Abstract
The dynamics of conflict (de)escalation by social movements or political opposition groups have attracted cross-disciplinary interest among social scientists, but there remain several knowledge gaps to be filled. On the one hand, there is already extensive research on the shifts from unarmed expressions of collective grievances to the adoption of violent strategies by oppressed constituencies or ‘minorities at risk’, as well as on the transition from armed insurgency to negotiations, demobilization, reintegration and conventional politics. However, there is scarce scholarship on the phenomenon of armed groups shifting their conflict-waging strategies from violent to nonviolent means, especially in contexts which cannot be resolved by force but are also ‘unripe’ for conventional de-escalation methods through negotiation and political integration. This article offers a first attempt to fill this conceptual and empirical gap, by investigating the nature and the drivers of transitions from armed strategies to nonviolent methods of contentious collective action on the part of non-state conflict actors. It focuses in particular on the internal and relational/environmental factors which underpin their decisionmaking process, from a change of leadership and a pragmatic re-evaluation of the goals and means of insurgency, to the search for new local or international allies and the cross-border emulation or diffusion of new repertoires of action. This multilevel analysis draws from past research on various self-determination or revolutionary movements which fit the scope of analysis (i.e. Nepal, Egypt, Palestine, West Papua, East Timor, Mexico and Western Sahara). The article also points to the need for more systematic enquiry on these cases through in-depth comparative empirical analysis.
Introduction
The revival of global interest in the phenomenon of nonviolent struggle since the 2011 Arab Spring offers a welcome opportunity to revisit the potential of unarmed resistance as an alternative pathway out of armed conflicts, in cases where neither military/counter-insurgency nor negotiated solutions have succeeded. However, to elucidate the interactions between the dynamics of violent and nonviolent interactions in macro-political conflicts, one needs to adopt an encompassing approach to conflict transformation, which is still severely lacking in the policy and academic worlds alike. Indeed, scholars and practitioners or activists in the fields of social movements, nonviolent action, political violence and conflict resolution seem to be largely evolving in parallel, often in relative isolation from each other. For instance, most security studies and conflict resolution experts are unfamiliar with the rich scholarship and empirics on civil resistance, given their narrow focus on armed conflicts and their termination through military means or negotiated settlements. In turn, most nonviolent scholars tend to hold oversimplified views on the dynamics and nature of armed struggle and warfare. Finally, social movement experts usually direct most of their attention to radicalization processes (i.e. shifts from peaceful to violent contentious action), at the expense of de-escalation (or demilitarization) dynamics.
In such a context, the purpose of this essay is to explore one possible avenue for linking up these rich disciplinary approaches, by offering a multilevel framework to analyse the possible factors that induce non-state insurgent groups to evolve from armed strategies to nonviolent methods of contentious collective action. In other words, it reviews what we already know about actors moving from armed to unarmed conflict, and identifies promising hypotheses and empirical cases that need to be further investigated through in-depth qualitative enquiry.
The article is organized as follows: section one clarifies the main elements of distinction between violent and nonviolent conflict behaviour; section two elucidates the dynamics of transitions from one to the other, and suggests four main types of shift from armed to unarmed resistance; section three presents some internal and relational/environmental factors which might explain or influence such transitions, with brief illustrations from various cases worldwide; finally, the conclusion stresses the need for further empirical research to assess the comparative relevance of these complementary factors. 1
Armed vs. nonviolent resistance within the context of contentious collective activism
Armed and unarmed resistance might be introduced as alternative methods of popular mobilization for collective, organized, non-institutionalized, contentious and coercive action. According to McCarthy (1990), the range of possible responses to situations of (objective, material) conflict can be classified along four main categories: non-conflictual responses (e.g. avoidance); conflict resolution (e.g. dialogue, negotiation); institutional procedures (e.g. legislation, litigation, lobbying); and collective action. The latter in turn can be further sub-divided into nonviolent action, material destruction and collective violence. Similarly, the Minorities at Risk dataset (Gurr, 1993) categorizes three main forms of protest representing increasing levels of scope and intensity: nonviolent protest ranges from verbal opposition through political organizing activity to small and large demonstrations, strikes and rallies; riot ranges from scattered acts of sabotage to serious and widespread rioting and armed attempts to seize power locally; and rebellion ranges from political banditry to protracted civil war in which rebel military units have base areas.
The two main categories concerned here are those of nonviolent/unarmed and violent/armed protest. They share a number of similarities that are worth spelling out before emphasizing their distinctiveness. Firstly, both belong to the arena of collective political action, as opposed to individual acts of political dissent. Secondly, they are employed by organized (vs. spontaneous) opposition movements,
2
representing an oppressed minority or disempowered majority, engaged in a struggle against the structural violence of the state – characterized for instance by an exclusionary regime or a foreign occupation army or administration. Thirdly, both are non-institutional, in the sense that they operate outside the bounds of conventional political channels (Schock, 2003). Moreover, both belong to the larger phenomenon of contentious or disruptive politics: they violate or bypass the routine conflict resolution procedures of a political system (Bond et al., 1997; Tilly, 2003). These complementary elements are well summarized by Tarrow (2011: 7): The irreducible act that lies at the base of all social movements, protests, rebellions, riots, strike waves, and revolutions is contentious collective action. Collective action can take many forms … Most of it occurs routinely within institutions … Collective action becomes contentious when it is used by people who lack regular access to representative institutions, who act in the name of new or unacceptable claims, and who behave in ways that fundamentally challenge others or authorities.
Beyond these similarities, the main difference between armed and unarmed activism lies in the use – or absence – of direct (behavioural) violence, defined as the intentional infliction of physical damage to persons or property (Bond et al., 1997). 3 The means of coercion at stake are therefore quite distinct: violence primarily relies on physical force, while nonviolent action employs a wider array of mechanisms, from discrete manipulation or public coercion to demonstrative appeals (Bond, 1994: 63).
Based on these distinctions, violent conflict-waging strategies might be defined as ‘all collective attacks within a political community against the political regime, its actors – including competing political groups as well as incumbents – or its policies’ (Gurr, 1970: 3–4), ranging from sabotage to guerrilla insurgency, quasi-conventional warfare and terrorist attacks. Moreover, this article is mainly concerned with strategies of armed struggle by resistance and liberation movements that are organized in hierarchical, accountable structures of command, bound by formal or informal rules and regulations (Dudouet, 2009).
For its part, unarmed resistance (also alternatively referred to as nonviolent action or civil resistance) 4 can be defined as ‘organized popular resistance to government authority which – either consciously or by necessity – eschews the use of weapons of modern warfare’ (Zunes, 1994: 403). One could also describe it as an active and sustained collective engagement in resisting violence in all its forms (whether behavioural, structural or cultural). As noted above, a specific emphasis will be placed here on formally organized social movements, as opposed to loosely connected networks or seemingly spontaneous ‘people power’ protests. Historically, nonviolent struggles have been waged with various methods of contentious direct action. In a widely cited study, Sharp (1973) documented 198 different forms of nonviolent action, classified into three categories according to their strategic function: the methods of nonviolent protest and persuasion intend to voice peaceful opposition to a policy or a law, or to persuade others (among the opponent or the grievance group) to particular views or actions; methods of non-cooperation deliberately aim to restrict, discontinue, withhold or defy certain existing relationships; and methods of nonviolent intervention involve direct physical obstructions to change a given situation, either negatively (by disrupting normal or established social relations) or positively (through creative actions forging new autonomous social relations).
In reality, however, the boundaries between these various forms of struggle are very fluid. Movements are not always exclusively and explicitly armed or unarmed – many involve a combination of both. Moreover, their members might not hold a differentiated understanding of these two dichotomic clusters of methods. Referring to the South African case, Seidman (2000: 165) argues that ‘most activists over the years viewed violent and nonviolent strategies as more intertwined and complementary than contradictory’. Indeed, from 1961 to 1990 the African National Congress (ANC) defined armed activities and nonviolent mass mobilization as complementary ‘pillars of struggle’, alongside the political underground movement and the international campaign to isolate the apartheid regime (Maharaj, 2008). Many other cases share similar patterns. During the civil war in El Salvador throughout the 1980s, the Farabundo Martí National Liberation Front (FMLN) emphasized the combination of armed struggle with the political struggle of the masses (under the motto ‘a sea of guerrillas and organized people’) (Alvarez, 2010). Even so-called terrorist groups which are seen as heavily militarized have also emphasized the value of unarmed forms of struggle. During armed campaigns by the Irish Republican Army (IRA) in Northern Ireland in the 1970s and 1980s, militants also resorted simultaneously to nonviolent resistance (e.g. hunger strikes or ‘blanket’ protests in prison to demand recognition of their status as political prisoners) and institutional (e.g. electoral) activities. In the Basque Country as well, when the militant separatist organization Euskadi Ta Askatasuna (ETA) was formed in 1958, it organized itself around six branches, including two dedicated respectively to military and mass actions.
Nevertheless, the fact that these actors might not perceive the value of differentiating violent from nonviolent strategies does not preclude the possibility for external observers to clarify the conceptual and empirical boundaries between various types of resistance in order to better understand the dynamics of contention (Shock, 2003: 709–710).
Multidisciplinary insights into the dynamics of conflict (de)escalation
Having clarified the similarities and differences between armed and unarmed conflict-waging strategies, this section turns to the nature and dynamics of transitions from one to the other.
Conflict escalation
In the conflict resolution field, much emphasis has been placed on the various steps describing the progression of conflicts – and their levels of contention – as they rise and fall in intensity over time. In particular, the label ‘conflict intensification’ refers to the transition from latent to overt (nonviolent) conflicts (Fisher et al., 2000), while ‘conflict escalation’ describes an increase in the intensity and frequency of coercive and violent behaviour directed at the other party (Mitchell, 2011). For matters of simplicity, this latter term will be used here in reference to the shift from peaceful to violent expressions of conflict behaviour on the part of the state challengers (usually in mutual interaction with the reciprocal behavioural patterns of the authorities).
Conflict escalation dynamics have also been subject to scrutiny by social movement scholars. For instance, Della Porta & Diani (2006) spell out four main thresholds in the gradual progression from the least to most extreme forms of protest: the transition from conventional to unconventional politics; a shift to direct action techniques; a turn to illegal, but nonviolent acts; and finally, the adoption of violent activities. When it comes to labelling this latter threshold, the term ‘radicalization’ which stems from the field of social psychology has become widely used in recent years in reference to ‘changes in beliefs, feelings and behaviours in directions that increasingly justify intergroup violence … in the context of non-state challenge to state authority’ (McCauley & Moskalenko, 2008: 416). However, the term is usually employed in the context of individual (as opposed to collective) trajectories, and most authors also seem to associate it primarily with a cognitive – rather than behavioural – shift (Bartlett & Miller, 2012: 2).
Conflict de-escalation
If transitions from unarmed to armed struggles are relatively easy to define and qualify, de-escalation dynamics are much more complex and multidirectional. At the individual level, one generally establishes a distinction between the processes of disengagement and deradicalization (usually in the context of research on terrorism). The former refers to a behavioural shift whereby a person ceases his/her participation in violent activities, whereas the latter points to a social, psychological or ideological decommitment from extremism and violence (Horgan, 2009; Reinares, 2011).
With regard to behavioural disengagements from violent strategies at the collective/group level, the literature on political violence has gathered evidence on various types of conflict de-escalation. For instance, it has been argued that the ‘decline and demise of terrorist campaigns’ might be pursued through six alternative trajectories (Cronin, 2009): decapitation (through the capture or killing of the leadership), success (achieving the objective), failure (implosion, backlash or marginalization), repression (through the use of force), negotiation (transition towards a legitimate political process) and, finally, a reorientation to other forms of violence (criminality, insurgency, major war). Hard security approaches tend to prioritize military, policing or judicial instru-ments in order to defeat, repress and/or decapitate armed groups – either by targeting their leaders or enticing individual members’ defections. However, empirical evidence seems to suggest that political conflicts cannot be resolved through force only. Indeed, a quantitative study conducted by the RAND corporation shows that among 268 identified ‘terrorist groups’ that ended their activities after 1968, only 20 (7%) were defeated militarily. By contrast, 107 (40%) were dismantled through policing (i.e. decapitation), and 114 (43%) joined the political mainstream. While policing was mostly effective in cases of small terrorist cells, the most common trajectory for groups above 1,000 members was a conversion to unarmed politics (Jones & Libicki, 2008).
Types and examples of transitions from armed to unarmed resistance
Conflict demilitarization: From armed to unarmed resistance
In comparison with the various conflict (de)escalation dynamics highlighted above, this article focuses specifically on organized resistance/liberation movements that have undergone a decisive shift from (primarily) armed to (primarily) unarmed methods of revolutionary, pro-democracy or self-determination struggle. The term ‘demilitarization’ might most adequately qualify this shift, as it rightly emphasizes behavioural features related to the tactics and strategies employed by militants, rather than their goals or ideology.
Among all the literature surveyed for this article, a couple of studies were found to refer directly to the transition from ‘terrorism to nonviolence’ (e.g. Garfinkel, 2007); however, their use of the term ‘nonviolence’ is quite misleading, as they depict individual disengagements from activism and a turn to peace and dialogue promotion, rather than a strategic shift to nonviolent contentious action. Only one study (Jones & Libicki, 2008: 26) was found to refer explicitly to the phenomenon of armed groups shifting to nonviolent activism. But the few empirical examples it cites are limited to very minor militant cells in Western Europe, apart from a brief reference to the 19th-century anarchist movement turning from violence to nonviolence (including the advocating of general strikes) as the working class became more active and governments more tolerant of worker protests.
Main hypotheses derived from the literature and initial empirical exploration
Factors of transition from armed to unarmed resistance
In recent years, social movement scholars have predominantly sought to explain the lulls and lumps of cycles of protest through the prism of ‘causal mechanisms’. According to Tilly & Tarrow (2007: 29), mechanisms refer to ‘a delimited class of events that alter relations among specified sets of elements in identical or closely similar ways over a variety of situations’. However, most of the existing research focuses on identifying mechanisms that might explain escalatory processes from unarmed expressions of collective grievances to the adoption of violent strategies – rather than the opposite shift. In order to bridge this gap, this final section attempts to identify possible factors which might entice armed militants to revert (back) to nonviolent means of struggle – along four main levels of analysis: intragroup, group–society, group–state and group–international (see Table II).
Intraparty factors
A first set of variables can be identified at the intragroup level. They might be classified along three main subcategories: idiosyncratic and cognitive factors (with a particular focus on the individual features and motivations of leaders, i.e. those leading the decisionmaking process); in-group (horizontal) dynamics and socio/organizational processes; (vertical) relations with the group’s constituency and social surroundings.
Idiosyncratic and cognitive factors: Role of the leadership
One segment of the literature on terrorism, inspired by psychological-oriented approaches, focuses on the factors explaining individual disengagement from ‘terrorist groups’ (e.g. Bjørgo & Horgan, 2009; Garfinkel, 2007). However, their analysis is not so relevant for the purpose of this article, which is primarily concerned with collective shifts directed from the top leadership as opposed to defections by individual (often rank-and-file) members. One might indeed assume that in cohesive and hierarchical organizations, group decisions over matters of primary importance (such as major strategic shifts) are taken by elites, and then carried out by members down the chain of command in a disciplined and coordinated fashion.
This implies, firstly, that shifts in the identity and attributes of individual leaders might play a decisive role in determining the behavioural patterns of their movements. Such shifts include a change of leadership (i.e. new members taking leading positions to replace deceased or ‘retired’ commanders), or a generational change (i.e. existing elites becoming older and undergoing a ‘learning curve’), as different generations can develop different tastes for specific forms of action (Della Porta & Diani, 2006). The dynamics of the self-determination struggle in East Timor can be cited as a case in point: the guerrilla war launched by the Falintil (the military wing of the political party Fretilin) from 1975 to 1978 was brutally crushed by the Indonesian military. The taking over of the leadership by one of the few surviving commanders, Xanana Gusmao (the current prime minister of East Timor), led to a transformation of the features of resistance, as he established a National Council made up of three wings, including a youth-led Clandestine Front responsible for organizing nonviolent resistance operations (Stephan, 2006). The self-determination struggle in Western Sahara also illustrates the importance of generational shifts in initiating strategic transitions. Whereas the exiled leadership of the pro-independence movement Polisario is rooted in the tradition of national liberation through an armed vanguard, younger activists within the occupied territories have emphasized the need for tactical innovation and advocate a strategy of expanding civil resistance (Barca & Zunes, 2009).
Secondly, behavioural shifts might also be caused by a reassessment of the means and/or goals of the movement in the light of a constantly evolving environment. Political scientists have indeed identified three broad categories of leaders: ideologues have a predetermined agenda which conditions all their decisions and their relations with their constituencies; strategists also have set goals but adapt the means of pursuing them according to the context (political timing) and what the constituents will support; pragmatists adapt their goals and agenda to the expectations of constituents and the situation (Hermann & Gerard, 2009). The pragmatist model points to leaders’ readiness to revise the primary objectives, guiding ideology and the discursive frame of their struggle, which might in turn influence the choice of action. Social movement scholars have stressed the importance of values in shaping the repertoire of contention modes (Della Porta & Diani, 2006), and thus a change in the elites’ belief system is likely to impact on their favoured strategy (Burgess & Burgess, 1994). One illustrative example for this factor is represented by the Egyptian jihadist movement Jama’a Islamiya. Following armed campaigns which lasted more than 20 years and cost thousands of lives, the group’s leaders suddenly declared a unilateral cessation of violence in 1997 and called upon their followers to embrace peaceful approaches to reach their goal of establishing an Islamic state. According to Al Hashimi & Goerzig (2010), this call was prompted by a reinterpretation of doctrinal texts by Islamist scholars heading the movement.
For its part, the strategist model mentioned above is consistent with the rational choice theory, according to which decisions are taken according to instrumental rather than ideological motivations, by considering the expected utility (i.e. potential costs and benefits) of various tactics, and seeking to minimize costs and maximize benefits (Jones & Libicki, 2008: 24–25). In other words, movement leaders select the primary means of struggle according to a careful calculation of the comparative effectiveness of armed vs. unarmed strategies; following a period of critical introspection, they might lose their belief in the efficiency of violence in terms of bringing about change and reaching their sociopolitical objectives. A telling example comes from Nepal, where the Maoist movement shifted its main strategy of struggle against the ruling oligarchy in 2006, from guerrilla warfare to mass-based nonviolent insurrection (while keeping its armed structure and arsenal intact), which eventually forced the regime to the negotiation table. When asked about the reasons for such a shift, the movement leaders cite their motto ‘firm with principles and flexible with tactics’ (Ogura, 2008). 5
Scholars in the field of nonviolent action have identified possible factors that might explain the strategic superiority of civil resistance over armed insurgency (Zunes, 1994; Chenoweth & Stephan, 2011). They include, for instance, the ability of nonviolent resistance to raise domestic and international sympathy, to foster broad-based participation, and to encourage dissent and defections in the opponent’s base of support, all of which intensify the pressure on the opponent. In turn, repressive measures by the regime or occupier are less easily justified and are more likely to ‘backfire’, to the benefit of the unarmed opposition. Finally, government agents may be less concerned about the consequences of a compromise with nonviolent insurgency (as they appear less threatening), while the activists themselves might be concerned about the counterproductive impact of militarism and violence on the levels of democracy, unity and development in the post-struggle society. All these arguments concerning the strategic or instrumental superiority of nonviolent resistance share a common assumption: they assume that opposition movements are unitary, rational and omniscient actors that have full knowledge of all available options and can freely and objectively choose the most effective ones. Alternative views should also be considered, including the role of internal dynamics within these sociopolitical entities.
In-group organizational dynamics
The rational choice approach highlighted above does not explain seemingly irrational acts and decisions, nor does it capture all intraparty dynamics that eventually contribute to strategic choices. The composite-nature approach championed by Wendy Pearlman calls attention to the problem of the propensity to refer to non-state (armed) groups or social movements as coherent entities ‘and to explain their behaviour as the outcome of that entity’s pursuit of specific goals’ (Pearlman, 2010: 197). She suggests instead that the conflict behaviour of a given group results from the interactions between clusters of members (i.e. subgroups). Such in-group dynamics might be observed both at the horizontal and vertical levels. For instance, the former refers to changes in the decisionmaking structure according to the internal balance of power between military and civilian/political entities (and leaders). Some resistance or liberation movements are organized around distinct dual structures which simultaneously allow armed and unarmed forms of struggle. In some instances, political movements predate the formation of a military branch, and both are kept largely autonomous in order to allow the political front to carry out public advocacy, pursue electoral politics or conduct negotiations. Other groups (such as most Latin American guerrilla organizations) are rather structured as ‘political-military organizations’ with a combined leadership, often dominated by military commanders. A power shift among militants from these respective branches might affect the overall strategies pursued by the movement. For instance, sociopsychological approaches to terrorism (e.g. McCauley & Moskalenko, 2008) point to the phenomenon of group radicalization through within-group competition for the status of being ‘most radical’: a reversed dynamic can also be envisaged, whereby members favouring unarmed methods might induce the ‘moderation’ of more radical or militarist comrades. The case of the Nepali Maoists might illustrate this point. One could assume that the evolving relations between the party chairman, Prachanda, and his deputy, Baburam Bhattarai (who was demoted from the party leadership in 2005 and reinstated several months later – he is currently Nepal’s prime minister), had an influence on the strategic shift to nonviolent struggle the following year, as the latter has always prioritized the democratic struggle against the monarchy as opposed to nationalist anti-India warfare (Ogura, 2008).
At the vertical intragroup level, Pearlman highlights the interactions between three clusters of members, namely, elites, aspirants and mass-level actors; she contends that ‘the form and intensity of a movement’s insurgent activity is the product of a combination of top-down and bottom-up dynamics’ (Pearlman, 2010: 210). Others also emphasize the crucial role of a movement’s constituency or support base in influencing demilitarization processes. For example, Al-Hashimi & Goerzig (2010) argue that armed groups that rely on a strong network of followers are more likely to undergo strategic shifts in the face of hardship and repression (e.g. mass imprisonment and torture) endured by these supporters, as it imbues them with a sense of responsibility and duty towards their constituency – as illustrated by the case of Jama’a Islamiya. Likewise, insurgencies that are deeply imbedded in their socionational community are more likely to be affected by (and responsive to) feelings of ‘war fatigue’ and popular demands for a new strategy within their constituency (Dudouet, 2009). In Chiapas (Mexico), the EZLN launched an armed struggle in 1994 on behalf of the indigenous population with the aim of resisting state military and corporate incursions, before progressively transforming into an exclusively nonviolent insurgency. This shift was strongly affected by the movement’s inclusive organizational structure (i.e. assembly-based Indian tradition) and strong connections with local social activists (e.g. indigenous church), who recognized the legitimacy of its demands but spoke vehemently against armed struggle (Rovira, 1994). Finally, for rebel groups affected by large-scale desertions and demobilization of key members or supporters, the turn to mass-based unarmed resistance might become an attractive alternative to gain new members, as it allows for a wider range of active participants – the physical, informational and moral barriers to mobilization are much lower (Chenoweth & Stephan, 2011).
Relational and environmental factors
This subsection shifts the unit of analysis from resistance/liberation movements and their internal dynamics to the mutual interactions between these groups and their external environment. As emphasized by Tilly (2003: 20), the repertoires of contentious politics employed by social actors are transformed over the course of physical and symbolic interactions and encounters, at the interpersonal or intergroup level, in a series of reciprocal adjustments. A wide range of studies point to various relational (interparty) or environmental factors which might influence the behaviour of opposition groups, from the ‘political opportunity structure’ analysed by social movement theory, to the reciprocal dynamics between terrorism and state counter-terrorism policies emphasized by security experts, the concept of ‘ripeness’ promoted by conflict resolution scholars, and the transnational geopolitical dynamics favoured by international relations specialists. These complementary approaches should be looked at in more detail, as they might contribute to explaining strategic shifts from armed to unarmed resistance, even if their authors were primarily concerned with other conflict dynamics such as negotiation, disintegration, defeat or violent escalation.
Societal level
Firstly, armed groups operate within societal environments inhabited by a multitude of civil society organizations and political actors, all of whom are competing and/or cooperating with one another for political influence vis-à-vis the state. In this context, three possible demilitarization mechanisms might be considered. The first one is inspired by the theory of outbidding, according to which groups competing for support and resources tend to radicalize their goals and strategies in an attempt to ‘out-bid’ one another (Bloom, 2004). A ‘reversed outbidding’ pattern might be envisaged, inducing armed groups to resort to nonviolent struggle in order to emphasize their difference from more violent groups, and thus increase their comparative support and legitimacy towards a shared sociopolitical constituency. A second mechanism, that might be called ‘mirroring’, refers to the decision by a movement to adopt a strategy that has proved effective (in terms of attracting support or effecting change) when used by another actor. This might be illustrated by the case of the Palestinian movement Fatah. Following several decades of armed resistance (mainly from exile) against Israeli occupation of the West Bank and Gaza Strip, in the early 1990s Fatah leaders adopted a combined strategy of negotiations with Israel and conventional party politics within the Palestinian Authority. However, after having failed to take the lead in various nonviolent campaigns across the occupied territories including the first intifada (1987–92), the party finally embraced nonviolent resistance more systematically after its 2009 General Congress, and its leaders are increasingly supporting local civil resistance initiatives across the West Bank. One might argue that this decision was partly induced by the proved effectiveness of strategic peaceful protests launched by popular committees in various Palestinian villages along the route of the ‘separation barrier’ (Qumsiyeh, 2011).
Finally, ‘coalition building’ can also be a strong motivating factor for turning to civil resistance, when a group realizes that it helps forge tactical alliances with other sociopolitical forces against a common opponent – whereas armed resistance tends to push these towards the opponent. ‘The greater the possibility of widening their range of alliances, the greater the attention social movements will pay to the preferences of potential supporters’ (Della Porta & Diani, 2006: 187). The case of Nepal mentioned above might be relevant here: through a few weeks of peaceful ‘people power’ protests waged in alliance with conventional political parties opposed to the king’s autocratic rule, the Maoist insurgents achieved what they had failed to accomplish by themselves through a decade of armed struggle, namely, the establishment of a republic and the ongoing constitutional reform process (Ogura, 2008). Another example worth mentioning is the shift undergone by the Basque pro-independence movement. Following three failed attempts at negotiated settlement, the armed group ETA has demonstrated since 2010/11 its complete rejection of violent methods to concentrate exclusively on a combination of international advocacy, political action and mass-based nonviolent mobilization of a ‘popular wall of the people’ against state repression. The main reason invoked for this transition has been, apart from the realization that violence did not bring the movement closer to reaching its political goals, an interest in coalition-building across Basque social and political mobilization sectors (Aiartza & Zabalo, 2010).
State level
In the social movement literature, civil society mobilization is regarded as a ‘reaction from below’ to macro-political events. Notably, the political opportunity structure theory seeks to explain the lulls and lumps in the ‘cycles of contention’ (Tarrow, 2011) as well as the repertoire of tactics adopted by social movements in their different stages of development (Meyer, 2004) by the ‘shifting institutional structures and ideological dispositions of those in power’ (McAdam, 1996: 23). In addition, the resource mobilization theory postulates that material resources constrain strategic choices (Della Porta & Diani, 2006). One can thus assume that the level and nature of the state’s response to peaceful or armed challengers strongly affects the resources which resisters need to deploy to sustain their mobilization. It has been argued, for instance, that one of the factors contributing to the overall decline in armed insurgencies in the Third World since the 1980s is the dramatic increase in human and financial costs resulting from state counter-insurgency strategies: liberation movements have realized that ‘the benefits of waging an armed insurrection may not be worth the costs’ (Zunes, 1994). According to terrorism research as well, the wider array of tools used by governments to ‘end terrorism’ (from intelligence to targeted killings, collective punishment and criminal justice mechanisms) aims to make terrorism ‘too costly for those who seek to use it’ (Alterman, 1999).
Scholars have indeed dedicated extensive empirical research to the mutual interactions between social movements’ mobilization and state repression (e.g. Davenport, Johnston & Mueller, 2005). One of the most sophisticated arguments concerning the impact of the latter on the strategic response of the former is the ‘theory of dissent strategy’ (Cunningham & Beaulieu, 2010), according to which dissidents analyse past state behaviour with regard to level of repression to a particular tactic, and how consistently it was used, in order to decide which methods of dissent (whether violent or nonviolent) to use. According to the authors’ ‘substitution hypothesis’, a consistent use of repression against violent tactics will induce dissidents to shift the primacy to nonviolent action, whereas an inconsistent state response will encourage further violent action. They illustrate this through the various episodes of armed and unarmed contention during the civil rights movement in Northern Ireland from 1963 to 1976.
Other strands of literature emphasize the dynamics of power relations between conflict parties, and their impact on escalation or de-escalation processes. As argued by Kalyvas (2006), violence in any civil war is a function of the distribution and degree of control of the contenders over the territory in conflict, and how it conditions the behaviour of civilians and authorities. As noted earlier, state actors and their challengers are more likely to engage in negotiations when they have reached a situation of relative power balance. One could then argue that in a situation of persisting (or increasing) power imbalance in favour of the state, insurgents might be enticed to experiment with innovative strategies in the search for tactical advantage (Tilly, 2008), leading to the adoption of nonviolent struggle as an alternative method of asymmetric conflict. For example, Palestinians’ shift from negotiation to ‘popular resistance’ mentioned above was partly prompted by the realization that power relations were detrimental to a negotiated approach, while violent strategies would further increase the prevailing asymmetry (as argued, for instance, in recent press statements issued by imprisoned senior Fatah leader Marwan Barghouti). 6
International level
Finally, geopolitical dynamics in the regional or international environment (such as cross-border relations or broader global trends) might affect the resource mobilization, and thus the strategic choices, made by conflict stakeholders. On the one hand, the sustainability of armed struggle might be heavily dependent on the existence of foreign support bases to raise financial and diplomatic support, or to be used as a safe sanctuary for fighters. Global events, such as the end of the Cold War and more recently the 11 September 2001 attacks and subsequent US-led ‘war on terror(ism)’, have resulted in a much more unfavourable terrain for violent insurgency strategies. While the global appeal of old revolutionary models (e.g. Cuba and Vietnam) is dwindling among liberation movements and transnational solidarity networks alike, the reconfiguration of global alliances has also caused a decline in state sponsorship for armed rebellions. Tilly & Tarrow’s (2007) mechanism of ‘decertification’ illustrates this phenomenon: it refers to an external authority’s signal of its readiness to withdraw its recognition and support from a political actor. In such contexts, opposition groups need to adopt strategies more likely to attract international sympathy and support. In Chiapas, the Zapatistas turned to civil resistance because their initial local supporters demanded it (as argued above), but also in order to deepen and broaden outside backing: nonviolent tactics such as media campaigns, demonstrations, and creating parallel sociopolitical institutions were particularly effective in attracting support from moderate peace, social justice and development NGOs both at home and transnationally (Bob, 2005: 148).
These movements might also be influenced by alternative models in their immediate or regional surroundings that demonstrate the effectiveness of nonviolent revolutions. According to Merriman & Duvall (2007), increased accurate coverage of nonviolent movements in global media might offer powerful incentives for other actors to adopt similar methods. For instance, it would be worth examining the effects of the Arab Spring on the dynamics of liberation struggles in neighbouring Palestine or Western Sahara, or the impact of the successful turn to civil resistance in East Timor’s independence struggle on other self-determination movements against Indonesian rule in Aceh or West Papua. At the global level, an evolution in international ethical or legal norms favouring nonviolent strategies of resistance might also have an impact on the strategic or tactical choice made by political opposition movements worldwide.
Finally, these dynamics might be also strengthened by the cross-border transmission of techniques and skills between activists. It is assumed that social movements tend to use the means with which they are most familiar (Merriman & Duvall, 2007); according to Della Porta & Diani (2006), existing repertoires of contention tend to be reproduced over time, as the forms of actions used in one protest campaign are recycled in subsequent ones. But innovation might be introduced through a combination of mechanisms such as ‘brokerage’ (the production of a new connection between previously unconnected or weakly connected sites), ‘diffusion’ (the spread of contentious performances from one site to another) or ‘emulation’ (deliberate repetition within a given setting of a performance observed in another setting) (Tilly & Tarrow, 2007). Nonviolent action training workshops or educational materials provided by organizations such as the Centre for Applied Nonviolent Action and Strategies (CANVAS), the International Center on Nonviolent Conflict (ICNC) or Nonviolence International aim to do just that: disseminate knowledge, tools and equipment so that successful examples of strategic nonviolent action might be emulated by other movements engaged in contentious collective action worldwide. For instance, in the West Papuan liberation struggle, which after 1998 underwent a dramatic transformation from a poorly armed and factionalized guerrilla warfare waged in the jungles and mountains to a numerically large unarmed civilian-based movement active in the cities and towns, activists have acknowledged the role of training and education workshops in learning from transnational experiences about the superior effectiveness of nonviolent strategies over armed activities (MacLeod, 2012). Such activities, as well as other forms of external support to nonviolent pro-democracy or self-determination movements, have been well documented and critically assessed in recent years (e.g. Dudouet & Clark, 2009; Johansen, 2010), but since such support has mainly been directed towards emerging social movements and civil society actors who never engaged in violent activities in the first place, the question of what impact it has had so far on armed opposition groups and their leaders remains to be answered by further research.
Conclusion
Based on working definitions of armed/violent and unarmed/nonviolent contentious collective action, this article has argued that the existing literature on social movements and civil war dynamics is mainly concerned with describing and explaining conflict escalatory dynamics from the latter to the former. By contrast, a reverse phenomenon has been stressed here, namely, the behavioural shift from armed to nonviolent resistance by revolutionary, pro-democracy or self-determination movements. Based on a preliminary study of key empirical examples, a number of plausible mechanisms or drivers of change have been identified in order to account for the occurrence and timing of such transitions. These include intragroup factors such as generational change or internal power shifts within the movements’ leadership; their pragmatic re-evaluation of goals leading in turn to a shift in the primary means of struggle; a demand for new strategies on the part of their support base (experiencing war fatigue and bearing the brunt of state repression); and the search for opportunities to expand the group’s membership. Additional factors have been presented at the relational (interparty) and environmental levels, including the search for sociopolitical allies against a common enemy, a critical (re)assessment of political opportunities and power relations vis-à-vis the state apparatus, the loss of external ‘patrons’ or the search for international supporters, and finally tactical innovation through cross-border emulation or the diffusion of new repertoires of action.
This multilevel framework should be seen as a list of research hypotheses which remain to be tested in a more thorough fashion. Indeed, in order to assess the empirical relevance of the listed factors, one should conduct an in-depth comparative analysis of the cases identified in this article, by interviewing leaders and members of the movements under scrutiny with regard to their subjective self-assessment of the primary motives that drove their decisionmaking process, and cross-checking this information through objective external analysis by local experts on additional factors and influences at stake. Beyond its intrinsic scholarly relevance, such research would also strongly benefit the policy world: its findings might be disseminated towards contemporary armed insurgency movements in order to counter their disbelief in the efficiency of nonviolent resistance and eventually persuade them to adopt more constructive political strategies; finally, it is also hoped that such research will help to break conventional stereotypes associated with the ‘war on terrorism’ and to identify avenues for constructive policy engagement with sociopolitical movements mobilizing around deep-rooted grievances.
Footnotes
Acknowledgements
The author would like to thank her research assistant, her collaborators for the forthcoming book on the same topic, and the anonymous reviewers for their very helpful contributions to this article.