On 15 February 1999, Abdullah Öcalan was arrested in Kenya by Turkish security agents and brought to Turkey. The Kurdistan Workers’ Party (PKK), founded by Öcalan, is one of the biggest challenges Turkey has faced. Until Öcalan’s arrest, the PKK insurgency resulted in nearly 30,000 deaths, and millions of civilians were displaced. Following his arrest, the PKK declared a ceasefire, marking what was interpreted as the end of the insurgency. The literature frequently mentions Öcalan’s arrest as an example of a successful and effective leadership decapitation strategy.Footnote 1
Despite the frequent use of leadership decapitation as a counterterrorism/insurgency strategy, the literature is far from having a consensus regarding its effectiveness and consequences. While some assert its effectiveness under certain circumstances,Footnote 2 others find it counterproductiveFootnote 3 or ineffective at best.Footnote 4 Some argue its effectiveness and consequences are context-dependent.Footnote 5 This study contributes to this debate by studying leadership decapitation in a fundamentally new way.
First, I argue that the literature focuses on the military implications of leadership decapitation (its implications for the organisation’s operational capabilities/ability to generate violence) at the expense of its impacts on negotiations between insurgents and the state. Second, I argue that leadership arrest and killing are analytically distinct categories of decapitation that can trigger different processes and outcomes. Relatedly, I argue that the potential role of an arrested leader from prison should be considered in the analysis since arrest alters the leader’s incentives, resulting in a new bargaining game between the leader, the state, and the armed organisation. This also suggests that leadership arrest can have different consequences depending on a leader’s ability to maintain their leadership from prison. Lastly, I argue that literature may benefit from tracing the process closely, considering the dynamic nature of conflicts without limiting the temporal scope of inquiry to assess the impacts of leadership decapitation. I illustrate these arguments by using Öcalan’s arrest as a theory-building case study through inductive iteration and identifying pathways leading to the different outcomes in the causal chain.
I show the broader relevance of my arguments by briefly discussing cases from various countries, including Israel’s recent assassinations of Hamas and Hezbollah leaders since 7 October 2023, and the recent statement by Devlet Bahçeli – leader of Turkey’s far-right nationalist party and de facto coalition partner of President Erdoğan – urging Abdullah Öcalan to announce the dissolution of the PKK and the laying down of arms in exchange for his release from prison.
This study has several contributions: first, it studies leadership decapitation’s impact on negotiations. Second, this study considers an arrested leader’s possible role in prison, and it specifically looks at the consequences of leadership arrest for negotiations between the state, the organisation, and the leader. All of these have been overlooked in the literature. In the language of variables, I offer both a new dependent variable (negotiations) and a new independent variable (leaders’ role from prison) of interest to the study of the consequences of leadership decapitation. It is important to note here that this is not the first study to disaggregate leadership decapitation (killing vs arrest).Footnote 6 However, these studies find no statistically significant differences or report mixed results when comparing the effects of killing versus capturing a leader and focus on their impact on targeted organisations’ operational capabilities. I also show that the arrest of a leader creates a new multi-actor bargaining scenario not only between the state and the leader but also between the leader and the organisation.
Moreover, although Öcalan’s arrest is often cited as a successful example of leadership decapitation,Footnote 7 these studies focus on the decline in the intensity of the conflict as evidence of its success. However, there is limited research on its impact on negotiations. Thus, by shifting our focus to negotiations from the military-operational capabilities of the PKK, I challenge/qualify the argument that sees Öcalan’s arrest as an example of a successful leadership decapitation strategy. I argue that Öcalan’s arrest created a window of opportunity for the resolution of the conflict in the short run, yet it further complicated the conflict in the medium and long runs by increasing the number of veto players. This suggests that scholars of decapitation may benefit from closely tracing the process without limiting the temporal scope of inquiry and focusing on the dynamic nature of conflicts, exploring how decapitation impacts the bargaining processes between the parties.
It is also important to note that this study does not aim to provide an account of the transformation of the PKK’s military-political strategy, ideology, or organisational structure following Öcalan’s arrest or of sources of the PKK’s resilience, as this topic has been addressed in other excellent works. For the Turkey–PKK conflict, the novelty of my analysis is to show how Öcalan’s arrest had different impacts on negotiations depending on the time horizon, which challenges the conventional wisdom that sees his arrest as a successful decapitation strategy. Moreover, relying on the broader conflict literature, I identify pathways and ‘formalise’ the descriptive evidence into analytical explanations,Footnote 8 which can be tested/traced in other cases. Relying on this literature, I also explain the motivation behind the actors’ actions. I show how his arrest created new incentives for Öcalan to offer concessions to the state and then to escalate violence and support the hardliners during the fragmentation of the PKK.
Methodologically, this study demonstrates how a single case study, selected for its intrinsic importance, can advance the literature and lead to theory development without any prior intention to do so,Footnote 9 broadly in line with the methodological pragmatism that emphasises ‘the efficient and efficacious production of useful knowledge’.Footnote 10
Lastly, the PKK–Turkish conflict has implications for regional politics and security in the Middle East. It is also one of the longest ongoing civil wars in the world in a NATO member state. Despite these facts, the PKK–Turkish state conflict rarely appears in the major journals as a theory-building case study.Footnote 11
It is important to note that the theory and causal chain were not developed through deduction but through an iterative process that involved going back and forth between the literature and the case study.Footnote 12 For clarity and readability, I have created an analytical narrativeFootnote 13 and separated the theoretical and the empirical sections.
The article is structured as follows. First, I review the literature on leadership decapitation. Then, I discuss the similarities and differences between leadership arrest and killing, and why we should study their impacts on negotiations. Third, I talk about the method and the case selection. Fourth, I present the theory/causal chain, and the pathways. Fifth, I narrate the events that unfolded after Öcalan’s arrest and present the pathways to particular outcomes. Before concluding, I discuss the study’s limits, avenues for future research, and the application of my framework beyond the case study.
Leadership decapitation
The literature does not have a consensus regarding the impacts and effectiveness of leadership decapitation. Several scholars argue that leadership decapitation is effective, at least under some conditions.Footnote 14 Others argue that leadership decapitation is counterproductiveFootnote 15 or ineffective at best.Footnote 16 Some argue that its effects are context-dependent.Footnote 17 This disagreement stems from several reasons.
First, the findings of case studies are hard to generalise, and in large-N quantitative studies, conceptualisation, and operationalisation of the variables differ.Footnote 18 Second, leadership decapitation is employed in various settings against different types of organisations.Footnote 19 Third, studies have different temporal scopes.Footnote 20 I argue that to assess the implications of leadership decapitation, one should closely trace the process without limiting the temporal scope through which pathways to different outcomes can be traced, since some ‘mechanisms have their effect only given enough time’.Footnote 21
Fourth, leadership decapitation may entail either the killing or the arrest of a leader. Scholars often find no statistically significant differences or report mixed results when comparing the effects of killing versus capturing a leader.Footnote 22 I argue that leadership arrest and killing should be analytically separated from each other, since they may have distinct implications, and that an arrested leader can maintain their leadership status from prison. This also suggests that leadership arrest can have different consequences depending on a leader’s ability to maintain their leadership from prison.Footnote 23 Fifth, I argue that the impact of the leadership decapitation depends on the state’s subsequent actions after decapitation. Thus, leadership decapitation should be analysed in relation to the state’s actions.
Sixth, it is unclear what effectiveness or success means. In the literature, effectiveness and success are operationalised in different ways, including the group’s ability to generate violence (how frequent and lethal the attacks are),Footnote 24 the nature of violence (the type of the attacks and its targets),Footnote 25 survival or longevity of the group,Footnote 26 and conflict termination.Footnote 27
Based on organisational and leadership theory,Footnote 28 current scholarship is primarily interested in the impacts of leadership decapitation on organisations’ operational capabilities, military effectiveness, and the ability to generate violence. Although this perspective sheds light on important questions, these studies mostly focus on the military-operational implications of leadership decapitation at the expense of its impact on negotiations and bargaining processes. I argue that we should theorise how leadership decapitation influences negotiations between the targeted organisation, the leader, and the state.
In the next sections, I explain the theoretical rationale behind this paper’s two main arguments.
Disaggregating leadership decapitation
In the discussion section, I provide a more comprehensive analysis of the potential implications of arrest for negotiations and mention them briefly here. Both arrest and killing trigger similar processes because they result in a loss of organisational skills and know-how, necessitate leadership succession, and signal the state’s superior capabilities.
However, it is crucial to recognise that leadership arrest can have different implications than leadership killing. The most apparent difference is straightforward: ‘Dead men tell no tales.’Footnote 29 Interrogation of a leader can yield valuable information for the state. Additionally, side-switching is a common dynamic in civil wars,Footnote 30 and an arrested leader is more likely to switch sides than a free leader, as the concessions they receive can include their freedom.
Moreover, in the case of an arrested leader, they can maintain their position as leader. Otto, Scharpf, and GohdesFootnote 31 find that 25 per cent of all rebel groups changed their alignment towards governments at least once, and most of these changes were anti-government organisations switching sides. Therefore, another potential outcome of leadership arrest is the defection or side-switching of the leader along with his faction. Fourth, a potential outcome is the co-option of arrested leaders by states to advance a pro-peace agenda or any other message the state wishes to promote. Fifth, the leader’s imprisonment, safety, and freedom can be significant to the negotiations.
Overall, the distinct implication of leadership arrest is that it creates a new bargaining dynamic between the leader, the state, and the targeted organisation by altering the incentives of the actors, particularly the arrested leader.
Bringing the negotiations back in
A long tradition in conflict studies sees the use of force as a means to achieve political ends, where political violence is instrumentalised by political actors to gain bargaining leverage over adversaries and to change their behaviour,Footnote 32 including terrorist groups.Footnote 33 An extensive literature has used formal and strategic models to study terrorism.Footnote 34 Moreover, the literature shows that terrorists strategically choose the timing of their attacksFootnote 35 and that voters punish incumbent governments for being unable to contain violence.Footnote 36 Thus, violence is costly for politicians, motivating them to seek ways to stop violence, including negotiations.Footnote 37
Moreover, most terror attacks are conducted by insurgent groups fighting civil wars, blurring the distinction between terrorist and insurgent groups.Footnote 38 A growing literature also studies various dimensions of terrorism in the context of civil war, including their effects on negotiated settlements.Footnote 39 Consequently, to fully understand the consequences and effectiveness of leadership decapitation, we must also consider its effects on bargaining processes and negotiations.
Method and data
I use Abdullah Öcalan’s arrest as a case study. Case studies are exploratory in nature and are especially suitable for developing new theories and frameworks.Footnote 40 As GerringFootnote 41 puts it: ‘Case studies may be more useful than cross-case studies when a subject is being encountered for the first time or is being considered in a fundamentally new way.’ This study studies leadership decapitation in a fundamentally new way using process tracing, a particularly useful method to build new theories and frameworks. It helps researchers make strong within-case inferences with particular attention given to the sequence of events, through which particular pathways to different outcomes can be identified,Footnote 42 especially in the analysis of decision-making at the individual, small group, and organisational levels.Footnote 43
The data comes from primary sources such as the publications and statements by the PKK leadership and former militants, Abdullah Öcalan’s books, and his lawyers’ notes taken during meetings with him in prison. Secondary sources include news articles, books, interviews given by the PKK leaders, and documentaries where state actors and journalist share their information and memories.
Given the high-stakes and non-public nature of the events discussed, I triangulated information from a broad range of sources, including publications and interviews from opposing factions within the PKK, statements from Abdullah Öcalan, and Turkish media outlets – each with distinct interests and viewpoints. I did not encounter any significant contradictions between these sources, except for political disputes among different political actors. The overall consistency of accounts supports the reliability of the findings presented in the article.
Next, since scholars sometimes present ex-post rationalisations for their case selection to follow the templates of deductive procedures that do not correctly describe their original intentions,Footnote 44 I need to be transparent about my case selection. I began this research to study how Abdullah Öcalan’s arrest impacted the dynamics of the conflict between the PKK and the Turkish state for its intrinsic importance. I had no intention of developing a new theory of leadership decapitation.
Indeed, many important questions about politics are what Levy calls theory-guided idiographic case studies, where the aim is to explain a single event using the established theories and frameworks to make sense of the case-specific evidence.Footnote 45 Through an iterative process between the case study and the literature, what initially started as an idiographic case study helped me to develop a new perspective to study the implications of leadership decapitation, as the literature overlooked the effects of leadership decapitation on negotiations and the role of arrested leaders from prison.Footnote 46
This process confirms Flyvbjerg’s insights about the difficulty of specifying the type of a case (critical, paradigmatic, etc.) a priori and the importance of case studies for producing generalisable knowledge.Footnote 47 For example, Mahoney argues that one of the challenges of finding a cause for a specific outcome in case studies is that ‘potential causes may include variables not yet even specified in the literature’.Footnote 48 In my case, the literature considered neither the leader’s possible role from prison – a distinct, independent variable – nor the potential impacts of leadership decapitation on negotiations – a distinct dependent variable.
Below, I present the case and the theory. I have separated the theory and the case for analytical clarity and readability. Yet theory development was an inductive process.Footnote 49 I have divided the process and the case into four sections, each identifying pathway(s) to a particular outcome in the causal sequence. In each section, I identify the pathway using broader conflict literature while discussing the incentives of the actors in general terms to create an analytical narrative.Footnote 50 I also visualise the pathways to each outcome to ensure the validity and usefulness of my analysis.Footnote 51 After that, I narrate the actual events to show how they worked out in reality in the next section.
The arrest of a leader and negotiations
Concessions (stage 1 – path 1)
Once a leader is arrested, there are several actions they may take. First, they may defect and provide the state with intelligence to address the insurgency. Second, they may refuse to cooperate with the state. Third, by using their influence over the organisation, they may try to negotiate with the state to get concessions. However, the arrest of a leader creates uncertainty. Both the leader and the state face an information problem. Neither knows whether the organisation and remaining leadership will follow the leader’s orders and continue recognising them as the leader. Therefore, the arrested leader faces what I call a problem of control and a problem of persuasion. First, they must be able to control the organisation. Second, to be able to negotiate with the state, they have to convince the state that they can effectively control the organisation.Footnote 52
A leader can overcome this uncertainty by offering risky concessions to the state since other senior leaders of the organisation may oppose them (Figure 1). Like state leaders,Footnote 53 rebel/terrorist leaders must consider ‘the domestic politics of their organization’, as there is a risk of factional splits and disengagement of the militants and the constituency. Additionally, supporting and engaging in terrorism/insurgency are high-risk actions, and replacing militants and the civilian support base can be particularly challenging. Thus, by making risky concessions to a state, an arrested leader can (1) test their ability to control the organisation from the prison and (2) signal to the state that they can effectively control the organisation.
Figure 1. Stage 1 – path 1.
Moreover, concessions serve a third purpose. A government likely sees an arrested insurgent or terrorist leader as an untrustworthy bargaining partner.Footnote 54 Therefore, the leader’s risky and potentially costly concessions (3) signal to the state that he is a trustworthy partner committed to negotiating (Figure 1). This is one way in which arrest can work in the short run.
Turmoil (stage 2 – path 2)
If the concessions lead to discontent and trigger organisational turmoil, this decreases the military capabilities of the organisation since it has to devote time to internal matters and loses its trained military force. This also decreases the bargaining leverage of the leader. This may be one of the ways in which the leadership arrest works, since the internal turmoil makes both (1) a military defeat and (2) a negotiated settlement more likely by making a leader more concessive (Figure 2).Footnote 55
Figure 2. Stage 2 – path 2.
Therefore, in the short term, the arrest of a leader presents a window of opportunity by incentivising the leader to offer concessions.
Escalation of violence (stage 3—paths 3 and 4)
Once a leader secures the organisation’s approval and signals that they are committed to negotiating, the next move is the state’s. If the state engages in negotiations, it may lead to the resolution of the conflict. However, if the state does not negotiate or negotiations fail, the leader must bring the state to the negotiation table by imposing costs. Thus, they have to (1) escalate violence, which also (2) gives them incentives to side with the hardliners within the organisation, who prefer escalation of the violence (Figure 3). Thus, in the medium run, leadership arrest becomes counterproductive by radicalising the organisation.
Figure 3. Stage 3 – path 3.
In such a situation, both arrested and free leaders may have incentives to escalate violence. Therefore, it is important to distinguish their incentives. First, free leaders can continue to exercise their freedom and govern their organisations. Second, in the case of an arrested leader, the leaders on the ground may secure and increase their hold on power over time. Thus, the arrested leader races against time and has to act as quickly as possible.
Therefore, the time between unsuccessful negotiation attempts and the escalation of violence can be shorter in the case of an arrested leader. In Grzymala-Busse’s terms,Footnote 56 the arrest condition accelerates the tempo of the process since the time between the two events (failed negotiations and the escalation of violence) is expected to be shorter in case of an arrest because the arrest condition alters the incentives of the leader and results in a new bargaining game not only between the state and the leader but between the leader and the organisation.
Second, intra-organisational dynamics and power struggles influence the dynamics and levels of violence. Should concessions result in turmoil within the organisation, this could give the arrested leader an additional incentive to escalate violence. First, violence can be used to unify the organisation.Footnote 57 Second, they might leverage violence to bolster public support by portraying themselves as possessing superior nationalist commitmentFootnote 58 or as the dominant force on the ground.Footnote 59 Thus, through violence, they can try to outbid other organisationsFootnote 60 (Figure 4).
Figure 4. Stage 3 – path 4.
Dual leadership (stage 4 – path 5)
Lastly, we may expect senior members of the organisation on the ground to increase/secure their hold on power vis-à-vis the arrested leader over time. This creates different incentives and interests for the arrested leader and leadership on the ground, and an interesting case of dual leadership emerges. During the negotiations, the state has to bargain both with the leadership on the ground and with the leader in prison, increasing the number of veto players and subsequently complicating the prospects for a negotiated settlement (Figure 5).Footnote 61 Thus, in the long run, leadership arrest becomes counterproductive.
Figure 5. Stage 4 – path 5.
The case analysis
Concessions: Ceasefire decision of öcalan and withdrawal from Turkey (stage 1)
Abdullah Öcalan left Turkey for Syria before the coup d’état of 1980 and started the PKK’s insurgency in 1984. He stayed there until he was forced out by the Syrian government in 1998. On 15 February 1999, he was arrested in Kenya by Turkey. After Öcalan was arrested, the PKK established a Presidential Council to govern the organisation. Yet, neither Öcalan nor the state knew whether the organisation would remain loyal to him.
After Öcalan’s arrest, the PKK started a massive campaign targeting the Turkish security forces and civilians as retaliation. Öcalan argued against the attacks, thinking that they undermined prospects for a peaceful resolution of the conflict. After several months, on the 6 May, the PKK leadership declared their loyalty to Öcalan and that they would follow his strategy.Footnote 62 Öcalan’s trial began on 31 May and ended on 29 June 1999. His first sentences in the trial are quite important to understand his concessionary words and the reactions to them:
In the line of the Democratic Republic, for peace and fraternity, I want to declare my willingness and determination to serve the state … I want to make a short statement for dear venerable families of martyrs (he refers to security forces who were killed by the PKK). I wholeheartedly share their sorrow and pain. I feel sorry for my part of the responsibility. I promise that I will do everything I can to end the bloodshed that is really caused by a social wound.Footnote 63
Öcalan declared that he would serve the state he had called fascists and fought against for years. Throughout his trial, he talked about his Democratic Republic project. He said that they respected Turkey’s territorial integrity and that they would leave the mountains if the cultural rights of Kurds were recognised.Footnote 64 Öcalan’s statements throughout the trial were concessive and created dissatisfaction among the PKK militants.Footnote 65 On 29 June9, Öcalan was sentenced to death. However, he did not believe he would be executed since this would undermine Turkey’s prospects for European Union membership, increase tensions in society, and intensify the conflict.Footnote 66 Subsequently, on 5 July 1999, Öcalan disclosed his intentions to withdraw PKK forces from Turkey to his lawyers, aiming to demonstrate to the state their readiness to seek a compromise:
If we stop the violence, this may be a beginning. A technique to leave the violent struggle behind can be planned. The Council can (he refers to the Presidential Council of the PKK) make a call. The relationship will be established. They will say that they are loyal. This will be a new step. In practice, the forces can retreat to the South (referring to Iraqi Kurdistan), behind the border. This will be a step further than a ceasefire.Footnote 67
This indicates that Öcalan was willing to give concessions to the state while signalling that he could effectively control the organisation and his willingness to start negotiations. On 2 August, Öcalan called for a ceasefire and for the PKK to retreat from Turkish territories. On 25 August, the Presidential Council announced its withdrawal from Turkey.Footnote 68 Moreover, Öcalan ordered several senior members of the PKK, including spokesman Ali Sapan in Europe, to turn themselves in to Turkey to signal their willingness to find a peaceful solution to the conflict. In two weeks, 15 senior members of the PKK returned to Turkey, and they were all arrested.Footnote 69 Sapan was sentenced to 10 years in prison.Footnote 70
These developments are consistent with my arguments in stage 1 – path 1 (Figure 1). In the first stage, leadership arrest creates uncertainty about the leader’s ability to control the organisation. Thus, the leader provides costly concessions (1) to test their ability to control the organisation, (2) to signal to the state that they can effectively control the organisation, and (3) to initiate negotiations by showing that they are a trustworthy partner.
Reactions to concessions within the organisation: Turmoil within the PKK (outcome 2)
Yet Öcalan’s concessions came at a price. His call for a ceasefire and retreat from Turkish territory was shocking for militants who risked their lives for years to fight in the mountains.Footnote 71 Thousands of militants left Turkey and became open targets for the Turkish army; hundreds of them were killed, and the PKK lost strategically important fields over which militants have fought for years. Moreover, several high-ranking members turned themselves in without any guarantees and were arrested. Some commanders and groups resisted Öcalan’s orders and did not leave Turkey.
Additionally, many militants could not make sense of this decision. This created tensions within the organisation since the Turkish army continued chasing the militants who were withdrawing. Close to 500 PKK militants were killed, and thousands of militants, including some of the founders and early members, left the organisation.Footnote 72 According to a militant at the time, ‘there was almost nobody, from the highest level to lowest, who thought this was the right thing to do, but no one had the courage to challenge Öcalan’.Footnote 73 Additionally, leaving the PKK was risky and often involved death and arrest, indicating a high level of internal turmoil.Footnote 74
These developments are consistent with my arguments in Figure 2, where I have argued that risky concessions would create internal turmoil and weaken the organisation militarily.
The PKK’s decision to the end the ceasefire (stage 3 – paths 3 and 4)
Öcalan’s concessions did not result in the developments he anticipated. The government was unwilling to negotiate and continued its military operations. Öcalan needed to put pressure on the government. On 4 September 2002, Öcalan called for the PKK to improve their capabilities for a ‘legitimate defence’, permitting the militants to enter Turkey. Then, he repeatedly called the Turkish army and the government to act to prevent the escalation of the conflict. On 15 July 2003, he declared that his peace mission had ended and called on the government to take action by 15 August.Footnote 75 In the summer of 2003, the number of PKK militants in Turkey reached 1,500.Footnote 76
The PKK held a congress between 7 October and 6 November 2003, where the divide within the senior leadership on the ground became apparent. Nizamettin Taş and Abdullah Öcalan’s brother Osman Öcalan – the leaders of the soft-liner group – became the deputy presidents of the PKK. Moreover, leading hard-line figures such as Cemil Bayık, Mustafa Karasu, Gülizar Tural, and Bahoz Erdal failed to get elected to the executive council, and most of the executive positions were dominated by the soft-liners.Footnote 77 Yet, after the congress, the soft-liners left the PKK-controlled areas, fearing assassination. On 9–10 February 2004, a meeting was arranged to find a compromise among the groups, but the soft-liner did not attend.
The divide further deepened as the two sides disagreed on which candidate to support for the upcoming municipal elections in Diyarbakır. The PKK declared that the soft-liners’ positions were halted and referred them to the discipline committee.Footnote 78 Responding to the announcement, Osman Öcalan blamed the hardliners for not following Abdullah Öcalan’s democratic ideals, arguing that they were the ones following the Leadership’sFootnote 79 path.Footnote 80
After these developments, Abdullah Öcalan requested an extraordinary congress to overcome the problems. Given the popularity of Nizamettin Taş and Osman Öcalan’s views, Öcalan probably wanted to prevent a possible disintegration of the organisation.Footnote 81 However, three days after the Congress, the leading figures of the soft-liners left PKK-controlled territories.Footnote 82 On 1 June, the PKK declared the end of the ceasefire. Available evidence suggests that Abdullah Öcalan personally ordered the end of the ceasefire despite the ceasefire being more popular in the organisation.Footnote 83 In an interview, Osman Öcalan stated that they wanted to leave the armed struggle fully. He further said that soft-liners’ views dominated within the PKK until Abdullah Öcalan started supporting the hardliners.Footnote 84 As argued in the theoretical section, Öcalan’s support for hardliners was not surprising since he needed to escalate violence to bring the state back to the negotiation table.Footnote 85
The comment of a Kurdish intellectual is also quite revealing: ‘The war decision stopped the arguments and opinions of the cadres who were left outside of the war; it rendered them irrelevant. The war decision played a uniting role that healed the division in the PKK ranks.’Footnote 86 In his long report that was written after interviews with politicians, state officials, and the highest-ranking PKK members of that time, Çandar concludes that ‘the urge to eliminate the internal cracks within the organisation played a decisive role in the escalation of the PKK’s armed struggle’.Footnote 87
These developments align with arguments presented in path 3 (Figure 3) and path 4 (Figure 4). In path 3, it is argued that the leader must escalate the conflict to compel the state to negotiate, incentivising support for hardliners who prefer escalation. Öcalan repeatedly urged the government to initiate negotiations while gradually preparing the PKK for violence. Additionally, to facilitate the escalation of violence, he backed hardliners within the organisation despite the popularity of soft-liners. In Figure 4, it is argued that organisational turmoil provides the leader with further motivation to escalate violence. This escalation, as observed by Kurdish intellectuals, helped Öcalan consolidate his power and maintain organisational integrity.
The peace process between the government and the PKK (stage 4, path 5)
After the PKK ended the ceasefire in 2004, the level of violence started increasing steadily. Beginning in 2006, the government started negotiating with the PKK behind closed doors.Footnote 88 State officials negotiated with Öcalan in prison and with the PKK in various locations while the fighting and occasional ceasefires continued. This process collapsed in 2011.
Subsequently, clashes intensified, and 2012 marked the most violent year of the conflict since Öcalan’s arrest in 1999. Nevertheless, back-channel negotiations continued. In December 2012, Erdoğan acknowledged the ongoing talks between Öcalan and the state. During this process, the government negotiated with Öcalan in prison and with the leadership in the Qandil mountains and Europe. The political party of the Kurdish national movement, HDP, facilitated communication between the PKK leaders on the ground, Öcalan, and the government. In March 2013, Öcalan’s message, calling for the PKK to disarm and withdraw from Turkish territory, was read in front of hundreds of thousands of people. This marked the beginning of the Peace Process, which collapsed in the summer of 2015.Footnote 89
These negotiations took place when the PKK leadership on the ground was arguably stronger than ever. Öcalan had been in prison for over 13 years, and 2012 saw the highest level of security force casualties since his arrest, while the state’s control was significantly challenged in border areas.Footnote 90 Moreover, the PYD (Partiya Yekîtiya Demokrat/Democratic Union Party),Footnote 91 closely linked to and receiving military and organisational support from the PKK, controlled a significant part of northern Syria bordering Turkey.Footnote 92 Despite these developments, Öcalan maintained his influence from prison, showing the importance and relevance of leaders long after their arrest.
Although this process does not allow me to argue the Peace Process failed because of the dual leadership within the PKK, the fact that the parliamentary members from the HDP and intelligence officers visited both Öcalan and the PKK leadership on the ground indicates the dual leadership situation. In Pathway 5 (Figure 5), I have argued that leadership arrest can result in a dual-leadership situation, complicating the negotiations and increasing the number of vote players, thus making a negotiated settlement less likely.
To summarise, I have shown that neither an arrested leader nor a state knows whether the arrested leader can control the organisation from prison. This incentivises the arrested leader to offer costly concessions (1) to test their ability to control the organisation, (2) to signal to the state that they can effectively control the organisation, and (3) to initiate negotiations. This makes a negotiated settlement more likely (Figure 1). Once the leader offers costly concessions, this can result in internal turmoil, weakening the organisation militarily. This makes both (1) a state victory and (2) a negotiated settlement more likely (Figure 2). If the state does not negotiate with a leader or fails to reach a negotiated settlement, this gives the leader an incentive to (1) escalate violence and (2) side with the hardliners in the organisation (Figure 3). Moreover, internal turmoil triggered by costly concessions gives the leader additional incentives to escalate violence to consolidate the organisation (Figure 4). Over the long run, the leadership on the ground increases its power vis-à-vis leadership in the prison, creating a dual-leadership situation and more veto players, which decreases the possibility of a negotiated settlement (Figure 5).
As the case narrative reveals, Öcalan’s arrest was productive for terminating the conflict in the short run. In contrast, it was counterproductive in the medium and long runs, suggesting that the literature may benefit from tracing the process closely, considering the dynamic nature of conflicts and the impact of decapitation on bargaining processes, without limiting the temporal scope of inquiry.
Finally, policymakers are aware of the political nature of leadership arrest, indicating the importance of studying its impacts on negotiations. When Öcalan was sentenced to death, the far-right Turkish nationalist MHP was a partner of the coalition government whose base strongly favoured and mobilised for Abdullah Öcalan’s execution. Despite this pressure, given its potential political costs, the government decided not to execute the death sentence. The politicians and bureaucrats, including Turkish intelligence, started a campaign to push public opinion against Öcalan’s execution. Intelligence officers met with journalists and briefed the media. Cevat Öneş, the former deputy secretary of intelligence, states: ‘In this process, whether Öcalan was to be executed or not, was not only evaluated from a security perspective but a political perspective. Its impacts on political processes were considered.’Footnote 93
In the next section, I discuss the study’s limitations, avenues for future research, and the application of my framework beyond the case study.
Discussion: Limits and application of the framework beyond the case study
The interactions between actors and pathways to outcomes only explain this case study and are causally connected. Thus, they are not easily generalisable. However, my broader arguments that the literature should pay more attention to
(1) The impacts of leadership decapitation on negotiations;
(2) The differences between leadership arrest and killing;
(3) Possible role of the leader from prison and the ability to maintain their leadership position;
(4) The new multi-actor bargaining situation that emerges after leadership arrest between the leader, organisation, and the state;
(5) Temporal variation in the impacts of leadership decapitation
are not case-specific and have important implications for studying the consequences of leadership decapitation.
Moreover, to overcome the problem of generalisability of case-specific dynamics, following Slater and Ziblatt’s advice,Footnote 94 I have framed the pathways to different outcomes, incentives, and interactions in general rather than context-specific terms and supported them with the findings of the broader conflict literature, i.e. I formalised case-study evidence into analytical explanations.Footnote 95 These pathways and interactions can be traced in other cases or can be slightly modified within the framework advanced in this study.
For instance, I have argued that a leader’s concessive behaviour may result in organisational turmoil. Yet, in another case, right after the leadership arrests, a dispute within the senior leadership over loyalty to the leadership may emerge, triggering rebel fragmentation and escalation of violence. If a negotiated settlement is reached after a leader’s concessions, the causal chain may end there, or the fourth outcome – dual leadership – can be observed even in the immediate aftermath of the leadership arrest.
One can also develop other pathways following the logic of bargaining, what leadership decapitation would signal to actors, the extent to which it damages the military capabilities of the targeted organisation, and how these would impact negotiations. For instance, one can argue that leadership decapitation signals that the organisation suffers from intelligence failures; this may make them more likely to negotiate with the government as it reveals their vulnerabilities. Moreover, a government might be more likely to renege on a deal after a decapitation now that the group is weakened, which the group would be aware of increasing commitment problems.
More pathways can be modelled and tested in other cases.
Further research can also comparatively study how relationship between the state and the arrested leader evolves over time and how the state’s treatment of prisoners and what opportunities the state offers to the arrested leader impact this relationship. For instance, a state can provide an arrested leader comfortable working conditions and provide opportunities to directly communicate with their followers, or in the opposite scenario, a state can even prevent an arrested leader from meeting their lawyers. These actions are likely to influence and be influenced by the negotiations and trust between the parties. For example, in Öcalan’s case, whenever there were negotiations between Öcalan/PKK and the state, Öcalan was allowed to be visited by his lawyers, politicians, and family members. Whenever fighting intensified, he was prevented from this opportunity. Moreover, leadership arrest operations require deeper intelligence penetration and are more risky for politicians and militaries as they are more likely to result in combat and fatalities; leadership killing, however, is often less risky for the states especially in the age of drones.Footnote 96 Thus, bargaining dynamics after leadership decapitation can also be a function of these intelligence advantages, and further research can look into these issues.
Overall, leadership decapitation can empower different factions with varying preferences for negotiations. Given its operational and military consequences, it can change the balance of power and bargaining leverage and reveal information about the capabilities and resolve of parties to the conflict, all of which can impact negotiations in various ways. Recent assassinations of Hamas and Hezbollah leaders by Israel and the analysis surrounding these events exemplify these.
For instance, after Hamas’s leader Ismail Haniyeh was assassinated in July 2024, some analysts suggested that this could make a deal between Hamas and Israel easier, as it gave the Israeli government an opportunity to represent itself as victorious, thus making a hostage deal politically more acceptable. Yet others suggested his killing decreased the probability of a negotiated settlement by creating mistrust between the parties and empowering hardliners, as Haniyeh was seen as representing the more moderate faction of Hamas.Footnote 97 Similarly, the assassination of Hezbollah’s leader Nasrallah, alongside other high-ranking Hezbollah leaders, was discussed in terms of its empowerment of different factions and organisations in Lebanon and within Hezbollah, of Israel’s resolve to escalate violence, and as a signal of Israel’s military/intelligence superiority vis-à-vis Hezbollah and Iran. Analysts have debated how these dynamics would impact potential negotiations between Iran, Hezbollah, Lebanon, and Israel.Footnote 98 Similarly, the killing of Yahya Sinwar, widely seen as the main architect of the 7 October attacks and successor of Ismail Haniyeh, sparked many debates over negotiations between Hamas and Israel. Indeed, some analysts suggested that this was a further proof of Israel’s resolve and would make Iran and Hezbollah more willing to reach a negotiated settlement with Israel, while others suggested this would only increase Hamas’s resolve, making a negotiation less likely.Footnote 99 Further research can more systematically analyse these pathways and conditions under which they are more likely to be observed.
When is leadership arrest and killing more likely to be relevant for negotiations? As the previous literature suggests, leadership decapitation might be especially consequential when targeting founding leaders, given their disproportionate impact on groups’ inner workings,Footnote 100 and when targeting leaders with highly personalised leadership styles in poorly bureaucratised organisations, as it would be harder to fill the leadership vacuum.Footnote 101
Second, focusing on the effects of leadership decapitation on negotiations instead of military implications/operational capabilities may be especially relevant in civil war settings when the organisation has at least some public support and legitimacy to engage in negotiations. My argument would be less useful for small and covert underground organisations that the states are unlikely to negotiate with.
Third, as argued, arrest and killing can have many different implications. One problem with focusing on the leadership arrest is that there are not many instances of the arrest of a top leader. Yet if we relax the conceptualisation of leadership to include senior leaders of the organisation, this framework may advance our understanding of leadership decapitation as a strategy even more. For example, Jordan identifies 7 top and 33 upper-level Hamas leaders that Israel arrested, in addition to 3 top and 38 upper-echelon leaders that were killed.Footnote 102 Moreover, in many organisations, senior leaders establish their network of supporters and control different factions.Footnote 103 Thus, the arrest of senior militants may result in various dynamics, including co-optation, side-switching, and the defection of the leader and their faction, which future studies can investigate.
Another way to apply my framework (negotiation implications vs military implications) is by differentiating how conflicts conclude – inactivity, negotiated settlement, or rebel victory/defeat – and examining how various forms of leadership decapitation influence these outcomes.Footnote 104 Moreover, different trajectories of arrested leaders can be traced and compared. For instance, even after 25 years of his arrest, Öcalan remains an important actor.
Lastly, Krause notes,Footnote 105 ‘the growing practice and analysis of leadership decapitation of insurgent groups are based on the idea that there is a difference between leaders and foot soldiers, although MST (his theory) Footnote 106 suggests that we should analyze whether that difference is based on individual characteristics or the incentive structure of one’s position’. This remark would be equally true for arrested leaders who maintain their leadership from prison and their successors. I do not claim that individual characteristics do not matter. This study, however, prioritised how leadership arrest alters the actors’ incentive structures and strategic calculations.
In short, leadership decapitation – both killing and arrest – can influence negotiations through its first- and second-order effects, and studies of the effects of decapitation should take these dynamics into consideration.
Possible effects of leadership arrest on negotiations beyond the case study
‘Dead men tell no tales.’Footnote 107 Why does this matter? The covert nature of their operations makes insurgent and terrorist organisations, especially those who do not control territory, extremely vulnerable to defection and information leaks.Footnote 108 An information leak becomes ‘a potential ticket to arrest, detention, death, or an entire operation being foiled’.Footnote 109 Thus, an arrest makes it possible to interrogate the leader for valuable information. For example, Israel was able to crack down on the entire upper echelon of Hamas leadership after interrogating its founding leader, Ahmet Hassan, and his son.Footnote 110
Second and closely related, defection and recruiting insurgents as informants is a common dynamic observed in many conflict settings, and rebels often switch sides to survive or maximise their gains.Footnote 111 Everything else being equal, side-switching in the case of an arrested leader has an extremely valuable pay-off: freedom. Even if we assume that an arrested leader has the same bargaining leverage and control over an organisation as a free leader, an arrested leader is more likely to side-switch since the concessions they receive would include freedom. In short, arrest gives a leader a strong incentive to defect. Therefore, another consequence of arrest may be the individual defection of a leader.
Third, according to Otto, Scharpf, and Gohdes,Footnote 112 25 percent of all rebel groups changed their alignment towards governments at least once in their life cycle, most of which became pro-government. Thus, another possible outcome of a leadership arrest is the defection/side-switching of the leader and their faction. This would give the government a larger pool of informants and an experienced military force. For example, in Afghanistan, many insurgent leaders rely on their tribal and local networks and frequently change sides depending on the balance of power.Footnote 113 In Chechnya, insurgent leaders were coerced by the pro-Moscow Kadyrov clan, whose leader Akhmad Kadyrov was a former rebel, to switch sides. To do that, the Kadyrov clan arrested relatives of insurgent leaders and threatened the leaders with the killing of their captured relatives.Footnote 114 Thus, an arrested leader may be compelled to switch sides with their network to retain at least some power rather than spend years in prison. The defection of a leader, together with their network, increases the military capabilities of the state even more than individual defection, and thus the government’s bargaining leverage.
Fourth, another possibility is the co-option of arrested leaders by states to promote a pro-peace agenda or to advance the government’s narrative. During the Peace Process between the PKK and the Turkish government, Öcalan’s approval of the process and pro-peace messages were crucial to create public support for the negotiations. In 2019, the government tried to use Öcalan to win the local elections in Istanbul when Öcalan called for the Kurdish national movement to remain neutral between the opposition and the government in the Istanbul local elections.
Lastly and relatedly, the leader’s imprisonment and/or release can be important parts of the negotiations. From the government’s perspective, the release of a leader can be a sign of goodwill and an initiator of negotiations. The release of a leader can be a red line for both the government and insurgent organisations, preventing a negotiated settlement. In Turkey, both violent and non-violent wings of the Kurdish national movement frequently engage in protests and demand Abdullah Öcalan’s release.
Most recently, in a highly unexpected turn of events, Devlet Bahçeli, the leader of the far-right Turkish nationalist party and de facto coalition partner of President Erdoğan, called on Abdullah Öcalan to announce the PKK’s dissolution and the end of its terror campaign at the Turkish parliament in exchange for his freedom/parole. President Erdoğan called Bahçeli’s proposal a historic window of opportunity. According to most analysts, this marked the start of a grand bargain between the PKK and the government, over territories controlled by the PKK’s Syrian affiliate, the PYD, Erdoğan’s term limits and constitutional amendments, and concessions to the Kurdish national movement in Turkey.
On 30 December 2024, lawmakers from the Kurdish national movement visited Öcalan in prison for the first time in nearly a decade. During the visit, Öcalan reportedly stated, ‘I have the competence and determination to make the necessary positive contribution to the new paradigm that Mr Bahçeli and Mr Erdoğan have empowered.’Footnote 15 However, significant uncertainties remain. Questions persist about how Turkey and the PKK will address commitment problems, the extent of government concessions to the Kurdish movement, and whether PKK leadership and rank-and-file members will align with Öcalan’s position. Nevertheless, these developments, 25 years after Öcalan’s arrest, underscore the importance of understanding the impact of leadership arrest on negotiations.
Moreover, my interpretation of statements and interviews from various actors since Bahçeli’s remarks reveals contrasting perspectives: the legal Kurdish national movement in Turkey appears more hopeful and optimistic about this new process, while statements from PKK leadership reflect great reluctance and scepticism.Footnote 116 This further highlights the significance of multi-actor bargaining processes as a consequence of Öcalan’s imprisonment.
Similar dynamics can be found elsewhere. Marwan Barghouti, who was the leader of Fatah’s armed wing before he was arrested in 2002, remains one of the most popular Palestinian leaders. His release has been an important component of various negotiations between Israel and Palestinian groups, including the most recent negotiations between Israel and Hamas after the 7 October.Footnote 117
Although the African National Congress (ANC) was primarily a non-violent organisation, it did have a violent wing and was considered a terrorist organisation by the South African government. In 1985, the South African government offered Nelson Mandela and other leaders a deal in which they would be released from prison if they renounced violence, which was rejected.Footnote 118 Later, Nelson Mandela’s release from prison showed the South African government’s commitment to negotiations. All of these suggest that the arrest of a leader creates a new bargaining game between the state, organisations, and leader.
I do not test when leaders are influential from prison. However, we can expect that arrested leaders will maintain significant influence even from prison, particularly if they wield ‘charismatic’ authority and are popular figures among the constituency, as Öcalan, Barghouti, and Mandela’s cases indicate. Moreover, we can expect leaders to be more influential if they have established a personalised mode of governance, especially if a cult of personality is cultivated around them within the organisation. Prominent examples include Abdullah Öcalan of the PKK and Abimael Guzmán of Sendero Luminoso.Footnote 119 Such leaders often foster strong emotional bonds with their followers, centralise the movement’s ideology around themselves, and eliminate rivals. Relatedly, as Price discusses in greater length,Footnote 120 we can expect the leaders of value-based armed groups, as opposed to profit-driven groups, to be more influential. Their abilities are vital for creating a shared identity, myths, and symbols and for mobilising and inspiring the masses.Footnote 121 Thus, the extent of a leader’s influence from prison is likely to be determined by the level of charismatic authority, popularity among the constituency, personalised mode of governance, and the extent of the cult of personality surrounding them.
Second, and relatedly, this may be especially true for founding leaders, who often have disproportionate influence over the organisation’s strategy and ideologyFootnote 122 and are often charismatic and popular figures. This may also be true for organisations that fight insurgencies, as these organisations often reflect the grievances of their constituency and require at least some level of civilian support to sustain themselves, often involving governance activities.Footnote 123
Yet these factors do not mean an organisation must have a single powerful or highly charismatic leader for a leader to have influence from the prison. Many armed groups have multiple high-level ranking leaders who control different factions, leveraging distinct networks to maintain their power.Footnote 124 In such a scenario, we might be especially likely to see co-optation and side-switching of the faction the arrested leader controls. Overall, these cases indicate that scholars must take the leader’s role from prison more seriously and identify its consequences for the broader conflict and negotiation dynamics.
Conclusion
I have argued that the leadership decapitation literature focuses on the military-operational implications of leadership decapitation at the expense of its impacts on negotiations between insurgents and the state. Second, I have argued that leadership arrest and leadership killing are distinct phenomena that should be addressed separately. Moreover, I have argued that an arrested leader’s possible role in prison should be considered in the analysis. I have supported the plausibility of these arguments with a case study of Öcalan’s arrest.
By focusing on the negotiations between the PKK and the state, I challenged and qualified the conventional wisdom that sees Öcalan’s arrest as an example of a successful decapitation strategy. By focusing on negotiations instead of military capabilities of the PKK, I showed that Öcalan’s arrest offered a window of opportunity for conflict resolution in the short term by giving Öcalan incentives to de-escalate violence and engage in negotiations with the state. Moreover, Öcalan’s concessions resulted in organisational turmoil, which damaged the organisation’s military capabilities and, thus, Öcalan’s bargaining leverage. Through these dynamics, his arrest was productive in terminating the conflict in the short term.
However, the state signalled that it had no intention to engage in meaningful negotiations with Öcalan, and his concessions resulted in organisational turmoil. These dynamics gave Öcalan incentives to escalate violence and side with the hardliners within the organisation to bring the state to the negotiation table. Thus, his arrest became counterproductive for ending the conflict in the medium run. Lastly, his arrest created a dual leadership structure within the organisation, increasing the number of veto players the state had to negotiate with. Thus, in the long run, his arrest had a counterproductive result by complicating the peace process. These conclusions show that the literature may benefit from not limiting its temporal scope and focusing on the causal chain that leadership arrest triggers.
Broadly, it can be said that the literature can greatly benefit from considering the dynamic and interactive nature of armed conflicts. This also indicates that leadership decapitation is not an isolated phenomenon, but its consequences are contingent upon the state’s subsequent actions. After all, conflicts are at least dyads. The leadership decapitation literature can greatly benefit from civil war and rebel fragmentation/fractionalisation literature to understand these dynamics.
Lastly, my framework has important policy implications. The consequences of leadership decapitation are varied. It surely disrupts the stability within the targeted organisations, but its exact direction and implications are hard to foresee and are prone to unintended consequences. Thus, an analysis of the implications of leadership decapitation requires us, first and foremost, to think about the end goal. What is the purpose of leadership decapitation, after all? Is it to resolve the conflict through negotiations? Is it to militarily overwhelm the enemy so that it can no longer carry out attacks? Is it just to militarily weaken them and drop the violence to a bearable extent?
Overall, my point is that the exact causal chain triggered by leadership decapitation is highly complex and context-dependent. Thus, when assessing the effectiveness and usefulness of leadership arrest and killing, policymakers may benefit from assessing the likelihood of various mechanisms that can be triggered, some of which I have tried to outline in the discussion section.
Acknowledgements
A very early draft of this article was written as a final paper for Zeynep Kadirbeyoğlu’s class, and I am grateful to her for her feedback. A preliminary version was later presented at the Political Science Colloquium at Boğaziçi University. I thank all the participants for their valuable comments and am especially indebted to Mert Arslanalp for his initial encouragement and insights. I also thank the reviewers and editors whose suggestions significantly improved the manuscript.
Open access
