Civil Wars: Escalation and De-Escalation

ABSTRACT

When we study civil wars and conflicts we tend to conceptualise them as occurring in stages: starting from domestic political disagreements, to demonstrations and protests escalating into violence and war. How armed conflicts end is often seen as the reverse process, moving from high intensity armed interaction, to a drawing down, war weariness, negotiations and termination, followed by a transition to peace. This contribution argues that this is a faulty understanding of conflict, which obscures rather than illuminates. More attention to the processes of aggravation of conflict and the many leaps and bounds of the use of pressure and coercion is warranted. Similarly, the drawing down of conflict is not necessarily linked to a linear progression of de-escalation. Sometimes armed conflicts end at the pinnacle of violence. This article focuses on the state of the art in the field of escalation and de-escalation in the study of civil war and conflict. Moreover, it will offer an invitation to scholars to focus more on these phenomena by outlining where our present knowledge and insights fall short.

Introduction

Mobilisation of the population and expressions of discontent have been core concerns for scholars who study contentious politics and conflict as dynamic processes. The field of Civil War Studies has, to date, had a strong focus on violence and its explanations, together with a preoccupation on how to address this via ‘liberal peace’, peacekeeping, and peace negotiations. A second major preoccupation has been looking at peace and its parameters, peace processes and transitional justice (during the final phase of armed conflict or after the signature of peace accords). Moreover, very valuable insights have been gained regarding the evolution of a conflict across time, giving witness to the many different trajectories that conflicts can experience.

However, the field has been less strong in its treatment of the concurrence of civil war both preceding and in simultaneity with other expressions of conflict (i.e., protests), which complicate conventional pictures of what civil war constitutes. Many scholars operate as if we could just focus on one expression of conflict, as this would be sufficient to present conclusive evidence in relation to such larger social processes. However, the fact is that different expressions of conflict coexist, ‘violent’ protests, ‘peaceful’ protests, riots, terrorism and insurgency, conventional war, all can and do occur simultaneously (Tschantret 2023) and are central to accounts of the emergence, conduct and demise of civil wars. This point echoes earlier claims in the debate that the binary distinction between what is war and what is peace is largely artificial in practice (e.g., Richards 2005, Barkawi 2016).

In this contribution, we focus on the way the field has studied the processes of escalation and de-escalation of conflict. As a phenomenon, civil war has never held the monopoly on how societies express their discontent but neither did the phenomenon emerge out of thin air. What we are interested in is the ‘missing middle’ of conflict studies. The field has tended to ignore the expressions of conflict that take place between war and peace and between peace and war, as well as the simultaneity of more and less violent expressions of discontent in many places around the world. Moreover, how do processes of escalation and de-escalation take place? Whereas research about civil war often mentions processes of escalation or de-escalation of conflict (implicitly or explicitly), an explicit conceptualisation of the mechanisms and factors explaining such processes is often missing, either in relation to a particular expression of conflict or relating to different expressions of conflict such as ‘war’ and ‘peace’ (Duyvesteyn 2021, Díaz Pabón 2022).

We argue that more intellectual and theoretical rigour is necessary and a distinct and structured approach to the study of escalation, and de-escalation of conflict would be welcome. In our discussion below, we find that the study of processes of escalation or de-escalation of conflict contributes to solving three interrelated challenges within the field of Civil War Studies that need to be addressed. Firstly, we tend to study conflict by looking at one particular expression of contention looking at one particular expression of contention, for example protests, armed clashes or terrorism; this is a form of conceptual autarky. Such an approach impedes our capacity to understand the causation and spillover effects between different expressions of conflict, leading us to misunderstand them. Secondly, there is a tendency to study conflict mostly as a binary, where either a particular category is observed or not (i.e., peaceful protests vs. violent protests). We see these as symptoms of the same set of underlying political problems. Finally, as we understand conflict as binary, this can partially explain the linearity in the assumption of how conflicts operate, leading us to try to fit ‘square pegs in round holes’. We argue that it is highly problematic to use a linear conception of contention when looking at the rich and diverse trajectories of conflict in history. These challenges oversimplify our conceptualisation of complex societal processes, in which for example, the absence of war should naturally lead to its ‘opposite’ - ‘peace’ (Duyvesteyn 2021).

Such a critique is not entirely new nor an absolute statement about the field. Previous literature on conflict and social revolutions has understood conflict as existing along a continuum (Lenin 1968), connecting rebellion and revolution, partially allowing for the framing of processes of contestation as including the description of processes of escalation and de-escalation of conflict.

Similar approaches have been used by other researchers looking at conflict beyond Marxist frameworks, as illustrated by how grievances and deprivation have been used to explain the emergence of armed conflicts (Gurr 2011); the occurrence of violent processes associated with revolutions in France, Russia, and China (Skocpol 1979); processes of collective action that can lead to violence (Olson 2009); the relation of conflict and its dynamics in relation to parallel processes of collective action (Tarrow 1993, McAdam et al. 2001), or the repertoires of discontent and their expression in different societies (Tilly 2006).

More recent scholarship has explored the emergence and effectiveness of non-violent uprisings vis a vis other expressions of conflict, such as violent resistance (Chenoweth and Ulfelder 2017, Chenoweth 2020); the role of inequalities and state strength in explaining the different stages of armed conflict (Bartusevičius and Gleditsch 2019); the factors that condition the escalation or de-escalation of genocidal violence (Chenoweth and Perkoski 2019); and the causal mechanisms connecting different stages of mass mobilisation (Shultziner and Goldberg 2019). With the recent emergence of more sophisticated datasets that account for different expressions of conflict, we could potentially account for the escalation or de-escalation between particular expressions of conflict. This could provide us with the empirical resources for exploring the relationship between different expressions of conflict and for the analysis of processes of escalation or de-escalation of conflict associated with civil war. This possibility depends on the existence of conceptual frameworks that enable the analysis of processes and the complexities of the processes of escalation and de-escalation of conflict.

However, and despite such approaches and possibilities, some recent strands of the literature tend to approach armed conflict as divorced from other expressions of conflict, such as mass mobilisation (Bodea and Elbadawi 2007, p. 3), so that

[w]hen we study […] war or genocide, or revolution, we assume that there is a dependent variable so to speak, whose value is either the presence or the absence of that phenomenon. You have war or you have peace. […] But in fact, […] these categories are connected on some sort of continuum (Kalyvas and Straus 2020, p. 9).

This point has also been raised by Bosi et al. (2019, p. 133), and by Tarrow (2021) calling for an understanding of common mechanisms across different expressions of conflict, or a better description of the mechanisms associated with the escalation or de-escalation of conflict (Duyvesteyn 2021).

We should heed these calls. Looking at conflict and its expressions in a relational manner, while examining examining at the relationship between different expressions of conflict, might help us to better understand past civil wars, and future internal conflicts, which are likely to increase due to the contention between global and local hegemons that will emerge. Also, it is likely to contribute to the understanding of conflict after peace negotiations, peace agreements, and political transitions, a topic that remains under-researched (Dudouet 2013, Wieviorka 2016, Della Porta et al. 2017, Duyvesteyn 2021).

Before we proceed to illustrate the structure of this paper, we would like to state our understanding of conflict. Whereas researchers tend to use the terms conflict and civil war interchangeably, following the definition by Bartusevičius & Gleditsch of civil conflict, conflict¹ here is understood as:

[…] an incompatibility over [power] and/or territory between two or more politically organized actors, […] that takes place primarily within the borders of one state and involves [institutional] and extra-institutional means of contention (Bartusevičius and Gleditsch 2019, p. 229).

This conceptualisation of conflict works as an umbrella definition, and enables us to frame the analysis of processes of escalation and de-escalation and also allows us to conceptualise the relationship between different expressions of conflict (i.e., how protests relate to the larger issues of contention), and of armed groups.²

In addition, it is important to define what we understand as an escalation and de-escalation of conflict. We propose to delimit escalation and de-escalation as follows: ‘The crossing of a threshold in [a conflict] that is empirically observable’ (Duyvesteyn 2021, p. 31). We view escalation in a broad manner. Traditionally, escalation was focused on two main parameters, horizontal and vertical escalations. Horizontally, conflict could escalate by involving increasingly more territory. Vertically, it could be aggravated by the employment of more or heavier weapons (Kahn 2012, pp. 4–6). This, as has been argued elsewhere, is too narrow because conflicts can escalate in at least seven dimensions: in the number of actors, the nature of the demands, the tactics and the means used, the targets selected, the extension of the geographical area where a conflict is being manifested and an extension in time (Duyvesteyn 2021, p. 39). De-escalation can be conceptualised as the opposite, as a lessening of conflict, similarly along these seven dimensions.

In this paper, we focus on the way the field has studied the processes of escalation and de-escalation of conflict. To do so, this paper first undertakes a big picture examination of the literature to map the insights of the research associated with processes of escalation of conflict. In section three, we proceed to analyse the under-researched processes of de-escalation of conflict. Then we proceed to critically evaluate this material and identify how we could further enable the study of processes of escalation and de-escalation in the context of conflict.

The Escalation of Conflict

Two substantial bodies of the literature are relevant when investigating escalation. First, social movement theorists, conflict studies scholars and those looking into protest policing have developed a very vibrant research line. This group of scholars largely subscribes to a distinct but narrow approach to escalation, by limiting escalation during a phase of protests, rather than within war. A second relevant body of work derives from the period of the Cold War and is based on work from scholars looking into nuclear escalation and theorisation about nuclear war. They have operated based on different understandings of escalation, namely escalation leading to war and within war.

Scholarship on social movements and protest policing has noted the importance of the relationship between mass mobilisation and armed conflict in processes of escalation. Recently, several researchers studying conflict, contestation³ and collective action have looked at the escalation of armed conflict from protests in different conflicts. Some examples are the Syrian armed conflict (Della Porta et al. 2017), the protests in Libya before the emergence of armed conflict (Paoletti 2011), and the mobilisations related to the Arab Spring. Moreover, scholars have studied the increase in protests and mobilisations across different countries (Chenoweth 2020); paying attention to the role of repression fuelling armed conflict (Blair and Sambanis 2020); the role of repression in protest escalation (Ellefsen 2021); and, more recently, the increase in social tensions caused across the world by the COVID-19 pandemic, and the collateral effects of the Russian invasion of Ukraine.

As they have noted, the boundaries between these different expressions of conflict are very fluid (Dudouet 2013, pp. 402–404). Moreover, this research has found preliminary evidence that supports the existence of a relationship between these different expressions of conflict. Still, there remains a conceptual and analytical challenge; ‘[d]ifferent forms of political violence are interlinked and are part of a continuum of repertoires of actions – rather than representing discrete and mutually exclusive types – and often [occur] successively or simultaneously during processes of conflict …’ (Bosi et al. 2019, p. 133). This has made the analysis of such processes more difficult, as it is assumed that processes of escalation follow a clean sequence in which different expressions of conflict substitute for preceding expressions of conflict in a step-by-step process.

A series of notable features stand out in this treatment of escalation. First, in such framing it is common to encounter narratives that speak of the emergence of armed conflict after protests, as if protests melt away as soon as armed conflict starts. It is based on an understanding that escalation is a linear process. Moreover, we argue that escalation emerges from a dynamic process of interaction, whereby the original intentions might unwillingly and unwittingly contribute to reaching increased levels of violence given the particular contexts, responsiveness of political systems and the opportunities for undertaking other forms of collective action. In particular, the role of the state and its approach to use the police and policing and repression contributes to aggravation of conflict, leading to armed conflict. Notably, for our discussion here, the process of escalation is limited to the phase of unrest, non-violent and violent protest but stops when armed conflict breaks out. It is rather odd that escalation towards violence is separated from violence within the context of civil war.⁴ This appears to be a totally artificial separation.

Second, the main explanation for escalation is in essence related to the counter-measures that the state and, in particular, the use of the police or repression mean to the commitment and the willingness to persist or escalate tensions by the contesting actors. Other potential escalatory factors including choices made internal to the groups, involvement of others, such as third parties, or changing contexts are largely left out of the equation. We will return to this issue below.

Apart from this discussion about escalation in the context of social movement radicalisation, a second major debate took place, which in a chronological perspective actually preceded the first. After the detonation of the atomic bombs in Hiroshima and Nagasaki in August 1945, a watershed occurred in the study of war and its associated phenomena. There was a recognition that with the advent of these powerful weapons a new era had begun. Theorisation of armed conflict took flight because the actual use of these weapons was too dangerous to contemplate. This created room for scholars with a background in mathematics and economics to use their modelling techniques to prepare for a Cold War. These Cold War scholars devoted substantive attention to the processes of escalation and the means and brakes to control them. Herman Kahn, Thomas Schelling and Richard Smoke are the most notable contributors to this debate (Holsti 1972, Smoke 1977, Schelling 2008, Kahn 2012).

They shared the view that escalation could be theorised as a logically stratified process in which rational calculations come into play. The commonly used image, originally offered by Herman Kahn of a ladder, is illustrative of this way of thinking. Kahn’s ladder had 44 rungs leading from crisis to a war of annihilation in which nuclear weapons were used in first and second strikes leading to the extermination of life on earth (Kahn 2012). Not only was the conceptualisation of escalation linear but also linked to the dominant approach of rational actor thinking. Cost–benefit calculations, not least informed by mathematical modelling, would inform behaviour to pursue escalation or de-escalation. The project of rational calculation would also include the opponent. The main challenge would be to outwit this opponent by being one step ahead, command more or stronger weapons and possess political will and commitment in order to acquire escalation dominance. Already at the time, critics questioned this stratified type of thinking (Wohlstetter and Wohlstetter 1965, Snyder 1977).

Similar to the contentious politics discussion, escalation is seen as linear and relational, based on a responding and active opponent. Where these discussions diverge is in the role of rational actor thinking. The contentious politics scholarship sees escalation as a product of interaction and path-dependence, with a dominant role for state responses and in particular protest policing. The Cold War scholars based their theorisation mostly on purely rational cost–benefit calculations, even though at a later stage in the discussion, the role of culture and strategic culture featured more prominently. Even though they focused on interstate rivalry, this scholarship has been very influential both based on their ontological and epistemological premises and in the subsequent theorisation. Via the ‘liberal peace’ paradigm, the interstate study of conflict transferred their insights and assumptions on how wars start and end and their assumed inherent linearity into the civil wars literature, leaving aside evidence that challenges this 'rationality' framework such as, for instance, the football war between El Salvador and Honduras in 1969.⁵ In short, these two main bodies of scholarship share the preoccupation with escalation as a linear process, contained in time and place with a large emphasis on the interactive aspects. Escalation, in these literatures, is a step-by-step process and its logic is derived from responsiveness to what the perceived opponent is doing or about to do. In comparison, the debate about de-escalation is a lot less developed.

The De-Escalation of Conflict

While there has been an incipient interest in the escalation within and between expressions of conflict and their critical junctures, there is no clear focus in any significant manner on the topic of de-escalation within and between different expressions of conflict in the armed conflict and Civil War Studies literature. Work on de-escalation within and away from armed conflict is sparse and lacks a solid conceptual framing. This in itself is a rather surprising reality, not only in the light of the prevalence of the use of the term de-escalation but also for its significance in discussions for those working on international efforts to end conflict. Studies on de-escalation are few and far between.

Often de-escalation is equated with armed conflict termination. However, there are many examples of armed conflicts ending when the violence is at a height. The deadliest bomb attack in the Northern Ireland conflict occurred in Omagh, when the peace negotiations leading to the Good Friday agreement in 1998 were underway. The civil war in Sri Lanka ended in 2009, with a conventional defeat of the Tamil Tigers (Diaz and Murshed 2013).

Another notable feature of the de-escalation discussion is some large categorical but unsustainable claims. In a RAND study, it is claimed that de-escalation subscribes to the same linear and wilful features as escalation and it is claimed that de-escalation is always deliberate and ‘accidental de-escalation is essentially unheard of’ (Morgan et al. 2008 p. 34). Another author similarly claims that clear linear patterns are observable in, notably, terrorist campaigns, leading to a situation in which ‘a consensus has emerged that most terrorist organisations conform to a pattern of a) an initial embrace of violence, b) an escalation period, and finally c) a period of decline’ (Becker 2017, p. 2). There is scant attention to shifts between violent and non-violent collective action in social movements (Schock and Demetriou 2018), or processes of de-escalation within the existence of armed conflict. This state of affairs leaves us rather empty handed.

Part of the challenge in understanding de-escalation is due to our erroneous assumptions about how non-violence or the absence of armed conflict are disjointed from armed conflicts (Germann and Sambanis 2021), or that mobilisation only takes place in violent form (Olson 2009, Bartusevičius and van Leeuwen 2022). The reality is that these assumptions condition our misunderstanding of what a conflict is and what it is not. As per our definition (see section above) - as the incompatibility of issues related to power and territory involving at least two politically organised actors operating largely within the confines of state borders and who use extraordinary means to express discontent - these different expressions of conflict (violent or non- violent) are thus a feature of all conflicts. Such assumptions have been challenged in the field by the non-violent resistance literature, which has positioned non-violence as part of conflicts (Chenoweth et al. 2022). This also forms the crux of the Clausewitzian approach to understanding war, as at its core a political contest of wills (1993). We reject the conceptualisation of irregular conflict or terrorism as being somehow ‘more political’ than other tactics or warfare practices. It is the core business of the strategist ‘to deploy the means they do in an attempt to attain their ends within the constraints of the environment in which they find themselves […] This, as Clausewitz intimated, is the most important strategic question of all’ and this puzzle is political at its very core (Smith 2003, p. 37, Biddle 2021).

Towards a Better Understanding of the Processes of Escalation and De-Escalation of Conflict

The challenges associated with the study of processes of escalation and de-escalation relate in essence to the following. First, the challenge is to account for the dynamic of escalation within and between different expressions of conflict and moreover to account for changes whose outcomes we cannot forecast ex-ante. The second challenge is the temptation to discuss conflict as a linear phenomenon.

Time and again the field has run into this insight without significantly tackling it. Three short examples illustrate this trend; in the 1990s, there was a lively research line focused on economic explanations for armed conflict. In no small part infused by the interests of major financial institutions, Paul Collier and others made a lot of the statistical correlation between the presence of natural resources and the risk of civil war (Collier and Hoeffler 2005). While only a correlation, detailed case studies later showed that the causal pathways turned out to be a lot more complex than the ‘resource curse’ literature would claim (Cramer 1997, Sambanis 2004). In the study of terrorism, a lot of time and energy has been devoted to finding out what the ‘root causes’ of this phenomenon could be (Bjorgo 2004). In no small measure influenced by 9/11, scholars had to eventually conclude that the pathways from deeper causes into violence were contingent on so many factors that this research line was eventually abandoned (e.g., Newman 2006). Closely related, investigations into individual radicalisation trajectories were all the rage not so long ago (McCauley and Moskalenko 2008). Not surprisingly, these are again often highly contingent (Neumann and Kleinmann 2013). A problematic aspect to this last discussion is that the shaky research provided input into a de-radicalisation industry involving large sums of money spent by governments on de-radicalisation programmes. In short, while linearity is intellectually highly appealing and immensely popular (Van Riper 2014, p. 6), it is also misleading. We propose that our categories will inevitably entail biases and might not possess universal validity. What we see and do not see in processes of escalation or de-escalation is fundamentally linked to the concepts and measurements we use. We will now dive deeper into possible pathways to engage with these challenging features in the study of processes of escalation and de-escalation of conflict.

Centring the Relational Nature of Different Expressions of Conflict

A starting point for such a discussion would be to separate the violence from the war (Kalyvas 2006, p. 20). This means that violence as a phenomenon infuses the discussion about conflict, but the interactions between the antagonists entail a lot more (Gutiérrez-Sanín and Wood 2017). These other aspects require more attention in order to appreciate what is going on. The conceptualisation of escalation as consisting of seven dimensions, not all involving violence, underlines this.

In the real world, we observe variations within expressions of conflict taking place in a given country (i.e., when the number of protests increases or decreases), or between (i.e., when protests are substituted or superseded by armed violence), as well as the coexistence between different expressions of conflict (Tarrow 1993, Chenoweth and Shay 2022). Not only do expressions of conflict change but we can also observe that specific organisations (police, armed forces and contesting organisations) and the demands from these organisations in different contexts change (Kang 2023). In addition, the organisational fluidity and the different repertoires organisations might choose can change as well - they respond to contextual and historical factors.

For example, in Colombia in the 1990s there was an expansion of the strength of the different armed groups, which led to an escalation of armed conflict and other expressions of violence (i.e., forced internal displacement and massacres). However, as this violence was increasing, the number of protests demanding peace negotiations and an end to armed violence increased as well (Durán 2006).

While the peace negotiations took place between 1998 and 2002, violence continued, as some organisations proved very fluid. Before this period paramilitary groups operated as a series of independent organisations, after 1997 paramilitary forces emerged as a federal organisation in the Autodefensas Unidas de Colombia (AUC) (Bejarano and Pizarro 2010, López 2010).

Building on the fine-grained historical approach, a multiplicity of escalation trajectories has been brought forward by Duyvesteyn (2021). Instead of the emphasis on geographical spread and the influx of more and heavier weapons, as the Cold War scholarship discussed above emphasised, escalation relates more often to the actors and their views and experiences as well as their ideas and discourses about the issues of contention, the core political dispute. Shifts in extremity, i.e., adopting more extreme positions as a result of internal group dynamics, as well as situational entrapment, spoiling behaviour, outbidding and divide and rule politics, caused escalation in real-world cases.

For example, in South Africa, the decision of the ANC to resort to armed violence and undertake armed conflict was made considering the repression of the state. It was also informed by the concern that if the ANC failed to lead in the opposition to the apartheid regime, it would lose political relevance. As other armed organisations such as the National Liberation Committee (NLC) in 1960, exemplified the fact that armed conflict was being considered by various organisations across the country (Díaz Pabón 2022, p. 176).

These are all related to group processes and have a stronger link to the conflict environment, rather than the perceived opponent. Escalation in a relational perspective brings forward issues of conflict saliency and commitment. Escalation can be caused by a changing political opportunity structure, counter-measures and considerations of organisational survival. These processes are highly dependent on the context and the specifics of the particular conflict. Adopting positions of increasing extremity and reinforcement of conflict saliency and commitment form the most significant causal pathways to conflict aggravation. Such processes are not immune to the role of agency, as the type of leadership structures, leaders, and the nature of the organisations can also influence processes of escalation and de-escalation of conflict. These insights form an invitation to further dissect the dynamic nature of the interaction between political opponents and dig deeper in order to understand escalation.

Leaving Aside Determinism: Making Unpredictability Explicit

Conflict is inherently unpredictable and does not observe strict sequences and linearity;

Although belligerents in [conflicts] generally act according to their objectives, they do so in a complex, multilateral, and interactive environment pervaded by uncontrollable external political dynamics and chance occurrences. [Conflicts are] shaped by not only the reasoned objectives and decisions of each actor, but also the complex nature of the political dynamics that their interaction produces, which are inherently unpredictable and irrational (Waldman 2010, p. 6).

Going back to Carl von Clausewitz, and in stark contrast to the stratified escalation ladder, prevalent in the literature discussed above, linearity is absent in war. ’[E]very war is inherently a nonlinear phenomenon, the conduct of which changes its character in ways that cannot be analytically predicted’. (Beyerchen 1992, p. 62) The presence of linearity as a framework relates to our conceptual frameworks, but also to the fact that small causes can have large and amplified effects that we cannot foresee (Beyerchen 1992, p. 78).⁶

When looking at de-escalation, non-linearity unintentionally comes into even starker contrast. Instead of a linear reversal of escalation and as a deliberate process, mistakes, accidents, the loss of sanctuary or fighters and legitimacy form the key ingredients in understanding the process. De-escalation is often unintentional and in practice messy and chaotic (Duyvesteyn 2021).

As we tend to conceptualise the processes of conflict de-escalation in a linear manner, we might be missing a description that highlights the role of chance, opportunity and ‘luck’. Cases such as in Colombia, where several armed conflicts took place at the same time, where the leader of the M-19 guerrillas (one of the groups that had signed peace accords) and two other presidential candidates were assassinated in quick succession in the early 1990s. The case of South Africa in the early 1990s illustrates this unpredictability as armed white nationalists broke into the venue where the peace negotiations were taking place while political violence was increasing (i.e., the assassination of Chris Hani - General Secretary of the South African Communist Party) (Douek 2020). Both cases are not exceptions and in fact illustrate how transitional periods away from armed violence are extremely uncertain and observe greater instability than what we usually account for in our descriptions (Díaz Pabón 2022).

Yet, there are two remaining inescapable challenges in embracing the complexity of escalation and de-escalation: working with more fine-grained understandings of the expressions of conflict could allow for a greater richness in the description of processes taking place, but would make the comparability of such processes difficult across different contexts and hamper attempts at theory formulation. Also, no single research undertaking will be able to account for all the expressions, actors and factors interacting in a given context. The diversity of contexts and conflicts transcends the capacity to account for and describes all-existing details associated with expressions of conflict. Until such time that we could potentially remedy this challenge, we have to account for this limitation. Presently, we need to compromise between embracing the variability between actors, contexts, expressions of conflict and time, and being able to conceptualise findings based on this richness so we can understand the evolutions of conflict over time in a traceable way.

However, we propose reconsidering our use of language and frameworks and looking at history as potential interconnected pathways for embracing these challenges as a way to challenge the conceptual autarkies, binaries and the assumption of causal linearity. We are proposing to take into account in a more profound manner the evidence that has been neglected in many accounts of history. History does not only present a framework to describe continuity and change, it also includes a use of language that is not bounded by the idioms of social movements, civil war and war literatures. As historical accounts do not have any disciplinary leaning to specific forms of conflict, they also account for conflicts without having disciplinary preferences about particular expressions of conflict (i.e., non-violent protests vs. civil war). It, therefore, has greater ontological range to account for the fullness of histories of conflicts.

Looking at the Conflict Studies and Civil Wars Literature as a Way Forward

We can and do need to change the way in which we frame and describe such processes. The language associated with processes of escalation and de-escalation of conflict has been a distinct one. Different terms have become prevalent in the field, such as ‘onset’, ‘outburst’, ‘emergence’ and ‘occurrence’. The field’s use of the term ‘conflict events’, instead of conflict processes and conflict dynamics leads to a discrete description of such processes. If we are to understand processes of escalation and de-escalation of conflict, we need to adjust our language so we speak of and emphasise processes, progressions, changes, and variations. Such use of the lexicon will better enable us to speak of and understand the de-escalation and escalation processes in different conflicts.

One possible lens to frame our understanding of conflict and the processes of de-escalation is looking at the conflict transformation literature (Lederach 1996, Galtung 1999). Within this literature, there is no assumption that conflicts cease to exist, and rather reflect on how conflicts are transformed across time. Such a lens could allow us to describe the conditions under which de-escalations of conflict take place within an expression of conflict (i.e., as when the intensity of an armed conflict declines) as well as a de-escalation of a conflict (i.e., in the case where armed violence is substituted for protests). This approach calls our attention to the perennial nature of conflict.

Not only do we argue for a different use of language and of our concepts associated with conflicts but to challenge the conflict binary by looking at different trajectories and sequences. For example, looking at the end of armed conflict, we need to be able to describe different paths to the end of armed conflict, one of them the role of military victories and appeasement (Luttwak 1999, Diaz and Murshed 2013). Another path would be studying successful peace processes and peace negotiations that end the violence of conflicts to unveil the different sequences that take place before the end of armed conflict. In some cases, we observe that the end of armed violence follows the pinnacle of violence and the reaching of a mutually hurting stalemate (Zartman 2000), whereas in other cases peace negotiations take place and successfully end armed conflicts due to the involvement of internal or external actors.

We argue for a reconsideration of the use of language, our concepts and prevailing binaries. Moreover, we need to be able to describe the different combinations of factors influencing these forms of de-escalation and their multiple causal paths and different outcomes in declines of violence (of which we know little) and in the end of armed conflict (of which we know more) (Duyvesteyn 2021). For the first – declines in conflict – we could study in more detail the history of ceasefires that attempt to bring an end to the conflict altogether (Clayton et al. 2022). This would also mean considering the analysis of detailed case studies associated with peace negotiations that could observe decreases in violence within ongoing armed conflicts as illustrated by Colombia (Karl 2018) or the evolution of protest movements and the decline in protests.

Such analysis should be cognisant of the risk of deterministic causation accounts of history, and the dominance of perceived linear connections between a particular set of factors and the tendency to describe the evolution of a conflict. To avoid this tendency, we propose to speak more probabilistically of social processes. We argue that using a Bayesian language for describing social processes, as used in process tracing analysis (Bennett and Checkel 2015) can serve us better to account for the complex realities and make our research more transparent. This can help us to better communicate our findings and importantly make explicit the uncertainty associated with our research methodologies and their findings (Breznau et al. 2022). We usually describe conflicts using deterministic statements stating that a particular variable makes conflict more likely (i.e., the presence of large natural resources cause civil war), but it would be more accurate to describe this as how factors are associated with the likelihood or the odds of a particular expression of conflict taking place (i.e., natural resources are associated with a higher probability of civil war taking place in a particular historical context). Stating the factors associated with the occurrence of different expressions of conflict as deterministic, forgetting that measurement, statistical analysis, and interpretation, ignores the significant margins of error – that are not negligible, and are inherent to this type of research. Acknowledging this uncertainty in our research will make our research more transparent and will enable greater validation and evaluation, as opposed to a language that could resemble one of the competing versions.

All this does not mean we should abandon theory as a way of understanding. It does, however, mean we should consider awarding greater importance to a historical understanding of conflicts and their complexities. Both the escalation and de-escalation of conflicts are path dependent, unique to their time and place, and such dependency can only be understood via a serious account of the history of conflicts. History and thick description can help us to account for subjective conditions, perceived and real capabilities and how they condition processes of escalation and de-escalation of conflict (Duyvesteyn 2021), they can also explain why different factors matter more at different stages of a conflict (Germann and Sambanis 2021). Such an account can be more enlightening to understand the underlying processes associated with different forms of conflict (Drury et al. 2020).

Similar framings have been presented, for example, by describing the de-escalation of violence within an armed conflict in Colombia, the escalation of armed conflict as a prelude to the end of armed conflict in South Africa (Díaz Pabón 2022), or how the loss of legitimacy of the warring parties relates to de-escalation of conflicts (Duyvesteyn 2021). This illustrates the possibility of how looking at history and rethinking our language and conceptual frameworks can serve us and bring about new conceptual and empirical settings that can challenge the autarky of concepts, the understanding of conflict as binary and the linear framings of conflict for bringing forward the complexities of processes of escalation and de-escalation of conflict.

Conclusion

Instead of being linear, stratified, structured or hierarchical, conflict is messy, chaotic, and opaque. The fog that, according to Clausewitz, affected commanders in wars can also affect the researchers studying war and conflict. In its study, we must work with concepts, definitions and delimitations to help us understand these important phenomena and their complex relations. This contribution has argued that our present demarcation of civil war and the most prevalent conceptualisations hamper our understanding of two significant but under-studied phenomena: escalation and de-escalation. We have made a case for not only a reconsideration of our language and definitions but also, more fundamentally, the prevalent approaches to understand what is going on in real-life conflicts and specific contexts. We need to move away from separate discussions focused on either demonstrations, violent protests, terrorism or armed conflict and instead look at these as connected in different and unique manners across time and space. Moreover, the binaries we have embraced so strongly in this field of study, of war-peace, violence-non-violence, prove impediments for understanding the real-world expressions of fundamental political contestation. We invite scholars in the field to reconsider these and attempt to move beyond them. Finally, and most importantly, we need to let go of linear conceptions of conflict and recognise that escalation and de-escalation cannot be described with deterministic lenses if we are to unveil the complex dynamics of any conflict. In practice, conflict is complex; perceptions, actors, relations, ideas, narratives, contextual factors and fictions shape the way many political disagreements unfold and morph across each contestation. If we are to describe reality (instead of prescribing it), we need to be able to account for these multiple escalations and de-escalations across multiple equilibria.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Fabio Andrés Díaz Pabón

Fabio Andrés Díaz Pabón is a research fellow in Sustainable Development and the African 2063 agenda, hosted by the African Centre of Excellence for Inequality Research (ACEIR) of the University of Cape Town, and a research associate at the department of Political and International Studies at Rhodes University. His research interests are related to politics, conflict, development, inequality and economics from a comparative perspective.

Isabelle Duyvesteyn

Isabelle Duyvesteyn is Professor of International Studies at the Institute of History at Leiden University in the Netherlands. She completed her PhD at the Department of War Studies at King’s College in London. Previously she has worked at the Royal Military Academy in the Netherlands, the Netherlands Institute for International Relations and the Department of History of International Relations at Utrecht University in the Netherlands. Her research has focused on the nature of war and peace in the developing world, terrorism and insurgency and the strategy of non-state actors. She has served as a member of the National Advisory Council for International Affairs assigned to advise the Netherlands government on issues of peace and security, a member of the Scientific Advisory Board of the Netherlands Defence Academy and a member of several book and journal editorial boards, notably the Journal of Strategic Studies, and Small Wars and Insurgencies.

Notes

1. Whereas Bartusevičius & Gleditsch speak of civil conflict in their definition, we use the term conflict without the adjective ‘civil’ – this given that the term civil conflict is associated with low-intensity civil war in the conflict-studies literature. We note that this association is strong but not exclusive. There are many examples of ‘civil wars’, which were fought in a conventional manner, e.g., the Spanish Civil War. We also adjust this definition to allow for the occurrence of conflict between non-state actors.

2. For an attempt to conceptually separate different expressions of conflict; Duyvesteyn and Fumerton (2009).

3. For the case of social uprisings, see Shaheen (2015); for the case of social movements, see Tilly and Wood (2016); for the case of how variables identified as causing civil war affect other forms of conflict, see Cunningham and Lemke (2014). For an analysis of the relationship between different forms of conflict in processes of state building, see Tarrow (2015); for an analysis of the escalation of conflict towards civil war, see Blair and Sambanis (2020); for an analysis of the trajectories of escalation of conflict in autocracies, see Rød and Weidmann (2021); for an analysis of the escalation of non-violent movements towards violence, see Ryckman (2020); for a simulation model to analyse the connection between different stages of political turmoil see Hegre et al. (2017); for a description of the different stages of processes of mobilisation see Shultziner and Goldberg (2019); for the processes of escalation of conflict in relation to political autonomy see (Germann and Sambanis 2021).

4. See for an example of this type of reasoning (Kocaman et al. 2023).

5. Tensions had been simmering between El Salvador and Honduras, related to the migration of Salvadorans to Honduras in preceding decades before. As land (whether legally or illegally occupied) by Salvadorans was expropriated, and rights of Salvadorans in Honduras were limited, this fuelled the animosity between countries. A series of soccer matches between El Salvador and Honduras (qualifier matches for the 1970 soccer world cup) were followed by violence between soccer fans, which led Salvador to find the justification to dissolve diplomatic ties and declare war with Honduras (Cable 1969). This illustrates how grievances associated with a soccer match, something that would not necessarily be associated with a ‘rationality’ framework, can be central in explaining the emergence of war.

6. An example is the discussion about the ‘strategic corporal’ (Krulak 1999).