. level 1 to 3), de-escalation from a higher to a lower level (e.g. level 5 to 4), and protraction when the level of the conflict remains the same. These levels of conflict show that, within the dynamics, one may distinguish between a proclivity to low-intensity tension or conflict and to high-intensity conflict and violence. And third, it studies the dynamics of peace and conflict processes in Madagascar, regarded as one conflict involving multiple episodes. The article does not study escalation or de-escalation of levels of conflict within an episode of conflict. It does focus on such stages in the years between subsequent episodes . 

In view of all this, this study uses Féron’s typology (2005:12‒13, see also Rihoux 2008:14). Each level in her typology corresponds to specific features and a given intensity. 

Level 1: peaceful and stable situation (no deaths or casualties), with a high degree of stability where people co-exist peacefully.

Level 2: political tension situation (1 or 2 killed), where there is an increase of social and political cleavages, but almost no human casualties.

Level 3: violent political conflict (up to 100 deaths), in which tension escalates to a political crisis and one can observe the national government’s political legitimacy opposed by an increasing number of violent political factions.

Level 4: low intensity conflict (between 100 and 999 deaths), which features open hostility, regime repression, insurgency and in many cases armed conflict among factional groups.

Level 5: high intensity conflict (1000 or more deaths) which in many cases features open warfare between rivals, and mass destruction.

Figure 1: Episodes of conflict in Madagascar between 1945 and 2016. 

This diagram illustrates the dynamics of conflict in Madagascar: the episodes, and the shifts in the levels of conflict-readiness in the intervening periods. The triangles represent the episodes of conflict (1947, 1971, 1972, 1975, 1985, 1991, 1996, 2002, and 2009) and the ellipses the shifts in conflict levels. The arrows show the escalation. For the pre-1971 period, the 1947 episode is indicated, but the timeline is not extended proportionally. No major episode of conflict broke out during this period, and there was no case of escalation or de-escalation. 

According to Féron’s typology, episodes of conflict in Madagascar can be grouped in the following way:

• Political tension: 1996
• Violent political conflict: 1972, 1975, 1991, and 2002
• Low intensity conflict: 1985 and 2009
• High intensity conflict: 1947 and 1971

When looking at the types of escalation in the intervening years, there are two variants:

Escalation A: level 3 (violent political conflict) to level 4 (low intensity conflict): from episodes 1975 to 1985 as well as episodes 2002 to 2009

Escalation B: level 2 (political tension) to level 3 (violent political conflict): from episode 1996 to episode 2002

This paper will explore the conflict-readiness of the parties who are preparing for conflict at low, medium or high levels of violence. But before doing so, it is important to shed light on the variants of de-escalation (which are not indicated in Figure 1 where the emphasis is on the cases of escalation): de-escalation A (high intensity conflict to violent political conflict), B (low intensity conflict to violent political conflict), and C (violent political conflict to political tension). These cases of de-escalation show proclivity towards peace as well as co-existence of peace and conflict, and success in building more peaceful situations and resilience. One element that these de-escalation cases have in common, is that the actors never resorted to an official external intervention of a third party to resolve the conflicts.  

4. Studying the interaction of peace and conflict processes and the factors involved

4.1 Approaching peace and conflict processes inclusively

A crucial caveat guiding the principles behind the author’s approach in this article is in opposition to the way peace and conflict processes have been studied so far: separating peace and conflict and studying each as an exclusive phenomenon. She rejects such an approach and instead adopts one in which peace and conflict are studied as involved together in a process.  They co-exist and co-construct each other. For example, there are situations where there is peace among the political elite (at the top), but conflict persists among the local communities (at the bottom).  This can be explained by the fact that the degree or quality of peace is influenced by the positive and negative contributions of various factors. 

4.2 Factors influencing the readiness for conflict or the promotion of peace

The author decided to group the factors into two categories: a) the methods the actors use, and b) the structural factors, also known as contexts or outcome-enabling factors. Each of them influences the quality of peace in a constructive (favouring peace) or destructive manner (prompting conflict). Note that this paper uses a great deal of narrative in this research. These are the stories actors use to recount their experiences or those of others, or to imagine the future. They can be a) metanarratives, the intangible ideologies or the universal truths that are deeply anchored in people’s beliefs, or b) local narratives, the modest and everyday ‘petits récits’ appearing in local contexts. For further discussions of the conceptualisation of these factors, see Razakamaharavo (2018:55) and Rihoux (2008:15‒20).

4.2.1 Methods developed by the actors 

• Repertoires of action. These are collective actions resorted to by actors in order to defend their causes. These can be conventional institutional actions (e.g. lobbying), or conventional non-institutional actions (e.g. petitions). The first abide by institutional rules and the second do not, but both types are generally accepted. Other actions are non-conventional actions with no violence against persons (e.g. destruction of property) and non-conventional actions with violence against persons (e.g. violent demonstrations). 
• Boundary construction of the perceptions of the self and the other. These are the constructions of image. For instance, the self considers the other as symbolically non-existent. Other situations display cases where actors consider one another as aliens creating conflict or confrontation. A third option favours collaboration and partnership where actors acknowledge the existence of the other. 
• Framings of the conflict. The origin of a conflict is imagined and interpreted by the actors involved in peace and conflict processes. They each have their own understanding of the conflict, also known as ‘conflict paradigms’. These understandings can be different (there is discordance), or the same (shared). They can change over time as they can be ‘reshaped or reframed’ (paradigm change). At other times, they remain the same throughout the stages of escalation or de-escalation (no change of paradigm). 
• Polarising actors. This unit of analysis features a situation where actors engage in polarising public opinion. They are usually highly motivated to mobilise others and have an obvious capacity to influence others by means of their ideologies and resources. 

4.2.2 Structural factors 

• Conflict dimensions. These are considered as the causes, issues or features of the conflict: cultural dimensions (concerning religion, identity, culture or language), socio-economic dimensions (e.g. conflicts based on Marxist ideologies), political dimensions (e.g. centre-periphery cleavage, security dilemma) and external dimensions (e.g. colonisation or decolonisation). 
• Accommodation policies. These feature initiatives addressing issues relating to the conflict; they can be mediation initiatives, negotiation, conflict management programmes or peace-building initiatives. 

5. Data collection and analysis

The author collected information from multiple sources: newspapers, video broadcasts, Ph.D. manuscripts, press releases or songs, and conducted 44 in-depth interviews. Once these pieces of information were gathered, she input and tested them with Qualitative Comparative Analysis (QCA) software (Tosmana and fsQCA). With the Boolean logic in QCA, she could assign values to the factors, called ‘conditions’ in QCA terms (but for the sake of clarity, this article uses the term ‘factor’). She used crispSet QCA (csQCA) in this task. It means that the data were dichotomised. For instance, for the accommodation policies, she assigned 0 to the absence of accommodation policy and 1 to its presence. The same with the perception of the self/the other: if the actors perceived each other as alien, she assigned value 1 to that characteristic and with other situations, she assigned 0. The results of the tests were then interpreted. QCA pairs up with other methods. She did conflict and policy analyses, and also interpretive analysis. 

6. Causal configuration

6.1 Principles behind Qualitative Comparative Analysis

The author decided to use the principles behind causal configuration in Qualitative Comparative Analysis (QCA).  It suggests that a configuration of factors (the independent variables) or a combination of factors explains the occurrence or non-occurrence of a given outcome (the dependent variable). Moreover, one or multiple paths can lead to the outcome (Rihoux and Ragin 2009:1‒17). Consequently, a given configuration can contain factors promoting peace or prompting conflict, depending on the situation. The following is an example of how causal configuration manifests itself in the proclivity to violence and conflicts in Madagascar.

Based on the results of the QCA tests, one of the variants of escalation shows that there are multiple causal configurations leading to its occurrence. These lead to a build-up of the intricacies of the co-existence and interaction of factors and contribute to the escalation of violence/conflict in Madagascar. For each path, a number of factors which can have positive or negative influence on peace work together explaining the escalation. For example, the escalation between level 3, violent political conflict, and level 4, low intensity conflict, features three paths in terms of structural factors:

• Among these is a combination of absence of accommodation policy, low saliency of external dimensions and high saliency of socio-economic dimensions.
• There is also a combination of presence of accommodation policy, high saliency of external dimensions, high saliency of socio-economic dimensions, and high saliency of cultural dimensions. 

The existence of these different paths might lead one to think that for a specific stage of escalation, each path might be qualitatively different. 

6.2 The factors in the causal configurations

Table 1: All the configurations from the QCA test. 

Configurations explaining the variants of the stages of escalation:

	Structural factors	Methods developed by the actors
Escalation A: from level 3 to level 4	Path 1: Combination of absence of significant accommodation policy, high saliency of socio-economic dimension, low saliency of cultural dimension  Path 2: Combination of presence of significant accommodation policy, low saliency of external dimension, high saliency of socio-economic dimension Path 3: Combination of absence of significant accommodation policy, high saliency of external dimension, high saliency of socio-economic dimension, high saliency of cultural dimension	One path: there is a combination of the following:  Both camps perceive each other as alien, the two camps disagree on the nature of the conflict and they change completely the framing of the conflict or add other framings to their understandings of the origins of the conflict
Escalation B: from level 2 to level 3	One path: Combination of presence of significant accommodation policy, low saliency of external dimension, and high saliency of cultural dimension	One path: there is a combination of the following: Both camps perceive each other as alien, there is more frequent use of non-conventional types of action, existence of polarising actors, the two camps disagree on the nature of the conflict, and they change completely the framing of the conflict or add other framings to their understandings of the origins of the conflict

Below is a list of all the factors present in the configurations, but categorised in accordance with their positive or negative effects on the quality of peace. This paper leaves out the details of some of them as it explores them in depth in sections 7 and 8.  

6.2.1 Factors influencing the quality of peace in a negative or destructive manner

The following, which are ascribed to the configurations of factors, have had destructive (negative) influences on the quality of peace:

• High saliency of socio-economic dimension: interwoven elements (see explanations below)
• High saliency of external dimension: the involvement of France and the United States (US) created anti-neo-colonialism and anti-globalisation
• High saliency of cultural dimension: federalism and autonomous provinces have been perceived as ethnic discourses and an attempt to divide the Malagasy (the two ethnic communities Merina and Côtiers)  
• Presence or absence of significant accommodation policies: more negative/destructive than constructive effects (see explanations below)       
• Presence of polarising actors: polarisation at the elite level (e.g. actors polarise over ethnic issues) inciting the population to use violence 
• Paradigm change: complete shift of framing and additional framings of conflict (higher level of complexity and thus more difficulty to resolve)
• Paradigm discordance (see explanations below)
• Construction of the image of the self and the other (see explanations below)
• More frequent use of non-conventional types of action: the more one is moving towards episode 2009, actions are more strategically organised and tactically planned, and more casualties result (repetition, reproduction and innovation of repertoires at play) 

6.2.2 Factors influencing the quality of peace in a constructive manner

• Low saliency of external dimension: this can be explained by the regular policies implemented by the state to address some of the conflict dimensions in former episodes of conflict. Or the state was more focused on experimenting with new ideologies and policies as well as programmes.
• Low saliency of cultural dimension: the same explanations above apply in this situation  
• Paradigm discordance: there was discordance but at the same time there was a shared understanding in terms of political dimension (political dimension is a constant in the research design) which opened up the possibility of collaboration among the conflicting parties

7. Conditions furthering the proclivity to violence and conflict

It is worth reiterating that, although this paper explores these factors separately, the QCA results show that for escalation to occur, multiple factors work together in the configurations.Ten out of the twelve factors (structural factors and methods developed by actors) have built up the different paths leading to escalation have been influencing the quality of peace in this country in a negative/destructive manner. 

7.1 Frustration and aggression caused by socio-economic dimensions

Féron (2005:25‒26) argues that socio-economic resources are sources of incompatibility in a conflict. In Madagascar, this condition is particularly relevant to the escalation from level 3, violent political conflict, to level 4, low intensity conflict. The Malagasy people have been experiencing constant structural violence in their everyday lives as result of successive policies including socialist and liberal ones. Socio-economic hardships created situations where different social groups felt they were socially and economically discriminated against. These feelings created situations of ‘us’ vs ‘them’, ‘the people’ vs ‘the elite’ and feelings of injustice and unfairness. One can say these have been traumatic experiences, some people were even ‘terrorised’ by these situations:       

I can only remember the 80s … This period of terror, the socialist propaganda … in the field of the economy there was great scarcity and shortage. There was constant queuing for basic necessities … the advent of a regime which aimed at being more liberal and democratic did not alleviate the poverty felt among the population (Express de Madagascar 2002a).

These socio-economic dimensions are interwoven elements concerning problems the actors who fought against the state denounced: pauperisation, social exclusion, economic imbalance between rural and urban areas or deprivation. These built up situations of frustration and aggression throughout the process, which explains the involvement of various social groups in the conflict, especially the poor population from the slums. Figure 2 maps out these main issues relating to the socio-economic dimensions.

Figure 2: Socio-economic dimensions in situations of escalation

Indeed, various interviewees from both sides of the conflicts discussed to a great extent these socio-economic dimensions and explained that they were among the reasons why conflict occurs and recurs in Madagascar. They especially referred to the cases showing escalation of levels of violence. For example, some politicians argued that the Malagasy ‘did not really experience political crises’; they ‘experienced injustice and poverty … human rights were not upheld and the central state did not care about the population, these were societal problems and crises of under-development’,⁶Andrianarivelo, Hajo. Interview with the author in June 2014, Antananarivo. while actors from the civil society organisations think that the Malagasy ‘have been in situations of crisis, the population does not have access to the resources that should be theirs’.⁷Razafindrakoto, Lalatiana. Interview with the author in March 2016, Antananarivo. Moreover, these viewswere reflected in the opinions of some of the interviewees who were members of the deprived population and the so-called delinquent ZOAM: ‘we all came from the same background: the poorest, the small people, we supported the popular movement not because of politics, we were all illiterate and dumb but we realised that the injustices were unfair … we said we were fed up with poverty, joblessness, imprisonment, the lack of jobs, and the bloodshed … give us jobs otherwise we will go to your house with crowbars and we will attack and loot your houses’.⁸ZOAM member. Interview with the author in March 2016, Antananarivo.

7.2 Explosive construction of images of the self and the other

The author also discovered two situations involving what she calls explosive constructions of images of the self and the other. One is when the two sides consider each other as alien and the other is when multiple conflicting parties are involved in peace and conflict processes and where multiple constructions of image as well as types of relationship are at play. 

First, the variants of escalation in Madagascar display the situation which Vetik (2007:8) considers to be the most explosive: when the two sides consider each other as alien. For instance, in the escalation between 1996 and 2002, one of the sides (the Ratsiraka clan) used ethnocentrism to construct its own image and accused the opposition of imposing a ‘nazi dictatorship’ and an ‘era of fascism’, but made it an excuse to use direct violence against the other, the enemy:

We have entered a phase of war. President Ratsiraka said ‘It is no longer enough to defend oneself … for us, the precept which says “attack is the best defence” has become our motto … If one attacks your family and if you have weapons, you must use these weapons, but it does not necessarily mean waiting to be attacked before using these weapons’ (Express de Madagascar 2002b. The last sentence is a statement Ratsiraka made, cited by one of his Counsellors).

Second, when looking at the escalation between episodes 2002 and 2009, a phenomenon the author calls ‘an interaction and succession of multiple boundary-construction types’ appeared and built up an explosive situation. Indeed, in 2009, at the beginning, the construction of the image of the self and the other was the ‘alien type’ (both sides saw each other as aliens). But during the peace process, the situation got even more complex as there were multiple constructions that were imposed by external third parties such as the United Nations (UN), the Southern African Development Community (SADC), and nations like France and the US. The actors in that episode were long-standing rivals. Many of the biggest and most important historical parties, who had pitted themselves against each other during the earlier episodes of conflict, united to support Rajoelina, the Mayor of the capital city, Antananarivo, and enabled him to topple President Ravalomanana in a military coup. 

During the negotiation and mediation period, the international community (IC) wanted to be inclusive in its approach and to take into account the identity dimensions involving the issues between the Merina and Côtiers ethnic groups. As a result, the IC increased the number of ‘mouvances’ or self-styled movements (Cawthra 2010:14), and included those of the two former Côtiers Presidents, Ratsiraka and Zafy. The other two mouvances (those of the two Merina, Ravalomanana and Rajoelina) were then obliged to recognise the existence of these actors. That situation gave birth to new perceptions. They had to construct new images of other actors in addition to those they had already built. Consequently, there were multiple perceptions of the self and the other (alien, partner etc.). These played an escalating role as relationships were difficult to manage and it was hard to identify who were enemies and who were allies. This can be seen in the many defections that took place within parties. Many joined other parties and multiple alliances were successively formed. Moreover, this played a role in lengthening the duration of the negotiation and the time it took to sign peace agreements.  

7.3 Accommodation policies creating negative rather than positive peace

Various accommodation policies were implemented during the cases of escalation of levels of violence.  For example, for the stage of escalation from level 3 to 4 (political tension to violent political conflict: 1975 to 1985 and 2002 to 2009), the actors resorted to negotiations as well as the mediation of internal and external third parties. These led to official signatures of agreements which in some cases resulted in international recognition of a given conflicting party and the setting-up of a government of transition. There were also situations that triggered sanctions from the so-called International Community or the imposition of policies such as the one that is called ‘ni-ni’ (neither-nor) which was suggested by SADC in the 2009 episode of conflict during which the main leaders of the conflicting parties were not allowed to run for the presidential election. Despite this the election was still considered to be ‘democratic’.    

For level 2 to 3 (political tension to violent political conflict: episodes 1996 to 2002), for example, there were accommodation policies involving the setting-up of a parallel government or the ‘unconstitutional’ transfer of power to the head of the High Constitutional Court. There were also mediation initiatives of external third parties which led to the official signing of a peace agreement, the commitment to consider an exit strategy, the organisation of a ‘democratic’ election, and the international recognition of a given party in the conflict.      

The situations of escalation of levels of violence in Madagascar show evidence countering theoretical expectations. Rather than promoting peaceful situations, the initiatives implemented by the actors to resolve the episodes of conflict were responsible for creating negative peace. Theoretically speaking, accommodation policies should play de-escalating roles and affect the quality of peace in a positive manner. Given the diversity of the accommodation policies and due to space restriction, the author will only use the peace agreements and the national reconciliation initiatives as examples in the following sections. 

The Dakar I agreement and the Dakar II document were central to episode 2002. These saw the light of day thanks to mediated bilateral negotiations supported by external third parties. In episode 2009, there were several series of multilateral mediation and negotiation initiatives, still involving external third parties, giving birth to multiple signed agreements and other accompanying documents: the Maputo agreements 1, 2, and 3, and the Maputo political agreement, the charter of the transition, the charter of values, the additional act to the Malagasy Charter of Transition of Addis-Ababa and the road-map for ending the crisis in Madagascar. Within that framework a Presidential election was organised in 2013, creating the Fourth Republic, while the implementation of these agreements continued its course after that election. Let us see what roles these play in escalation.

7.3.1 Scraps of papers: Political and elitist agreements

The majority of the interviewees and much of the written work on some of these agreements bluntly categorisedthem as complete failures (Bachelard and Marcus 2012:364). There were various reasons why these agreements were considered as such. For instance, they were highly political and elitist as many of their provisions were concerned primarily with institutional arrangements,⁹Ramananarivo, Stéphane-Pakysse. Interview with the author in June 2014, Antananarivo. power distribution and solutions addressing mainly the problems of the elite. With regard to the latter, for instance, some of the provisions mostly concerned the issue of the amnesty of exiled politicians and of those the government and the president had sent to jail. These were among the central, deal-breaker elements as well as the stumbling blocks in the process. Moreover, these agreements were highly political, they were designed to address specific issues of the moment:¹⁰Andrianirina, Fetison Rakoto. Interview with the author in March 2016, Antananarivo. the political issues that the third parties or the politicians involved in the process considered to be priorities during mediation and negotiation:

In mediations and discussions, we are in the short-term. Historical analysis, the causes of the conflicts, it is important to discuss the causes … these are rarely discussed, these are more academic … we should negotiate institutional agreements, power distribution, the agreement on the ni-ni, the causes of the crisis will be dealt with later.¹¹Châtaigner, Jean-Marc. Interview with the author in June 2016, Paris.

These agreements can also be more or less considered as simply scraps of paper in the sense that they were signed but not fully implemented. There have been important discrepancies between what was written and what was implemented, and these were serious impediments to the peace processes. The agreements signed in episode 2002 are evidence of these discrepancies. None of their provisions was taken seriously by the conflicting parties and there was no implementation of the plan suggested by the external third parties. For instance, there was a provision for the cessation of hostilities. However, both sides continued to commit direct violence after the signing of the agreement. Another example is what happened during the implementation of the agreements in episode 2009. A provision in the roadmap stated that land grants would not be allowed during the transition period. However, the winners of the conflict, the Rajoelina clan and consorts, did not respect that agreement and signed contracts with large Transnational Corporations (TNCs) which allowed large-scale land grabbing.¹²A former civil servant and diplomat serving under the Ratsiraka regime. Non-attributable comment made in March 2014, Antananarivo. Thus, it is clear that there was no accountability set up for insuring that follow-through took place, despite the existence of institutions such as a team conducting evaluation, a follow-up procedure for the Maputo and Addis Ababa agreements, and later the establishment of a joint-liaison office of the African Union and SADC. 

7.3.2 Perceived sources of failure of the Malagasy peace agreements

The author tapped into local perceptions of the reasons why these agreements failed and found that they were thought to be short-term solutions, not capable of solving the root problems, and addressing only the immediate ones. Moreover, the interviewees said that these were signed without sufficient background work, thus there was no link to the needs of the population.¹³Employee of Search for Common Ground Madagascar. Interview with the author in March 2014, Antananarivo. Others thought they were only made to legitimise conflicting parties and calm down the situation.¹⁴Rakotondranaivo, Lalatiana. Interview with the author in March 2014, Antananarivo. As far as the provisions were concerned, many said there was a strong focus on personal conflicts among the actors rather than on national solutions. 

Many criticisms were about procedures and technical issues. For instance, there were reservations regarding the involvement of external third parties who were perceived as biased and only protecting their national interests.¹⁵Randrianarisoa, Guy Rivo. Interview with the author in June 2014, Antananarivo. Furthermore, when the mediators included the former Côtiers presidents in the process in a spirit of inclusivity, many thought this approach revived old wounds among the Malagasy. As a result, the third parties were seen by some as negotiating with actors perceived as ‘terrorists’. As far as the internal third parties, like the Churches or the local Kings, were concerned, they were thought to be impotent and incapable of helping the Malagasy resolve their conflicts due to their own particular bias. The interviewees also said that some of the agreements were written in a hurry, within just one night, and there was a lack of transparency. Most importantly, there was no communication with the population, explaining what had been decided/negotiated. 

Who are you kidding? They imagined articles, no transparency … the content of the paper was not bad but the system through which politicians explained it to the population was bad, not transparent.¹⁶Ramananarivo, Stéphane-Pakysse. Interview with the author in March 2014, Antananarivo.

Added to this, there was neither a clear procedure for concretisation nor a follow-up of the implemented initiatives. Finally, the ni-ni approach was severely criticised as it was not inclusive, in that it forbade the main conflicting parties to run for election. Last but not least, one of my interviewees said that during the peace processes, the ideas of the population were taken, stolen, but not returned to them which means that there was no ownership.¹⁷Rakotondranaivo, Lalatiana. Interview with the author in March 2016, Antananarivo.

7.3.3 Failed national reconciliation: Lack of local ownership

National reconciliation was a pivotal element in the implementation of accommodation policies in the variants of escalation between 1996 and 2009. Yet, it played more of an escalating role leading towards more conflict in the process. The issue of national reconciliation is divisive in Madagascar. It is complex as it involves many dimensions which have evolved over time and have thus become progressively part of the framework: amnesty and reparation, autonomous provinces and federalism, ethnicity, truth-seeking and forgiveness, centre-periphery issues, trust between the government and the population, and violation of the constitution.

It took a long time for the idea of national reconciliation to be officially concretised (official concretisation here means having a legal framework and being institutionalised). An entity called the Comité pour la Réconciliation Nationale (CRN, Committee for National Reconciliation) was independently created by President Zafy in episode 2002.¹⁸Zafimahova, Serge. Interview with the author in March 2016, Antananarivo. It was a timid movement with little capacity to mobilise, and taxed with being ‘racist’, given the reputation of its leader as a Merina-hater and his identity as Côtier. The idea resurfaced in episode 2009, was intensively publicised, and in the opinion of the author was generally a trendy concept among the actors. Therefore, after the first official round of mediation of external third parties, the Conseil National pour la Réconciliation (CNR, National Council for Reconciliation) was set up. It is another entity espousing the same concept as the CNR, with the exception that the CRN, back then when it was established, was against the African Union’s ‘interference in national sovereignty’. It was in favour of reconciliation but excluded what it considered to be ‘negotiation with the terrorists’.¹⁹Ramananarivo, Stéphane-Pakysse. Interview with the author in June 2014, Antananarivo. The definitive concept was concludedin Maputo. It was only in episode 2009 that national reconciliation was institutionalised, and the road-map became the legal framework. The controversial 2010 constitution of the fourth republic institutionalised the idea of setting up The Filankevitry ny Fampihavanana Malagasy (the Council for Malagasy Reconciliation, FFM) as a state institution with a role focusing onnational reconciliation (but this constitution is not valid in the eyes of a number of actors).

Such a concretisation took a long time to realise as there have been various stumbling blocks throughout the process. Among these were the identities of the actors implementing initiatives (their history and their membership of a specific ethnic group) and the lack of local ownership and understanding of the concept. It was also perceived merely as an idea and concept from the outside (e.g. from France) and, most importantly, many saw national reconciliation as an issue that only concerned the elite and as a way to favour impunity. Finally, once the FFM was established as a state institution, it was perceived as biased.

8. Factors influencing the quality of peace in a constructive manner

Only two out of ten of the factors building up the multiple configurations explaining the occurrence of escalation of violence in Madagascar influenced the quality of peace in this country in a constructive manner. 

8.1 External third parties’ solutions: A temporary recipe for peace

The presence of significant accommodation policies can be both escalating and de-escalating, depending on the situation. Two examples of this were the mediation of the external third parties and the ‘ni-ni’ approach mentioned earlier. 

The multilateral mediation and negotiation initiatives led by external third parties in episode 2009 were criticised by many of the interviewees and, as explained earlier, their effects had many shortcomings. However, it is through these multiple initiatives and very lengthy processes (which took years), without mentioning the successive reneging from all sides, that functioning institutions could be established. The involvement of the external third parties, although severely criticised, was de-escalating. Indeed, given that the internal third parties completely lost their legitimacy and even the ability to convene the actors around a table, the external actors resorted to their commitment to democracy as well as the protection of human rights as guiding principles and thus supported the process for the establishment of a more peaceful situation.

For us, for me as US Ambassador, the most important thing was the constitution … it was the American policy … The only people who could convene them around a table and make them stay was us, the international community … at one point we thought there would be a civil war in Madagascar … What we discovered was that the civil society was very underdeveloped in Madagascar.[mfnMarquardt, Niels. Skype interview with the author in May 2016, Geneva.[/mfn]

The election of 2013 was a fruit of these initiatives and, most importantly, the product of the so-called ‘ni-ni’ approach. In December 2012, SADC suggested that neither of the leaders of the main conflicting parties (Ravalomanana and Rajoelina) should run for President. The rationale behind this was to avoid escalation. Given the persisting political instability during that period; the results could be contentious (Connolly 2013). In the view of the conflicting parties, it played an escalating role. The ni-ni was rejected by many.²⁰Randrianarisoa, Guy. Interview with the author in March 2016, Antananarivo. As it was the idea of the mediators, the approach was seen as imposed from the outside. Instead of being inclusive, the decision to discard the actors who were mostly linked to the episode of conflict is against the peaceful approach of inclusiveness, which is a general, positive premise in most conflict resolution processes. Despite that lack of inclusiveness, the decision brought to a certain extent a more stable situation as the election was a gateway to the setting-up of the new Republic and the restoration of Madagascar’s relations with the outside world, especially the donors. Indeed, after the military coup, the country was suspended from regional organisations, donors withdrew their support and many well-known personalities were sanctioned.

8.2 Political dimension as shared framing: A stabilising element

Finally, all the variants of escalation display that ‘paradigm discordance’ was at play. The table below contains all the types of framing in some of the recent episodes of conflict. 

Table 2: Mapping out of framings in situations of escalation

Episodes of conflict	Conflict dimensions	Framing of the conflicts: Actor A (the state)	Framing of the conflicts: Actor B (the opposition)
1975	Cultural dimension External dimension Political dimension	External dimension Cultural dimension	External dimension Cultural dimension
1985	Socio-economic dimension Political dimension	Political dimension	Socio-economic dimension
1996	Political dimension	Political dimension	Political dimension
2002	Cultural dimension Political dimension	Cultural dimension	Political dimension
2009	Socio-economic dimension External dimension Political dimension	Political dimension Socio-economic dimension Cultural dimension	Political dimension

When the actors interpret differently the origins of the episodes of conflict, such a situation plays an escalating role because the conflicting parties disagree over the issues that need to be resolved. Despite such discordance over time, during episode 2009 the actors did share the interpretation of the conflict as located in the political dimension. The possibility for collaboration and for striking a deal can be based on such shared framings, and the elitist and political aspects of the accommodation policies in episode 2009 can indeed be understood from this shared interpretation of the origins of the conflict. And most importantly, that shared understanding of the conflict in terms of its political dimension opened up the possibility of collaboration and negotiation.

9. Dynamics of proclivity to violence and conflict in perspective 

These different effects of the various factors in the causal configurations on the quality of peace demonstrate that, throughout the processes of escalation of levels of violence in Madagascar, peace co-exists alongside conflict, and that the two are not mutually exclusive, but co-construct each other. For instance, the escalating, negative/destructive effects of the high saliency of a given conflict dimension interact with the de-escalating effects of some of the accommodation policies. There are gradual transformations of the factors. Each factor  could become transformed in a positive/constructive or a negative/destructive way, and the pace at which a transformation takes place can differ. Note that the co-existence of peace and conflict has already been discussed by some scholars working on politics of non-violent actions or in security studies in general (for example through Sharp’s concept of ‘political Jiu-Jitsu’ 1973: 657‒703; also see the concept of grey zone in Brands 2016), but has not yet been pursued enough in peace or conflict studies. 

It is also important to reiterate that, in the case of Madagascar, throughout the variants of escalation, if one looks at the result of the QCA tests there were more factors in the configurations having a negative rather than a positive effect on the quality of peace. However, despite this situation, Madagascar succeeded in keeping a certain semblance of peace. This shows that it is not only a matter of escalation built upon unresolved grievances (as is generally argued in much of the literature in conflict studies) but also of de-escalation built on factors that have the capacity to affect the quality of peace in a positive way. The existence of the factors favouring peace in the configurations can explain the fact that the intensity of most of the episodes of conflict in this country did not reach level 5. The exceptions were the first two: episodes 1947 and 1971. Here there was no escalation as the level of conflict remained the same. And episode 1947 did not involve the state. 

This research allowed the author to work extensively on accommodation policies and the involvement of external third parties in processes of escalation of conflict levels. In episode 2009, for example, without the intervention of external actors, the conflicting parties would not have agreed to gather round the table. These mediation and negotiation initiativessucceeded in creating functioning institutions affecting positively the quality of peace. That being said, why is it that situations of escalation occurred and that there is a significant proclivity towards violence and conflict? 

The nature of accommodation policies themselves had a great influence on those dynamics. When examining the peace agreements and national reconciliation, it becomes clear that liberal peace was the main element guiding the shaping and implementation of these initiatives. Unfortunately, the democratic principles and the spirit of inclusion that were presentthroughout these processes created institutions that have been establishing negative peace and affecting the nature of peace in a negative manner. Negative peace, because some of them were not able to address many of the root causes of the conflict: conflict dimensions (cultural and socio-economic), the relationship between the actors (the perception of the self and the other), and even the framings of the conflict. The result is a semblance of peace. 

Most importantly, liberal peace is elitist and, indirectly, it promotes exclusion when the choice of the tools to regulate conflicts lies solely in the hands of the powerful elite. The elitist and political agreements signed to resolve the Malagasy episodes of conflict, as well as the ni-ni approach, are evidence of that. Even the initiatives implemented within the framework of national reconciliation, which were supposed to be bottom-up initiatives, did not gain local ownership as many still perceived them to be imposed by external third parties. In addition, it is important not to forget that, on the whole, the involvement of external third parties in the process was rejected by many local actors. Therefore, this research confirms the hypotheses of Nepstad who calls for more research into the fact that in some initiatives international actors can cause harm (2011:135).

10. Conclusion   

The key findings in this article point towards an under-researched nature of proclivity towards peace or violencein Madagascar. Despite the constructive transformations at play in the dynamics explored in this research, liberal peace has been playing a significant role in creating negative peace. It has reinforced and accentuated symbolic and structural violence which in turn builds up the conflict-readiness of the parties who are preparing for conflict at low, medium or high levels of violence. 

Despite the accommodation initiatives implementedwith the principles behind liberal peace, which the author recognises has initially had noble and peaceful goals, the Malagasy still face the risk of conflict recurrence. There have been no specific mechanisms addressing the adverse effects of interventionist and autocratic liberal peace. The actors blindly adhere to its principles without considering whether the local population at grassroots level want this model and would appropriate, recognise, and own such a solution. The condition in which a majority of the Malagasy live in constant exposure to structural violence creates a context which incitesthem to be conflict-ready and engage in various types of violence, including direct physical violence, even if they do not ‘rationally’ intend to do so.

The co-existence of peace and conflict discussed in this article, and the specific focus on the proclivity towards violence and conflicts, show the intricacy of building a long-lasting and sustainable peace. There is no one-size-fits-all solution to address the origins of the problems regarding, for instance, conflict dimensions, the framing of conflicts, the repertoires of actions, and the negative or destructive effects of the accommodation policies. This means that, in the case of Madagascar and other countries, it is necessary to come up with solutions to the problems concerned and to involve various actors in the process. 

Most importantly, it is necessary to give more power to the local people and allow them to do bottom-up work building peace with the collaboration, but less intervention, of the political elite. The timing of the implementation of these activities is important and must not be linear (e.g. conflict management, mediation, peace deal, and peacebuilding). The actors resolving the conflicts must be aware that the consequence of their solutions can create a situation where peace and conflict interact and co-construct each other. This requires a thorough understanding of the local context, the causes of the conflicts, andthe local infrastructures of peace. 

These conclusions lead to the necessity of identifying further opportunities of seeking to understand the conflict in Madagascar and beyond. A first limitation of this article was that it did not have enough space to discuss in depth the variations of structural factors and the methods of the actors that would differentiate the different types of proclivity towards violence and conflicts. Doing this would yield more accurate modelling of conflict prediction that would allow the identification of better and more responsive conflict resolution mechanisms capable of addressing the multiple elements of a given configuration. Second, although this article took into account how liberal peace works in Madagascar, it could not undertake an in-depth study of its coexistence with local methods and actors, for instance, indigenous peacebuilding and traditional third parties. Thus, further research should be done on how the interaction between these two approaches affects the proclivity towards violence and conflict. And third, research in a larger comparative design, for example, cross-national studies on the dynamics behind such a proclivity in recurring conflicts, should yield richer results.

Sources

• Bachelard, Jerome and Richard Marcus 2012. Madagascar. In: Dizard, Jake, Christopher Walker and Vanessa Tucker eds. Countries at the crossroads 2011: An analysis of democratic governance. Lanham, MD, Rowman and Littlefield. pp. 363‒382. 
• Blum, Françoise 2011. Madagascar 1972: L’Autre indépendance : Une révolution contre les accords de Coopération (Madagascar 1972: The other independence: A revolution against the cooperation treaties). Le Mouvement Social, 236 (3), pp. 61‒87.
• Brands, Hal 2016. Paradoxes of the gray zone. Available from: <http://dx.doi.org/10.2139/SSRN 2737593> [Accessed 16 September 2019].
• Cawthra, Gavin 2010. The Role of SADC in managing political crisis and conflict. Friedrich Ebert Stiftung: Peace and Security Series. Maputo. Available from: <https://library.fes.de/pdf-files/bueros/mosambik/07874.pdf> [Accessed 20 April 2019].
• Connolly, Lesley 2013. The troubled road to peace: Reflections on the complexities of resolving the political impasse in Madagascar. Available from: <https://accord.org.za/publication/troubled-road-peace/> [Accessed 10 August 2019].
• Express de Madagascar 2002a. Lettre Ouverte à Mr Didier Ratsiraka (Open Letter to Mr Didier Ratsiraka). By ‘Andrianjaka R.J.M. Université d’Antananarivo’. Express de Madagascar, 4 February, p. 16.
• Express de Madagascar 2002b. De retour Lundi Soir à Toamasina. Le Président Ratsiraka fait savoir que « la Meilleure Défense c’est l’Attaque » (Back to Toamasina on Monday evening. President Ratsiraka advised that ‘Attack is the best defence’). Express de Madagascar, 12 June, p. 8. 
• Féron, Élise 2002. Irlande du nord : La paix en haut, la guerre en bas ? (Northern Ireland: peace at the top, war at the bottom?). Critique Internationale, (3), pp. 24‒30. 
• Féron, Élise 2005. Community conflicts in Europe: A review of the literature. PEACE-COM. Unpublished manuscript. Available from: <https://cdn.uclouvain.be/public/Exports%20reddot/spri/documents/Deliverable_1_PEACE-COM.pdf> [Accessed 15 March 2018]. 
• Nepstad, Sharon E. 2011. Nonviolent revolutions: Civil resistance in the late 20th century. New York, Oxford University Press.
• Ralambomahay, Toavina 2011. Madagascar dans une crise interminable (Madagascar in a never-ending crisis). Paris, L’Harmattan.
• Ratsimbaharison, Adrien M. 2017. From the failure to consolidate democracy to the recurrence of violent political conflicts: The case of Madagascar under the Third Republic (1992‒2009). Available from: <http://dx.doi.org/10.2139/ssrn.2915621> [Accessed 12 September 2019].
• Razakamaharavo, Velomahanina T. 2018. Processes of conflict de-escalation in Madagascar (1947–1996). Peace and Conflict: Journal of Peace Psychology, 24 (1), pp. 54‒63.
• Rihoux, B. 2007. The key determinants of community conflict escalation and de-escalation in Europe: A systematic comparative analysis (QCA). PEACE-COM (Peace processes in community conflicts: From understanding the roots of conflict to conflict resolution). Unpublished report prepared for the European Commission. Available from: <https://cdn.uclouvain.be/public/Exports%20reddot/spri/documents/Deliverable_8_PEACE-COM.pdf> [Accessed 15 March 2018].
• Rihoux, Benoît and Charles Ragin eds. 2008. Configurational comparative methods: Qualitative Comparative Analysis (CSQCA) and related techniques. ThousandOaks, Sage.
• Sharp, Gene 1973. The politics of nonviolent action. 3 vols. Boston, Porter Sargent.
• Vetik, Raivo 2007. Testing a semiotic model of ethnic conflict: The case of Estonia. Paper presented at the European Consortium for Political Research Joint Sessions, 7‒12 May, in Helsinki, Finland.