On the Divergent Trajectories of African Islamism: Explaining Salafi Non-Radicalisation in Zanzibar

Abstract

With the rise of Jihadist groups like Boko Haram and Al-Shabaab, the theory-oriented exploration of the causes and dynamics of militant Islamist mobilisation in sub-Saharan Africa has become an important research endeavour. Existing explanatory frameworks highlight the causal relevance of multi-faceted Muslim grievances, pre-existing histories of non-Islamist political violence, and a lack of institutional regulation of Salafism in the first decades after independence. Examining the deviant case of Zanzibar, the article investigates why, despite the fact that all these conditions have been present on the islands, local Salafi activism has so far not gravitated towards Jihadism. It shows that the deeply entrenched Zanzibari two-party conflict between Tanzania’s long-standing authoritarian ruling party Chama Cha Mapinduzi and the oppositional Civic United Front has persistently blocked the opening up of political space for the development of sizeable militant Islamist activity. The article thus contributes to area-centred theory-building by demonstrating that vigorous electoral competition can effectively forestall the rise of Jihadism.

Introduction

With the rise of Jihadist groups like Boko Haram and Al-Shabaab, the theory-oriented exploration of the causes and dynamics of militant Islamist ¹ mobilisation in sub-Saharan Africa has become an important research endeavour. While it is beyond doubt that Jihadism’s emergence as a regional security challenge is linked to Salafi ² revivalism, it is also clear that an ideologically homogeneous Salafi sphere does not exist in Africa. For instance, dismissing a “blanket equation between Salafism and militancy,” Østebø (2015: 2) points out that while in settings like northern Nigeria and Somalia Salafi quests for religious purity paved the way for the formation of violent movements, and in other countries (e.g. Niger or Ghana), Salafi activism has partly developed a more conciliatory stance towards established Sufi ³ Islam and the secular state. Consequently, he advocates a careful assessment of “the actual ideological contours of each movement; the origins, sources of inspiration, and local appropriations” (Østebø, 2015: 20).

The experience of Zanzibar, semi-autonomous member state within the Tanzanian Union, supports Østebø’s negation of an inevitable Salafi propensity for radicalisation and violence. In the largely Muslim archipelago, ⁴ Jihadist political entrepreneurs have so far not risen to prominence. Against this background, alleged militant Islamist activity has remained confined to a limited number of incidents, most notably the killing of a Zanzibari Catholic priest in February 2013 and an explosives attack on Stone Town’s Darajani mosque which killed one person and injured several others in June 2014 (LeSage, 2014: 6). ⁵ These events have been frequently linked to the mobilisational efforts of the Jumuiya ya Uamsho na Mihadhara ya Kiislam (Organisation for Islamic Awareness and Propagation – JUMIKI), a Salafi non-governmental organisation (NGO) which in 2012 attracted international attention by campaigning for the dissolution of the Tanzanian Union (see, e.g. Bergmann, 2013; Le Sage, 2014: 12). Yet JUMIKI leaders have vehemently denied any involvement in violent activities (JUMIKI, 2013). In fact, in a 2008 study of Zanzibari Islamic reformism, Turner finds that the organisation has promoted a decidedly democratic “language of Human Rights and Good Governance” (Turner, 2008: 11). In a similar vein, a recent International Crisis Group (ICG) report on Jihadism in East Africa classifies JUMIKI as belonging to a group of NGOs “that express political grievances but do not champion violence” (ICG, 2018: 17). Based on these findings, this article contributes to the evolving theorisation of the causes and dynamics of militant Islamist activity in sub-Saharan Africa by investigating why in Zanzibar, Salafi activism has not gravitated towards Jihadism. It takes as a point of departure the pioneering work of Sebastian Elischer and Caitriona Dowd, who have developed two different but not incompatible explanatory frameworks, both informed by comparative intra-area research agendas. ⁶

Focusing on the historical interactions between African states and Islamic civil societies, Elischer (2015, 2019) contends that the application of different postcolonial state management approaches has crucially shaped the course of Islamist mobilisation. Specifically, he puts forward the hypothesis that in Africa’s Muslim majority societies, the prospects for the demobilisation of nascent Jihadism(s) have been best in settings where postcolonial states have historically resorted to the strategy of institutional regulation by establishing monopolistic Sufi-dominated Islamic associations. By contrast, Elischer identifies governmental inaction as well as indiscriminate repression against Salafis as enabling conditions for collective radicalisation.

In Dowd’s theory, the popularisation of militant Islamist ideologies is primarily linked to pervasive perceptions of marginalisation among Muslim populations and pre-existing histories of non-Islamist violence. Emphasising the relevance of multi-faceted horizontal inequalities between sociocultural groups, she promotes the premise that “collective grievances contribute to a wider social context in which militants can evade capture among a supportive population, and in which populations refuse to cooperate or share information on militant activity with security forces” (Dowd, 2015: 514).

As will be shown below, the major drivers of Islamist radicalisation and violence identified by the two authors have been present in Zanzibar. Yet so far, militant Islamism has not turned into a substantial security threat. Addressing this anomaly, I argue that the deeply entrenched Zanzibari two-party conflict between Tanzania’s long-standing authoritarian ruling party Chama Cha Mapinduzi (Party of the Revolution – CCM) and the oppositional Civic United Front (CUF) has persistently blocked the opening up of political space for the genesis of sizeable militant Islamist networks. Specifically, I will show that based on its popular subnationalist agenda, the CUF repeatedly prevented the escalation of state–Salafi confrontations by effectively (re-)shifting local oppositional politics and the focus of state repression into the domain of non-Islamist multi-party competition. In doing so, the party provided a strong incentive for Salafi leaders and frustrated Muslim youth to remain committed to the electoral process.

By demonstrating that vigorous electoral competition can effectively forestall the rise of Jihadism, I provide a counterpoint to Elischer’s theoretical focus on the state. Furthermore, I complicate Dowd’s framework by showing that pre-existing histories of non-Islamist conflict can also work against the expansion of militant Islamist activity. In line with the “best practices” of process tracing identified by Bennett and Checkel (2015: 19–31), my analysis is based on diverse and independent streams of evidence, including grey literature, newspaper articles/interviews, JUMIKI, and CUF press statements and online videos of political speeches and public rallies. Combined with the comparatively rich conventional secondary literature on Zanzibar’s political and religious history, these sources enable me to adequately piece together the temporal sequences and causal mechanisms underlying the islands’ anomalous trajectory of Salafi politics. ⁷

The article proceeds as follows. In a first step, the two existing theoretical frameworks tracing the causes and dynamics of violent Islamist mobilisation in sub-Saharan Africa are briefly presented and integrated. Zanzibar is then identified as a deviant case vis-à-vis these models. Finally, the absence of substantial militant Islamist activity in the archipelago will be explained.

Grievances, State Management of Religion, and Militant Islamism in Sub-Saharan Africa

Even though in countries like Uganda and Nigeria, the rise of modern Jihadism can be traced back to the 1980s (see, e.g. Kayunga, 1994; Loimeier, 2012), it is only thanks to the recent exploratory efforts of Dowd and Elischer that theory-oriented research on militant Islamism in sub-Saharan Africa has gained traction. While the two authors prioritise different explanatory factors, their approaches can be considered cornerstones of one evolving macro-political framework.

To begin with, Dowd’s qualitative account of the genesis of Islamist security threats is based on research on subnational developments in Nigeria, Kenya, and Mali. Making use of the disaggregated data provided by the Armed Conflict Location & Event Data Project (ACLED), she contends that the escalation of Islamist violence in different subnational contexts is causally connected to pervasive perceptions of political and economic disenchantment among Muslim populations (Dowd, 2015: 512–515).

To substantiate her claim, Dowd cites Afrobarometer survey data from pre-escalation periods which show that areas characterised by high levels of Islamist violence also tend to exhibit higher than average rates of political and economic grievances among their Muslim populations (Dowd, 2015: 515–520). Moreover, comparing the genesis and evolution of Boko Haram and the Lord’s Resistance Army, Dowd and Drury emphasise that irrespective of their cultivation of different ideological narratives (Jihadism vs. Christian fundamentalism), both insurgent groups have their origins in disadvantaged regions (northern Nigeria and northern Uganda) where “historical episodes of rotation of control over central state power between northern and southern elites have further shaped access to power and perceptions of marginalisation and exclusion” (Dowd and Drury, 2017: 141).

However, given the prevalence of grievances across the Muslim world, Dowd also posits that perceptions of political and economic marginalisation only work as necessary but not as sufficient conditions for Islamist violence to become a significant security threat. Thus, in her model, grievances are complemented by two additional causal factors: prior histories of non-Islamist violence and key triggering events such as the death of Boko Haram leader Mohammed Yusuf in police custody in 2009 or Kenya’s invasion of southern Somalia in October 2011 (Dowd, 2015: 520–526; Dowd and Drury, 2017: 137).

Emphasising that “the very language and targeting of Islamist violence cannot be divorced from domestic politics and historical violence in the state,” Dowd (2015: 521) finds that Jihadist political entrepreneurs capitalise on and reconfigure pre-existing forms of non-Islamist conflict and contestation such as the Tuareg conflict in northern Mali, secessionist tensions in coastal and north-eastern Kenya, or violent inter-communal rivalries in northern Nigeria. Against this background, Dowd observes that “all areas which […] experienced high levels of Islamist violence are also areas in which high rates of non-Islamist political violence previously occurred” (Dowd, 2015: 520). She therefore cautions against overstating the importance of the ideological and cultural dimensions of militant Islamism.

While in Dowd’s theory, the state exclusively acts as a decisive “facilitator” of Islamist radicalisation and violence, Sebastian Elischer has developed a historical–institutionalist state management model wherein governments either fuel or undermine the spread of militant Islamist activity depending on the specifics of their strategic engagement with the religious sphere.

Conceptually, Elischer emphasises the prominence of Salafism within the spectrum of Islamist thought. He draws upon Wiktorowicz’s (2006: 219) influential distinction between three different types of Salafis: “purists” who repudiate any active political engagement and concentrate on Islamic education, “politicos” who opt for Salafi participation in partisan politics, and “jihadis,” that is, militants working towards the violent subversion of existing sociopolitical orders (Elischer 2015: 583). Acknowledging the relevance of this typology, Elischer identifies five state strategies regarding the regulation of Salafi activism: “concession,” “toleration,” “cooptation,” “repression,” and “institutional regulation”, that is “allowing some religious or doctrinal practices while banning others” (Elischer, 2019: 201).

According to Elischer, the type of strategy chosen by governments and rulers is determined by their perceptions of Salafis as either political allies or challengers and by Salafism’s overall leverage within the religious sphere. Moreover, highlighting the causal relevance of historical path dependencies, he also argues that “the initial choice about strategy sets countries on pathways that affect the selection of subsequent strategies” (Elischer, 2019: 200). Against this background, Elischer contends that in Africa’s Muslim majority societies, the prospects for the demobilisation of incipient Jihadism have been best in settings where postcolonial states have historically resorted to the strategy of institutional regulation by establishing Sufi-dominated national Islamic councils. Comparing different trajectories of Salafi activism and state management in the Francophone Sahel, he finds that government-controlled national organisations like the Association islamique du Niger (AIN) and Chad’s Comité islamique effectively undermined the spread of Salafism in the 1970s and 1980s. Furthermore, he credits these associations with having crucially facilitated Salafism’s later integration into the religious establishment (Elischer, 2019: 212).

Elischer describes, for instance, how in 1974 Niger’s new military government created the AIN in its endeavour to forge an alliance with the country’s Sufi establishment after growing Arab/Libyan influence in the religious and educational sphere had triggered protests by secularly minded students and civil servants. He also depicts how, as part of a corporatist state machinery, the AIN subsequently succeeded in effectively monitoring Islamic life by denying Salafi preachers access to state-owned Friday prayer mosques. Finally, Elischer highlights that in the aftermath of political liberalisation, the AIN (with partial success) worked against the registration of Salafi civil society organisations and, in the wake of growing inner-Islamic tensions, served as the “cognitive template” for the creation of a new national association, the Conseil islamique du niger, in 2000 (Elischer, 2015: 584–592).

According to Elischer, in the absence of institutional regulation, it is more difficult for governments “to distinguish between the Islamic sphere as a whole, the Salafi community and potential challengers within that community” (Elischer, 2019: 212). For instance, he shows how in Mauritania, the postcolonial state contributed to the emergence of a domestic Jihadist movement by facilitating Salafi preaching up to the 1980s and by responding to the subsequent genesis of Islamist opposition groups in the 1990s with random crackdowns on mosques and clerics. Elischer (2019: 212) thus arrives at the conclusion that “both indiscriminate violence and toleration lead to the same outcome: the emergence and spread of political and security challengers.”

Regarding his model’s potential compatibility with other approaches, Elischer (2019: 213) explicitly declares that he does not consider state management the only variable accounting for different manifestations of Salafism. Most importantly, his argument does not challenge Dowd’s assumption that collective Muslim grievances and pre-existing histories of non-Islamist conflict are conducive to Islamist radicalisation. Rather, Elischer provides a valuable counterpoint to Dowd’s emphasis on the state’s contribution to the emergence of Islamist security threats. Consequently, the two authors’ explanatory frameworks can be treated as two sides of the same coin.

Zanzibar as a Deviant Case

Deviant case study designs are devoted to the investigation of “observed empirical anomalies in existing theoretical Propositions, With The Aim Of Explaining why the case deviates from theoretical expectations and in the process refining the existing theory and generating additional hypotheses” (Levy, 2008: 13, emphasis in original). With respect to Dowd’s and Elischer’s modelling of militant Islamist activity, the application of this logic of inquiry becomes a relevant exercise when considering the characteristics of the subnational case of Zanzibar. As will be shown in the following subsections, the major drivers of Islamist radicalisation and violence identified by the two authors, that is, the prevalence of multi-faceted subnational/Muslim grievances, a lack of institutional regulation of Salafism in the first decades after independence, a pre-existing history of non-Islamist political violence, and potential triggering events revolving around heightened state repression, have been present in the archipelago. Thus far, however, alleged Islamist violence has been confined to a limited number of incidents, and Jihadist political entrepreneurs have not risen to prominence. The article uses this deviance to complement Elischer’s theoretical considerations on the “preventability” of collective Jihadist radicalisation and to qualify Dowd’s conceptualisation of pre-existing histories of non-Islamist violent conflict as enabling conditions for the expansion of militant Islamist activity. It does so by showing that in Zanzibar, a resilient two-party confrontation has persistently blocked the opening up of political space for the rise of Jihadism.

The Genesis of Subnational Grievances

To begin with, substantial parts of Zanzibar’s population have historically exhibited a tangible sense of political and economic marginalisation linked to the question of the islands’ statehood. In fact, the design and the operation of the Union were surrounded by controversy from the outset.

In the aftermath of the Zanzibari revolution of January 1964, US and British intelligence agencies became increasingly worried about a Communist takeover of the archipelago’s new government which they feared could destabilise the whole East African region (Shivji, 2008: 72–76). Their concern was shared by Tanganyika’s left-leaning but anti-communist President Julius Nyerere, who eventually managed to bring about a merger of his country and Zanzibar by openly threatening Zanzibari President Abeid Karume with the immediate withdrawal of the islands’ Tanganyikan police contingent (Hunter 1966: 118–130). On 22 April 1964, the two heads of state signed the Articles of the Union, a unification treaty drafted by Nyerere’s legal advisers in a highly secretive process without any Zanzibari participation. It provided for the dissolution of the Tanganyikan government and transferred the jurisdiction over eleven areas (e.g. defence, foreign affairs, external trade, finance, and immigration) from Zanzibar’s Revolutionary Council to the new Tanzanian administration (Shivji, 2008: 76–82).

While Karume, who had signed the treaty without seeking legal advice, was under the impression that he had agreed to the formation of a federation of two sovereign states (Wilson, 1989: 79), the mainland leadership apparently operated under the assumption that it had created what Sheriff (1990: 154) has called “a half-way house to total assimilation of Zanzibar into a unitary state.” In the aftermath of the merger, Nyerere, a dedicated Pan-Africanist, immediately started to work towards the steady reduction of Zanzibari autonomy in line with a vision of “one country, one state, one party and one ideology” (Shivji, 2008: 155). For instance, in 1965, the mainland-dominated Tanzanian parliament adopted an interim constitution that increased the number of official Union from eleven to sixteen, adding to the list sensitive issues such as the management of mineral oil resources. Moreover, the constitution explicitly envisaged the fusion of the mainland’s ruling party, the Tanganyika African National Union and its Zanzibari counterpart, the Afro-Shirazi Party (ASP) (Shivji, 2008: 124–129). Assisted by Karume’s successor, Aboud Jumbe, Nyerere eventually succeeded in realising this project in 1977 when the CCM was created (Bakari, 2001: 113–114). A permanent Union constitution adopted the same year turned the new party into the supreme ruling body of Tanzania and Zanzibar, thereby further diminishing the islands’ autonomy (Shivji, 2008: 163–181).

In the years following the party merger, the Union became increasingly unpopular in Zanzibar. The failure of import-substituting industrialisation and the war against Uganda in 1978 plunged Tanzania into a deep economic crisis. Shivji (2008: 206) highlights that against this backdrop, “for the first time […] anti-Union sentiments began to be expressed openly from virtually the whole spectrum of Zanzibari society.” Trying to capitalise on this dissatisfaction, the islands’ President Jumbe now started to work towards more Zanzibari autonomy. After staging media attacks against the Union and secretly ordering the preparation of a court case demanding the establishment of a three-tier federation, he was forced to resign by the CCM’s National Executive Committee (NEC), a party organ comprising a majority of mainland members, in 1984 (Shivji, 2008: 202–225).

Another severe intra-party disagreement arose in 1992 when Zanzibar’s fourth President, Salmin Amour, arranged the islands’ entry to the Organization of the Islamic Conference (OIC) without consulting the Tanzanian government. Sparking sharp criticism from the CCM’s mainland elite, including a personal intervention by the party’s now retired founding father Nyerere, Amour finally backtracked and revoked Zanzibar’s OIC membership after only nine months although it had been highly popular among large parts of the local population (Fengler, 1997: 81–82).

At the same time, political liberalisation paved the way for intensified public discussions on the Union question. In 1992 and 1997, Constitutional Review Commissions (CRCs) headed by the former judges Francis Nyalali and Robert Kisanga, respectively, recommended the establishment of a federation but remained unheard by incumbent CCM governments. In December 2010, Tanzania’s fourth President Jakaya Kikwete eventually launched a far-reaching participatory constitutional reform initiative in the aftermath of his re-election. Six months after the adoption of the Constitutional Review Bill in November 2011, he appointed former Tanzanian Prime Minister Joseph Warioba as head of a thirty-member CRC which between July and December 2012 held more than 1,000 meetings at the district and ward level, seeking the opinion of citizens and civil society organisations before producing a draft constitution for later discussion by a Constituent Assembly (Branson, 2015: 3–4).

Drawing upon the conflictive history of the Union, Zanzibar’s JUMIKI activists used the extensive constitutional review process as an opportunity to mount a concerted secessionist campaign accentuated by the popular slogan “Tuachiwe Tupumue” (Let us breathe). In 2012, they regularly managed to organise large public rallies by cultivating a politico-religious discourse depicting the fusion of Tanganyika and Zanzibar and the Union government’s increasing authority over the islands’ affairs as a sustained project of anti-Muslim suppression launched by Tanzania’s long-standing Catholic President Nyerere (Fouéré, 2014: 11–13; Poncian, 2014: 175–176). ⁸ Against this background, JUMIKI leader Sheikh Farid Hadi Ahmed demanded the immediate termination of the constitutional review process and the holding of an independence referendum (Mwananchi, 2012).

The Rise of Political Salafism

JUMIKI’s religiously imbued secessionist activism unfolded within a widening sphere of Salafi influence that has characterised East African Islam since the mid-twentieth century. On the islands, the spread of the Salafi creed was crucially facilitated by the revolution of 1964. Imprisoning and forcing into exile the members of a prestigious scholarly elite linked to the Alawiyya brotherhood, the new regime clamped down on the local Sufi establishment, perceived as a political threat by President Karume (Loimeier, 2009: 461–468). ⁹

In the early 1970s, Aboud Jumbe’s government addressed the ensuing vacuum in the Islamic sphere by recruiting Salafi teachers from Sudan and Egypt who trained a new generation of Zanzibari scholars at the islands’ former Muslim Academy, shut down since 1965 and reopened in 1972 under the name Chuo cha Kiislamu (Islamic school – CCI) (Loimeier, 2009: 117–118). ¹⁰ Moreover, from the 1970s, Zanzibari students were increasingly sent to study at modern Islamic universities in Saudi Arabia, Egypt, and Sudan. Together with the CCI’s graduates, many of them would later be provided with formal positions in public institutions and thus turn into loyal representatives of an official bureaucratic government Islam (Loimeier, 2009: 118–124; Turner, 2008: 18–19).

However, among the returnees from the Arab world were also several politically minded activist clerics who, organising independently of the state, took an oppositional stance vis-à-vis the ruling CCM elite and embarked on a mission of religious purification. Branding the culturally embedded practices of established Sufi Islam as un-Islamic innovations, they have come to be known as Ansar Sunnah (group of defenders of the Sunnah) or, ironically, as “watu wa bidaa” (the innovation people) (Loimeier, 2009: 126; Turner, 2008: 9).

Loimeier (2009: 115) has pointed out that the Ansar Sunnah movement emanated from a circle of disciples around Zanzibar’s former Deputy Qadi Sheikh Farsy who himself had left the islands for Kenya in 1967 because of the new government’s hostility towards him and his family. Educated in the local tradition of the Alawiyya, Farsy, together with other East African scholars active in the colonial era, had also become inspired by North Africa’s modernist Salafiyya movement (Loimeier, 2009: 109–112). Based in Unguja’s Kikwajuni quarter, his intellectual heirs engaged in the construction of mosques and adopted new modes of politico-religious public preaching, known as “mihadhara” (lectures) (Turner, 2008: 11).

In the 1980s, Sheikh Nassor Bachoo became the leading figure within the Ansar Sunnah milieu. A follower of Sheikh Farsy, Bachoo established himself as Imam of Unguja’s Kikwajuni juu mosque and initiated the formation of JUMIKI. ¹¹ His rise provides a prime example of how the post-revolutionary government fostered the propagation of activist Salafism as in the 1970s, he had been trained by the “new” Egyptian and Sudanese lecturers at the CCI (Gilsaa, 2015: 46).

In line with Elischer’s typology of state management strategies vis-à-vis the Islamic/Salafi sphere, the overall approach pursued by the islands’ CCM regime up to the late 1980s can be classified as co-optation with toleration/without institutional regulation. While on the one hand, the government integrated several Salafi university graduates into the public administration, on the other hand, apart from expelling a number of Sudanese clerics they had recruited themselves (Loimeier, 2009: 129), the authorities did not resolutely counteract the emergence of an assertive Salafi civil society composed of networks unwilling to become affiliated with “official” Islam. This non-interventionist path was only abandoned when Islamist activists started to initiate large public protests against the Tanzanian and Zanzibari ruling elites.

In the late 1980s, the islands’ Ansar Sunnah Sheikhs spearheaded a growing informal opposition movement. For instance, after the chairperson of the CCM’s women’s wing, Sophia Kawawa, called in May 1988 for the abrogation of the Islamic Marriage Act due to its discrimination against women, Sheikh Bachoo organised a religious demonstration with strong anti-Union overtones (Lacunza-Balda, 1989: 304–305, 328; Sheriff, 1990: 156). Vigorously cracking down on the protest in Unguja, the Tanzanian police killed two demonstrators and temporarily detained forty-six people, including the islands’ leading Salafi preachers (Bakari, 2001: 90; Purpura, 1997: 377).

In 1989, Islamic stickers reading “bismillahi al-rahmani al-rahim” (in the name of God, the benevolent, the merciful) began to appear all over Zanzibar’s Stone Town as signs of pious opposition to the CCM establishment (Purpura, 1997: 378–379). According to Seif Sharif Hamad, who would become the islands’ opposition leader, the dissemination of this protest symbol had been initiated by Islamist Zanzibari expatriates from Dubai and prompted the security apparatus to arbitrarily detain people it suspected of being involved (Hamad in Burgess 2009: 261–262). About three years later, in late 1992, Salafi activists staged their second large public demonstration, protesting against a government family planning programme and claiming the islands’ right to become a member of the OIC (Parkin, 1995: 200–201; Purpura, 1997: 37). However, by then oppositional politics had already largely shifted out of the Islamic sphere in a context of increasing demands for political liberalisation.

Facing the prospect of growing donor pressure, in February 1992, the CCM’s National Congress eventually paved the way for the end of the single party era by passing a resolution in favour of the reintroduction of multi-party competition at the Union level and in Zanzibar. On the islands, this political opening would revive a tradition of crisis-laden electoral polarisation.

Pre-Existing Forms of Violent Confrontation

On the path to independence during the late 1950s and early 1960s, politics in the Zanzibar Protectorate were marked by fierce, racially imbued multi-party competition between the ASP and an electoral alliance between the Zanzibar Nationalist Party (ZNP) and the Zanzibar and Pemba People’s Party (ZPPP). The ASP was predominantly backed by poor Shirazi ¹² peasants and mainland-born labourers, and its activists cultivated an aggressive “Africanist” rhetoric depicting members of the islands’ Arab minority population as foreign intruders. By contrast, promoting a decidedly anti-mainland version of nationalism that championed the idea of a distinct (Muslim) Zanzibari identity, the ZNP-ZPPP coalition drew most of its support from the Protectorate’s landed Arab oligarchy as well as from a multi-racial (Arab and Shirazi) middle peasantry based on the island of Pemba. Against this background, highly contested elections to the colonial Legislative Council were accompanied by violent clashes between party supporters which culminated in the revolution, that is, the overthrow of the ZNP-ZPPP post-independence government and a massacre against the Arab population (Lofchie, 1965: 183–257; Sheriff, 2001).

In the years following the revolution, Abeid Karume established a despotic regime notorious for its brutality and economic mismanagement. In doing so, he could count on the support of a small but entrenched group of mainland-born former ASP youth activists who had been the driving force behind the events of January 1964 (Burgess, 1999: 34–35). Commonly referred to as “hardliners” or “liberators,” these militants were to dominate Zanzibar’s new supreme ruling body, the Revolutionary Council (Triplett, 1971).

In the late 1980s, increasing infighting within the ruling party culminated in the expulsion of the island’s former Chief Minister Seif Sharif Hamad who had become a major threat to the liberators’ dominance. While Hamad was in detention from May 1989 until November 1991, ¹³ several of his allies formed the Kamati ya Mwelekeo wa Vyama Huru (Committee for Multi-Partyism – KAMAHURU) to agitate for political liberalisation (Bakari, 2001: 175). As the political reforms of 1992 only allowed for the creation of national parties, the KAMAHURU leadership decided to merge its organisation with the Chama cha Wananchi (People’s Party), a small mainland party, to form the CUF (Bakari, 2001: 173). Nevertheless, the question of Zanzibar’s autonomy would become the core issue for the new party, which was dominated by its island wing. Regularly organising huge rallies in the run-up to the 1995 general election, Hamad and his allies started to openly campaign for the transformation of the Union into a full, three-tier federation with a weak central government. Moreover, following Zanzibar’s withdrawal from the OIC in 1993, they demanded an immediate re-entry (Fengler, 1997: 78–79).

Not surprisingly, the CUF rapidly succeeded in establishing itself as a serious contender to challenge the ruling party’s local hegemony. During the 1995 Zanzibari election, CCM incumbent Amour only managed to stay in power by instructing the Zanzibar Electoral Commission (ZEC) to change the result of the presidential vote in his favour after it had turned out that CUF candidate Hamad had won by a margin of 1.5 percentage points (Bakari, 2001: 227–228). To a significant extent, this outstanding electoral performance was based on Hamad’s popularity on his rural home island Pemba where the CUF won more than 80 per cent of the vote (Bakari, 2001: 239). Among Pemba’s largely Shirazi population, the NEC’s 1985 decision to deny Hamad the Zanzibari Presidency in favour of the Unguja-based “African” liberators – notorious for their hostility towards the smaller island ¹⁴ – had generated widespread disillusionment with the Tanzanian party-state nomenklatura (Sheriff, 1990: 156). However, far from becoming a party of Pemban regionalism, the CUF’s subnationalist rhetoric also appealed to a substantial share of the electorate in urban Unguja, the heart of Salafi reformism. For instance, the opposition won both Stone Town constituencies and was only narrowly defeated in Kikwajuni and Mlandege amid claims of electoral fraud (Bakari, 2001: 227, 323).

In the following months and years, the polarisation of the island’s political landscape further intensified. While state security forces violently looted opposition-dominated neighbourhoods, CUF supporters carried out a number of minor attacks on government buildings and properties owned by members of the ruling party (Bakari, 2001: 258–262). Furthermore, in 1997, several CUF officials were arrested and charged with plotting the overthrow of the Tanzanian and the Zanzibari government (Anglin, 2000: 45–46).

After an inter-party accord brokered by Commonwealth Secretary-General Emeka Anyaoku had been largely ignored by the CCM, the ruling party reinforced its violent crackdowns on the opposition in the aftermath of the 2000 general election which it had won under controversial circumstances. In January 2001, Tanzanian security forces responded to CUF-organised demonstrations by killing at least 35 and injuring more than 600 protesters in cooperation with local youth militias (Human Rights Watch, 2002).

Considering Dowd’s conceptualisation of pre-existing cleavages of conflict and contestation as a fertile breeding ground for violent Islamist mobilisation, one would expect the escalating tensions between CUF and CCM to have finally paved the way for the rise of Jihadist activists advocating violence as a legitimate response to the repressive authoritarianism of an entrenched “Christian-controlled” ruling party. ¹⁵ However, in stark contrast to this scenario, the islands’ two-party antagonism turned out to be a strong barrier to collective politico-religious radicalisation.

Violent Party Conflict as a Barrier to Islamist Radicalisation

Sparking sharp international criticism, the violence of January 2001 was followed by an inter-party dialogue that culminated in the signing of a second accord, the Mwafaka ¹⁶ II. Overseen by a newly established Joint Presidential Supervisory Commission, its implementation entailed the creation of a permanent voter register and the enactment of legislation entitling the CUF to two of the seven seats on the ZEC (Matheson, 2012: 98).

While the tensions between CCM and CUF eased, state–Salafi confrontations resumed. In 2001, the Zanzibari government created the office of the Mufti, a presidentially appointed religious bureaucrat tasked with mediating intra-Muslim disputes and administrating mosques, Quran schools, and Islamic holidays (Loimeier, 2009: 121). However, rather than perceiving him as an impartial broker, the islands’ Salafi activists rapidly accused the Mufti of serving a governmental project of religious control (Turner, 2008: 13). Their first major dispute with the new office arose during Ramadan in 2001 when the police arrested more than twenty leading JUMIKI members who had started the holy month one day before the Mufti’s official announcement. While the Salafis followed the Saudi timing, the Mufti, Sheikh Harith bin Khelef, sighted the moon locally (Loimeier, 2011: 13–14). A similar conflict occurred in February 2003 when followers of the Ansar Sunnah celebrated Eid al-Adha “too early.” Responding to this disobedience to the Mufti’s authority, the security forces entered a Salafi mosque in Stone Town using tear gas and plastic bullets to disperse the praying crowd. Six people were temporarily detained and charged with contravening Khelef’s rulings (BBC, 2003). One week later, an unauthorised protest rally against this police operation was violently dispersed as well (Loimeier, 2011: 131).

In March 2004, JUMIKI held another “unlawful” anti-Mufti demonstration – permission would have required the Mufti’s approval – which gathered hundreds of Ansar Sunnah supporters and led to the brief arrest of more than thirty people (BBC, 2004; Tanzanian Affairs, 2004). During the weeks following this incident, it seemed as if the islands’ Islamist movement, headed by figures like Sheikh Farid Hadi Ahmed and Sheikh Azzan Khalid Hamdan, both long-time companions of Sheikh Bachoo, was about to enter a militant path. In April, a number of minor, unsophisticated bomb attacks took place in Unguja, damaging several properties, including the Mufti’s residence, the house of Zanzibar’s Minister of Transport and Communications, and a police car (Tanzanian Affairs, 2004). Furthermore, a grenade whose pin had not been pulled was thrown into a tourist bar (Loimeier, 2009: 134). However, rather than signalling the ultimate escalation of the hostile interactions between the state and the Salafi civil society, these incidents primarily contributed to an increasing re-intensification of the CCM-CUF conflict ahead of the 2005 general election. Attempting to internationally discredit the CUF by depicting it as an Islamist party, high-ranking CCM officials, including Tanzanian President Benjamin Mkapa, implicated the opposition in the bombings (Loimeier, 2009: 134; Tanzanian Affairs, 2004; Turner, 2008: 26). The CUF leadership, which has repeatedly highlighted its secular orientation (Hamad in Burgess, 2009: 274–275; Turner, 2008: 26), responded to these allegations by demanding an independent international investigation, pointing out that the attacks might have been staged by the state in its endeavours to delegitimise any form of Muslim dissent (Loimeier 2009: 134; Tanzanian Affairs, 2004).

By May 2005, political violence and contestation had largely reshifted into the domain of multi-party competition. During the voter registration process that started in November 2004, youth militias linked to the CCM, the so-called Janjaweed, regularly attacked and intimidated residents in neighbourhoods believed to be CUF strongholds (Bakari, 2011: 248–251; US Embassy/Tanzania, 2005a). Moreover, in March 2005, clashes between CCM and CUF supporters at Daraja Bovu in Zanzibar Town in the aftermath of two parallel party rallies left several people injured (Bakari, 2011: 250). In the wake of these incidents, the Zanzibari government called in the Field Force Unit (FFU), Tanzania’s riot police, which launched Operation Dondora. ¹⁷ Under the pretext of keeping peace in Unguja’s Urban West Region, the FFU arbitrarily arrested and mistreated about thirty people, releasing most of them after a few days in detention (Bakari, 2011: 250; US Embassy/Tanzania, 2005a). Reacting to this police intervention, Zanzibar’s JUMIKI activists publicly condemned the harassment of innocent citizens (US Embassy/Tanzania, 2005a). In fact, in May 2005, the US Embassy in Tanzania reported that “Uamsho ¹⁸ has turned to human rights advocacy; specifically, calling the attention of the press and foreign embassies to abuses committed against presumed opposition supporters during Zanzibar’s voter registration period” (US Embassy/Tanzania, 2005b). This clearly shows that the resurgent CCM-CUF confrontation not only moved the focus of state repression away from the Islamic sphere but also drew Salafi leaders into conventional electoral politics.

More than five years later, state–Salafi tensions heightened again following the formation of a government of national unity between CCM and CUF and the launch of Tanzania’s constitutional review process. ¹⁹ With the CUF in government, it seemed as if Zanzibar might move into a new era of oppositional politics. Writing in the aftermath of the large JUMIKI-organised public rallies against the Union, LeSage (2014: 8) claimed that “the Government of National Unity created a new type of political vacuum in Zanzibar” and that “as a result, interest in alternative, non-governmental, and more radical leaders has grown”. Similarly, in late 2012, the CUF’s then Deputy Secretary General Ismail Jussa Ladhu declared that “by the time we woke up, we found ourselves engulfed by this religious group” (The Economist, 2012). For instance, after JUMIKI’s charismatic new leading figure Sheikh Hadi Ahmed went missing in October under mysterious circumstances, urban Unguja was rocked by two days of uncontrolled youth riots accompanied by the killing of a policeman and the arrest of more than fifty protestors (BBC, 2012a). When Ahmed eventually reappeared after three days, claiming to have been abducted by the security forces, he was detained on sedition charges together with six other JUMIKI representatives (BBC, 2012b).

Tensions between the Tanzanian/Zanzibari state and the islands’ Salafi community were clearly about to escalate. However, at the same time, the CUF leadership also rapidly readopted its confrontational stance vis-à-vis the CCM. In a public speech he delivered at a party rally a few days after the arrest of the JUMIKI activists, Hamad, then Zanzibar’s First Vice-President, blamed aggressive police crackdowns on JUMIKI’s followers on the island’s new President Ali Mohammed Shein as well as on the Tanzanian government. Furthermore, he openly questioned the militant image ascribed to the Islamist organisation by the CCM elite and decried the continuous deployment of ruling party-sponsored youth gangs (Kwanza, 2012a: 16:26–27:18 min; 2012b: 0:01–10:23 min).

Effectively absorbing the separatist sentiments inflamed by JUMIKI, the CUF used the constitutional review process as a platform to call for the establishment of a loose, merely treaty-based union accommodating two sovereign states (Hamad, 2013; The Citizen, 2014). This stance, accentuated by subnationalist party slogans like “Zanzibar kwanza” (Zanzibar first”) or “Mamlaka kamili” (full authority), was vigorously rejected by the local CCM whose cadres had come to value the current Union structure as the military as well as symbolic pillar of their hold on the Zanzibari Presidency (Paasch, 2015).

In the end, Tanzania’s CCM-dominated Constituent Assembly adopted a constitutional proposal that provided for the maintenance of the two-tier status quo (Branson, 2015). ²⁰ Against this background, the CCM-CUF coalition collapsed in the wake of another highly controversial presidential election held in October 2015. When it turned out that Hamad would defeat CCM incumbent Shein, the CUF leader declared himself the winner prior to any official ZEC announcement. However, ZEC chairman Jecha Salim Jecha annulled the vote, claiming that it had been marred by fraud (Throup, 2016). A rerun held in March 2016 was boycotted by the CUF, allowing the CCM to capture most of the seats in the House of Representatives and enabling Shein to win the presidential race with 91 per cent of the total vote (Ng’wanakilala, 2016).

Due to the rapid post-2010 repolarisation of Zanzibari party politics, the salience of the local state–Salafi antagonism declined significantly even though the JUMIKI clerics arrested in October 2012 have been held in custody on the mainland ever since. Rather than significantly fostering militant Islamist activity, it appears that the controversial detention of Sheikh Ahmed and some of his fellow activists was primarily capitalised on by CUF functionaries who repeatedly criticised the Tanzanian authorities for denying the preachers bail and for mistreating them in prison (CUF, 2014; Mtanzania, 2018). Overall, the CUF’s resilient subnationalist project reliably channelled religiously imbued disenchantment with Tanzania’s and Zanzibar’s CCM establishment into secular electoral politics.

Conclusion

This article has sought to contribute to the evolving theorisation of the causes and dynamics of militant Islamist activity in sub-Saharan Africa by providing a deviant case study of Zanzibar. Overall, the examination of the Zanzibari “anomaly” offers two major insights.

First, it qualifies Dowd’s theoretical propositions by indicating that whether or not pre-existing histories of non-Islamist violence act as enabling conditions for the expansion of militant projects of Islamist mobilisation depends on the precise nature of the conflict formations they entail. Where, as in Zanzibar, political violence is historically attached to electoral contests characterised by an extraordinarily high degree of mobilisational capacity – voter turnout in the island’s multi-party presidential elections has constantly exceeded 89 per cent (African Elections Database, 2012) ²¹ – the sustainable development of any alternative agenda of collective political action is blocked.

Second, the deviant case study complements Elischer’s state-centred thoughts on the “preventability” of collective Jihadist radicalisation by showing how electoral polarisation acted as a barrier to the emergence of militant Islamist networks in a subnational environment marked by a lack of restrictive institutional regulation of the Islamic sphere in the pre-liberalisation period. In the light of this finding, future theory-oriented research could explore how secular multi-party competition has shaped Salafi politics elsewhere in sub-Saharan Africa.

Critics might argue that rather than stemming from electoral mass mobilisation, the absence of substantial militant Islamist activity in the small Zanzibari archipelago is primarily the result of the significant local presence of the Tanzanian security apparatus ever since the creation of the Union in 1964. However, the recent political history of nearby coastal Kenya (see, e.g. ICG, 2018: 8–13; Mwakimako and Willis, 2014) clearly underlines that while strong coercive state capacity may prevent Jihadist movements from launching fully fledged insurgencies, its authoritarian use against the Islamic sphere is among the factors enabling the rise of such movements in the first place. In fact, it cannot be ruled out that in a regional setting troubled by persistent Al-Shabaab recruitment efforts (see ICG, 2018), Zanzibar will eventually turn into a more typical case with respect to the two explanatory frameworks discussed in this article. This is because since the controversial 2015/2016 Zanzibari general election, the local opposition leadership has been facing an existential threat to its political authority. In August 2015, the CUF’s mainland chairman Ibrahim Lipumba announced his intention to quit, only to declare his letter of resignation void several months later. Sparking a bitter disagreement with CUF Secretary-General Hamad and his followers, who insisted on the irrevocability of Lipumba’s resignation and accused him of acting as a “CCM puppet,” this move eventually resulted in the Hamad group’s defection to the Alliance for Change and Transparency-Wazalendo (ACT-Wazalendo) in March 2019 (The East African, 2019). Several days after this dramatic move, Tanzania’s Registrar of Political Parties was already threatening to deregister Hamad’s new party on the grounds that several Zanzibari members had allegedly burned CUF flags (DW Kiswahili, 2019).

If the Tanzanian authorities should succeed in blocking the Hamad group from participating in Zanzibar’s forthcoming general election, it would become extremely difficult for the islands’ former CUF leadership to survive as a serious opposition force. Under these circumstances, increasing anti-CCM youth radicalisation would become a realistic scenario.