Abstract
Research has shown the potential of nonviolent civil resistance in challenging autocratic state regimes (e.g. Sharp, 1973; Chenoweth and Stephan, 2011). Yet, little is known about its applicability in jihadist proto-states, that is, territories governed by militant jihadist groups. We argue that civil resistance is more likely to occur when jihadists impose a rule that local populations perceive as alien and when organizational structures capable of collective nonviolent mobilization are activated. We develop this argument through a comparative analysis of three jihadist proto-states: one in which manifest and organized civil resistance occurred (Islamic Emirate of Azawad in Mali in 2012), and two in which it did not: the Islamic State of Iraq (2006–2008) and the Islamic Principality of al-Mukalla in Yemen (2015–2016). Whereas the former was met with mainly armed resistance (the Sunni Awakening campaign), the latter saw neither armed nor unarmed organized and collective resistance by locals under its rule. We demonstrate how variation in the jihadists’ governing strategies (especially the degree of adaptation to local conditions) as well as in the social structures for mobilization (i.e. whether opposition was channeled through civil society networks or tribal networks) created different conditions for civil resistance. This study adds to a growing research discussion on civil resistance against rebel governance (e.g. Arjona, 2015; Kaplan, 2017). More broadly, our study is an innovative first attempt to bridge research on terrorism, rebel governance, and civil resistance, three fields that have been siloed in previous research.
Introduction
Recent years have seen a significant rise in the number of radical Islamist groups with territorial aspirations who seek to establish so-called “jihadi proto-states” (Lia, 2015: 31), or “terrorist semi-states” (Honig and Yahel, 2019: 1210). After the Arab Spring, many jihadist groups have expanded into the civilian realm, transforming themselves (at least partly) from underground terrorists into socially embedded insurgent groups. They have moved to declare proto-states on an unprecedented scale: Between 2011 and 2016, jihadist groups created more proto-states than in the preceding 20 years (Lia, 2016: 81). The creation of such proto-states is part of a larger trend of a dramatic increase in armed conflicts over transnational Islamist claims (Svensson and Nilsson, 2018). The Islamic State in Iraq and Syria (ISIS) and its diffusion to other areas of the world indicate that jihadist proto-states are a serious contemporary security issue.
In this study, we suggest that there is space for civil resistance against jihadist proto-states, something that has received little attention in policy-making and research. Where jihadists try to govern, they need civilians to accept their rule. Civilian populations can, however, withdraw their consent, mobilize, and resist jihadist rule. A fundamental insight from the study of civil resistance is that power ultimately relies on dependency relationships outside the power-holders’ control (Sharp, 1973). Yet, this insight remains to be incorporated in the study of jihadist proto-states. A growing body of work has shown the potential of nonviolent resistance to challenge unjust rules and authoritarian systems. Nonviolent means outperform violent means (Chenoweth and Stephan, 2011) and increase the chances for democracy in the longer term (e.g. Karatnycky and Ackerman, 2005). After previous research predominantly focused on civil resistance against established state regimes, there has been a growing interest recently in civil resistance against rebel governance (Arjona, 2015; Barter, 2014; Hallward et al., 2017; Idler et al., 2015; Kaplan, 2017; van Baalen, 2020). Yet it has largely avoided the study of civil resistance against jihadist rebel governance projects. Thus, with a few exceptions (Aarseth, 2018; Revkin, 2016, 2020; Stephan, 2015), there is next to nothing when it comes to empirical research on civil resistance against jihadist proto-states. These exceptions are important as they have identified the potential of civil resistance (Stephan, 2015), demonstrated how ISIS’s governance impacted civilian populations (Revkin, 2016, 2020), and described how civilians in Mosul engaged civil resistance against ISIS in acts of noncooperation (Aarseth, 2018), but they have not examined conditions for civil resistance.
Our study addresses this research lacuna—located in between the study of terrorism, rebel governance, and civil resistance—and is the first to ask why civil resistance occurs against some jihadist proto-states but not others. We advance an argument in which the interaction between the jihadist groups’ governance strategies, and which (if any) of the local social structures for political mobilization that are activated, account for the popular response to jihadist proto-states. More specifically, we suggest that civil resistance against jihadist proto-states is more likely to occur when the jihadists impose a type of rule perceived as alien by local populations, thus generating grievances and motives for resistance, and when civil society organizations capable of collective nonviolent mobilization are activated to mobilize this potential for resistance.
We demonstrate the feasibility of our argument by conducting a comparative analysis of three jihadist proto-states. We compare the case of Mali, where there was manifest and organized civil resistance against the Islamic Emirate of Azawad in 2012, with two cases without civil resistance: the Islamic State of Iraq (2006–2008) and the Islamic Principality of al-Mukalla in Yemen (2015–2016). Whereas the former was met with mainly armed resistance (the Sunni Awakening campaign), the latter saw neither armed nor unarmed organized and collective resistance by locals under its rule. In line with our theoretical explanation, we argue that civil resistance against the Islamic Emirate of Azawad can be explained by the active role that Malian civil society, including women organizations, took in mobilizing the grievances that arose from the jihadists’ forceful implementation of their rule, which included strict sharia law and hudud punishments. al-Qaeda in Iraq’s (AQI) rule was equally strict, but civil resistance did not occur because preexisting social structures for mobilization was mainly tribal networks, whose gender norms, customs for dispute resolution as well as access to weapons, readily mobilized militias, and extensive external support favored violent resistance over nonviolent resistance against the jihadists. By contrast, al-Qaeda’s more careful approach to governing in Yemen, which rested on indirect (rather than direct) control through a ruling council staffed with local dignitaries and bureaucrats in Mukalla, prevented the rise of strong motives and networks for resistance. One reason why they were able to do this was that they were more locally based and thus better able to adjust their governance to local conditions. Also, they were able to learn from their own previous ruling experience.
Research on civil resistance against jihadist proto-states could shed light on the role of civilians living under jihadist rule, thus taking civilian agency seriously in these situations. Before we can evaluate whether civil resistance could be an untapped potential for countering jihadist rule, we first need to understand why it occurs. Moreover, although jihadist proto-states in many ways resemble other examples of rebel governance, there are important differences between these types of rebel governance, which motivates the study of jihadist rebel governance as a particular subcategory of rebel governance. Jihadist rebel governance, as all rebel governance, seeks to establish control over territory, uses violence to assert power, depends on at least some degree of popular consent, and varies with regard to commitment to ideology and effective governance. Jihadist proto-states are different primarily because they are local manifestations of a global jihadist movement with transnational territorial ambitions. For example, as will be seen below, Abu Musab al-Zarqawi and later leaders of AQI saw Iraq as one arena for restoring the caliphate and spread global jihad. This feature distinguishes jihadist proto-states from most other rebel governance projects (such as the Revolutionary Armed Forces of Colombia (FARC) in Colombia or the Moro Islamic Liberation Front (MILF) in the Philippines) whose territorial ambitions are typically confined to a particular subnational region and this particular feature of jihadist rebel governance is, we surmise, important to take into account when understanding conditions for civil resistance to occur.
Theoretical framework: explaining civil resistance in jihadist proto-states
There is a growing scholarly debate on why civil resistance occurs in some cases, but not in others (Butcher and Svensson, 2016; Karakaya, 2018; Schaftenaar, 2017). Previous research has identified four clusters of factors that help explain outbreaks of nonviolent conflict: grievances, political opportunities, modernization, and resource mobilization (Chenoweth and Ulfelder, 2017). Our argument draws mainly on two of these factors: grievances and resource mobilization.
Grievances provide motivations for dissent and may stem from longer-term structural discrepancies between expectations and performance of a political and economic system. Such discrepancies can lead to “relative deprivation,” or to more immediate triggers of resistance, when civilians experience a behavior from a power holder as ethically and politically outrageous, leading to a moral shock (Cederman et al., 2013). Previous research has shown that repression of political opposition can backfire and escalate grievances, thus strengthening rather than weakening motives for mobilization, depending on the nature of repression as well as the ability of the opposition to delegitimize the rulers and maintain resilience (e.g. Francisco, 2004; Davenport, 2007).
Yet, grievances are not enough to explain resistance, as not all actors with grievances decide to collectively resist. Dissenting voices also need to have the capacity to resist. Human, economic, or other types of resources conducive to mass mobilization have thus been highlighted as other important explanations for why civil resistance occurs (Tilly, 2003). Of crucial importance are the social structures and the existence of a mobilization infrastructure, which can help to overcome collective action problems, pool resources, and share information (Lichbach, 1998; McAdam, 1986). Movement entrepreneurs who can utilize preexisting organizational structures are in a better position to organize popular mobilization against power holders (Butcher and Svensson, 2016). Here, civil society institutions can play an important role by providing “safe spaces” that help to overcome the coordination and collective action problems (Nepstad, 2011).
Drawing on a grievances and resource mobilization framework, we argue that civil resistance against jihadist proto-states can be explained by the interaction between the ruling jihadist group’s governance strategy and the type of social structures for political mobilization that are activated in a certain locality. More specifically, we argue that civil resistance against jihadist proto-states is more likely to occur when two interacting factors are present. First, civilians will tend to resist when the jihadists forcefully impose their rule on local populations, thus generating grievances and motives for resistance. This grievance is greater the larger the ideological and cultural distance that exists between jihadists groups and the preexisting local customs (a perceived foreignness). Second, civil resistance is more likely when civil society organizations are the main social structures that are activated to mobilize the potential for resistance. By contrast, our argument implies that when jihadists are able to decrease the perceived foreignness and adapt to local customs and practice or when other social structures, particularly networks with access to arms and with norms of violent conflict management, are channeling the resistance, civil resistance is less likely to occur. In line with previous research, our argument thus suggests that grievances provide motives for resistance whereas civil society organizations turn those grievances into actual resistance.
Jihadist armed groups forcefully impose their rule when they employ a nonpluralist approach to governance that removes power from previous power holders and concentrates it in the hands of the jihadists, and when their implementation of sharia law and hudud penalties contradicts local traditions. Kalyvas (2018: 44) highlighted that “their rule is often highly interventionist, clashes with established local norms and practices, [. . .], and generates considerable popular opposition and resentment.” Indeed, jihadist armed groups often struggle to gain local legitimacy for their globalized, revolutionary ideology (Drevon, 2017).
One reason why jihadist groups are sometimes perceived as alien relates to the global nature of the jihadist movement. As described by Kalyvas (2018: 42), “A key feature that sets jihadist groups apart from many other rebel groups is their transnational dimension: they are part of a broader transnational social movement.” Many contemporary jihadist groups are part of this movement in which they exchange ideas, resources and fighters (Hafez, 2003; Hegghammer, 2010; Moghadam, 2008). Although there are internal differences and disagreements, some commonalities exist within the movement regarding, for instance, the legitimization of violence and the principles of governance as based on strict sharia law (Crenshaw, 2017; Gerges, 2005; Mendelsohn, 2016).
As argued by Lia (2015), jihadist proto-states typically share four traits. First, they are committed to governance: in addition to their military activities, they spend significant resources on delivering various civilian services. Second, they are deeply ideological projects. They follow the ideological imperative to apply sharia law and wage jihad, which often takes the form of vigorously oppressing minorities, implementing physical punishments, and razing “un-Islamic” shrines, tombs, and monuments. Third, they are local manifestations of an essentially international movement. Their leaders try to attract foreign fighters, endorsement from foreign religious clerics, as well as funding and material support from external constituencies. Fourth and related to the third point, disregarding current international borders and aiming to create a new world order, they are aggressive toward neighboring states and seek territorial expansion (Lia, 2015).
The global appeal and reach of the jihadist movement has been its main asset. It has allowed the movement to transcend national borders and create unique synergies between local theaters of conflict. Militarily, the ability of groups and individuals to resurrect themselves in other parts of the network has made the movement extraordinarily resilient (Gerges, 2016; Toft and Zhukov, 2015). Yet the globalized character of the jihadist movement can also create problems, by generating tensions in the local contexts where the revolutionary, globalized ideology manifests itself in the jihadists’ institutions, governance, and social practices. Thus, as demonstrated in the case of Chechnya, the transnational character of Islamist groups can also be a weakness, as it risks alienating the groups from local customs (Bakke, 2014). Jihadist rebel groups are stuck in a governance dilemma: on the one hand, a strict implementation of their radical program would appeal to their international constituencies, potentially generating economic and others forms of support. On the other hand, it can alienate the locals living under their rule.
So why do jihadist groups vary in their governance strategies? Acknowledging that there are broader discussion on what explains different governance strategies in de-facto states (e.g. Florea, 2020; Stewart, 2018), we here suggest that this largely is a function of their internal learning process. Groups that have more extensive local networks, are more locally anchored, and that have previous experiences of governance would overall be more likely to be attentive to local concerns. Local adaptivity is a process unfolding over time and driven by their reliance on local pillars of support.
The second tier of our argument deals with the type of social structures that are activated to mobilize collective resistance. Although there are many social groups and structures that can be used for mobilization, we focus on two main forms: civil society and tribal networks. We suggest that an active civil society increases the propensity for civil resistance because it provides the necessary precampaign mobilization infrastructure. We define “civil society” in line with previous research: “Civil society is understood to comprise organizations that take voluntary collective action around shared interests, purposes, and values and that are distinct from those of the state, family, and the market” (Paffenholz, 2014: 70). While some definitions of civil society would include traditional community networks, such as tribes in the concept of civil society (for a conceptual overview, see Spurk, 2010), we instead see them as a different type of social structure based on family linkages and boundaries, thus falling outside civil society. Civil society organizations provide the essential qualities needed for mobilization, in terms of awareness building, coordination, information sharing, and experiences of the strategies of civil resistance. Societies with stronger civil society structures are in general better equipped for nonviolent mobilization (Butcher and Svensson, 2016). Previous studies have suggested that nonviolent uprisings require organizational structures for mobilization, as civilians need them to coordinate and organize nonviolent action (Chenoweth and Ulfelder, 2017: 305). Local level civil institutions can help explain group-level nonviolent responses to armed actors (Kaplan, 2013). Second, civil society structures are less gender-biased and thus tend to appeal to a broader, more inclusive pool of people. Previous research has linked higher degrees of gender equality (Schaftenaar, 2017) and inclusive gender ideologies among organizations with the choice of peaceful forms of resistance (Asal et al., 2013), whereas gender inequalities are associated with higher risks for civil violence (Caprioli, 2005; Melander, 2005). Because more vibrant civil societies are generally associated with more inclusive gender norms, there are greater pools of potential activists and lower barriers to inclusion. Civil society organizations therefore have specific advantages for nonviolent mobilization of resistance. Other forms of social organization can also be potent forms of mobilization, but may also shape the nature of such resistance. The most relevant alternative in the contexts we examine here are tribal networks, and these tend to have features that increase the relative prospect that resistance will take violent forms. Compared to civil society structures, tribal structures commonly use practices of violent conflict management. The tribal value system, while also including justice and forgiveness, does rely on vengeance as a central component (Fraihat, 2016: 251). Tribal networks, because of their traditional role, also tend to provide a rather limited role for women, a group of the population that tends to have a critical role in civil resistance (e.g. Chenoweth, 2019). And importantly, tribal networks often have access to weapons, especially in situations where the state is weak, and thus have a higher capacity for violence.
Whether resistance is mobilized through civil society channels depends on the relative influence and strength of this type of social organization vis-a-vis other forms. If other forms of social organizations are empowered, for example, by external actors, then the comparative feasibility of civil society organizations to organize resistance will be smaller.
Research design
Civil resistance is understood as organized and collective manifest nonviolent action aimed at challenging a status quo or defending against a violent attempt to overthrow an existing order (Sharp, 1973). Civil resistance, in this article, thus excludes action that is private (or of a more hidden nature) rather than public and manifest, individual-based rather than collective, and violent rather than nonviolent (Nepstad, 2011; Schock, 2013). Examples of included events are large-scale public protests, acts of disobedience, demonstrations, and strikes. In line with previous research, we require that there must be a series of such events rather than one isolated protest (Chenoweth and Lewis, 2013). We also require that the civil resistance aims directly at the jihadist proto-state, rather than at a government for its lack of action against the jihadists. The other central concept here, jihadist proto-states, is defined as territories that have been declared as emirates, Islamic states, or caliphates, and that are controlled and governed by militant jihadi groups (see Table 1). Our terminology and definition draw on the work of Lia (2015).
In this study, we pursue a most-similar system approach and case selection is a critical part of this type of comparative analysis (e.g. George et al., 2005). Departing from the data collected by Lia (2015) on 19 jihadist proto-states since 1989, we selected three cases for comparison: one jihadist proto-state that faced civil resistance and two that did not. Additionally, we wanted the two cases that did not experience civil resistance to display one each of the other two possible sub-variations in the dependent variable. That is, we selected one case that saw violent collective resistance and one case that saw no collective resistance. This allowed us to capture all three possible variations of resistance, although the main outcome of interest here is the onset of civil resistance in jihadist proto-states.
In order to identify our cases, we first established three scope criteria, all coded by Lia (2015). First, to ensure a minimum level of establishment, the proto-state had to have lasted for at least 6 months.1 As a result, proto-states such as the Caucasus Emirate that lasted only 1 month (October 2007) were excluded. Second, the jihadist groups had to have had territorial control over their proclaimed proto-states. This criterion ruled out, for example, The Islamic Republic of Imbaba run by al-Gama’a al-Islamiya in Egypt, whose territorial control was too limited. Third, we included only proto-states that had civilian institutions, in order to analyze cases that bore as much similarity as possible to “normal” states. As a result, cases such as The Islamic Emirate of Kunar in Afghanistan (1989–1991) and the proto-state of Fatah al-Islam in Lebanon were excluded.
These scope criteria left us with 12 possible cases; for each, we conducted a search of several leading news outlets using Factiva in order to determine whether they faced civil resistance as defined in this study.2 Table 1 provides an overview of these cases (Al-Qaeda in the Arabian Peninsula [AQAP] appears twice because the group declared two separate proto-states in Yemen). If there were more than 10 separate events of nonviolent protests, we specified this as Extensive; if less than 10 but more than one, Occasional. From this distribution, we then selected one case from the first category and two from the other that bore as many similarities to each other as possible. We thus selected the proto-states proclaimed by al-Qaeda’s different regional branches: al-Qaeda in the Islamic Maghreb (AQIM; including its governing partners Ansar Dine and Movement for Oneness and Jihad in West Africa [MUJAO]), AQI, and AQAP, because of their similarities in terms of underlying ideology through their ties to one central organizing actor, al-Qaeda Central. Several of the other groups (e.g. Taliban, al-Shabaab, and al-Nusra) also have links to al-Qaeda Central, but in a more loose and autonomous way, whereas some (such as GIA and Ansar al-Islam) are unaffiliated with al-Qaeda and more stand-alone groups.
Having selected these al-Qaeda branches, we consulted a range of qualitative sources to determine which two cases displayed the clearest subvariation in terms of not having had civil resistance. Here we eventually settled on AQI and AQAP/Ansar al-Sharia (2015–) as the reference cases for our main case of interest, the Islamic Emirate of Azawad in Mali. These two cases stood out because the former faced collective violent resistance, the so-called Sunni Awakening, whereas the latter faced neither armed nor unarmed collective resistance. We decided to focus on the second proto-state by al-Qaeda in Yemen because the first lacked a clear outcome on the dependent variable. Although locals set up a council in the city of Jaar to negotiate with AQAP and Ansar al-Sharia (Amnesty International, 2012), these initiatives were instances of negotiation rather than resistance as such. We thus decided to exclude this case because it was too difficult to place it in one of the categories of civil resistance, armed resistance, and nonresistance.
Because of their structural similarity, this case selection enables us to hold constant four explanatory factors for civil resistance against rebel governance. First, in rebel governance literature, territorial control is often seen as a key dimension of rebel governance (e.g. Arjona, 2014). In all three cases studied here, the jihadist rebel rulers had territorial control (this is part of our scope condition, as outlined above). Second, civilians could potentially have more space to raise their voice when there is competition between armed actors over the control of the population (Kalyvas, 2006: 13). Yet that explanation would have challenges to explain the variation in outcome, since all three cases occurred in a context of rivalry amongst different rebel groups. Third, political opportunity theory would lead us to expect that lower levels of repression would increase the chance of civil resistance (e.g. Tarrow, 2011). However, this will not help to explain the variation in civilians’ strategy: both in Iraq and Mali, the levels of brutality were high, but the strategies selected different, and the case with the lowest level of repression experienced the least of civil resistance. Fourth, previous civil resistance research has argued that probability of civil resistance is largely a function of societies’ degree of modernization (e.g. Lipset, 1959), globalization (Karakaya, 2016), and industrialization (Butcher and Svensson, 2016). Clearly, this modernization argument cannot help explain why civil resistance occurred in the low-developed country of Mali, but not in similarly low-developed Yemen. Nor can it explain why civil resistance did not occur in Iraq, the most modernized, globalized, and industrialized of the three cases. We acknowledge that while we can hold constant these four explanatory factors on the structural level, the degree and type of territorial control and rebel rivalry can still hold explanatory power, something we return to in the discussion.
Empirical analysis: explaining variation in resistance to jihadi rule
In this section, we provide brief backgrounds on the cases and then turn to the comparative cross-case analysis. The sources for our analysis are triangulated, so in extension to the news-based reports, the analysis here also is based on reports from international NGOs and secondary sources. We argue that civil resistance occurred in Mali because local civil society organizations were able to turn strong grievances into collective political action. Thus, in addition to grievances, there were sufficient resource mobilization capacities. In Iraq, communities had strong grievances too. Yet, there, mobilization of dissent occurred through tribal structures whose customs for conflict resolution and availability of arms were more conducive to violent resistance. Finally, in Yemen, AQAP’s relatively accommodative and effective governance seems to have prevented the rise of strong grievances and thus clear motives for resistance.
Islamic Emirate of Azawad, 2012–2013
In early 2012, the jihadist Ansar Dine group and the secular MNLA took control over Kidal, Gao, and Timbuktu in Northern Mali. But MNLA soon split from this coalition as Ansar Dine allied with AQIM and AQIM splinter group MUJAO. Together they proclaimed the Islamic Emirate of Azawad in the conquered territories in March 2012 and introduced harsh sharia law and hudud punishments (Lebovich, 2013; Lia, 2015; Roggio, 2013; Stuster and Noble, 2014). Women were forced to wear the hijab, otherwise they would be whipped. Individuals accused of theft and looting were amputated or executed, and unmarried couples stoned. The jihadists also banned alcohol, tobacco, music, television, dancing, and sports. These laws were in line with Ansar Dine’s goals, which were to enforce sharia throughout Mali (Flood, 2012). AQIM and MUJAO had transnational ambitions. AQIM’s goal is to overthrow regimes in several North and West African countries and install a caliphate in the region, based on sharia law (Chivvis and Liepman, 2013). MUJAO’s objective was to “expand jihad in West Africa” (Flood, 2012: 5).
It is thought that around 300,000 Malians fled their homes during this period (Gaasholt, 2013; Stuster and Noble, 2014). Infamously, the militants also destroyed ancient and sacred shrines in the areas they controlled, most notably in Timbuktu (Lebovich, 2013). The enforcement of strict sharia caused resentment among locals, who for centuries had lived under a more moderate form of Islam (Roggio, 2013; Stuster and Noble, 2014). Civil resistance took place in several occupied Northern Malian towns, and in the capital, Bamako, between May and October 2012. Many local imams and other prominent individuals refused to cooperate with the extremists, and numerous displays of disobedience took place, including the organization of forbidden sports and cultural events (Gaasholt, 2013; Lebovich, 2013). The first major display of civil resistance was in Gao in mid-May, when hundreds of residents joined a protest against the ban on playing football and watching TV. On that day, locals reportedly also formed a protective belt around ancient tombs, preventing their destruction by the jihadists (L’Obs, 2012). Gao also saw civil resistance in early August, as hundreds of residents blocked the entrance to the town square where the Islamic police were to amputate the hands of alleged thieves (Human Rights Watch, 2012).
In Kidal, in June, at least 150 people, mostly women and youth, protested the application of the sharia. Also in June, in Timbuktu, the jihadists’ infamous destruction of ancient tombs caused people to march in the streets (Hirsch, 2012). Moreover, parents took their children out of school and other establishments in Timbuktu, after boys and girls were put in separate classes, and girls were forced to wear full body robes. Tuareg women protested after Ansar Dine decided they must wear the veil and banned them from participating in economic ventures (Haddadi, 2012). In mid-July, locals in the town of Goundam near Timbuktu prevented the jihadists from entering the town mosque to perform their Friday prayers (Times of Malta, 2012). Goundam also saw youths setting up roadblocks in protest against Ansar Dine’s beating of a woman who, allegedly, had not covered her baby’s head with a veil (Hirsch, 2012).
Jihadi rule and motives for popular resistance
The jihadist coalition in Mali was considered mostly foreign, with only Ansar Dine being able to claim domestic roots. Witnesses said the majority of commanders and many trainers in the armed groups were from Mauritania, Algeria, Western Sahara, Tunisia, and Chad (International Crisis Group, 2013). They brought with them an alien version of Islam and radically different ideas about social structures and cultural life, which they tried to violently enforce upon the local populations. The disconnect between the foreign jihadists and their interpretation of Islam, on the one hand, and the local social and cultural context, on the other hand, seems to have been an important factor in generating potential for mobilization among locals. Thus, the perception amongst civilians that the jihadists were alien to their local social and cultural context, including the traditional role of women, was one of the key factors driving people to manifesting dissent.
Most Malians adhere to a form of Islam influenced by Sufism, which includes mystical beliefs and ancestor worship. They celebrate artistic expression, especially music, and the liberal philosophy of the exiles. Although women have little access to the public political sphere, they are respected and play an important role in the social order (International Crisis Group, 2012; Lackenbauer et al., 2015). The jihadists’ introduction of strict sharia law and hudud punishments and the destruction of sacred mausoleums triggered strong grievances among local communities. For instance, as locals prevented MUJAO militants in Gao from amputating the hand of one of their fellow group members who stood accused of stealing, one resident said: “We don’t want to know what this young man did, but they are not going to cut his hand off in front of us” (Agence France Presse, 2012b: para. 6). Thus, locals felt obliged to intervene although the accused was a MUJAO member. And related to a protest in May 2012, a radio host in Gao expressed the prevalent public frustration: “Women, children, the young, everyone is outside and demands the departure of the armed groups. We are no longer afraid. Too much, it is too much” (L’Obs, 2012: para. 8, translated from French).
In Timbuktu, the large-scale destruction of sacred mausoleums, which have great religious, historical, and cultural value to Malians, by Ansar Dine and AQIM militants in May and June 2012, caused particularly strong tensions between the locals and the jihadists. As expressed by a man who witnessed the destruction of the tomb of a Sufi saint: “As they broke the tomb, yelling ‘Allah hu Akbar’ for all to hear, hundreds of us were weeping both inside and out” (Human Rights Watch, 2012). These events represented a breaking point for the population in Timbuktu, said Alpha Sanechirfi, director of the Malian Office of Tourism in Timbuktu: “When they smashed our mausoleums, it hurt us deeply. For us, it was game over” (Callimachi, 2013a). Accordingly, angry residents came out in large numbers to protest. City mayor Ousmane Cissé said: “They are protesting against the destruction of our culture and abuse of our local residents” (Hirsch, 2012). Similar developments took place under Ansar Dine and MNLA in Gao, where the added involvement of MUJAO, which claimed the city on 5 April, increased tensions between the locals and what they perceived as an alien force (Galy, 2013).
Social structures and the type of resistance
The jihadist proto-states in Mali and Iraq saw collective and manifest resistance, albeit in different forms. In Northern Mali, it was civil society organizations, with substantial participation by women, that provided the necessary networks and resources for nonviolent mobilization. The political debate in Mali has been open and extensive since the end of military rule in the early 1990s, and freedom of assembly and association have been respected. Many civic groups and NGOs have been able to operate actively without interference from the state, although their activities in the north have been restricted during periods of deteriorating security. The constitution has also guaranteed workers the right to form unions and to strike (Freedom House, 2012). Civil society played a critical role in bringing about peaceful settlement of previous armed conflicts (Lode, 1997). The fact that civil society in Mali had had time and space for growing, organizing, and acting, could thus explain why organizational resources and social networks were in place that were activated in the crisis of 2012–2013.
As the first protests broke out in Gao in mid-May 2012, the Collective of Inhabitants of the North (Colléctif des Ressortissants du Nord Mali, COREN), a Bamako-based umbrella for civil society organizations in Gao, Kidal, Mopti, and Timbuktu, released a statement assuring “all populations of the north of its support and its determination to back them up” (Agence France Presse, 2012a: para. 3). Accordingly, socioanthropologist Lalla Mairam Haidara, a native of Timbuktu and a specialist on women’s rights in Mali, said that against the background of government inaction toward the northern occupation, COREN and other northern civil society organizations “responded through convictions, sit-ins, and protest marches” (as cited in Colekessian, 2012: para. 11). And in Bamako, COREN played an active role in keeping the issue of the northern occupation in the public eye, on occasions gathering thousands in stadiums and protest marches in the capital (Lecocq et al., 2013; Saddier, 2017).
Women’s rights groups played an important part in the nonviolent mobilization in Northern Mali. In the words of Haidara: “Dozens of women from the Collective of Northern Women and national women’s groups are at the forefront of the conflict, challenging what they believe to be a lack of interest” (as cited in Colekessian, 2012: para. 13). Indeed, women led the protests against the MNLA and the jihadists, which was partly the result of structural factors, partly of situational circumstances. On the one hand, in Malian society, age outranks gender in social hierarchies, which granted older women indirect agency through their position in the hierarchy relative to younger men, by mobilizing them to protest (Lackenbauer et al., 2015: 23, 59).4 On the other hand, women gained episodic power, and thereby more direct agency, because many men either left or stayed at home (Lackenbauer et al., 2015: 60).
The association Women of the Azawad organized protests in Kidal on 6 June and in Khalil on 7 June, cities 500 km apart, in response to the violent imposition of sharia. The latter protest was allegedly triggered by Ansar Dine’s violent repression of the protest in Kidal on the day before, which indicates the links in the network of women’s organizations in the north and the role they played in resisting the jihadists’ rule (Lecocq et al., 2013). In Timbuktu in October, an estimated 100–300 women gathered to protest against Ansar Dine’s harsh rules on women’s clothing. As one participant said: “Life has become more and more difficult with these people. . .We are tired. They impose veils on us and now they are hunting us like bandits for not wearing them” (Reuters, 2012). Women also wrote critical songs and poems against the jihadists, which they distributed via social networks and cell phones (Lackenbauer et al., 2015; Lecocq et al., 2013).
While the subsequent French military intervention did eventually remove the jihadist rulers, there was no external troop involvement in the Malian conflict in 2012. The French only intervened after the Malian government requested them to do so, once the jihadists were approaching the south of the country. There was thus no increased militarization of local actors, such as tribal networks, which left civil society as the main actor through which mobilization against the jihadists could manifest itself.
Islamic State of Iraq, 2006–20085
On 13 October 2006, the Mujahideen Shura Council (MSC), an umbrella organization for Salafi-jihadist groups in Iraq dominated by AQI,6 proclaimed The Islamic State of Iraq (ISI), the predecessor of ISIS. The claimed territory of the proto-state covered Anbar, Baghdad, Diyala, Kirkuk, Niwana, Salah al-Din, and Babel provinces in Iraq’s Sunni Triangle (al-Tamimi, 2015; Gerges, 2016). In reality, however, AQI’s territorial control was weak, but the group was still able to carry out nonmilitary activities, especially in terms of implementing sharia law (Lia, 2015). In Anbar province, where AQI was most consolidated, the organization set up Islamic courts that banned cigarettes, imposed a strict dress code for women, and prohibited female drivers (al-Tamimi, 2015; Jones and Libicki, 2008).
AQI did not hesitate to carry out bloody attacks against their fellow Sunnis, including prominent tribal, religious, and resistance figures. They brutally killed nonsupporters, in particular those working with the Iraqi government or the US forces, and also targeted symbols of Iraqi nationalism, by decapitating tribal leaders and members of the Iraqi Security Forces (ISF). Where the organization governed, it forced families to provide subsistence and shelter to the jihadists and marry their daughters to suicide bombers (Al-Jabouri and Jensen, 2010; Lister, 2014). These acts caused public support for AQI from ordinary Iraqis, Sunni tribal leaders, and even other insurgent groups, to wane (Al-Jabouri and Jensen, 2010; Gerges, 2016). In short, whatever support AQI had enjoyed earlier began disappearing quickly as the group imposed an ideology and rule that were perceived as alien and a severe threat to incumbent power holders.
AQI’s declaration of an Islamic state in October 2006 further accentuated the ideological divide between the group and its surrounding communities. Late AQI leader Abu Musab al-Zarqawi and his successors saw Iraq as one arena in a larger transnational project to restore the caliphate and spread global jihad. This transnational ideology did not resonate well with Iraqis who doubted that the group had their best interests in mind. Moreover, AQI’s religious ideology clashed with the greater Sunni nationalist resistance (Al-Jabouri and Jensen, 2010; Gerges, 2016). Tribal leaders in particular were wary of a theocratic ideology that left little room for tribal authority. It was a direct threat to their power, aggravated by the jihadists’ intrusion on their traditional smuggling routes (Fishman, 2009; Jones and Libicki, 2008). Finally, although some other major insurgent groups in Iraq at the time also adhered to Salafist ideology, none of them prioritized the restoration of the caliphate. The divide between them and AQI eventually grew even more as AQI attacked and killed an increasing number of their leaders (Jones and Libicki, 2008).
Jihadi rule and motives for popular resistance
Just like in Mali, the violent and alien elements of AQI rule created mobilization potential for the local Iraqi communities. Perhaps most importantly, AQI posed a serious threat to tribal power, as it was cutting into the tribes’ lucrative banditry and smuggling business, thereby undermining their key revenue sources. One of the affected tribes was the Albu Risha, which had lost control over parts of the Baghdad-Amman road, where extortion of traders and travelers had been a key income. The challenge to tribal livelihoods and revenues was a source of growing resentment toward AQI and has been cited as the most important reason to why the tribes finally stood up (Jones and Libicki, 2008; Long, 2008; Malkasian, 2017; Simon, 2008). Other contributing factors were AQI’s transnational and fundamentalist ideology, its harsh sharia regime, violent attacks on fellow Sunnis, and the harassment of civilians, including property theft and the kidnapping of women. From the perspective of tribal leaders, AQI’s ideology was a threat to their authority and clashed with their more local or national agendas. Civilian populations, on the other hand, were alienated by AQI’s extreme demands and violent rule. By September 2006, AQI had lost much of the legitimacy it enjoyed early in the war, and the Sunni triangle, previously an important AQI base, was becoming an increasingly hostile place (Andersen, 2017: 94–95; Long, 2008: 77).
Social structures and the type of resistance
In Iraq, unlike Mali, resource mobilization occurred through tribal structures. Although, in 2006, freedom of association and assembly were recognized in law and NGOs could in theory operate without restrictions, in practice, their activities in several regions remained limited due to insecurity (Freedom House, 2006). Furthermore, Iraqis had only recently come from decades of dictatorship under Saddam Hussein, whose regime restricted freedom of assembly to progovernment gatherings, whereas freedom of association extended only to progovernment political parties or civic groups. There were also restrictions on the right to strike and no independent unions were allowed. Thus, civil society organizations in Iraq lacked the strength and influence to mobilize collective civil resistance against AQI. This type of agency was instead reserved for tribal structures, which, in the absence of strong central authority, have endured as the primary mechanism of societal organization, with tribal leaders administering resource and conflict management and law enforcement (Khan, 2007: 3).
Organized through tribal structures, the main response to AQI was violent. In Anbar province, Sheikh Sattar had gathered 25 of the 31 tribes to set up the Anbar Salvation Council, whose primary strategy was to ramp up the police forces of Ramadi and other Anbar towns to retake control. In mid-2007, the strategy was bearing fruit: the Anbar police force had grown to 24,000 members (Jones and Libicki, 2008). Seeing this development, other tribes beyond Anbar soon formed similar councils, and less than a year after the formation of the Anbar Salvation Council, the Sunni Awakening was born, an armed resistance movement with an estimated 80,000 militias (Gerges, 2016).
As the Sunni Awakening spread across Iraq in mid-2007, the US initiated a program in which it paid Sunni tribesmen and former resistance fighters to man security checkpoints in AQI areas of operations, gather intelligence, and engage in combat (Al-Jabouri and Jensen, 2010). The US also offered protection to tribes for building up local police forces. Between 2007 and 2008, the support for local armed resistance networks was a key element of US anti-AQI strategy (Gerges, 2016; Jones and Libicki, 2008). The US also strengthened its unilateral actions by inserting thousands of new combat soldiers into Iraq and initiating daily raids on jihadi hideouts by special operations teams (Byman, 2015).
There are three main explanations for why the resistance through the tribal structure in Iraq was violent rather than nonviolent. One is related to tribal customs for conflict resolution (Carroll, 2011).7 Khan points out that customs among Sunni tribes in Iraq require male members to protect the family honor, display their masculinity and courage in battle, and avenge the death of other members of the tribe. The latter of these customs refers to the concept of “blood feuds,” which is especially important in single families within a tribe. Blood feuds are settled by either killing the responsible person of another tribe or arranging financial compensation for the dead. Khan tells of running blood feuds between AQI and several tribes, which resulted from AQI assassinations of tribal leaders (Khan, 2007: 4–5). The fact that tribes were bound to certain customary practices that centered on violence rather than nonviolence could help explain the violent nature of the resistance against AQI’s proto-state. Second, and this is closely tied to the first point, mobilization through the tribal networks meant that there was an appeal to a more limited pool of members, primarily males, excluding an active role for women in the resistance. Women are marginalized in traditional tribal structures and “their ability to influence or interact with the tribal structure without a male intermediary is extremely limited” (Bobseine, 2019: 11).
A third factor was the presence of arms. The tribal society had a preexisting resource structure in terms of decentralized access to arms as well as through the networks of local tribal militias. After 1991, Saddam Hussein had allowed tribal sheikhs to build up their own armies equipped with light and heavier weapons such as rocket-propelled grenades and mortars, to allow the sheikhs to police their respective areas. Moreover, in 1996, as part of a broader effort of retribalization of Iraq, Saddam set up a high council of tribal chiefs, to grant the tribes political privilege, weapons, and land (Long, 2008; Simon, 2008). The fact that the tribal society was armed implied that it possessed the means for violent resistance.
The Iraqi tribal society’s violent response must also be understood in light of the pivotal role played by the US, both in terms of supporting the Sunni Awakening and strengthening its own unilateral actions, as laid out above.
Islamic principality of al-Mukalla, 2015–2016
AQAP has a long-standing presence in Yemen. The group was formed in January 2009 under the leadership of Nasir al-Wuhayshi, but its Yemeni roots can be traced back to the early-mid 1990s. With the start of the popular uprising against then-President Ali Abdullah Saleh in 2011, AQAP and its local affiliate Ansar al-Sharia took control over several towns in Abyan and Shabwah Governorates in May 2011, where they declared an Islamic emirate and implemented sharia law. Having ruled the areas for a year, a joint military assault by the Yemeni army and local militias eventually pushed AQAP back to their mountain hideouts in May 2012 (International Crisis Group, 2017; Stuster and Noble, 2014). On a mission to Yemen in Summer 2012, Amnesty International documented a wide range of human rights abuses perpetrated by the groups during their rule. These abuses included summary killings, amputations, and public flogging of alleged criminals, harassment and detainment of community activists, violations of women’s rights, disruptions of education and health care, and the destruction of “idolatrous” tombs and shrines (Amnesty International, 2012).
In the context of ongoing civil war and state collapse, AQAP reverted to guerrilla warfare and asymmetrical attacks. Nevertheless, shortly after the Saudi Arabia-led international coalition began attacking the Houthi-Saleh forces, AQAP again captured territories in Southern Yemen. In early April 2015, the group established control over large areas along the coastline of Hadhramaut Governorate, including Mukalla, and proclaimed the Islamic Principality of al-Mukalla (Bawzir, 2016). Upon its arrival in Mukalla, the fifth largest city in Yemen, AQAP freed several hundred inmates from a prison (including some senior AQAP operatives). They also looted the city’s central bank and took control over its ports. This strategy helped the group to consolidate its hold of the city and install civilian institutions to govern the lives of locals (Horton, 2015). Instead of establishing direct rule, however, AQAP created a ruling council of local dignitaries and prominent non-al-Qaeda members, the Hadhramaut National Council (HNC), and handed it a budget to pay salaries, import fuel, and hire staff to pick up garbage (Hubbard, 2015; International Crisis Group, 2017). The group remained in charge of security, military operations, and dispute resolution, but left everyday governance to local bureaucrats under the supervision of the HNC.
Jihadi rule and the lack of motives for popular resistance
Unlike in Mali and Iraq, AQAP adjusted to the local context, which was partly due to the nationality of most of its members: AQAP is a Yemeni organization with mainly Yemeni members (International Crisis Group, 2017). It has an extensive social and family network in Hadhramaut Governorate and other parts of the country, and therefore focused on working within the limits of local norms and sensibilities (International Crisis Group, 2016). For example, one report by the International Crisis Group argues that AQAP’s pragmatism and effective governance were key factors to secure public acceptance. They prioritized security and basic service provision, and set up a judicial system for resolving grievances, but refrained from the strict application of sharia (International Crisis Group, 2017). Although they introduced religious courts, a religious police force, and an unpopular ban on selling qat, they adjusted their strategies to the local context. The courts were seen as fair and swift in contrast to the slow system of the state, locals were hired to the religious police force, and the ban on qat (a flowering plant, that contains a stimulant and is often chewed) was only sporadically enforced. Adding to this, AQAP allowed women to stay outside after dark, interfered little with dress norms, did not force people to pray or pay a religious tax, and made no effort to ban smoking, music, or television (Horton, 2015; Hubbard, 2015; International Crisis Group, 2017). AQAP’s softer, more accommodative approach may have prevented the rise of strong local mobilization potential against the organization, and could thus help explain why AQAP seemingly faced neither nonviolent nor violent manifest and collective resistance.
AQAP’s governance strategy was a function of the group’s ability to learn from and adapt to earlier experiences and local context. AQAP learned from its experience of ruling in Abyan and displayed improved governing skills once they were in control of Mukalla. They arrived in Mukalla with a softer approach, socializing with residents and avoiding the introduction of draconian rules. They created the HNC instead of ruling directly, launched infrastructure projects, and provided social services such as the distribution of food and medical supplies. They even organized community events and street festivals. Further indication of AQAP having learned its lesson came in a letter from AQAP leader al-Wuhayshi to the leaders of AQIM, after AQAP had to withdraw from Abyan, in which he advised them on how to govern well. For example, he urged them to refrain from using Islamic punishments as much as possible (Callimachi, 2013b; International Crisis Group, 2017). This learning experience, partly a function of their local basis, may well be the best explanation to why AQAP rule in Mukalla did not create strong grievances and collective resistance against the jihadists.
Conclusion: understanding civil resistance against jihadist proto-states
In this study, we argued that civil resistance against jihadist proto-states can be explained by the interaction between the jihadist governance and the social structures for political mobilization that are activated in a certain locality. More specifically, we argue that civil resistance occurs when jihadist rulers impose an ideology and a rule that are perceived as alien by local populations, and when civil society organizations are activated to mobilize collective nonviolent action. By contrast, when mobilization occurs through tribal or other traditional authority structures, rather than through civil society organizations, resistance is more likely to be violent. Variations in the nature and degree of resistance against jihadist proto-states are an analytical focal point that can help bridge insights from research on terrorism, rebel governance, and civil resistance. In Mali, the jihadist rulers tried to enforce an alien and violent Islamist ideology, thus generating strong motives for resistance, which civil society organizations, including many initiatives by women, pounced on to mobilize people for collective nonviolent political action. In Iraq, AQI used a similar approach to governance, generating grievances against the jihadists, but civil resistance did not occur because dissent was instead mobilized through tribal structures and manifested itself as a large-scale violent resistance movement, which was empowered by US support. Lastly, nonresistance in Yemen can be explained by AQAP’s more accommodative and efficient rule, relative to its counterparts in Mali and Iraq, which created better conditions for relative acceptance among local communities. Tribal networks existed in all three cases and thus the key difference should not be found in the presence of tribal structures per se, but rather in how the jihadist groups related to the tribal structures, challenged them in the case of Iraq (which is one of the reasons for why they were activated for resistance) and sought to cooperate with them in the case of Yemen.
A comparative analysis of a few strategically selected cases, as has been conducted in this study, does not have the analytical leverage to rule out all alternative explanations. We recognize that the variation in civilians’ responses may be due to other factors. For example, a strand in previous research emphasizes political opportunities that civil resisters capitalize on (Tarrow, 2011). Political opportunities for nonviolent mobilization are thought to exist when activists see low costs of mobilization and expect that mobilization can be successful. One element of the political opportunity framework is elite divisions, which indicates that power holders are less in control over the situation than previously and that there may be possibilities for nonviolent challenges to be effective (Skocpol, 1979). Whereas all three cases occur in a context where there were rival rebel groups, it is still possible that the deeper internal splits within the local jihadist proto-state in Mali created political opportunities for civil resistance, which activists were able to capitalize on. Further, the absence of civil resistance in Iraq and Yemen might have been because AQI and AQAP were more unified as organizations. Still, we do not find this to be a particularly convincing explanation, as it has been argued that it is only when armed actors are in complete control over an area and its population, that they can allow dissent to be raised (Kasfir, 2005). Thus, elite divisions amongst jihadist groups could as well undermine the opportunities for civil resistance.
Another alternative factor explaining the variation in popular responses may have been the extent and nature of the service provision of the jihadist group. Arjona (2015) suggests that this may be a critical aspect in explaining resistance. We saw in the case of AQAP’s governance in Mukalla that they took care in providing services to the locals, which could have made the popular appetite for resistance weaker. It is possible that the variation in resistance is best explained by the level of state services before the jihadis’ arrival as well as the nature and level of provision of state services by the jihadist groups. While some types of institutions were established in all three cases, it may be the case that the degree to which those institutions were effective in providing services to locals may have influenced the probability of resistance. Whereas this may form part of an alternative explanation as to why the Yemenis did not resist when Malians and Iraqis did, this explanation seems less capable of explaining the difference between violent and nonviolent resistance, as there are no evident differences in service provisions between the Islamic Emirate of Azawad in Mali and the ISI. Moreover, while we were able to hold structurally constant territorial control and inter-rebel rivalry, we also recognize that the degree and the forms of territorial control and the nature of the inter-rebel rivalry was not exactly the same in the three different contexts. Thus, these factors cannot be completely ruled out. Overall, we recommend future research provide more comprehensive tests as to which explanation is the best to account for the variation in resistance strategies and the onset of civil resistance against jihadists.
Our screening of civil resistance in jihadist proto-states provides a first comprehensive overview of civil resistance in these contexts, and allowed for a systematic identification of cases for comparative analysis. We recognize, however, that some data-related issues—the reliance on secondary sources in particular—represent a limitation of this study. Importantly, it limits our control over the information-gathering process. We acknowledge this and recommend future studies on this topic to complement our approach by using surveys and interviews, as more first-hand information would be of great value to the field (see [omitted due to anonymous review process, Revkin, 2020]). Unfortunately, we have been unable to collect primary data ourselves for this study. Another issue concerns possible differences in data availability across the three studies cases, which could potentially affect the way we measure the outcome in the dependent variable, civil resistance. International media may pay more attention to violent than nonviolent events, and thus, our focus on news articles to determine the presence versus absence of civil resistance could be a potential problem (Dorff, 2019). It is possible that our material leads us to miss out on important events of civil resistance that occurred in the midst of violent conflict and terrorism. Perhaps nowhere more so than in Iraq in the mid-late 2000s, where civil resistance against AQI may easily have been marginalized in reporting on Iraq due to the ongoing violent unrest, especially because there was such heavy international military involvement. On the other hand, one could argue that even nonviolent resistance would attract international attention once it reaches a certain magnitude, that is, when it gathers large numbers of people in streets and squares, especially in a country like Iraq at the time, which was in the spotlight of international news media. While we have tried to mitigate against this potential bias by using other sources, again, we do encourage future research to analyze civil resistance using other empirical strategies.
This is the first comparative study of civil resistance against jihadist proto-states, yet it does not take all possible nuances into account. We recognize that further analysis should be done that takes intra-case variation into account. For example, whereas there has been a relatively stable support base for Jabhat Fateh al-Sham (formerly al-Nusra Front) in Northern Syria, there have also been areas in which civilians have protested against their rule (Haid, 2017). Moreover, our study focuses on onset of civil resistance, and so more research is needed on the dynamics of civil resistance once it has been initiated, and the conditions under which it can be successful.
Even though our study is limited to civil resistance against jihadist proto-states, our findings could be applicable to a wider universe of cases. In other words, the conditions that explain the onset of civil resistance against jihadist proto-states could be generalizable to civil resistance against rebel governance more generally. On the other hand, jihadist groups do possess certain characteristics that distinguish them from other types of rebels. In particular, they aspire to goals and territories that transcend the local and the national. Due to this transnational ideology, it may be that jihadist rebel governance creates unique conditions for civil resistance. These noted differences provide rationale for the study of (popular resistance against) jihadist proto-states as a phenomenon in its own right. Still, whether jihadist proto-state is a distinct phenomenon is an open empirical question, which calls for comparative analysis. Systematic comparison between jihadist rebel governance and other forms of rebel governance, and how they affect the prospects for civil resistance, would therefore be a particularly promising avenue for future research.