Legacies of the authoritarian past: religious violence in Indonesia's Moluccan Islands (1).

In January 1999, sudden and surprising violence broke out between Christians and Muslims in Indonesia's Spice Islands. Previously seen as a stable province, Maluku (2) quickly became the site of devastating inter-religious strife. Thousands of people were killed in a spiral of violence over the following years. Little known in the rest of Indonesia and mainly neglected under the Suharto regime, Maluku became a core preoccupation of the Habibie, Wahid and Megawati governments.

This article focuses on the conditions that increase the potential for violent conflict to erupt. (3) Among these conditions, three are relevant to the case of Maluku: unresolved questions over principles of the nation; institutions that reinforce rather than defuse group identities, such as patrimonial relations under authoritarian rule; and rapid democratic transition. The particular confluence of these factors in Maluku created heightened tensions and uncertainties, and was compounded by the relative group size that was almost equal regionally but unequal nationally. These conditions made severe violence likely but not inevitable, nor necessarily lengthy and widespread.

I argue that the conflict in Maluku has its roots in the mid-1980s and 1990s, when Suharto shifted to Islamic groups for political support. Seen as a departure from the previous marginalization of Islamic groups under Suharto's New Order regime, it was widely interpreted as a significant policy change. This change led to heightened tensions between Christians and Muslims because of the uncertainty it created about the future role of Islam in the polity. A civic form of nationalism had been adopted as the basis of the Indonesian state, as institutionalized in the 1945 constitution and the Pancasila ideology, but it was imposed by authoritarian decree to terminate debate over the question of an Islamic state. (4) The New Order's policy change threatened to rekindle this debate.

Patrimonial features reinforced group identities and exacerbated tensions. Under the New Order's authoritarian institutions, few channels could be used to advance group interests. Powerful and lucrative positions in the civil service were major sources of material resources and protection. In Maluku, such competition increased divisions between Christians and Muslims, especially since Muslims had been disadvantaged. (5) In combination with the uncertainty surrounding the future role of Islam, tensions grew to unprecedented levels as Christians saw their interests potentially threatened and Muslims saw possibilities of redressing past imbalances.

Rapid democratic transition in this context was also destabilizing. The resignation of President Suharto in May 1998 began a democratic transition after more than thirty years of authoritarian rule. His successor, B.J. Habibie, initiated a process of institutional change leading to the country's first free and fair elections since 1955. This change had been quite sudden and people were unprepared for it. No one knew the extent to which the patrimonial features of the New Order system would be dismantled or whether the Pancasila ideology, which had maintained a quasi-secular orientation for the state, would continue to define the principles of the Indonesian nation. As a result, Muslims and Christians in Maluku feared losing their relative positions and access to resources. Muslims were concerned that Christians might reassert their regional dominance, while Christians feared an erosion of their status in a state more strongly inclined to favour its large Muslim majority.

Factors that would otherwise not have been sources of tension fed group fears because of this particular context. Spontaneous migration from Sulawesi and policies designed to encourage migration had been occurring for decades. They had created few frictions between Muslim migrants and local Christians, but this situation changed with fears of Islamization. As Muslims became the largest group in Maluku, where Christians and Muslims were previously equal in numbers, the disruption in the numerical balance of the communities in the region was seen as a threat for Christians. Well before the violence of 1999, therefore, religious identities in Maluku were highly politicized and the communities were prone to conflict.

Nationalism, ethnicity and conflict

Conflict between religious groups can be explained partly as a result of the national institutional context. The absence of certainty regarding the secular orientation of the nation can heighten tensions between religious groups. Furthermore, conflict can arise from political institutions that intensify group identities, particularly if these identities become the basis for networks of access to positions and resources.

In most states, an implicit or explicit conception of the nation defines inclusion, exclusion and the terms of inclusion for various groups. The literature on nationalism has often classified these conceptions as either "civic" or "ethnic" nationalism: (6) The former defines the nation's citizenship in the broadest inclusive terms, based primarily on birth or long-term residence in a territory; conversely, the latter limits citizenship to members of particular ethnic groups. Although admittedly a heuristic device, this dichotomy nevertheless clouds some of the most important elements of debate along the spectrum between pure civic and ethnic models. Issues such as representation of minorities, official languages, special representation and autonomy can occur in predominantly civic or ethnic models and can be sources of violent conflict as well as solutions for peaceful relations between ethnic groups. It is precisely in second-level conditions, in the negotiations about the particular mix of civic values and ethnic representation embedded in national conceptions and institutions that the sources of peace and conflict lie. I suggest, instead, that an analysis of how and why national conceptions evolve and change the balance of representation between various groups can shed most light on the sources of tension and conflict. The explanation resides with the negotiation and complexity of solutions found in the details of civic-ness and ethnic representation.

Formal political institutions reflect conceptions of the nation. Constitutions and special laws define the terms of citizenship, enshrine principles of representation, and give meaning to the structure and process of institutions. In the extreme ethnic case, citizenship is extended only to one particular ethnic group while it is denied or restricted for ethnic minorities, who are also given limited or no political rights and representation. (7) Further along the spectrum, some institutional arrangements explicitly recognize ethnic group identities and provide power-sharing arrangements, veto powers and other special measures that give equal political weight to each major ethnic group in a polity. (8) Other institutional arrangements give special emphasis to individual rights and duties, and follow principles that prevent the institutionalization of ethnic identities. (9) In such contexts, political representation is exclusively defined along territorial lines or proportional to population, with no special pre rogatives for ethnic groups.

In many cases, however, strategies are designed to enhance civic values and inclusiveness, while recognizing the complexities arising from ethnic diversity. A wide variety of institutional forms can be adopted. Horowitz mentions electoral incentives to foster inter-ethnic coalitions that reduce the political significance of ethnic identity and warns against the adoption of majoritarian formulas that risk giving recurrent majority power to a particular ethnic group. (10) Linz and Stepan suggest the creation of "state-nations" that combine the "collective rights of nationalities or minorities in a multinational, multicultural society and state, with the rights of individuals fully protected by the state." (11) While a useful critique of attempts at state-led nation-building, their concept of state-nations still requires a determination of the terms of inclusion of nations or cultural groups in the state, negotiation and compromise over collective rights, and institutional designs to reflect these understanding s. The outcomes of these debates and negotiations constitute in themselves a new conception of a nation that is meant to replace or re-imagine the previous national imaginations in a given territory. (12)

The political role of religion can raise problems of defining group rights relative to a conception of the nation that includes civic principles. This question has been most debated in Islamic countries, where the question of Islamic law and its application confronts principles of secularism and liberal rights embedded in the modern concept of the civic nation. There is no a priori reason, however, why the choice must lie between the adoption of an Islamic state and a concept of nation that is secular and civic. The pursuit of state instruments to enhance a Muslim way of life for Muslims can be compatible with democratic principles, protection of individual rights, and guarantees for religious and other minorities. (13)

Debate and negotiation over the state's role in promoting Islam can create fears among non-Muslims about restrictions of their rights. Without guarantees of individual and collective rights, and specific provisions about religious freedom, there is little reassurance that debate about Islam may not result in the development of a conception of the nation that excludes non-Muslims or reduces them to second-class citizens, subjected to rules and laws following Islamic principles.

Such guarantees can only be developed within a democratic framework. Without an institutional context that favours a broad debate, consensus, and agreement over conditions of inclusion in the nation, there is little scope to reassure religious minorities that they will be included as equal citizens. Authoritarian institutions and practices tend to exacerbate tensions, especially when combined with patrimonial features. In the latter case, restricted access to power and resources creates incentives to use informal networks, especially ethnic and religious ones. (14) Although not likely to be the cause of ethnic or religious identification in itself, the use and reinforcement of these networks for patrimonial purposes often intensify ethnic or religious divides. Furthermore, an authoritarian solution to the definition of the nation, the terms of inclusion of minorities and, particularly, the role of religion in the polity are likely to be maintained only as long as repressive instruments can prevent the emerge nce of alternative visions. Even though in some cases authoritarian institutions can allow for compromises to be reached between conflicting groups, thereby providing stability and consensus when institutions change, (15) more often groups whose visions of the nation have not been heard will mobilize as soon as authoritarian institutions weaken.

This makes rapid democratic transition particularly prone to violence. A more gradual transition can allow groups to reach a certain understanding about a conception of the nation that is sufficiently inclusive to prevent ethnic or religious mobilization with the removal of authoritarian controls. When transition is rapid, however, the removal of authoritarian controls opens up channels to pursue group interests. In particular, groups may feel threatened by a democratic context that can narrow the inclusiveness of the nation or set new terms of inclusion to the disadvantage of minorities. (16) In addition, patrimonial channels of access to power and resources are challenged and often targeted for elimination when democratic institutions are established. These combined factors can create heightened tensions between religious and ethnic groups if they already have tense relations and compete with each other.

In sum, three factors related to institutional change create conditions for religious conflict. First, uncertainty over a secular definition of the nation leaves open the question of an Islamic state and the terms of inclusion of non-Muslims. Second, patrimonial features of authoritarian regimes often reinforce religious and ethnic networks, thereby intensifying group competition and creating vulnerabilities in the face of institutional change. Third, rapid democratic transition fuels ethnic instability when combined with the two previous factors. It reopens unresolved questions over the secular conception of the nation, with new channels to express differences, whereas previously authoritarian instruments could be used to repress group demands. It also creates sudden vulnerabilities for patronage networks, thereby increasing group competition and uncertainty. The rapid nature of the transition prevents inter-ethnic compromise before a political opening and fuels the uncertainty of outcomes.

Ambon and Maluku in the Indonesian Republic

The ambiguity of the role of Islam in the Indonesian nation and patrimonial relations sustaining the New Order regime reinforced divisions between Christians and Muslims in Ambon and Maluku. The regime's policies and manipulative use of religion for political support had negative consequences, particularly since Dutch colonial policies had already created a divide among the religious communities.

Ambonese Christians were favoured during the colonial period, while Muslims were marginalized. The city of Amboina (Ambon) became one of the most important towns in the Dutch East Indies: it was the administrative centre and major port of the Spice Islands. The Dutch required the use of natives to staff positions in its colonial administration and to serve in the colonial army. Many Ambonese converted to Christianity, as opportunities for work in the colonial service were open only to Christians. They became one of the groups most loyal to the Dutch. They occupied staff positions in the administrative centre and were the principal group in the Dutch colonial army. Meanwhile, Muslim Ambonese were marginalized and isolated. They pursued their livelihoods in villages across the region with little involvement in the Dutch colonial institutions. (17)

When the Indonesian republic declared its independence in 1945, many Ambonese Christians resisted, as they feared inclusion in a new country with a Muslim majority. The new republic fought the Dutch for four years when they returned to regain possession of their colony after the Japanese defeat in 1945. An interim settlement led to the creation of a federation of independent states, many of which were partially or completely under Dutch control. When the Dutch conceded their defeat and left Indonesia, the republic returned to a unitary view of the state and dissolved the federation by political and military means. Bolstered by some support among the Christian Ambonese community, disgruntled officers of the former Dutch colonial army declared the independence of the South Moluccan Republic (Republik Maluku Selatan -- RMS). The Republic of Indonesia rejected this new state, especially since some Christian Ambonese had participated in the nationalist struggle against the Dutch and supported the republic. The In donesian National Army (Tentara Nasional Indonesia -- TNI) therefore was sent to the island to quell the resistance. (18)

As a result of the RMS incident, the position of the Ambonese in the republic remained ambiguous. Some nationalist Christian Ambonese played important roles under the regime of President Sukarno as Cabinet members and prominent leaders of the Protestant Party of Indonesia. Yet, after being the centre of the Dutch East Indies, Ambon and Maluku became remote areas of the archipelago. The region was tagged as "rebellious" and later, under the New Order, was kept under a tight security noose. More often than in other provinces, the positions of governor and regent (bupati) were filled by military officers and non-locals. Governor Akib Latuconsina, appointed in 1992, was in fact the first civilian governor and the first Moluccan to hold such a post under the New Order. The region remained relatively calm and passive since critics of the government were accused of supporting the RMS. (19)

For decades Christian Ambonese nevertheless benefited from the favourable orientation toward Christians and moderate Muslims under Sukarno's Old Order and Suharto's New Order. Proponents of an Islamic state opposed a nationalist vision based on modern principles and inclusiveness of all religious groups, but they were defeated. Old and New Order leaders used authoritarian means to quell demands for an Islamic state and sought support for their nationalist orientation among Christians and moderate Muslims. (20)

Nationalists had constructed a model based on a modified civic nationalism. The draft constitution of 1945 and the Pancasila ideology eventually became the written forms that spelled out the core elements of this nationalism. (21) The Pancasila ideology upheld the civic values of modern nationalism. Four of its principles appealed to modern values that nationalism was meant to enhance. In its original formulation, nationalism was its first principle, with the explicit meaning of the nation as represented by all the people living within the recognized territory of Indonesia (the former Dutch East Indies). Internationalism or humanitarianism was the second principle, which implied peaceful relations and respect between nations. The third principle was repre-sentative government, as the basic principle for the formation of political institutions. Social justice and prosperity constituted a fourth principle. The fifth principle, belief in one God, was primarily aimed at appeasing Muslims who wanted the establish ment of an Islamic state. The principle was sufficiently ambiguous to enshrine a principle of religious tolerance and recognition that all religious groups could freely practice their respective religions, while conceding that the nation would not be purely secular. (22)

Indonesia had a brief period of liberal democracy during which the nationalist vision competed with that of the Islamic state. Shortly after its independence and the final departure of the Dutch in 1949, the Constituent Assembly was mandated with the task of drafting a new constitution that would be ratified democratically. By the mid-1950s, a crisis had emerged over its difficulties in reaching a sufficiently large consensus over the basic principles for a new constitution. Islamists strongly opposed Pancasila and the nationalist proposals. Faced with this deadlock, President Sukarno suspended the assembly and the national parliament, and replaced the regime of liberal democracy with that of Guided Democracy. (23)

Guided Democracy, and its replacement by the New Order regime, marked Indonesia's period of authoritarian rule. Sukarno decreed a return to the Constitution of 1945 and Pancasila as the state ideology. Suharto solidified the authoritarian regime after he rose to power in 1965. He raised the spectre of communism to justify the need for more political control and use of the military. He reaffirmed that the Pancasila would be the basis of the Indonesian nation and the 1945 Constitution would remain. He moved against Islamic groups that were identified as potential threats to the nation based on Pancasila. After a few years, he imposed Pancasila as the sole ideology of all socio-political organizations. (24)

The New Order regime drew its support from secular nationalists, Javanist Muslims and Christians. (25) Pancasila and the Constitution of 1945 protected Christians from a Muslim state in return for their loyalty and support for the regime. Javanist Muslims constituted its core support group, while santri Muslims, divided among traditionalists and modernists, were generally marginalized. The modernists were particularly sidelined since they had been among the strongest supporters of an Islamic state. Even if few supported an Islamic state by the 1970s, many wanted to see greater representation for Muslims and policies consistent with the fact that Indonesia had a large majority of Muslims. (26)

Suharto kept control over religious groups through repression, political and bureaucratic control, and patrimonial relations. He significantly weakened political parties in the mid-1970s by forcing their amalgamation into three political parties, including the government party Golkar. Military personnel occupied a large number of positions at all levels of government and the bureaucracy was streamlined to reinforce central control and ensure order even in remote areas of the country. Elections were held regularly but Golkar's dominance was maintained by structural measures, intimidation and vote manipulation. The media was censured and freedoms of expression and association were curtailed. The regime also co-opted leaders of various organizations and groups by providing positions and financial opportunities in return for maintaining control over their constituents. Suharto even attempted to control religious organizations through these means, although with less consistent results. (27)

Through these patrimonial linkages, Christians were favoured under the New Order regime. They were particularly well represented in cabinet positions and in the military. Major economic positions were held by Christian technocrats, and Christians were well represented in top positions of Golkar, the military and the bureaucracy. The early 1980s seemed to mark a high point in the strength of Christians, especially under armed forces commander Benny Murdani, one of the most powerful commanders the Indonesian armed forces ever had.

In the meantime, Muslim groups were mostly marginalized from the political process and their political activities curtailed. Nahdlatul Ulama, the country's largest Muslim organization, abandoned its political role in the Unity Development Party (PPP) in the early 1980s since there was little space to advance its interests from within the political system. A low point in the regime's relations with Muslim groups was reached in the 1984 Tanjung Priok events, when soldiers opened fire on a crowd that protested the military's lack of decorum in a Muslim prayer house. Tensions had been running high as Muslim groups were demonstrating against the regime's anti-Muslim policies. (28) From the late 1950s to the mid-1980s, Suharto repressed Islam as a set of political principles, a source of political mobilization, or a basis for political organization.

In Maluku, the effects of these policies reinforced divisions between Muslims and Christians. In other areas of Indonesia, divisions between Muslims were often more significant than those between Christians and Muslims, but in Maluku, the latter were most important, especially since they were almost equal in numbers and religious divisions had been created under colonial rule. The military maintained a strong presence and controlled the top positions in the province. Below these top positions, military authorities allowed competition between Moluccan Christians and Muslims. Patrimonial networks were formed and could be used to control each community.

Christians benefited from the Pancasila state in which they were well represented and the patrimonial system extending to the region. Because of their strong role in the colonial civil service and the dominance of missionary schools, Christians were well positioned to play a stronger role than Muslims. They maintained their educational advance and overwhelming representation in the regional bureaucracy. Furthermore, the Ambonese prized civil service positions, which were highly regarded under the Dutch, so they sought these positions over any other occupation. Below the regional command that was firmly held by military outsiders, Christian Ambonese maintained a strong position regionally and promoted the interests of their community.

Over the years, however, it was expected that Muslims would increase their political weight. The introduction of a secular system of education allowed Muslims to catch up and expect positions previously dominated by Christians. Since religion had already been politicized as a result of Dutch policies and the RMS incident, religious identity played a strong role in the competition for positions. The state sector became divided into sections controlled by each group. A particularly interesting example was the University of Pattimura (UNPATTI). The powerful Education Faculty (FKIP), one of the largest faculties in the university, was almost exclusively staffed with Christians well into the 1990s, while other departments included more Muslims. Within the regional and municipal bureaucracies, such tendencies were common. Christians resented the growing presence of Muslims in areas they previously controlled, while Muslims saw their advancement as a just redress since they had been previously marginalized in the re gion.

Local historical factors therefore set the stage for tensions to grow along religious lines. The Ambonese Christian role under the Dutch had created a political divide between Christians and Muslims, even among Ambonese with common origins, and it remained salient under the independent republic. The RMS issue gave the New Order state the means to repress opposition and maintain strong military presence, while allowing elites to benefit from patronage networks. Policies designed to contain potential uprisings and the marginalization of the region under the New Order regime allowed religious networks to be a source of access to government positions and to protect respective communal interests.

Islamization and the struggle for government positions in Maluku

Tensions between Christians and Muslims rose steadily when Muslims became favoured under Suharto's change of policy in the late 1980s and early 1990s. Previously favouring Christians and Javanist Muslims in his government, Suharto began to seek support from Muslim groups, especially modernist Muslims who had been previously marginalized. With the founding of the Association of Indonesian Muslim Intellectuals (ICMI) in 1990, Muslims throughout Indonesia saw growing opportunities for advancement, while Christians at all levels of government were replaced.

Suharto's approach to Muslim groups changed in the late 1980s. He began to project an image of a good Muslim, with actions such as a well-publicized pilgrimage to Mecca in 1991. He also began to support policies favourable to Muslim groups, including the relaxation of restrictions on schoolgirls wearing the jilbab headgear and the establishment of Indonesia's first Islamic bank. This change occurred in the context of a parallel side-lining of general Benny Murdani, his allies in the military and Christians from top positions in the armed forces. Benny Murdani was discarded after becoming critical of the business activities of Suharto's children and apparently suggesting that Suharto step aside. (29) The process took a few years, as many high ranking generals were close to Murdani and their sudden removal might have been too disruptive.

Meanwhile, Suharto sought support from previously marginalized Islamic groups and gave his approval to the creation of IGMI. The latter provided a new political vehicle for modernist Muslims who had been shunted aside under the New Order. As part of his strategy to create new patrimonial linkages, Suharto reduced links with his former support group among Christians and Javanist Muslims and removed Christians from cabinet positions, which culminated in the 1993 cabinet, with its overwhelming representation by IGMI members and only token positions for Christians. Top positions in the military went to generals with strong credentials within the Muslim community. An apparent "greening" (for the colour of Islam) of the New Order had taken place.

In Maluku more than elsewhere, these policies had the effect of heightening tensions between Christians and Muslims. Moluccan Christians became defensive, as they were worried about a trend toward the Islamization of government and a loss of their dominance in the region. As one Christian leader noted: "With Pancasila (30), [Muslims] can't change the state, this is fundamental. For the younger generation of Muslims, if they can't change the state then why not try to change society. That's what happened with the creation of ICMI: a deliberate process of Islamizing society." (31) Moluccan Muslims, on the other hand, found a political opportunity to break Christian regional dominance and establish a more equitable balance between the two communities in government positions.

The governor of Maluku, Akib Latuconsina, a Muslim, faced many difficulties in his attempt to redress past imbalances in the allocation of positions and resources. Shortly after his nomination, he accelerated the appointment of Muslims to the higher echelons of the bureaucracy, thereby displacing Christians from former strongholds. These nominations further politicized religious identities. (32)

Between 1992 and 1996, the governor managed to place Muslims in all of the most important positions within the regional bureaucracy. All of the regents were Muslim in 1996, even in Central and Southeast Maluku where Christians are a majority. Regional heads of national departments, who were previously almost all Christians, were replaced with Muslims, as, for example, in the education and health departments. By the mid-1990s, Christians felt that they were left with very little control over top positions. (33)

Furthermore, most new teachers hired by the government were Muslims even though most of the graduates from local education faculties were Christian. Christians had long dominated education and 90 percent of teachers were Christian. The FKIP at UNPATTI produced most of the new teachers and almost all were Christian. The new head of the education department, a Muslim from Southeast Sulawesi, had decided to hire mainly Muslims and FKIP saw its share of new teachers decline to 10 percent. Many new teachers originated from outside of Maluku since there were insufficient numbers of Muslims being trained in the region. (34) These kinds of policies were increasingly resented by Christians, even though they were designed to create a fairer balance between Christians and Muslims.

Struggles for two key positions were especially crucial in the mid-1990s. The positions of mayor of Ambon and rector of UNPATTI were seen as Christian strongholds and the Christian community strongly resisted attempts to fill these positions with Muslim candidates. A military officer, Yohanes Sudiyono, a Javanese Catholic close to the Protestant community, held the position of mayor. During his tenure, he had promised that Christian Ambonese positions would be guaranteed. He had even hired Christians, who then represented about 90 percent of employees in the city district office. Governor Latuconsina, who sought more Muslim representation in the city government, was unhappy with Sudiyono's approach. When the time for renewal came about, the Democratic Party of Indonesia (PDI) representatives in the local assembly (DPRD-II) proposed that Sudiyono be returned to his position. Latuconsina, however, favoured Diponegoro, another military officer but a Muslim. The PDI and Christian Ambonese rejected the candidate and mounted a strong resistance to the governor's choice. (35) In the end, the governor chose not to further alienate the Christian community and selected a Christian candidate, Chris Tanasale.

Governor Latuconsina met with another strong protest from Christians in his attempt to appoint a Muslim as rector of UNPATTI. As a former head of the Economics Faculty, Latuconsina apparently had tense relations with the rector of UNPATTI, Jan Nanere. The latter was accused of favouring Christians in his sponsorship of staff being supported to pursue graduate studies abroad. The governor asked Nanere to step down from his position, citing the age limit as the official cause. His proposed replacement was the local head of ICMI, a candidate who was supported by the governor but was not on the list proposed by UNPATTI's senate. The deadlock lasted eight months. During this time, 230 UNPATTI professors threatened to resign and student demonstrations were organized on several occasions. In the end, Secretary of State Moerdiono resolved the issue and a Christian, Mus Huliselan, was finally appointed. Muslims had been rectors in the past, but at a time when religion was not as politicized. In the national context o f the 1990s, this kind of event stirred high emotions among Ambonese Christians and Muslims. (36)

The allocation of government positions assumed a heightened symbolic value that, interpreted within a context of growing religious tensions, could be particularly conflictual. As pointed out by many Ambonese, it was normal to accept a gradual increase in the number of positions for Muslims to redress the imbalance that had favoured Christians since Dutch colonial times. Furthermore, Muslims in powerful positions had appointed Christians to high-level positions and Christians had appointed Muslims: "[even the former governor] a Catholic, gave Muslims important positions in government...since [the Muslims] represented more than 50 percent of the population, it was important to start giving them important positions." (37) But as the context changed, and Christians became displaced on a national basis, the effects in Maluku were multiplied. Every position lost became perceived as a threat to the position of Christians in the region and heightened the fear that Islamization was occurring across Indonesia. "If you look at Java, it was not impossible that a Christian be appointed as camat (district head) or bupati (regency head). In the past there was no problem but now it is impossible. In the army there are still some Christians, although less. And ministers, from 6 or 7, there are now 1 or 2." (38) In this kind of political context, the rapid inclusion of Muslims at the top of the bureaucracy created concern. Even those, such as Sammy Titaley, the head of the Protestant Church of Maluku (Gereja Protestan Maluku, GPM) who supported the governor's attempt to create a balance between Muslims and Christians, and who thought that the threat to Christians was overstated, still expressed some worry:

I reminded the governor that Christians also need to be taken care of. It is right to promote the interests of Muslims to create an equilibrium but I tell him: "Don't forget you're not the governor of Islam but the governor of Maluku, of all Moluccans." One recent example is the regional head of the Department of Education and Culture. Now, he is a Muslim and caretaker until a new person is found. There is a Christian who is quite good and most capable of being in this position. If the governor takes a Muslim from Java and promotes him as a candidate and does not promote the Christian, then he is not respecting Pancasila and he has ulterior motives in mind, because this is a clear example of where an Ambonese Christian is competent and able to assume the responsibility. This would be a clear sign that he is not respecting the need for equilibrium between Christians and Muslims. (39)

This line of argument was one of the most moderate among Christian Ambonese leaders. Many others expressed views that more directly accused the governor of pursuing the interests of Muslims in general and ICMI in particular.

Even some Muslim leaders acknowledged the growing impact of ICMI and the national context on regional politics. Abdullah Soulissa, the head of Ambon's al-Fatah Mosque Foundation, and one of the most respected Muslim leader in Maluku, argued that it was difficult to contain the flow of Islam in a country with 85 percent Muslims. He thought that it should be easy to create a balance in Maluku, however, since the Christian and Muslim communities were almost equal in numbers. Yet, he acknowledged that he understood Christian fears about government positions in Maluku and, furthermore, about the increasing amount of Islamic programming on television that, he thought, was excessive and should be more sensitive to ensuring good Muslim-Christian relations. (40)

Therefore, the allocation of government positions was being interpreted within the context of perceived Islamization of Indonesia. The result was an intensification of the political value to both religious communities of gaining positions in top levels of the bureaucracy. Muslims sought a redress for past imbalances and discrimination in favour of Christians, while Christians feared not only a loss of their dominance but also a loss of a proportionate representation and role in the state's institutions. At stake was access to resources and patrimonial networks, as well as fears that Christian inclusion in the Indonesian nation would be reduced by Islamization.

The impact of migration

These tensions were compounded by rapid migration that changed the balance between Christians and Muslims. From a roughly 50 percent divide between Christians and Muslims, migrants from Sulawesi and Java significantly increased Muslims' advantage relative to Christians in the region. Furthermore, some of the Muslims who migrated to Maluku, such as the Bugis or Butonese, gained increasing control of the local trade and established successful small businesses. As a result, Christians became concerned that "outsiders" were more successful economically and were also threatening the balance between Christians and Muslims that had been maintained for several decades.

The transmigration program attracted many Muslims to Maluku and triggered some accusations against the government. Through official sponsoring of migration from overpopulated areas, the government sought to alleviate pressures on scarce land. The program targeted Java in particular and Maluku was chosen as a recipient area because of its vast land and relatively small population. Given that Java and the rest of Indonesia is overwhelmingly Muslim, it could be expected that the vast majority of transmigrants would be Muslims as well. "Before, the majority was Christian [in Maluku] but with transmigration, Muslims are a majority in the province... [Given the population problems], it makes sense to have transmigration, which is necessarily going to be mainly Muslim since 90 percent of the population is Muslim. Locations with population problems are mainly Muslim so of course Muslims are the ones who go to Burn and Seram," the main transmigration areas in Maluku. (41)

Yet, some Ambonese were suspicious of the government's intentions. They feared that there was a deliberate attempt to prevent Christians from migrating to Ambon. Mgr. Andreas Sol concurred with this view: "Almost all of the transmigrants are Muslim. If they write down that they are Catholic or Protestant, they don't have permission to transmigrate. Sometimes, they write down "Muslim" and, once they get here, they see the minister/priest and say that they declared themselves as Muslims just to transmigrate." (42) This example suggests that, at the very least, rumours abounded linking Muslim transmigration to the government's turn to Islam:

Ambonese Christians were therefore viewing the increase in the number of Muslims as a deliberate government policy. But the evidence supporting the argument of the government's deliberate discrimination against Christian transmigration is not very strong and may not, in fact, have affected the overall numbers, given the large number of Muslims in Java.

Transmigration is mainly Muslim and, as a result, there are more and more Muslims in Ambon and more mosques being built. There is a limit to this. Eastern Indonesia won't accept this....In East Timor, many Bugis were encouraged to establish commercial ventures and all their shops were burnt down. They had to leave. This is a sign that the Timorese won't accept this. It's th e same in Irian....This is a sign that Eastern Indonesia is not willing to accept Islamization on that scale. (43)

Instead, most migration to Maluku was spontaneous. While the influx of Javanese resulted from the transmigration program and affected mainly Burn, Seram and parts of Halmahera, most of the migration came from Sulawesi. The Bugis and Butonese had long migrated to Maluku in search of economic opportunities. Along with economic development in Indonesia, many were attracted to Ambon and other areas of Maluku to develop commercial enterprises.

The increasing presence of Butonese and Bugis had an impact on the economy. "Moluccans have difficulty competing with people from outside. So all the middle economy (commerce) is controlled by Butonese and people from other regions. The big problem is that the Ambonese and the Moluccans don't have the capacity to compete in the area of commercial enterprise. They are only interested in salaried jobs, especially civil servants." (44) As the government has opened up more areas by building roads and infrastructure in transmigration and other areas, the Bugis and Butonese have moved in and seized control of commerce. So, as a result, the Ambonese and Moluccans saw Muslim outsiders, who increased their wealth relative to the Ambonese, dominate these areas of the economy. (45)

The Butonese and Bugis could increase their wealth while economic opportunities were diminishing for Christian Ambonese. The civil service offered fewer opportunities for the Christian Ambonese in urban areas, whereas traditional agriculture was on the decline. Villagers in many parts of Central Maluku relied on the production of spices as their main livelihood. As the price of cloves and other spices plummeted in the 1990s, they found themselves deprived of an important source of revenue. The fisheries industry also encountered difficulties as large-scale over-fishing in the region contributed to an important decline in fish stocks. (46)

Economic differences between the Butonese/Bugis and Christian Ambonese raised tensions along religious lines. Christians noted the effect on economic distribution of wealth but also, more importantly, the impact on the proportional numbers of Christians and Muslims in Maluku and the effects on Christian-Muslim tensions. While the ethnic Chinese were also heavily involved in the commercial economy, they were rarely mentioned by Ambonese Christians as a problem. The fact that they were Christian may well have contributed to the relatively lesser sense of threat from their strength in the middle economy. Furthermore, tensions in Ambon City were mainly between Christian Ambonese and Muslim migrants. Frequent fights occurred between youths in the predominantly Christian neighbourhoods of Mardika and the neighbouring one of Batu Merah, with a high concentration of Butonese and Bugis migrants. (47)

Migration to Maluku became a source of tension when relations between Christians and Muslims deteriorated. Migrants from Southeast Sulawesi and other regions had come to Maluku for decades without significant effects on political stability, but with growing tensions between Christians and Muslims, the religious identity of migrants became a politicized issue. Tensions were so high that Christians feared a deliberate government policy of Islamization through transmigration, despite little evidence to support this claim. Nevertheless, with more Muslims migrating to Maluku and their relative success in local commercial ventures, migration became an additional irritant to religious relations. Small violent events had already occurred between Ambonese Christians and Muslim migrants in Ambon city, where large-scale violence would later erupt. While there was no a priori reason why migration should contribute to conflict, in the context of growing polarization between Christians and Muslims in the region, migration became one of many factors that would lead to violent conflict.

The eruption of violence

The eruption of violence and its intensity were a direct outgrowth of the tensions between religious groups. The violence began with a small incident that degenerated into a large-scale riot and ethnic war. Initially contained to Christian Ambonese and Muslim migrants, the conflict rapidly included Muslim Ambonese against Christian Ambonese as well, thereby showing underlying tensions. The riots spread to other areas of Central Maluku and then to South and North Maluku, involving large sections of the population in a continuous stream of rioting.

The explanation for this violence is complex. There is no doubt that unique factors caused initial riots to degenerate into prolonged ethnic warfare. (48) Among these, one can point to the possible role of provocateurs and the involvement of security forces. The use of unknown provocateurs and the bias of security forces are not uncommon in Indonesia's history. Violence can be explained in part by the habitual use of extra-judicial forces, such as gangs and thugs, by the authoritarian state and their continued use after the demise of Suharto. (49) Yet these factors, in turn, occurred in a rapidly changing institutional context. Rapid democratic transition followed thirty years of authoritarian rule under President Suharto. This transition created much uncertainty about the future of Islam in the polity and changed the patrimonial structures that had sustained authoritarian institutions. (50) In addition, policies to decentralize power to the regions intensified the religious dimension of local competition, as was evident in the spread of conflict to North Maluku.

The following description of the conflict's evolution shows that the changing institutional context tipped an already tense relation between Christians and Muslims into violent conflict. Provocation and the involvement of security forces had such repercussions because of the already high regional tensions, which were explained as an outgrowth of fears of Islamization, attempts to redress past injustices toward Moluccan Muslims, unintended effects of migration patterns, and competition for resources and government positions. Rapid democratic transition tilted these tensions into conflict by raising the potential costs associated with the new Indonesia that would result from this institutional change.

The first large-scale riot erupted in January 1999. It began on January 19, Idul Fitri, the last day of the holy Muslim month of Ramadan. A minor incident involving a local minibus driver and local youths turned into three days of bloody rioting. Accounts of the incident vary according to different sources, with some blaming the Christian driver of the minibus of wounding his Muslim attacker with a knife, while others recount that the driver was attacked and had to flee. Nevertheless, supporters of both the Christian driver and the Muslim youth began to spread word of the incident and emotions ran high in their respective communities, the Muslim quarter of Batu Merah and the neighbouring Christian quarters of Amantelu and Mardika. Again, there are contradictory accounts of which group attacked dwellings in the others' quarter and began the subsequent rioting. Groups began rampaging through the area. The destruction subsequently spread to other areas of Ambon City. When news began to spread that churches and mosques had been burned down, the violence escalated further.

The violence continued during the subsequent three days. (51) It spread to the communities of Batu Gantung, Waringin, Benteng Karang, Passo, Nania, Wailete, Kamiri, Hative Besar and others. Violence was also reported in the village of Hila several dozen kilometres away from Ambon City. It appeared that news of the violence in Ambon City had spread rapidly and caused other communities to mobilize. In most cases, Christians tended to target Muslim migrants, the Butonese and Bugis. They burned down many of the markets and shops dominated by these groups, whereas they spared the Chinese, who were mainly Christian. As for the Muslims, they attacked houses and districts where Christians were majorities and several of these areas were burned to the ground. (52)

The incident resulted in several deaths and widespread destruction across Ambon City. At least forty-eight people were killed, and several churches and mosques were burned. Buildings, cars and motorbikes were also destroyed. (53) Several hundred people sought refuge in local churches, mosques and government offices. The city of Ambon was left with an unprecedented scale of fear and destruction.

The original incident and subsequent brawl between local residents were not unprecedented. Local Ambonese recall regular outbreaks of violence between the Christian residents of Mardika and the neighbouring Batu Merah, mainly populated by Muslim migrants from Sulawesi. (54) Usually, fighting would break out between youth gangs from both quarters and would remain contained. A similar fight had broken out in Hative Besar on December 12, when a soldier tried to attend a village dance party held by Christians. A fight broke out between Christian youths and soldiers, with subsequent rioting leading to the destruction of a few houses. Again, such incidents were not uncommon since Muslims, both local migrants and personnel from other areas in Indonesia, often disapproved of Christian parties that involved ballroom dancing and drinking. The incident was resolved through traditional appeals to religious and inter-cultural tolerance. (55)

The greatest surprise was the spread of the violence to other parts of Ambon and the rapid increase in its scale. This development was a departure from past occurrences. While Muslim-Christian relations had been tense in the past, they were contained through dialogue between Muslim and Christian leaders while the military and police kept a lid on violence.

This time, the violence erupted when Muslim-Christian tensions were at a high level and a rapid democratic transition was under way. President Suharto had resigned in favour of his vice president, B.J. Habibie, in May 1998. Habibie began a process of transition by announcing the removal of restrictions on political parties and the media, allowing more freedom of expression, and announcing parliamentary elections for the following year. Much uncertainty surrounded the outcome of these changes: from a complete transformation of Indonesia's political institutions to a return of authoritarian rule. Certainly, the military maintained a strong role and many groups remained supportive of Suharto's authoritarian regime. Some pro-Suharto groups were suspected of attempting to destabilize the country by provoking riots and communal violence. (56) There was evidence of such provocation in incidents of inter-religious violence in Ketapang and Kupang only a few months before. Provocation and prior events had an effect in Ambon but they were compounded by the regional sources of tensions between Christians and Muslims. The Ambon incident triggered conflict because conditions were ripe for violence.

After a very short respite in Ambon City, the conflict began to spread to neighbouring islands in central Maluku. On February 3, Christians armed with sharp weapons attacked Muslims whom they had initially invited for a peace initiative in Kairatu, on Seram Island. The following day, Muslim Ambonese and Butonese burned Christian houses in the Waitasu hamlet. On another island close to Ambon, Saparua, Christians burned down a Muslim dormitory, leading to a mobilization of Christian and Muslim communities that fell short of open violence when local leaders successfully calmed their followers. (57) On February 14, at least twenty-three people were killed in a fight between Muslim villagers of Pelauw and Christians from the neighboring hamlet of Kariu, which was almost completely burned to the ground. Many of the deaths were apparently a result of gunshot wounds inflicted by police and army units attempting to stop the rioting. (58) The violence involved local Muslims, with some support from Butonese migrants an d Muslims from neighbouring islands who rushed to Haruku upon hearing about the beginning of hostilities. Christians and Muslims both made accusations that security forces had assisted the other camp. These were the first reports that security forces appeared to be taking sides rather than restoring order, and that they were willing to use live ammunition against local residents. (59)

Continuing violence around Ambon Island soon fed renewed rioting in Ambon City. The city had remained very tense after the January trauma. Isolated incidents continued to occur and small indications of possible renewal of conflict made crowds panic. The Haruku incidents spread fear among residents in Ambon City who stayed at home and closed their shops. Small groups attacked a few houses, including that of the head of the AlFatah Mosque Foundation, Abdullah Soulissa. (60) On February 23 several Butonese people were attacked and houses were destroyed by rioting mobs. Armed conflict broke out in the villages of Waai (Christian), Tulehu (Muslim) and Liang (Muslim), where residents from the villages became engaged in an "all-out war." (61)

By early March, Ambon city was turned into a "war zone." Violent clashes on March 1 set Christians from Ahuru hamlet against people from the Muslim Rinjani hamlet, with contradictory reports about the instigators of the violence. Again, police opened fire against the rioting crowd. At least thirteen people were killed and nine others wounded. The incident angered Muslims even beyond Maluku because of rumours that the initial attack had taken place in a mosque. Although these reports were highly contested, they had the effect of increasing the Muslims' sense of outrage. Similarly, an attack close to Ambon's Silo church on March 6 outraged Christians. Barricades were set up by local residents to protect their respective Muslim and Christian districts, as well as local checkpoints. (62) The violence was settling in and escalating.

Much of the initial reaction to the violence focused on finding the provocateurs of the January incident. The media, military and police personnel, as well as many politicians, especially the leader of Nahdlatul Ulama, Abdurrahman Wahid, pointed to evidence of provocation in the series of violent conflicts between religious groups in Ketapang, Kupang, and then Ambon. Public accusations were made against Yorrys Raweyai, a leader of the Pancasila Youth, an organization that had been known to use illegal and often violent methods to support the interests of Suharto's New Order. Although not in Ambon himself, Yorrys was suspected of fuelling the violence in Ambon, in coordination with a disgruntled former mayor of Ambon City, Dicky Wattimena. (63) Some pointed to the events in Ketapang, which had involved a gang of Ambonese criminals who were subsequently rumoured to have gone to Ambon shortly after the conflict in Jakarta in December 1998. (64) The press dissected the details of the events to prove the theory o f provocation. Army and police personnel also engaged in the speculation while never publishing any names of suspects. It was common among the security forces to blame provocation since it provided a means of containing the conflict and avoiding the confrontation of more serious issues that could underlie such violent outbursts. Yet, despite months of speculation about provocation and demands for the instigators of the Ambon violence to be prosecuted, only local rioters were arrested and charged, while no strong evidence emerged to condemn any of the suspected provocateurs.

When it became apparent that the conflict was spreading and intensifying, the public focus shifted to concern about the management of the conflict. By March, Christians and Muslims were accusing the local security forces of taking sides in the attacks. Much evidence was presented to suggest that police and army personnel were responsible for many of the deaths that occurred. Troops from the elite army strategic reserve, Kostrad, were sent to Ambon after the January incident to support the local army command. Many of the soldiers, however, were Bugis and Makassarese, which only raised the suspicions of local Christians. Moreover, Christians were worried about the partiality of troops, given that the head of the Ujung Pandang regional command was a Muslim Ambonese. (65) Under pressure from these accusations and from failure to control the violence, Armed Forces Commander General Wiranto ordered the removal of the Maluku regional police head. Wiranto admitted that some security personnel had participated in the conflicts and taken sides. (66) A marine corps from Java was sent to Ambon as reinforcement and as a force that could be perceived as more neutral in the conflict.

These responses came amidst rising emotions across the country about the conflict in Maluku. There were cries of outrage that people had been attacked in a mosque in Rinjani on March 1. Groups demonstrated in many major cities, including Jakarta, Semarang, Surakarta, Pekanbaru, and others. They demanded that the government take rapid action to stem the violence in Maluku and protested against the massacre of their Muslim brethren. Most of the demonstrators were Muslim student groups and youth organizations. In Jakarta, the Indonesian Committee of Muslim Students (Komite Mahasiswa Muslim Indonesia, KAMMI) organized a large demonstration at the Al-Azhar mosque in Jakarta. Many demonstrators called for a jihad (holy war) to protect Muslims in Ambon, and organizations registered volunteers to be sent to Ambon for the war. Organizations such as Furkon (an Islamic youth organization), the youth wing of the PBB party (Partai Bulan Bintang, an Islamic political party), KISDI (Komite Indonesia untuk Solidaritas Dunia Islam, the Indonesian Committee for World Islamic solidarity), and the FPI (Front Pembela Islam, Front for the Defense of Islam) all called for volunteers to defend Islam in Maluku. (67) While none were sent at the time, they illustrated the high sensitivity of the religious conflict.

In the following months, the conflict began to spread to other areas of Maluku. In April, violent clashes set Catholics against Muslims in Kei, Southeast Maluku, located far away from Ambon. Initially triggered by an incident involving the exchange of insults between two youths on March 27, a subsequent fight between religious groups degenerated into violence and destruction of property for the next six days. The violence spread to more than thirty villages and to the local capital city of Tual. Eighty people were killed, more than sixty people were wounded and hundreds of houses were destroyed. Explained by the local police as an outgrowth of a local conflict, the violence in fact mirrored the clashes in Ambon, with Muslims adopting the label "White group" against the Christians identified as the "Red group, as they had done in Ambon. (68)

With the conflict spreading beyond Ambon and Islamic groups mobilizing across the country, the government sought solutions to the conflict. In March, the government sent a delegation from the military, composed of prominent military officers of Ambonese origin, who had spent much of their life in the region. This delegation met with various groups in Ambon to try reconciling the differences. The latter's efforts were met with more bombs and violent outbreaks, in part because of the local population's growing resentment of the armed forces' role in the killings of the past months. (69) After previously removing the chief of the regional police, the government then created a new military command for Maluku (Kodam XVI Pattimura), with a Moluccan, Brigadier-General Max Tamaela, at its head. This measure was taken to remove the role of troops from Sulawesi, many of whom were Bugis, and restore confidence in the armed forces. Furthermore, it could enhance the control of the armed forces in the region. Ironically, on the day the new command was inaugurated, renewed violence exploded between groups from Batu Merah and Mardika. They had been brought together for the annual celebration of the regional Moluccan hero, Pattimura, but the celebrations turned into another wave of violence and destruction instead of an occasion for peaceful reconciliation. (70)

The last months of the Habibie presidency were marred with a continued recurrence of violent incidents throughout the region. In July, violence broke out in the village of Sirosiri on Saparua Island and in the town of Saparua after Muslims discovered that a Christian from a neighbouring village was intentionally damaging the village's clove trees. (71) As tension rose again in the region, Christians and Muslims fought each other in Poka and Rumah Tiga in the outskirts of Ambon and the violence eventually spread to Ambon City. For many days in July and August, both sides unleashed almost unprecedented destruction and fury. The whole city was engulfed in violence, hundreds of shops and houses were destroyed and fighting broke out in villages all over Ambon Island. (72) The armed forces opened fire once again on Muslims in front of the Al-Fatah mosque. By the end of the month, Ambon City and surrounding hamlets were severely damaged and an increasing number of people were living as refugees in the Al-Fatah mosq ue and Silo church, as well as other places of worship around the city. It was the most severe violence since the January outbreak, with a higher number of casualties.

Ambon was left in a state of disarray after this second wave of large-scale violence. Thousands more people began to leave the city and the area. Refugees who had returned in March with expectations of a more peaceful future left with no intention to return. Violent incidents continued to erupt on a regular basis and normal activities could not be restored. Provincial and district departments were relocated to other regions of Maluku, as employees could not safely reach their offices. Local markets and schools were segregated along religious lines and opened in territory secured by each community. Permanent outposts and borders were set up between Christian and Muslim quarters. (74)

The armed forces continued to demonstrate its inability to control the violence and to prevent its members from taking sides. Various reports implicated different units present in Ambon. The Mobile Brigades (Brimob) of the regional police were accused of favouring Christians during clashes in September, while the elite army strategic reserve Kostrad forces were accused of supporting the Muslim side. There were even reports of clashes between members of these forces, although such incidents were not clearly confirmed. Nevertheless, there was enough evidence to force the regional military commander Brig. Gen. Max Tamaela to admit to some irregularities among the armed forces. Pressures began to rise for Tamaela's removal from his command. (75)

The crisis in Ambon began to take on a dynamic of its own but reflected the deep distrust between Christians and Muslims. Maluku was ripe for conflict as inter-religious riots that had risen in the last few years of the Suharto regime intensified after his resignation. Local competition for resources and positions, the polarization of religious identity, fears over the future opportunities and threats to each community provided a fertile ground for violence to erupt. Rapid democratic transition increased the uncertainty of outcomes over these struggles that had already created high tensions between Christians and Muslims well before Suharto's resignation.

The Wahid presidency and the spread of violence

The conflict escalated within the first month of President Wahid's accession to power. Previously confined mainly to Central and Southeast Maluku, the conflict spread to North Maluku with some of the worst violence to date. In November 1999, twenty people were killed when violence erupted in the cities of Ternate and Tidore. In late December and early January, violence reached unprecedented levels in North Halmahera, a predominantly Christian area of North Maluku. The violence began in Tobelo on December 26 and several hundred people were killed in subsequent days in Tobelo, Galela and Jailolo. According to the official military account, 907 people were killed between December 26 and January 7, while less conservative estimates reached more than 2000. (76)

According to the sociologist Thamrin Amal Tomagola, a native and specialist of the region, local issues were at the heart of the violence. The violence actually began in August 1999, shortly after a government decree was issued to create a new district, Malifut -- located at the border between North and Central Halmahera. Malifut included a majority of Muslim villages inhabited by Makianese, who had been resettled in 1975 when their place of origin was threatened by volcanic eruption. Other local villages populated by Christians rejected the creation of this district and initiated the violence. They viewed the new district as a loss of territorial control for Christians and an attempt to prevent the spread of Christianity to Central Halmahera.

Tomagola also blamed a competition between the sultans of Ternate and Tidore for some of the violence. The Sultan of Ternate, drawing his source of support in part from Christians in North Halmahera, had long been a rival of the Sultan of Tidore, whose basis of support was greater among the Muslims in Central and South Halmahera. When the Habibie government decided to create a new province in North Maluku, the Sultan of Ternate vied for the position of governor, while an ally of the Sultan of Tidore, Bahar Landily, a career bureaucrat, was also in the running. Tomagola viewed the ensuing violence as a result of competition for the governorship, as well as the destabilizing factor in Malifut. Finally, the violence was also caused by the Australian exploitation of a local gold mine, the proceeds of which were the subject of highly contested claims by Christians and Muslims allied with the two sultans. (77)

In the same analysis, however, Tomagola recognized the importance of "contextual" factors in explaining the violence in North Maluku. He noted that the area of the new district of Malifut had been contested for a long time. The territorial competition between Christians and Muslims in Halmahera stretched back to the initial settlements and conversions by the Dutch in North Halmahera. This competition was particularly significant in the 1990s when Christians were worried about the political gains of modernist Muslims after the creation of ICMI and the 1993 elections, in which Muslims obtained a significant share of seats in the parliament and cabinet. (78)

It is not coincidental that the violence erupted after months of conflict between Christians and Muslims in other parts of the region. These events could only raise the tensions and stakes for both religious communities in the North and therefore make local issues appear much more significant. News and contacts with the rest of Maluku intensified the sense of threat felt by Christians in Ambon and other local areas long before the violence first erupted. The former rector of Pattimura University, Nanere, was at the centre of one of the growing controversies between religious communities in Ambon during the mid-1990s. Since Nanere was a native of North Halmahera, this controversy would have an impact on Christian communities. Thus local issues, while important, provided the trigger and the filter through which tensions at the national and regional level were expressed.

The subsequent reverberations show how conflicts were not only local but also fuelled tensions among Muslims and Christians across the region and elsewhere. In Ambon, the violence resumed at roughly the same time. On Christmas day, a new wave of rioting was triggered by a minor incident. One of the city's major churches, Silo church, was burned down. The attack on this important Christian site intensified the violence. News of an "attack on Christians" spread to other regions of Central and Southeast Maluku and triggered similar rioting. In Tanimbar, Southeast Maluku, Christians attacked several mosques in retaliation. Violence erupted on the island of Burn, which had been spared so far (79) Violent outbursts were now closely linked to events in other parts of the Moluccan archipelago. Furthermore, their eruption during Christmas and Ramadan provoked particularly intense emotions.

Reports of these events enraged Muslims and Christians throughout Indonesia as it also became evident that the security forces failed to remain neutral. The security forces were taking sides along religious lines and senior military officers admitted that the troops were firing on rioting crowds, thereby increasing the number of casualties. The new head of the armed forces commander, Admiral A.S. Widodo publicly acknowledged that the armed forces were involved in the conflict and that soldiers were taking sides again. He justified these acts by arguing that they intervened to protect their families. Despite past complaints about the role of security forces, it became widely reported that the Mobile Brigades (Brimob) of the police, who were 75 percent Christian, were siding with Christians, while the overwhelmingly Muslim elite strategic reserve forces (Kostrad) were siding with Muslims. (80)

Groups began to mobilize as a result of the increased violence. Two hundred thousand people demonstrated in Jakarta to denounce the government's inaction in Maluku. The demonstrators were mainly supporters of Islamic parties and organizations. In other parts of Indonesia, especially Yogyakarta and Makassar, more recently formed Islamic organizations, such as the Front for the Defense of Islam (Front Pembela Islam) and the People's Communication Forum for the Province of North Maluku (Forum Komunikasi Masyarakat Provinsi Maluku Utara -- Forkap-Malut) began recruiting volunteers for a jihad to Maluku. Several hundred had already been recruited by mid-January. (81) A jihad force formed in Ternate had already been sent to North Halmahera to defend Muslims, as more deaths were recorded in Tobelo and Galela. (82)

By April and May 2000, jihad volunteers began to leave for Maluku. A group calling itself the Ahlus Sunnah Wal Jama'ah jihad.fighters staged demonstrations in front of the presidential palace and the parliament. They called for a holy war in Maluku. Three thousand of them had been training in Bogor, near Jakarta, and were ready to depart. Although they disarmed themselves under government orders and ceased their training sessions, many left for Ambon at the end of April. The armed forces allowed them to enter Maluku, as long as they came strictly for humanitarian purposes, which they claimed to be their primary mission. Yet, by mid-May, there were almost 4000 reported in Ambon and their presence was blamed for a resumption of violence. (83)

At the end of June, the Wahid government declared a state of emergency. For months, Wahid had resisted this course of action, which was favoured by former armed forces commander Wiranto and the commander of the Pattimura military region, Max Tamaela. Wahid feared that military repression would only intensify the violence and fall to provide adequate solutions. Instead, he had espoused dialogue and calls for various parties to negotiate a compromise. In several instances, he repeated that only Moluccans could find a solution to the crisis. (84) He probably meant that it was best to build peace from below rather than impose a military solution from above, such as had been often the case under the Suharto government. But with increasing pressure to end the violence, he accepted the emergency solution. Although the violence diminished in the following months, the conflicting parties have not reached any kind of agreement to forge a new peace.

Conclusion

Violence between Christians and Muslims in Maluku was unprecedented. When rioting broke out in January 1999, no one could have suspected that it would spread from Ambon's Mardika plaza to numerous other locations in Maluku. It occurred during a time of increasing conflict across the archipelago but few would have expected a large-scale conflict in the region. Furthermore, it was the first time in Indonesia that Christians and Muslims were pitted in such ethnic warfare.

Two institutional factors were conducive to the polarization of religious identities and heightened tensions, both of which were legacies of authoritarian rule. The authoritarian resolution of the role of Islam in the state created uncertainties. Christian and Javanist Muslim dominance in the early New Order perpetuated Christian dominance in Maluku, and reinforced Muslims' sense of regional disadvantage. Suhoarto's later use of Islamic groups for political support re-opened uncertainties about the role of Islam, and gave new means of access for modernist Muslims in the Indonesian state and Muslims generally in Maluku. In turn, these changes raised fears of marginalization among Christians in the region. The authoritarian resolution of definitions of the nation, and particularly the terms of inclusion of religious groups, therefore contributed to fears of marginalization and exclusion in both communities.

Patrimonial networks also contributed to the polarization of religious identities. Under authoritarian rule, patrimonial linkages became the primary means of access to state positions and resources. In Maluku, patrimonial networks were formed along religious lines that were drawn during colonial times. Each group vied for control of various state sectors, with Christians maintaining a regional advantage until the mid-1990s. Once the creation of ICMI and policy changes in support of Muslim groups led to the displacement of Christians and Javanist Muslims at the national level, a similar trend in Maluku allowed Muslims to reduce their disadvantage regionally and seek greater control.

Rapid democratic transition, in this context, made violence increasingly likely. Uncertainties, fears of marginalization, and loss of relative position were already strong in the region. Institutional change at the national level opened up the field of possibilities, thereby further polarizing the communities. Conceptions of the nation were open for renegotiation and the means of access to positions and state resources were likely to change. In the absence of any clear institutional means of guaranteeing the protection of group interests, the potential for violent conflict was much stronger.

These institutional factors provided the context for violence to erupt, but cannot explain its intensity and scale. Combined with local factors, such as the historical division between Christians and Muslims from colonial rule, the almost equal numbers of both communities and the effects of migration, they made Maluku particularly prone to violence. Provocation may well have been a factor in triggering violence, where it was ready to occur, but certainly could not explain why it had such devastating consequences. An analysis of reasons why the first riots degenerated into ethnic warfare would require an examination of the cycle of revenge, further provocation and participation by security forces.

The violence in Maluku illustrates the degree to which authoritarian rule can exacerbate ethnic relations. Authoritarian solutions to solidify national unity actually undermined it over time. This analysis shows the fallacy and danger of maintaining stability by imposing solutions from above and creating patrimonial linkages to support the authoritarian system. Once the basis of support changes and, even more so when democratic transition occurs, few institutional channels exist to renegotiate terms of inclusion in the nation or means of access to state resources and power to replace patrimonial networks. Under these conditions, the potential for violence is high.

(1.) The research for this article was funded by the Social Sciences and Humanities Research Council of Canada. I am grateful for comments by many people on various drafts presented in different forums. I would especially like to thank Bill Liddle, Bob Hefner and Sidney Jones. I would also like to thank those who attended the panel in which this version of this paper was presented at the annual meeting of the Association of Asian Studies, Chicago, 22-25 March 2001.

(2.) The province of Maluku encompasses the Moluccas Islands, often known as the Spice Islands, once at the heart of the Dutch East Indies' spice trade. The Indonesian "Maluku" is used throughout the paper but "Moluccan" is used to describe its inhabitants or as an adjective.

(3.) This article therefore does not explain the intensity, form or scale of the conflict. Much of the literature on ethnic conflict has emphasized conditions leading to conflict, rather than explaining why conflict is expressed as protest, riots, long and intense warfare, or terrorism. See Donald L. Horowitz, Ethnic Groups its Conflict (Berkeley: University of California Press, 1985) for an explanation based on group psychology; Ted R Gurr, et al., Minorities At Risk: A Global View of Ethnopolitical Conflict (Washington, D.C.: United States Institute of Peace Press, 1993) for an analysis of the role of political, economic and cultural discrimination; Paul R Brass, ed., Ethnic Groups and the State (London: Croom Helm, 1985); and Crawford Young, The Politics of Cultural Pluralism (Madison: University of Wisconsin Press, 1976) for an emphasis on the role of elites in mobilizing ethnic identity for instrumental purposes. In a more recent study, Horowitz explains the specific phenomenon of ethnic riots and analy zes factors influencing their intensity and scale but notes that "[s]tudents of violence agree...that its various forms are convertible from one to another under certain conditions....But no one has specified the conditions that turn an ethnic riot into protracted civil war or terrorism." See Donald L. Horowitz, The Deadly Ethnic Riot (Berkeley: University of California Press, 2001), p.4.

(4.) Herbert Feith, The Decline of Constitutional it Democracy in Indonesia (Ithaca, NY: Cornell University Press, 1962).

(5.) For an article that found similar local dynamics in Maluku but placed more emphasis on these patrimonial networks, as well as other factors, see Gerry van Klinken "The Maluku wars: Bringing Society Back In," Indonesia, vol. 71 (April 2001).

(6.) Liah Greenfeld, Nationalism: Five Roads to Modernity (Cambridge, Mass.: Harvard University Press, 1992) and Jack L. Snyder, From Voting to Violence: Democratization and Nationalist Conflict (New York: Norton, 2000).

(7.) Particularly violent examples include attempts to impose such regimes through genocides in Yugoslavia and Rwanda.

(8.) The most prevalent form of this arrangement is "consociationalism." See Arend Lijphart, Democracy in Plural Societies (New Haven: Yale University Press, 1977).

(9.) For an interesting critique of this perspective in the context of Latin America, see Deborah J. Yashar, "Democracy. Indigenous Movements, and the Postliberal challenge in Latin America," World Politics, vol. 52, no. 1 (October 1999). She argues that indigenous movements in Latin America pose a challenge to the liberal emphasis on the individual as the primary political unit and the idea of a homogeneous nation corresponding to the state.

(10.) Horowitz, Ethnic Groups. pp. 628-51.

(11.) Juan J. Linz and Alfred Stepan, Problems of Democratic Transition and Consolidation: Southern Europe, South America, and Post-Communist Europe (Baltimore and London: Johns Hopkins University Press, 1996). See also Yashar, "Democracy", that shares a critique of nation states, and an emphasis instead on collective rights for cultural groups, along with individual rights.

(12.) I believe that the nuance is more than semantic. States with several nations within their boundaries cannot hope to fully maintain their territorial integrity without providing a new foundation for a common, imagined community that becomes the relevant locus of sovereignty.

(13.) For a discussion of the Indonesian case and the theoretical basis for such compatibility, see Robert W. Hefner, Civil Islam: Muslims and Democratization in Indonesia (Princeton, NJ: Princeton University Press, 2000).

(14.) On patronage and its role in ethnic mobilization, see Snyder, From Voting to Violence, p. 72 and Horowitz, Ethnic Groups, Pp. 128-29. For an analysis in the African context, see Michael Bratton and Nicolas Van de Walle, eds., Democratic Experiments in Africa: Regime Transitions in Comparative Perspective (Cambridge: Cambridge University Press, 1997).

(15.) For examples and a discussion of the possibility, see Donald L. Horowitz, "Democracy in Divided Societies," in Nationalism, Ethnic Conflict, and Democracy, eds. Larry Diamond and Marc F. Plattner (Baltimore; London: Johns Hopkins University Press, 1994), p. 36; and Anthony W. Marx, "Apartheid's End: South Africa's Transition From Racial Domination," Ethnic and Racial Studies, vol. 20, no. 3 (July 1997), pp. 489-92.

(16.) Snyder, From Voting to Violence.

(17.) Richard Chauvel, Nationalists, Soldiers, and Separatists (Leiden: KITLV Press, 1990), especially chapters 2 and 3. See also pp. 168-69.

(18.) Chauvel, Nationalists, Soldiers, and Separatists.

(19.) Jan Nanere, rector of Universitas Pattimura (UNPATTI); interview by author; Ambon, May 1996. P.J. Siwabessy, Vice-Rector of UNPATTI; interview by author; Ambon, May 1996. Jopi Papilaja, head of the Democratic Party of Indonesia (PDI), Maluku and professor of economics, UNPATTI; interview by author; Ambon, May 1996.

(20.) On the issue of the Islamic state, see B.J. Boland, The Struggle of Islam in Modern Indonesia (The Hague: Martinus Nijhoff, 1982).

(21.) I am simplifying the complexity of this period for analytical purposes. In fact, the Constitution of 1945 was adopted in haste in the dying days of the Japanese occupation. It was a poorly written draft with many loopholes. A federal system was adopted in 1949 with a draft constitution that lasted seven months, following an agreement between the Netherlands and the Republic of Indonesia. See George McT. Kahin, Nationalism and Revolution in Indonesia (Ithaca, NY: Cornell University Press, 1952), pp. 446-51. Shortly thereafter, another constitution (1950) Iaid the basis of the liberal democratic system under an unitary state, which was in force throughout the 1950s. See Feith, Decline of Constitutional Democracy, pp. 92-99.

(22.) On Pancasila and its original formulation, see Kahin, Nationalism end Revolution, pp. 123-27.

(23.) For an analysis of the period of liberal democracy, see Feith, Decline of Constitutional Democracy.

(24.) On Guided Democracy, see Daniel S. Lev, The Transition to Guided Democracy: Indonesian Politics, 1957-1959 (Ithaca, N.Y.: Modern Indonesia Project, Southeast Asia Program, Dept. of Asian Studies, Cornell University, 1966).

(25.) The distinction between santri at and abangon (Javanist) Muslims was originally formulated by Geertz. The santrirefer to those Muslims who are more rigorous in their following of Islamic teachings, including prayer five times a day, regular attendance at the mosque and, if possible, performing the holy pilgrimage to Meeca. The term abangan has been used to denote Javanese Muslims who follow a more syncretic type of Islam, which has kept many characteristics of its Hinduist past. See Clifford Geertz, The Religion of Java (Chicago: University of Chicago Press, 1960).

(26.) For an excellent account of these divisions and their political ramifications, see Hefner, Civil Islam, especially pp. 58-127 on political Islam in the 1990s.

(27.) On the early years of Suharto's New Order and particularly the role of the military, see Hamish McDonald, Suharto's Indonesia (Blackburn: Fountain Books, 1980) and David Jenkins, Suharto and His Generals: Indonesian Military Politics, 1975-1983 (Ithaca, NY: Cornell Modern Indonesia Project, Cornell University Press, 1984). A more recent book that tackled many aspects of the late New Order years is Adam Schwarz, A Nation in Waiting: Indonesia in the 1990s (Boulder and San Francisco: Westview Press, 1994). See R. William Liddle, Leadership and Culture in Indonesian Politics (Sydney: Asian Studies Association of Australia in association with Allen & Unwin, 1996) for a compilation of articles written at various periods of Suharto's New Order and with good arguments about the role of Suharto himself as the key political force in crafting and maintaining the institutions of the New Order. See Jacques Bertrand, "False Starts, Succession crises, and Regime Transition: Flirting with Openness in Indonesia," Pac ific Affairs, vol. 69, no. 3 (Fall 1996) for a discussion of a brief period of openness in the New Order and early signs that Suharto was attempting to build a power-base for Habibie to succeed him. On the role of patrimonialism and leadership in the New Order system, see R. William Liddle, "Soeharto's Indonesia: Personal Rule and Political Institutions," Pacific Affairs, vol. 58, no. 1 (1985). For its application in the realm of religion, see Hefner, Civil Islam, pp. 153, 171-74.

(28.) John Bresnan, Managing Indonesia: The Modern Political Economy. (New York and Oxford: Columbia University Press, 1993), pp. 218-26; Douglas E. Ramage, Politics in Indonesia: Democracy, Islam, and The Ideology of Tolerance. (New York: Routiedge, 1995), pp. 37-38.

(29.) Liddle, Leadership and Culture, pp. 196-97

(30.) The national ideology, Pancasila, contains five principles. The first principle, belief in one God, has long been interpreted as a rejection of an Islamist orientation to the state and, instead, as an obligation for the state to promote religiosity while maintaining religious tolerance.

(31.) Confidential interview by author; Ambon, May 1996.

(32.) Nicolas Radjawane, former head of the Protestant Church of Maluku (GPM), former rector of the Christian University of Indonesia, Maluku (Universitas Kristen Indonesia Maluku, UKIM); interview by author; Ambon, May 1996.

(33.) Sammy Titaley, head of the Protestant Church of Maluku (GPM); interview by author; Ambon, May 1996. Abraham Soplantila, former head of the GPM; interview by author; Ambon, May 1996. Mgr. Andreas Sol, former Catholic Bishop of Maluku; interview by author; Ambon, May 1996. Siwabessy, interview.

(34.) Papilaja, interview.

(35.) Titaley, interview. Papilaja, interview.

(36.) Titaley, interview. Pailaja, interview. Nanere, interview.

(37.) Sol, interview.

(38.) Sol, interview.

(39.) Titaley, interview.

(40.) Abdullah Soulissa, head of Ambon's Al-Fatah Masque Foundation; interview by author; Ambon, May 1996.

(41.) Titaley, interview.

(42.) Sol, interview.

(43.) Confidential interview by author; Ambon, May 1996.

(44.) Soulissa, interview.

(45.) These trends were confirmed by interviews with Sammy Titaley, Nick Radjawane, and Jopi Papilaja.

(46.) Papilaja, interview.

(47.) Based on fieldwork observation and personal communications; Ambon, Feb-May 1993.

(48.) It is beyond the scope of this article to analyze the conditions under which riots degenerate into ethnic warfare. See footnote 3.

(49.) For an interesting argument along these lines see Benedict R. O'G. Anderson, Violence and the Stale in Suharto's Indonesia (Ithaca, NY: Southeast Asia Program, Cornell University, 2001).

(50.) For a different explanation, which also draws on the context of political change and its effects on competition for resources and jobs, see John T. Sidel, "Riots, Church Burnings, Conspiracies," in Violence in Indonesia, eds. Ingrid Wessel and Georgia Wimhoefer (Hamburg, Germany: Abera, 2001).

(51.) Accounts in the two leading newsmagazines, Tempo and Gatra, present different accounts of the incident between Jopie Saiya, the Christian minibus driver and the Muslim youths, as well as which group, Christian or Muslim, first attacked the other community. See Tempo, no. 17, 26 January-1 February 1999 and Gatra, 30 January 1999.

(52.) Human Rights Watch, "Indonesia: the Violence in Ambon," Human Rights Watch Report (New York: Human Rights Watch, 1999); Suara Pembaruan, 4 February, 1999.

(53.) Tempo, no. 17, 26 January-1 February 1999.

(54.) Personal communication; Ambon, May 1996.

(55.) Kompas, 15 December 1998 and Suara Karya, 15 December 1998.

(56.) See Sidel, "Riots" and Robert W. Hefner, "Islam and Nation in the Post-Suharto Era," in The Politics of Post-Suharto Indonesia, eds. Adam Schwarz and Jonathan Paris (New York: Council on Foreign Relations Press, 1999).

(57.) Human Rights Watch, "Indonesia: the Violence in Ambon," pp. 22-23.

(58.) Gatra, 27 February 1999; Republika, 15 February 1999.

(59.) Human Rights Watch, "Indonesia: the Violence in Ambon," pp. 23-25.

(60.) Republika, 16 February 1999.

(61.) Human Rights Watch, "Indonesia: the Violence in Ambon," p. 26.

(62.) Human Rights Watch, "Indonesia: the Violence in Ambon," p.27-28; Tempo, no. 23, 9-15 March 1999.

(63.) Tempo, no. 23, 9-15 March 1999

(64.) Tempo, no. 18, 1-8 February 1999.

(65.) Human Rights Watch, Indonesia: the Violence in Ambon," p. 29.

(66.) Kompas, 4 March, 1999.

(67.) Republika, 6 March 1999; Kompas, 4 and 6 March 1999; Forum Keadilan, no. 25,22 March 1999.

(68.) Tempo, vol. 28, no. 6, 13-19 April 1999.

(69.) Forum Keadilan, no. 1, 11 April 1999.

(70.) Temp, vol. 28, no. 11, 17-24 May 1999.

(71.) Gatra, 31 July 1999.

(72.) Gatra, 3 August 1999.

(73.) Tempo, vol. 28, no. 22, 2-8 August 1999 and no. 24, 16-22 August 1999.

(74.) Gatra, 4 September 1999

(75.) Gatra, 31 August, 1999; Forum Keadilan, no. 20, 22 August 1999; Forum Keadilan, no. 22, 5 September 1999 and no. 25, 26 September 1999; Gatra, 16 October 1999.

(76.) Gatra, 20 November 1999; Gatra, 22 January 2000; Gatra, 5 February. 2000.

(77.) Thamrin Amal Tomagola, "The Bleeding Halmahera of North Moluccas," paper presented at a workshop entitled "Political Violence in Asia," Centre for Development and Environment, University of Oslo, 5-7 June 2000, pp. 8-10.

(78.) Tomagola, "The Bleeding Halmahera," pp. 5-6.

(79.) Tempo, vol. 28, no. 44, 3-9 January 2000.

(80.) Tempo, vol. 28, no. 41, 13-20 December 1999; Tempo, vol. 28, no. 42, 20-26 December 1999; Gatra, vol. 6, no. 10, 22 January 2000.

(81.) Gatra, 22 January 2000.

(82.) Gatra, 29 January 2000.

(83.) Jakarta Post, 14 April 2000; Tempo, vol. 29, no. 7, 17-23 April 2000; Tempo, vol. 29, no. 12, 22-28 May 2000.

(84.) Tempo, vol. 28, no. 42, 20-26 December 1999.

JACQUES BERTRAND is Assistant Professor of Political Science at the University of Toronto, specializing in the politics of Southeast Asia -- Indonesia, in particular. His research has focused on issues of ethnic and religious conflict, nationalism, democratization, and peasant resistance. He is currently writing a book entitled Nationalism and Ethnic Violence in Indonesia. He has published articles in Pacific Affairs, Asian Survey and Comparative Politics. He received his Ph.D. from Princeton University. Email: <[e-mail]>