The Political Economy of Mountain Java

An Interpretive History

Robert W. Hefner

UNIVERSITY OF CALIFORNIA PRESS
Berkeley Los Angeles Oxford

Introduction: Mountain Java in History and Social Theory / 1
Politics and Community in Premodern History / 31
Agricultural History: Intensification and Degradation / 52
The Green Revolution in Mountain Agriculture / 81
Relations in Production: Social Change in Land and Labor / 113
Consumption Communitites / 159
Politics and Social Identity: The 1965-66 Violence and Its Aftermath / 193
Conclusion: Economy and Moral Community / 228

SEVEN

Politics and Social Identity

The 1965-66 Violence and Its Aftermath

Since Hindu-Buddhist times the Tengger highlands have been a subordinate territory within a state whose political and cultural heartland lay elsewhere. In the late eighteenth century the area, which had been peripheral to the rising Muslim principalities of the sixteenth and seventeenth centuries, was decisively integrated into what would eventually become the dominant political entity on the island, the Dutch colonial state. Though a regionalist tradition survived, highland identity never achieved the dimensions of ethnic separation characteristic of upland populations in much of Southeast Asia. Politically and culturally, the Tengger mountains were too much part of larger Java.

Indonesia’s independence in 1949 seemed to promise to further integrate the highlands into the politics and culture of the surrounding lowlands. In fact, however, this did not occur. As national parties penetrated the countryside, the cleavage between the two regions, if anything, only hardened. The Muslim lowlands became the strongest center in all of Java for Nahdatul Ulama, the party of Muslim traditionalists. The Javanist highlands, by contrast, were dominated by the Indonesian Nationalist Party (PNI) and, to a lesser degree, the Indonesian Communist Party (PKI).

A similar process of polarization along sociocultural lines occurred in other areas of Java at this time (Mortimer 1982, 60; Liddle 1978, 190; Jay 1963). Religious symbols “came increasingly to be identified with ideas of the definition of the nature of the state and with the way social and material resources should be distributed” (Lyon 1970, 37). In rural Pasuruan the cleavage between Muslims and Javanists took an extreme form because it coincided with long-standing tensions between upland and lowland society. As we shall see, highlanders regarded the conflict that culminated in the bloodletting of 1965-66 as but the latest phase in a struggle pitting them against the stronger, more hierarchical Islamic lowlands.

The causes and consequences of that violence are the focus of this chapter. Though in some ways distinctive, the Pasuruan example provides insight into the dynamics of political conflict throughout rural Java before and after the bloodshed of 1965-66. More generally, it presents an unusually rich opportunity to reflect on the role of class and community in comparative political development. The example argues against the reduction of political violence to such polar alternatives as class struggle or cultural values. In their place, it demands a model that can recognize the cultural dimensions of class and the politics and economics of ethnicity, status, and religion. What is needed, in other words, is an interactionist model of politics and culture that acknowledges the multiple bases of allegiance and power. Such a recognition is at the heart of a non-economistic understanding of political-economic change.

The example also sheds light on the social forces that are realigning politics and identity in Java today. Driven by broad changes in the Indonesian state, recent developments have linked village to nation in a way unanticipated by observers of the old regime. As in much of the developing world, regional traditions are giving way to national culture. In so doing, they are creating the symbols and instruments for a standardized social hierarchy and the conditions for a broader struggle over the form and meaning of modern community.

POLITICAL COMMUNITY IN THE PRE-1965 PERIOD

The regency of Pasuruan has long been one of the most important centers of Islamic traditionalism in all of Java. Founded in East Java in 1926, Nahdatul Ulama, the party of traditionalist Muslim clerics, quickly came to dominate lowland political life. Its ascent was related to the area’s earlier history of conflict and colonialism, which had swept aside Javanese traditions and laid the groundwork for the expansion of Islam.

Islam’s rise in this region goes back to the eighteenth century. At that time—the dawn of the colonial era—the Pasuruan area was a “cauldron of warfare” (Elson 1984, I). The most sustained violence pitted an alliance of eastern Javanese forces under the rebel Surapati against the combined armies of Mataram and the Dutch (see ch. 2; Kumar 1976). Even after their defeat in 1707, Surapati’s followers continued to put up resistance for well over half a century, which suggests that the aristocracy had an only weak hold on the countryside. Eventually, however, the Dutch-supported rulers consolidated their control over the rich rice lands around Bangil and Pasuruan (Elson 1984, 12). No sooner was this done, however, than events again conspired to undermine their authority. With the indigenous population depleted by warfare, the colonial government encouraged a migratory influx of Madurese and Javanese. Fleeing the exactions of their own native aristocracy, animated by a commitment to Islam, and more proudly independent than their Javanese counterparts (de Vries 1931,131; Elson 1984,15), the Madurese who settled in the Pasuruan region were little inclined to defer to comprador Javanese aristocrats. The migration created a frontier society, with a more “freewheeling and open” social structure (Elson 1984,16), and a less abiding respect for hierarchical ways.

In time, as population grew and the Dutch implemented their program for the forced cultivation of export crops, the lowland frontier dosed. The earlier fluid social structure eventually gave way to a rigidly hierarchical colonial society, characterized by high rates of landlessness and a privileged rural elite (Elson 1984, 90-94; Alexander and Alexander 1979, 31). While strengthening state power, colonial programs did not enhance the independent authority of the gentry (priyayi). On the contrary, the gentry’s role was thoroughly dependent on European might, and their actions were subordinate to European interests. As Sartono Kartodirjo (1972, 88) has observed of Java as a whole, the cooption of native priyayi only further estranged them from the mass of rural villagers, creating a vacuum of authority in popular society. In coastal areas around Pasuruan, some of the chief beneficiaries of this crisis were Muslim religious teachers (kyai). Untainted by association with the Dutch, they exercised influence across ethnic and territorial boundaries, becoming, in effect, the cultural brokers for a new rural order with ideals and commitments distinct from those of a discredited Javanese aristocracy.

Muslim influence was abetted by the expansion of religious schools (pesantren) in the lowlands during the last half of the nineteenth century. Mechanisms for the propagation of a more orthodox Islam, these schools were also central to lowland politics and social structure. Marital alliances linked wealthy families to pesantren teachers. Affluent villagers were expected to make gifts of money and land to religious teachers in support of school operations. With the pilgrimage to Mecca, sponsorship of Muslim activities became the preferred index of social standing, replacing the slametan festivals earlier prized in Javanese communities. In the process the mosque replaced the dhanyang spirit shrine as the ritual center of the village.

This ascent of the crescent over the banyan tree, to borrow a phrase from Nakamura (1983), represented more than the advance of an Islamic world view. It signaled the emergence of a new, less localized, social order based on interregional political ties, an alliance of religion and wealth, and a commitment to Islam that transcended Javanese and Madurese ethnicity.1

It is this historical background that explains why first Sarekat Islam (the Islamic Union; SI) and later Nahdatul Ulama (NU) grew so rapidly in the Pasuruan countryside at the beginning of this century. Sarekat Islam swept through Pasuruan in the mid teens, attracting a following among a peasantry hard pressed by declining circumstances (Benda 1983, 42; Elson 1984, 196). While it did not formulate an overarching political program, SI did serve briefly as a bureau for peasant grievances. Its vilification of the infidel government and priyayi officials (Benda 1983, 43) resonated strongly among Pasuruan’s orthodox Muslims. Conversely, its promotion of “a purified and more devout practice of Islam” (Elson 1984, 199) ensured that its appeals were less sympathetically received in the Javanist highlands. Less hard pressed by economic decline, the latter region remained unmoved by SI’s appeals.2

In the early 1920s Sarekat Islam’s influence waned. Here, as in other parts of Java, this was in part the result of repression by priyayi officials, who deeply resented SI’s charges that they were mere lackeys of the infidel government. The party was also racked by infighting, however, as orthodox Muslims vied with secularly oriented left-wing nationalists for control (Benda 1983, 44; Ricklefs 1981,164-67). In Pasuruan, SI remained solidly in the hands of conservative Muslims. In nearby Malang, however, the left-leaning “red” faction won control of the organization (Elson 1984, 2o6). Long after SI’s decline, in fact, Malang remained the one stronghold of left-wing and “secular” nationalist activity in this eastern region of Java. The contrast between a secular-nationalist Malang and an NU-dominated Pasuruan remained a characteristic of regional politics up through the 196os.

Although the Great Depression had a strong effect on Pasuruan’s export economy, the political climate in the regency remained relatively temperate. Beneath the calm surface, however, NU had quietly begun to expand its base. The combined repression of Dutch and priyayi officials had taken its toll on the leadership of other political parties, and there were few large demonstrations of anticolonial sentiment during this period (Elson 1984, 247). In the face of this repression, and in the aftermath of SI’s decline, NU adopted a nonconfrontational strategy. Rather than directly challenging the government, it exploited its ties to rural religious teachers so as to promote its own style of religious politics. Santri students were sent into remote areas of the countryside, where they organized classes in Qur’anic study {pengajian), and laid the groundwork for new party chapters.

In characterizing NU’s strategy in Java as a whole, Harry Benda has described it as a “consciously non-political course” (Benda 1983, 55). This may be true in the conventional sense of politics. But the initiative brilliantly reflected the organization’s political priorities: not to foment immediate anticolonial resistance, but to disseminate orthodox Islam, winning the hearts and minds of a still weakly orthodox population. Unlike Indonesia’s secular leaders, the ulama-teachers in the party leadership were untutored in Western ways and as yet unmoved by the vision of a nationalist Indonesia. They spoke not of nation and independence, but of moral regeneration in a countryside afflicted, in their eyes, by the twin evils of infidel colonialism and irreligious Javanism.

The quiet fruits of this strategy were already apparent in the late 1930s. Having begun with a base in towns and lowland villages, NU chapters spread to the lower fringes of the Tengger highlands, moving into communities where the contest between Islam and Javanism was as yet unfinished. Attacking dhukun ritual specialists, dhanyang spirit cults, and the ritual exchange system (Hefner 1987c), the party often shifted the balance of power to local Muslims. It was during this period, for example, that many villages at the lower periphery of the Tengger mountains discarded their Javanist ritual heritage. Through this mixture of proselytization and party building, NU expanded its rural base at a time when other political parties were in decline.

The Japanese occupation only served to strengthen NU’s advantage. In Java as a whole the Japanese at first courted Muslim support more actively than they did that of nationalists or priyayi-aristocratics (Benda 1983, 201). As a result Muslims got a head start in establishing organizations that, after the war, would play a leading role in building Islamic political parties. Japanese support was crucial, for example, in the formation of Masyumi, a powerful federation of all-Indonesia Muslim organizations. Nahdatul Ulama executives were accorded a leading role in Masyumi’s operation, and it eventually became a de facto part of the government through its involvement in the Japanese-created Bureau of Religious Affairs. In 1944 these offices became operative in all residencies of Indonesia (Benda 1983, 161).

The final months of the Japanese occupation intensified the competition between Muslims and nationalists. A nationalist youth corps was established in mid 1944 and a rival Muslim one was founded a few months later (Benda 1983, 178). These became the core of Indonesia’s first auxiliary military units. Rather than coordinating their activities, however, the two groups competed for control of the rural population. Organized in early 1945 as a military arm of Masyumi (Kahin 1952,163), Hizbullah, the Muslim militia, launched a bold drive for mass membership in Pasuruan (and other parts of Java) during the first months of 1945, just prior to the Japanese surrender. Its success guaranteed Muslim dominance in lowland Pasuruan. The nationalists faltered in the face of this Muslim advance. The only large centers of population still outside NU’s control were Javanist strongholds in the Tengger highlands. But these were so remote that the nationalists at first saw little advantage in attempting to organize a following there. For all intents and purposes, by the end of the Japanese occupation, Pasuruan was an NU stronghold.

The rivalry was complicated, however, by the struggle for national independence during 1945-49. By early 1946 Dutch patrols were moving unopposed through urban Pasuruan; by late in the year they made regular forays into the countryside. The military occupation forced republican resistance south, into the Tengger highlands and, beyond them, into rural areas along the southern coast in the regencies of Malang and Lumajang, where Dutch influence remained weak. The geography of resistance suddenly put Pasuruan’s Muslim organizations at a disadvantage. The southern terrains were predominantly Javanist and had long resisted NU appeals. The situation thus presented the nationalists with a rich opportunity to expand their influence. Long castigated by lowland critics as primitive heathen, highlanders responded eagerly to nationalist appeals. Their role in the resistance laid the foundation for nationalist dominance in the region after the war. NU’s advance had finally been checked.

Independence

Independence at first saw a relaxation of tensions in the regency, as villagers turned to the difficult task of rebuilding a shattered economy. Until 1953, in fact, Indonesian politics remained largely an elite affair, focused on parliamentary intrigues; little effort was made to draw the rural populace into the contest (Feith 1957, ro; Liddle 1978, 173). Party rivalries intensified, however, with the approach of the first general elections in September 1955. In the months prior to the elections, the major parties extended their recruitment drives into rural areas in an effort to drum up support, politicizing the countryside to a degree never before seen. In much of Java, the new mass organizations followed the lines of preexisting religious and cultural cleavages (aliran). Orthodox Muslims were drawn to Masyumi and Nahdatul Ulama, and Javanists rallied around the Indonesian Nationalist Party (PNI) or the Communist Party (PKI) (Jay 1963; Geertz 1965a, 148; Feith 1957). By comparison with other areas of Java, however, the political geography in Pasuruan was relatively simple, polarized between the NU-dominated lowlands and the nationalist uplands.

As Feith has observed (1957, 13), the 1955 electoral campaign was strongly colored by religious issues. Masyumi attacked the communists as atheists, while Nahdatul Ulama criticized Masyumi for its modernist reforms.3 In Pasuruan certain factions within NU also attacked the PKI as an anti-Islamic organization, though some NU leaders were less concerned about the communists than they were about their reform Muslim rivals. The PKI and its peasant affiliate, the BTI (Barisan Tani Indonesia), also raised the banner of land reform at this time. For a variety of reasons, however, the issue did not catch on in Pasuruan. It was overshadowed by the question of whether Indonesia should become an Islamic state (negara Islam).

There were several reasons why religious issues were especially prominent in Pasuruan. Among Muslims and nationalists, first of all, there was a strong sense that a Muslim triumph was near. Nahdatul Ulama so dominated the local political scene that this party and its rivals had an exaggerated sense of its national influence. Campaign events also helped to push religious issues to the forefront. In several well-publicized incidents, Masyumi leaders and radical kyai from (the otherwise nonradical) NU demanded that upland ritual practices be outlawed. Allied with the nationalists in opposition to NU, village leaders in the midslope highlands responded warily to these attacks. They knew all too well that if Indonesia became an Islamic state, orthodox restrictions would fall especially heavily on them.

Hindu villagers in the upperslope highlands had reason to be concerned as well. Muslim party leaders in the lowlands flatly rejected the highlanders’ claims that they were non-Muslim and thus should be spared restrictions imposed on Muslim believers. Citing the Hindus’ practice of circumcision (regarded by most Indonesians as a Muslim rite) and their occasional participation in Muslim festivals (Hefner 1985), these lowland critics insisted that the highlanders were really Javanist Muslims like those in the midslope region. They too, then, would be subject to the restrictions of an Islamic state.

As the election approached, tensions escalated, as did the tactics of some of the parties. Nationalist officials organized rallies in which they underscored the threat to Javanist religious traditions. Urban-based nationalist intellectuals created an umbrella organization for the defense of midslope traditions. In several of its leaflets, they denounced orthodox Islam as “un-Javanese” and called openly for a return to the non-Muslim “religion of Majapahit” (Hefner 1987c). The nationalist intellectuals also worked with midslope village chiefs to sponsor political rallies at several of the highlands’ most famous dhanyang spirit shrines. Not surprisingly, NU leaders regarded these actions as a deliberate insult to Islam.

Just prior to the 1955 election, tensions reached a climax when a group of Muslim activists (loosely linked to radical factions in NU) invaded one of the most famous upland dhanyang, located amidst the ruins of a fourteenth-century Hindu water temple. The guardian spirit of this shrine was revered by midslope Muslims and upperslope Hindus alike, and pilgrims from the entire regency regularly flocked to its grounds. Under cover of night, the militants entered the shrine, smashed the faces of its old Hindu icons, and hauled off statues for disposal in a nearby river. Word of the attack spread terror through the highlands, but no counteraction was taken. Highlanders feared that it was but a small sample of what was to come.

In the end Muslim hopes for a national mandate were disappointed. The Islamic parties failed to win a clear majority of the national vote. The PNI and Masyumi emerged as the largest parties, capturing 22.3 and 20.9 percent of the vote respectively, while NU and the PKI won 18.4 and 16.4 percent (Feith 1957, 58). Nowhere in Java were Muslims more surprised than in Pasuruan, where the sheer dominance of NU had blinded Islamic party leaders to the national balance of power and caused them to expect a Muslim landslide.

Despite the disappointment at a national level, in Pasuruan itself the election results confirmed the unqualified dominance of Nahdatul Ulama. The party won an astounding 61.0 percent of the vote—the highest proportion in any of Java’s regencies. Its main rival, the PNI, won only 22.1 percent, while the PKI and Masyumi obtained a mere 8.7 and 4.4 percent respectively (Alfian 1971). NU’s share of the vote was almost as high in neighboring Probolinggo. The only region of Indonesia where NU’s achievement was more stunning was in Madura, where its share of the vote was even larger than in Pasuruan. By comparison with other areas of East and Central Java, the Pasuruan results were also notable for the poor showing of both the PKI and Masyumi. The 1955 elections confirmed what was already apparent to many people: that the contest in Pasuruan was essentially a two-party affair, pitting the powerful, lowland-based Nahdatul Ulama against the much smaller, upland-based PNI.4

For Pasuruan’s Javanist highlanders the election results meant that the most serious threat to their religious heritage was for the moment deferred. Subsequent political events reassured them even further. From 1957 on, the national influence of the most militantly reformist Muslim party, Masyumi, declined, with the dismantling of parliamentary democracy and the transition to “Guided Democracy” (Lev 1966; Ricklefs 1981, 243). During this same period, conversely, PKI influence grew. The 1957 elections for provincial councils resulted in substantial PKI gains in East and Central Java, largely to the detriment of the PNI (Lev 1966, 84).

Parallel to the PKI’s rise, however, there was another development of equal importance. Alarmed by secessionist rebellions and the growing influence of the communists, the Indonesian military began to take a more direct role in national politics. Threatened by the military advance, but unsympathetic to the political leadership of the PNI, President Sukarno found himself looking more and more to the communists as a counterweight to the armed forces. Meanwhile, the national economy continued its precipitous decline. Not yet recovered from the war, rural commerce was disrupted in 1959 by an army order to expel all aliens (primarily Chinese) from rural areas (Ricklefs 1981, 255). Shortly thereafter, inflation began to increase, averaging 100 percent per annum from late 1961 to 1964 (Arndt 1971, 373; Ricklefs 1981, 259). As Mackie (1971) has noted, the impact of inflation was highly regressive, favoring those who controlled productive resources like agricultural land. This meant special hardship for Java’s poor, who depended on patrons for access to land and on the market for rice.

Pressured in part by poor peasants hurt by the economic decline, but also hoping to seize the moment from its political rivals, including the military, the PKI responded to the national crisis by trying to expand its rural base. In late 1963 it announced a national campaign for implementation of the land-reform provisions of the 1960 Agrarian Law. In early 1964 PKI-organized peasants launched “unilateral actions” (aksi sepihak), seizing the land of large landowners and staging demonstrations in support of squatters and sharecroppers. The initiatives provoked strong reactions on the part of landlords and Muslim party supporters. Rural violence quickly escalated to new heights (Mortimer 1974, 309-28; Huizer 1974, 125).

There is no question that class issues were central to much of this conflict. But their influence was qualified by the complex play of interests and organizations in the struggle. The PKI itself, for example, often protected landlords with party ties, while vigorously attacking landlords linked to its rivals (Utrecht 1969, 83). Muslim party leaders saw the campaign as a direct challenge to their authority, and thus characterized it as an assault on Islam rather than an economic struggle. Anti-Muslim outbursts in several well-publicized PKI actions (Walkin 1969, 828) lent credence to their charges.

The PKI was caught off guard by the strength of the countermobilization, and it was forced to rein in its followers. In late 1964 a compromise on implementation of land reform was achieved. For a while it looked as if some measure of calm might return to the countryside.

Tensions had risen to unprecedented levels, however, and the conflict left Muslim organizations both embittered and better-organized. In the months that followed, Muslim leaders felt further threatened by President Sukarno’s continuing support for PKI initiatives. The die was thus cast for the events of 1965-66. In the aftermath of a failed left-wing officers’ coup in Jakarta, the army and Muslim organizations moved against the PKI, rounding up and executing several hundred thousand of its supporters. These events laid the foundations for the New Order government.

PASURUAN: PRELUDE TO VIOLENCE

In 1964 East Java was the site of some of the most violent confrontations in all Java between Muslims and communists (Walkin 1969). Political conflict in Pasuruan’s lowlands, however, was surprisingly restrained. There was some PKI activity around large sugar estates. There were also a few odd agricultural villages where the PKI was strong, especially in communities adjacent to the sugar estates where landless or land-poor laborers lived. Even at the height of the 1964 unilateral actions, however, the PKI in lowland Pasuruan found itself unable to mobilize support across broad segments of the rural populace. Its failure was owing to the fact that it faced the most powerful NU organization in all Java. Thus, despite their poverty and pervasive inequality, Pasuruan’s lowlands saw fewer landlord-peasant confrontations than in neighboring Malang and Lumajang, where the influence of Muslim kyai was weaker and the PKI more powerful.

Faced with this impasse, the Pasuruan PKI focused its attention on the uplands, the one area of the regency where it looked as if it might be able to mobilize support. NU’s influence was weak or nonexistent in most of the region, except in a few communities with Madurese residents. In addition, land-reform issues were beginning to attract more attention in the area. Squatters still occupied some 250 hectares of formerly European land, over which they had yet to be given legal rights. Though the PKI had not effectively exploited land-reform issues in the 1955 elections, they took on greater importance after 1957, when the government finally invalidated the land tides of Dutch landholders. Regency-level officials began to talk of returning the lands to the public domain, and rumors spread through the squatter community that they were about to be expelled from their holdings, breathing new life into the PKI’s land-reform campaign. For the first time the PKI found itself able to rally poor uplanders to its cause.

Party efforts intensified after 1963 as the PKI began preparations for its national campaign. Its activities were concentrated in a handful of villages near midslope Puspo and Nongkojajar, and around Tosari in the upperslope region. The greatest portion of the squatter land (two-thirds of the total) was scattered around the Nongkojajar area, where Europeans had operated their largest estates. It was in this region too that the party attracted its largest following. The area of squatter land was smaller near Puspo and Tosari, comprising forty to fifty hectares around each village. In these communities, squatters made up only about 12-15 percent of the total farm population, and the PKI’s following was more modest. By itself, this was clearly not enough to build a coalition capable of winning control of local government.

In the highlands the PKI faced a rival entirely different from the one it confronted in the lowlands. Local government was solidly controlled by PNI-supporters, not Nahdatul Ulama. PNI cadres were drawn from the ranks of the traditional Javanist elite, not a Muslim leadership. Wedded to village hierarchy in this way, the nationalists did not bother to build an independent party structure, but merely appropriated local reins of power. Their intravillage organization remained nominal.

Nahdatul Ulama had a smaller, but more vocal, following in several midslope communities, primarily recruited from the ranks of the Madurese minority. Inevitably, the NU minority controlled the village mosque, and this gave them a visibility beyond their numbers. They used the mosque as a base for criticizing PNI tolerance of spirit cults and such Javanist aesthetic traditions as tayuban dancing (Hefner 1987b.) PNI leaders, meanwhile, championed the local cult of guardian spirits, using services at village spirit shrines (dhanyang) to symbolize their opposition to the religious policies of the NU minority. Religious officials still today acknowledge that the politicization of religion during this period resulted in a precipitous decline in Friday mosque attendance (Hefner 1987b.)

Not surprisingly, given its small following, the highland PKI adopted tactics that were moderate by comparison with other areas of East Java. There were no militant seizures of private land and no calls for a full-scale reform of village landholding. Instead, the communists limited their appeals to a demand for fair distribution of the European holdings.

The party’s unwillingness to call for broader land reform was not just a matter of organizational timidity. Under recently passed national legislation, the limit for private holdings of rainfed land in this region was 8-10 hectares per household, which only a handful of households in the highlands exceeded. In the face of this fact, the PKI wisely confined its efforts to supporting squatters. As a result, however, the party had few grounds for appealing to people in villages where there was no squatter land. In such communities its following remained negligible.

Where it could exploit the squatter issue, and where it was able to develop a fledgling party structure, however, the PKI eventually extended its influence beyond the squatter community. For example, in several villages the party attracted an unusual following among rich farmers and traders, some of whom eventually went on to become cadre leaders. For these people the party was not primarily an organization of class struggle, but an alternative to a local PNI-elite regarded as conservative and backward.

Observers elsewhere in Java have also noted that the PKI sometimes attracted a more affluent, urban-oriented following on similar leadership grounds. In assessing the background of PKI leaders in Javanist villages around rural Pare (in the Kediri region), for example, Robert Jay (1969, 434) has commented that many of the leaders of the PKI, BTI (the peasants’ union), and other communist affiliates came from the ranks of the “well-to-do.” Analyzing the 1955 elections, Herbert Feith (1957, 35) notes that “unaccommodated youth” sometimes used party competition to legitimize their “conflict with the old hierarchy or hierarchies of the village.”

Not coincidentally, the villages in which the PKI was most active in the Tengger mountains were, by upland standards, relatively cosmopolitan. The most important centers were located near upland market sites; by comparison with most of the highlands, these villages had an unusually large resident community of traders, craftsmen, and outsiders. They also had a larger number of young people who had completed elementary school and gone on to high school in the lowlands. These were communities, then, in which urban ideas had a greater hold than elsewhere in the highlands. They were the only villages in which a better-educated, well-traveled, and thus moderately detraditionalized “unaccommodated youth” had a noticeable presence. It was from the ranks of these relatively cosmopolitan villagers that the local PKI recruited its most prominent followers.

This fact was particularly clear in one small community near upperslope Tosari that I will call Sesari. There the PKI leaders were known as “merchant communists” (bakul komunis) because they included in their ranks some of the most successful vegetable traders in the whole highlands. These men came from families that had pioneered the vegetable trade in the 1920s. Despite their wealth, most of these traders were not members of the elite families that had traditionally dominated upperslope politics. Entrance to leadership positions was restricted to people who could trace descent back to earlier village chiefs or Hindu priests (Hefner 1985, 82). Having achieved their status through unconventional occupations, the merchant communists challenged the village establishment’s monopoly of power and its traditionalist principles of exclusion. They also complained that the PNI elite was old-fashioned and inept. It did nothing, they said, to promote education, new forms of agriculture, or any of the other measures needed to yank the highlands into modernity.

The situation was more complicated in the midslope region. There, the most prominent “merchant-farmers” of the 1910s and 1920s had not been indigenous, but lowland immigrants. Most had also been devout Muslims. In the 1950s and 1960s the descendants of these men became leaders in the midslope region’s small NU faction. In this area, too, however, there were youths from the ranks of affluent families of Javanist persuasion who resented the stranglehold of the conservative PNI on local government and looked to the Communist Party as an enlightened alternative. Because of this region’s historic antagonism to orthodox Islam, it was inconceivable that they would turn to Muslim parties to express their antiestablishment sentiments. For them and other disaffected youth in the region, the PKI was the only alternative to the old-fashioned PNI.

Modern political parties, however, are complex national institutions. No matter how parochial they may appear when examined at a local level, their popular dynamics can never be reduced to village forces alone. This simple truth was dramatically apparent in the political struggles that swept Pasuruan in the 1960s. As party rivalries at the national level intensified, the strategies of the local-level PKI also changed. The focus of their attention shifted from what had been a largely intraelite rivalry to a contest involving a wider range of villagers. The issue around which the new lines of conflict were cast was the still unresolved matter of what to do with the old European lands.

The squatters had grown anxious about their situation, and with good reason. After officially invalidating European land titles in 1957, regency officials instructed subdistrict and village-level leaders to conduct a survey of the now-nationalized properties. They were to determine the precise area of the affected lands and the identity of their current occupants. The survey was carried out in a casual, haphazard fashion, however, and ultimately nothing came of it. The squatters continued to occupy their lands. In late 1963 and early 1964—apparently in response to the PKI’s national campaign for land reform—regency officials issued similar directives, accompanied by instructions that each subdistrict and village form a committee to oversee disposition of the lands. Throughout Indonesia the constitution of such committees was a key element in the land-reform process (Utrecht 1969, 77). Not surprisingly, selection of their membership was often a hotly contested issue, since the committee was ultimately responsible for deciding who was to receive land, and how large an area.

In mountain Pasuruan the new directives caused great confusion. In announcing committee formation, PNI officials made no statement as to just what was to be done with the lands. In several instances, in addition, they packed the reform committees with their own followers in a brazen attempt to award themselves the lion’s share of lands. By early 1964, therefore, alarm was spreading among the squatter population as rumors circulated that they were going to lose their lands. The situation was particularly tense in one midslope village, where half of the squatters were landless peasants from outside the region, mostly from villages at the lower fringe of the highlands. They had migrated to the area during the independence struggle hoping to obtain a plot of land from among the European holdings. The outsiders’ presence had always been a source of tension, since many highlanders felt that the European lands were rightfully theirs. The local PNI leadership exploited this fact in an effort to divide the squatter population and seize lands for themselves.

In this one village, PNI leaders devised an ingenious plan to accomplish their goal. They announced that all outsiders on European land should surrender their holdings for redistribution to local villagers. The village chief then approached each of these families privately, offering them a small cash payment and house plot of one-tenth of a hectare if they complied with the decree. This offer, they were told, was purely the largesse of the village chief, and should it be refused, the regency government was under no obligation to provide further compensation. Meeting publicly with villagers, subdistrict officers (all PNI) backed up the village chief and hinted the outsiders might be arrested if they resisted. Eventually, about half of the outsiders relinquished their holdings. While some of this land was redistributed to local villagers, a large portion was quietly sold to lowland investors or given to regency officials for private use.

News of this treachery enraged the squatter population in neighboring communities and galvanized the PKI leadership. After several strategy sessions with their lowland superiors, the upland PKI called for mass protests against PNI corruption. Finally, it seemed, the party had found a cause to give it an appearance of strength in a regency where it was notoriously weak. The PKI announced it would support the cause of the upland squatters by bringing in several hundred lowland followers for demonstrations. By the standards of upland society, this was an unprecedented move. It violated the long-unspoken rule that, no matter how serious, local problems were best handled by local people. Even some people sympathetic to the PKI were alarmed by this flagrant violation of village tradition.

Ultimately there were three PKI rallies, and for each one, protestors were transported in from around the regency. The show of force was overwhelming: more demonstrators showed up than there were adult residents in any of the upland villages. The initiative succeeded brilliantly in its goal of intimidating PNI village chiefs. It is important to remember that at the village level the PNI really had no independent organization, having been grafted on to the existing social hierarchy in a relatively ad hoc fashion. Similarly, unlike their counterparts in many lowland communities, village chiefs could not call on a reserve of sharecropping dependents to come to their aid. Nor could village leaders stigmatize their opponents as anti-Muslim heathens, since the PKI and PNI shared the same Javanist views. Nor, finally, could the imperiled PNI chiefs expect support from landowners or party leaders in nearby NU communities. After years of abusive exchanges between the upland PNI and the lowland NU, the PNI chiefs were in no position to turn to Muslim neighbors for help. In this instance, therefore, religious antagonisms prevented the rural elite in the two regions from organizing along class lines. “Economic” interests alone were insufficient to bring the two together. The inability to organize along (lateral) class lines because of countervailing ties of religion and party has been widely noted in discussions of the PKI and the poor (Mortimer 1974, 327). It has been less widely appreciated that the same problem plagued rural elites, who were also split by religious and political commitments.

In the end the PKI won a notable, if local, victory. Although, in the midslope region, squatters who had already lost their land received no further compensation, the PNI land scam otherwise came to a halt.5 Local communists were given seats on the land committees, and the reform program proceeded smoothly. Many squatters had to resign themselves to receiving smaller landholdings than they had initially controlled. But no more people lost their land, and many poor people who had previously been excluded from the old European lands received a share.

Class and Power in the Pre-1965 Contest

Between the mid 1950’s and 1965 the upland PKI moved from being the instrument of a progressive, if still affluent, counterelite to being a mass party mobilized around economic issues. With its affiliates, the PKI eventually counted several hundred supporters in the mountain region, with a core party cadre (concentrated in a handful of villages) of about one hundred people. Until the end, however, its leadership was primarily recruited from the ranks of traders and affluent peasants. Although it promoted issues of interest to the poor, in organizational terms it remained a complex alliance of people from varied class backgrounds.

Other evidence similarly confirms that class was not the only influence at work in the conflict, even where the issues in question were apparently economic. Some large landholders, for example, deeply resented the effort by PNI leaders to steal squatter land—so much so, in fact, that several threw their support to the PKI. Village officials, after all, were not acting in the interest of these wealthy villagers. The officials had no intention of sharing the land they seized with other large landowners; they were in cahoots with subdistrict- and regency-level bureaucrats. The primary basis of their power, then, was not class or property as such, but their privileged role in a different mechanism of power, the state. Its machinery provided village officials with access to the disputed European lands and with influence beyond that based on control of the means of production. “Certain forms of political power can be generated independently of class power,” we are again reminded, “and can indeed annihilate the power of social classes” (Parkin 1979, 140). There was no better illustration of this important truth than the political behavior of these PNI elites.

The PKI’s influence provides a similar lesson about the complex interplay of organization and allegiance in the political struggle. The decisive element in the party’s competition with the PNI leadership was not its class base among the squatters, but its ability to appeal to an extraregional party organization in an area heretofore characterized by a marked lack of interest-group mobilization. This organizational advantage, then, was not reducible in any direct way to control of the means of production. It also depended on the greater initiative and skills of the communists relative to the less dynamic upland PNI.

In appreciating the dynamics of this political contest, then, we must avoid the romantic urge to discount the independent role of party organization and to overemphasize class. As Theda Skocpol (1982, 169) has noted, this error is commonplace in the literature on peasant political resistance. In class terms, after all, the local Communist Party was really an alliance of people from varied class backgrounds, even though poor people predominated in its lower ranks. In part because of this, the upland PKI never presented its effort as a pure class struggle. The central demands in its revived campaign were for an end to corruption and the removal of all functionaries implicated in misappropriation of lands. The party made no demands for the wholesale redistribution of property, or anything else that might have aggressively challenged local class structures.

This “nonclass” emphasis was owing in part, perhaps, to the preponderance of wealthy villagers among the PKI leadership. In part, too, however, it may have reflected the very correct sociological assessment that the real issue here was not just one of production and class, but also the abuse of office. In an area where villagers still retained strong regional allegiances, the village chiefs had violated public trust by turning their backs on their neighbors and throwing their cards in with district-level bureaucrats. They thereby made clear their primary allegiance. From this perspective, whether intentional or not, the PKI’s emphasis on corruption, and its reluctance to portray the campaign as a class struggle, correctly underscored the fact that the rights and responsibilities at issue here were not merely those of property and class, but of state and society.

Though class influenced the political contest, therefore, it was not the sole basis for the conflict of interests, and it was not the sole source of power for its resolution. Village-level PNI leaders were nothing without their links to PNI officers in district and regency government. These were not lines of class per se, but relations specified by structures of state and party. In bringing popularly based extravillage institutions to bear on local rivalries, the Communist Party had also introduced a form of political organization that was not simply derivative of class. Unwittingly, however, its appeal to extravillage allies also anticipated the kind of power that would presently be fatally used against the party itself. All players in this local contest were soon to be reminded that their fate was all too much bound up with the affairs of larger Java.

THE VIOLENCE

The violence that swept across Java following the abortive left-wing officers’ coup in Jakarta the night of September 30, 1965, caught the Pasuruan PKI by surprise. The coup and subsequent leftist army rebellions in Central Java were crushed in a few days (Mortimer 1974, 413-17; Crouch 1978, 97-157). Shortly thereafter, the armed forces began internal purges of leftist officers. Not long after this time, Muslim organizations in Central Java launched bloody attacks on PKI strongholds (Huizer 1974,135). Muslim youth groups quickly followed suit in East Java. There ANSOR, the youth wing of Nahdatul Ulama, was reported to be at the forefront of the killings. But modernist Muslim organizations also supported the slaughter. The modernist social and educational organization Muhammadiyah, for example, issued a fatwah referring to the extermination of the PKI as a religious duty (Boland 1982, 145-46). Other non-communist organizations also volunteered or were compelled to participate in the campaign (Anonymous 1986, 136).

In the Pasuruan lowlands, youth organizations associated with militant NU leaders initiated the attacks even before the army or local government was in a position to coordinate or restrain their actions. The first targets of the violence were labor organizers, fishermen, estate workers, and others directly associated with the lowland PKI. At this stage of more “spontaneous” or uncoordinated violence, the attacks were undisciplined and brutal, destroying whole families and involving widespread mutilation and torture. Heads, sexual organs, and limbs were displayed along the side of the main road outside of Pasuruan. Canals were choked with bodies. In the Pasuruan regency it was more than a month before the armed forces were able to move in to effectively focus the violence. Even then, it is said, gangs of lowland youths continued to operate independently of military guidance, so much so that army commanders in more than one instance intervened and detained religious leaders whom they accused of undisciplined killing. By the time the bloodshed was brought under military control, most of the lowland PKI cadre had been killed.

In the area of the regency where the party had had its most solid base, however—the highlands—there was no violence during this uncertain time. Not yet grasping the enormity of the bloodshed, most uplanders regarded the killings as a temporary, if terrifying, breakdown of law and order. National news was at first unclear, but villagers expected that calm would be restored in a few days. Indeed, during this first phase of the bloodshed, the upland PKI had the temerity to stage several demonstrations, demanding an end to the killings and punishment of their perpetrators. As the violence continued, however, party members realized that something far more serious than they had initially realized was occurring. Demonstrations stopped. Key PKI officials went into hiding. Several who were not native to the upland region took flight, hoping to find refuge in their natal homes. The party structure slowly dissolved as uplanders began to comprehend the scope and ferocity of the violence.

PNI leaders in the villages were anxious too. This was Pasuruan, after all, home of the strongest NU organization in all Java, and NU— longstanding rival of the smaller upland PNI—was at the forefront of the violence. Reports from the lowlands were sketchy for the first weeks. Rumors circulated that PNI Javanists were among those being rounded up for execution; others bluntly warned that after the PKI had been destroyed, the PNI was next. The rumors were fueled by the fact that ANSOR youth in the lowlands called their actions a jihad, or holy war. Its purpose, they proclaimed, was not just the destruction of the PKI, but the creation of a Muslim state. Rumors of anti-Javanist violence appeared confirmed when, early on in the bloodletting, several PNI members active in anti-Muslim religious organizations were captured and executed (Hefner 1987c).

Such anti-Javanist attacks struck terror into the upland population. In midslope villages, services at dhanyang spirit shrines were quietly suspended, and attendance at Friday mosque services swelled. In the upperslope Hindu communities, there was even greater confusion. People here could not make the least pretense of professing Islam. Some villagers argued that the only defense against the massacres was mass conversion to Islam. Others replied that this could never be done; if that were the only choice, then everyone should prepare to die. Most people, however, were simply immobilized with terror, awaiting the bloodshed they now knew was to come.

The violence did finally reach the mountain region. However, it had little of the spontaneity or anarchic indiscipline of the earlier lowland purges. It was carried out in a planned, coordinated fashion; with only a few exceptions, its victims were confined to communist ranks.

The timing and organization of the violence reflected its difference from the killing that had originally swept lowland Pasuruan. Its agents here were neither members of the local landed elite nor the local Muslims who had long done battle with PKI neighbors. In fact, they were not local people at all. The key players were a momentary alliance of regional and national organizations, dominated by NU and its affiliates, but armed and directed by representatives of the East Javanese military. With the exception of two small incidents near Nongkojajar, none of the violence against the PKI, astonishingly, was initiated by the local population.

This fact highlights the organizational dynamics of different phases of the anticommunist campaign throughout Java. While encouraged and in some instances even armed by agents of the state, lowland activists in the first stages of the killing acted with considerable independence of the government or military. The second phase of the violence, by contrast, resembled the later killings in other parts of Java. It was less spontaneous and more dependent upon the support of the military and civilian arms of the state. Its targets were Javanist strongholds where anticommunist groupings were unwilling or too weak to carry out the purge alone. Outsiders had to be brought in to get the roundup going, and threats of violence were used to secure cooperation from a population that was often unwilling. In the end, then, the violence in highland Pasuruan was not in any simple sense a product of local class or religious cleavages. It was thoroughly regulated by agents of the state and included in its ranks representatives of a variety of nongovernmental organizations, especially Nahdatul Ulama.

Days of Terror

It had been almost two months since the coup when Muslim groups from the lowlands, armed and accompanied by a smaller number of army supervisors, arrived in the region to begin carrying out the blood purge. Three weeks before their arrival, village officials had been ordered to place local-level PKI members in detention, shaving their heads to identify them. Village officials who refused to cooperate with this order were threatened with execution on grounds of hiding communists. In the face of such threats, there was little resistance to the directive. By the time the screening teams arrived, local communists were under house arrest, guarded by a terrified local militia.

There were no Javanists, Hindus, or, for that matter, village officials from the mountain region on the coordinating committee for the purge. It consisted instead of lowland youths from student action groups (see Boland 1982,142; Crouch 1978,165), NU, and its ANSOR youth affiliate, all under the technical supervision of army officers. Local officials and villagers were required, of course, to aid in the identification of PKI leaders. Nonetheless, village leaders were not regarded as willing allies in the purge. Indeed, there was some tension between NU leaders, who were hostile toward the local PNI, and the army coordinators, who were more interested in flushing out communists than condemning anti-Muslim Javanists. The NU representatives wanted nothing to do with the PNI leadership, and in several well-publicized instances they called for the execution of PNI leaders linked to Javanist organizations. Ironically, one such village chief had himself been the focus of a PKI demonstration; he had been one of the most notorious players in the PNI land scam. In the end, the man was spared, after army enquiries confirmed that he was not a communist at all, but merely a Javanist of considerable renown.

Here again, the decision in this case provides a dramatic example of the changing nature of the violence. During the more spontaneous killings several weeks earlier, an individual like this Javanist village chief would have been executed without a moment’s thought. ANSOR youths had been able to execute several leaders of a prominent anti-Islamic organization known as “Javanese Buddha-Visnu Religion” (Hefner 1987c). Its headquarters was located in a Javanist village in an area otherwise dominated by NU, at the foot of the Tengger mountains. In this instance, ANSOR youths had taken the initiative, without securing clearance for the action from the government or military. Now, however, as the violence was extended into areas where NU was too weak to act on its own, the support of the military was more important. The price for such support, however, was that stricter guidelines be used in targeting victims; a man’s having been a vociferous promoter of Javanism no longer sufficed to condemn him.

Nahdatul Ulama’s calls for holy war and the highly charged presence of its leaders in the purges still gave many uplanders the clear impression that what was really at issue in the conflict was religion. Clearly, from an organizational perspective, this was a misperception. Nonetheless, given the long history of regional antagonisms, the prominence of NU leaders in the roundup, and Muslim declarations that the violence was intended to usher in an Islamic state, it was easy for upland villagers to come to this conclusion.

Nowhere was this sense of primordial antagonism more clearly felt than in the Hindu upperslope region. In some nearby mountain communities, the screening teams required local militias to carry out the executions, but this option was ruled out in the Hindu region, apparently because the situation was considered too volatile. It was not that the PKI was stronger there. In fact it was smaller, and its cadres had always been rather timid in their actions. Its leadership, after all, was thoroughly dominated by the “merchant communists,” many of whom came from some of the region’s most distinguished families. Army leaders concluded, however, that the prominence of lowland Muslims in the purge of a non-Muslim region risked antagonizing the local population to the point of rebellion. Reportedly they also feared that the Hindus had dangerous magical powers that could be used against their enemies. Hence the military coordinators of the roundup decided to remove the local communists as quickly and quietly as possible and complete their task in a non-Javanist community below Puspo.

In some ways, the army leaders’ assessment showed an astute perception of local tensions. Even more than the people of the midslope region, the upperslope Hindus had reacted to the violence with communitywide terror. During the first days of lowland killings, several PNI village officials had taken the unusual step of calling on their people to prepare to defend themselves from Muslim attack. Grossly misperceiving the organization of the violence, one PNI village chief (who had been active in the Republican guerrilla army) made a personal appeal to army officers in the lowlands to intervene and protect his threatened village. In another community a PNI village official who had long opposed the PKI nonetheless ordered that the small communist cadre in his village be protected from ANSOR violence. When screening teams entered his village a few weeks later, he paid for this directive with his life. These and other incidents help to explain why many uplanders saw the bloodshed as an act of Muslim violence against highland Javanists and Hindus. It was, many people said, ngare lowlander against Tengger highlander.

As the screening teams prepared to move into Hindu villages, whole communities were swept by incidents of spirit possession. Individuals were seized by ancestral and guardian spirits—the spirits invoked in every major ritual, and regarded as responsible for protecting the well-being of family and community (Hefner 1985, 70). The spirits spoke through the possessed, vainly urging reconciliation and an end to the violence. Such traditional religious instruments for dispute resolution may have been well suited for village feuding, but they were utterly ineffectual in the face of the machinery of state. Overwhelmed by a sense of impending doom, other villagers were found possessed along the sides of roads or outside their homes, shaking or crying catatonically. They remained possessed until friends, neighbors, and the village priest presented offerings to distraught ancestral spirits. No one could recall any such incident of mass possession before this period, and none has happened since.

The communists were finally moved to Sesari, where they were interrogated, and those deemed leaders were separated from the rest. When the day finally came for taking the prisoners away, a small crowd gathered quietly in the main square of the village where the PKI men were held. Guarded by anxious militia youths, two trucks (confiscated from one of the merchant communists) stood ready to carry away the forty suspects. As the men appeared, heads shaved, blindfolded, thumbs tied behind their backs, someone in the crowd let out a low moan, and then someone else another. Soon a handful of villagers began sobbing, shaking, and crying out furiously, their bodies stiff with the force of possessing spirits. Fearing violence, other villagers tried to restrain them and restore order. Alarmed by the outbreak, the screening-team leaders ordered local militia to push back the crowds. There was no resistance, only the gasping voices of ancestral spirits calling for peace and the release of loved ones. The trucks left without further incident.

For weeks afterward, a wave of collective possession swept neighboring communities as the souls of the ancestors cried out against the violence that had taken away relatives and loved ones. The prisoners who had been removed were never heard from again. Several days after their departure, the clothes of a few were quiedy returned to relatives by a midslope villager who had witnessed the killings lower down the mountainside. The trucks had proceeded out of the highlands to an NU stronghold just below Puspo. There the men had been unloaded, and forced to dig a large pit. Then, one by one, they were beaten with bamboo dubs, their throats were slit, and they were pushed into the mass grave.

Many highlanders feared that this was just the beginning of a larger assault on their region. For weeks after the killings, rumors circulated that an ANSOR army was moving up the mountainside arresting and executing anyone who did not swear allegiance to Islam. The rumor, of course, was unfounded. No such ANSOR army materialized. For many people, nonetheless, the events of 1965-66 seemed to suggest a sad, but unavoidable, conclusion. It was the end, they said, the end of an upland way of life. A contest of generations had reached its conclusion. The people of the mountains had lost.

THE CHANGING BASES OF RURAL POLITY

Highlanders’ assessment of their desperate situation ultimately proved inaccurate. Whatever its religious dimensions, the conflict was not in any organizational sense exclusively or even primarily about religion. The role of the state, or certain branches of it, was pivotal, as were an assortment of nongovernmental organizations.6

In the end too, the upperslope Hindus were not obliged to convert to Islam. They were required to establish formal ties with nationally recognized Hindu organizations, however, and to accept ritual and doctrinal reforms that reflected new state policies on religion (Hefner 1985, 247-65; cf. Lyon 1980). Similar changes occurred in the midslope highlands. Just two years after the bloodshed, Javanist Muslims in that region let up on mosque attendance as fears that they would be condemned as communists diminished. Later developments in this same area, however, undermined Javanism. Required by law to profess an officially recognized religion (of which Javanist Islam was not one), many young people in the mountains, especially those from affluent backgrounds, came to see little point to traditional customs. Less attracted to the ways of the village, they were drawn to the alluring images of modern Indonesia. An important aspect of their new cultural citizenship was the profession of an official religion; for them that meant Islam. Ironically, the late 1970s saw a home-grown movement for Islamic orthodoxy take shape right at the heart of this Javanist stronghold (Hefner 1987c), another sign of the post-1965 diminution of antagonisms and realignment of social allegiances.

The emerging pattern of politics and community remains a complex one, but already some trends are apparent. Developments since 1965-66 have systematically altered the economic organization of upland society in a way that has battered down barriers to the outside. With roadbuilding and agricultural intensification have come new consumer goods, increased travel, and electronic media, all of which have provided highlanders with increased exposure to outside ways. Education is also on the rise, although in a fashion highly stratified by class. More young people than ever are looking for employment opportunities outside the region or to economic activities without precedent in the earlier, more solidly agricultural, era. These developments have blurred the distinction between uplander and lowlander and weakened the force of in-group allegiance (see ch. 6). They have also changed the form and bases of power in rural society.

Lured by new goods, blessed by growing income, and backed by a more powerful state, affluent villagers have responded enthusiastically to these changes. Today they depend more for their standing on extravillage institutions and less on the good regard of their neighbors. Not all members of the elite have turned away from the village; in the upperslope region, in particular, adherence to a minority religion presents a barrier to wholesale appropriation of outside ways of life. Nonetheless, what James Scott (1985, 177, 314) has noted of rural society in contemporary Malaysia is broadly true here too: the affluent are less concerned to show off and make a big name in the village because they spend more of their political and economic lives elsewhere.

The National Framework for Rural Change

These local changes in community and power have been facilitated by national developments. Early on in its rule, the New Order government launched a variety of programs designed to restrict political activity and weaken and redirect old allegiances. The first and most basic of these measures was the destruction of the Indonesian Communist Party. Its elimination silenced a virulent, if often compromised, opponent of agrarian inequity. This ensured that future government initiatives would not be bothered by mass-based leftist foes.

But the government’s political reforms extended further. First banning a number of the old parties, it then moved (shortly after the 1971 elections) to consolidate the remainder into two umbrella parties. One party (the “Party of Unity and Development”) united all moderate Muslims under a label and platform stripped of any explicitly religious program (Samson 1978, 210). Muslims who had worked with the military to crush the PKI were disappointed to learn that they were as much subject to the new controls as everyone else. The other party amalgamated followers of the nationalist, Christian, and regional parties under a similarly domesticated platform. Candidates for leadership positions in these parties had to win government approval for nomination. The government exercised this right vigorously, intervening directly in party congresses and manipulating selection of party chiefs (Hering and Willis 1973, 6-9). Above both parties stood Golkar, the purportedly nonpolitical party of government functionaries. From the nation’s capital to the most remote villages, officials were required to profess “monoloyalty” to this bureaucratic party. This policy effectively ended political pluralism within government (Ward 1974, 32-71; Emmerson 1978, 100-106).

In rural areas the guiding principle of these political reforms has been the doctrine of the “floating mass,” according to which the rural population should not be distracted from the task of economic development by the intrusive activities of political parties except during those brief periods when electoral campaigning is allowed (Ward 1974,189). Political parties are forbidden to establish branches below the regency level, restricting their operations to urban areas. In ex-PKI strongholds, village chiefs have been replaced with retired military commanders. Nationally, the doctrine of dwifungsi (dual function) legitimates such direct military involvement in political affairs by asserting the permanent responsibility of the armed forces in both national security and social development (Ward 1974, 28).

One consequence of these changes is clear. The political parties that penetrated the countryside in the 1950s, and that Western observers often regarded as the foundation of “a fundamental social reorganization” of rural life (Geertz 1965a, 127), have, for the moment, become massively irrelevant to agrarian political culture. In appreciating the significance of this, one must remember that it has occurred against a backdrop, not of rural quiescence, but of sweeping social change. The political organizations of old, in other words, have disappeared precisely when rural society has been most forcefully integrated into national structures. Village institutions provide little help in such radically detraditionalized circumstances. Neither can the political parties of pre-1965 days. Whereas the 1950s saw economic stagnation and a flowering of political organizations, the 1970s and 1980s have seen far-reaching social and economic change with few popularly based organizations to respond to its challenge.

As events in the Pasuruan highlands show, the one important exception to this general pattern is religion. It has now been a quarter century since the violence of 1965-66. All of the militantly anti-Islamic organizations that once operated in this mountain area, promoting the repudiation of Islam and a return to “the religion of Majapahit,” have long since been outlawed. A generation has come of age that has little or no first-hand knowledge of the political antagonisms that pitted uplander against lowlander and Javanist against santri Muslim. It is a generation that, after 1966, was required to undergo compulsory education in one of the five religions recognized by the government: Islam, Catholicism, Protestantism, Hinduism, or Buddhism.

For Javanist highlanders these new strictures meant they no longer had the option of withdrawing their children from Islamic classes as formerly allowed (Boland 1982, 196; Noer 1978, 37). The related arrangement, once common in the uplands, of allowing Javanists to serve in schools as teachers of religion was also suppressed. The teachers who now provide religious instruction in midslope schools are graduates of Islamic teacher-training schools, institutions that have experienced significant expansion in the New Order period (Boland 1982, 197). The result has been that the quality and orthodoxy of Islamic instruction in midslope schools have increased significantly (Hefner 1987c).

Regarded under the old regime as a wing of Nahdatul Ulama, since its staff was recruited from that party’s ranks, the Department of Religion has also come to play a more interventionist role in upland religion. Early on in the New Order, the department was “depoliticized.” Its cabinet minister was recruited from the ranks of Golkar rather than Nahdatul Ulama, and, as with all government employees, its staff were required to swear their allegiance to the government party (Emmerson 1978, 96). Despite its “Golkarization” the department’s budget has swelled, and its representatives have not let up on their promotion of adherence to recognized national religions. In East Java, especially, the department has earned a reputation for zeal in its promotion of dakwaah. (Islamic revitalization). Not surprisingly, the department has targeted Muslim areas of the Tengger highlands as one of several “weak zones” for which it allocates extra resources (Labrousse and Soemargono 1985, 222).

Ironically, the department’s depoliticization may only have enhanced its effectiveness in promoting Islam in areas like the midslope highlands. A generation ago villagers openly defied the department’s directives on religious education, since they viewed it as a tool of Nahdatul Ulama. Now when it urges villagers to attend mosque services or engage in Qur’anic study, as it has done here, it speaks with the authority of government. Its message is received by villagers who hear of the importance of religion in schools, the media, and village meetings. They hear this at a time, moreover, when the immensity of recent changes has convinced many people that the time has come to put village ways behind them and develop more modern social attitudes. The department’s diatribes against dhukun ritual specialists and wasteful slametan festivals have struck a particularly responsive chord among middle and affluent farmers attracted to new forms of status and investment.

The time seems ripe for religious change, and it is occurring. After dedining in the late 1960s, attendance at mosque services in midslope villages is again growing. “Javanist” villages now organize collection of the annual zakat (alms). Bright new prayer houses (musholla) can be seen in hamlets where years earlier no one performed daily prayers. Evening classes for Qur’anic study—rare in the midslope region prior to the 1970s—are training large numbers of mountain youths in ritual ways once contemptuously dismissed as ngare-lowlander.

Political and Religious Islam in Disjunction

These developments suggest that there may be a disjunction between the role of Islam in national politics and in popular religious culture. Nationally, the government’s restrictions on party activity, and its requirement that Muslim parties renounce any plan of working for the establishment of an Islamic state, suggest that the organized influence of Muslim parties has declined. Similarly, it is well known that many high-ranking civilian and military leaders are avid enthusiasts of Javanist mysticism and deeply suspicious of Islam (Ward 1974,123; Sundhaussen 1978, 77; Emmerson 1978, 96).

Although some observers view these facts as proof that Islam is everywhere in decline, other information suggests a more complicated picture. In some areas, after all, Golkar officials have worked closely with organizations promoting Islamic orthodoxy (Ecklund 1979, 260; Ward 1974, 82; Labrousse and Soemargono 1985). As Tamara (1985) has emphasized, the late 1970s and early 1980s have seen Islamic revival in large sectors of Indonesian society, and certain governmental bureaus have actively supported the effort (cf. Raillon 1985, 249).

Given the “bureaucratic pluralism” (Emmerson 1983) of the Indonesian state, of course, it would be a mistake to attribute a uniform practice to government officials. No state is a single agent, and, despite the best efforts of some despots, none is ever wholly monolithic. Here in Indonesia, it appears that some officials in the Department of Religion have promoted policies the long-term consequences of which might be objectionable to others in the government. For the moment, at any rate, this much is clear: Islam as an organized political entity may be in decline, but as a popular religious force it shows clear signs of good health.

All this is but one more sign that the “integrative revolution” (Geertz 1973b, 260) redefining the ties of village to nation has taken a somewhat different course from that widely expected thirty years ago. Although some political groups still promote alternative definitions of what the nation is to be, the range and intensity of debate have, for the moment, been decisively restricted. Today, the only popularly based nongovernmental organizations active in rural areas are religious ones. At the same time, rural society has been opened to a wide range of political and market forces, wreaking havoc with village traditions and creating tastes for more “national” life-styles. The changes in religious culture, then, are not happenstance events occurring independendy of political and economic developments. Both have undermined the authority of the village at the same time that they demand, and provide the vehicles for, a new and more national identity.

It is not what one might have expected in the heyday of Indonesian parliamentary politics. In rural areas the promoters of the new economy and the new religion are drawn from the ranks of an elite that in an earlier era might have channeled its leadership energies into rival national parties. Now, if they play a formal role in government at all, chances are it is as members of the ruling party. If they are active in religious organizations, at least in the midslope region, it is in Islamic ones. In Muslim and Javanist areas alike, the members of this elite find themselves drawn to the same Honda motorcycles, television sets, and urban house styles. Strongly differentiated by class, this economic culture is less clearly marked by religion than was the case two generations ago. It bridges communities once separated by different status and ritual goods. In muting communalist differences, then, the new economy makes those of class all the more apparent. This, too, seems a distinctive feature of the emerging national culture.

The diminishing divisiveness of religion is reflected in other ways. In the midslope highlands former nationalist foes of the Muslim parties today speak loudly of the importance of Islam. Golkar-supported village chiefs promote an Islam purged of a Javanism that their counterparts two generations ago did their best to defend (Hefner 1987c). The Islam of which they speak, of course, is a ready-to-wear version that owes less to local ritual traditions or pious santri Muslims dian it does to a generalized sense of national identity, and the importance of religion within it. It is also an Islam that at the moment wants little of political activism. Nonetheless, NU leaders in the nearby lowlands welcome these signs of Muslim piety on the part of their upland neighbors. They see it as proof that, whatever its setbacks in national politics, as a religious force, Islam continues to grow. For them, one must remember, Islam was never merely a means to a political end. It was a valued end in its own right.

In the upperslope highlands, these trends toward a more national religious culture have been less certain because of the local population’s identification with Hinduism and continuing fear of Islam. Nonetheless, the response by leaders of this community has not been to close in on themselves, but to look outward to Hindu brethren in other parts of Indonesia (Hefner 1985, 247-65). The result is that even here a new, more “Indonesian” religiosity is slowly taking shape. It emphasizes belief in a supreme being, the need to replace “wasteful” ritual festivals with simple acts of devotion, and stricter bureaucratic controls over rural religion.

Some people refer to this new style of religiosity as agama pembangunan, “development religion.” Like the changes in consumption shaking the modern highlands, this curious creation points to the “rationalization” of popular culture. As with those changes in consumption, however, the key to this process lies not in the abstract advance of an independent reason, but in changing state-society relations and, more particularly, the moral and political challenge of incorporation into the modern nation state.

CONCLUSION: POLITICS AND SOCIAL IDENTITY

During the 1950s and 1960s observers of the Indonesian scene spoke of the emergence of new forms of political integration, identified with the term aliran (from the Indonesian word for “current” or “stream”). Popularized in Indonesian studies by the anthropologist Clifford Geertz, the aliran concept emphasized that political organizations in independent Indonesia were built on preexisting allegiances of religion and community. Rather than grouping people by class, these cut vertically across social strata, distinguishing people along lines of Javanism, Islam, and, more generally, political ideology. It is important to emphasize that the aliran were not pure ideological entities for Geertz, as some of his critics have suggested. Each aliran segment had its own leaders, party structure, and associated social organizations, and it was their institutional intercon-nectedness, not ideology alone, that gave the aliran their social force (Geertz 1963, 15-17; 1965a, 127-29).

As Herbert Feidi emphasizes (1982, 47), the concept of aliran was useful because it provided “the first really valuable tool for understanding the ways in which the ideas of national parties related to cultural patterns at the grassroots level in any part of the country.” In the Javanese context in particular, the aliran idea seemed to make sense of the cleavage between, on one hand, Muslim parties, with their base in santri-dominated areas, and, on the other, the PNI and PKI, with their support in Javanist regions.

In Geertz’s early work the alimn notion was also linked to a peculiar vision of nation-state development, strongly influenced by the seminal ideas of the sociologist Edward Shils (Shils 1957, 1960). Geertz saw the aliran as a “transformation of the santri, abangan, and prijaji traditions into modern universalistic ideologies.” Recast in this more generalized format, the aliran, Geertz speculated, would eventually provide the symbolic framework for a “fundamental social reorganization” of rural allegiances (Geertz 1965a, 127). In short, the aliran were an integral part of an emerging national culture. It was to be built on a “civic sensibility” that superseded parochial ties of language, religion, and ethnicity, allowing for a broader, more consensual political dialogue (Geertz 1965a, 126).7

In recent years the aliran model has been widely criticized for its purported neglect of class (Anderson 1982, 79), confusion of ideological constructs with actual social structure (Kahn 1982, 92), and failure to recognize that patron-client exchanges often superseded ideological loyalties in determining aliran allegiances (Wertheim 1969, 10). Though the substance of some of these criticisms is debatable—Geertz most certainly did not think of the aliran as purely ideological or wholly independent of class, but argued that they were ideologically marked forms of social organization that incorporated and cut across class lines—these criticisms are right to demand that we reflect on the causes and consequences of the 1965-66 violence. In comparative and theoretical terms, the debate requires us to rethink the role of class and community in political conflict. A quarter century after the bloodshed, these issues must be at the heart of any effort to understand what Java was, and what it has become.

Class, Community, and the Sources of Social Power

The political history of Pasuruan provides insight into these complex issues. As we have seen, prior to 1965, political organization here did coincide in broad outline with long-standing cleavages of region and religion. NU dominated the Muslim lowlands, while the PNI and PKI vied for support in the Javanist uplands. To acknowledge that there was such a coincidence of religious and political affiliations, however, is not to imply that religion was their sole motive or organizational basis. Social groups often identify themselves in terms of shared cultural criteria such as religion, ethnicity, or ideology. But this does not mean that such criteria alone explain why actors come together as a solidary group. To emphasize that politics in pre-1965 Java was often played out along religious lines, then, is not to say that religion was the sole or even the primary cause of the conflict. As Michael Moerman (1967,167) has argued for the study of ethnic groups, to understand the dynamism of status groups, one must look at them not only in terms of how they see themselves (though this too is a critical element), but also in terms of what they do, the interests they serve, and their relationship to other lines of social cleavage.

To rethink the history of the Javanese violence, then, we must clarify just what we mean when we say that the lines of political conflict sometimes coincided with religious community. “Religious solidarity” is almost always a misnomer or sociological synecdoche, inasmuch as what it involves is never just spiritual belief or ideology, but involvement in and identification with a whole community. As the history of highland Pasuruan shows, the compellingness of that identification is never just a spiritual or ideological fact. It emerges from participation in a social world, with its distinctive patterns of morality, class, and power. Religion in this broader sense, therefore, is always an index of more than itself. However rich its meanings, its authority is never just ideological. It depends, too, upon the social and political structures in which it is embedded and their ability to reproduce commitments to its ideals and life-ways.

This point is essential for an understanding of events in pre-1965 Pasuruan. From the 1950s on, religious issues—invariably in conjunction with party rivalries—exacerbated tensions between highlanders and lowlanders. These were not mere surrogates for an interclass struggle for control of the means of production. The social issues with which these disagreements were concerned—on the nature of religious community, the proper forms of worship, the rights of villages to regulate their own customs, the public display of female sexuality in dance and dress, and so on—were compellingly real in their own right.

To acknowledge the irreducibility of these cultural issues, however, is not to say that they were free of economic interest or consequence. Within each region the party structures that took shape “along religious lines” were often used to advance the economic interests of competing actors. Once in place, moreover, these political organizations achieved a measure of power and influence in their own right. Lowland landlords used Nahdatul Ulama’s youth corps as a vehicle to defend their rights to land. Nationalist leaders in the uplands used their ties to district government to seize a portion of the disputed European lands for themselves. Though economic interests were at issue in both cases, the power wielded to advance them was not based on class organization alone, but on political structures of a more complex sort.

Similarly, even where class position might appear to have specified a common interest, political and religious allegiances often prevented members of the same class in different regions from working together. From the point of view of class, after all, it would have made better sense in 1964 for elites in both highland and lowland Pasuruan to put aside their religious differences and make common cause against their PKI rival. Clearly, this would have been to the particular advantage of the upland PNI leaders. Facing a vigorous communist challenge, they found themselves unable to rally around (lateral) class interests and call upon Muslim landowners for help.

The lesson is simple but important. Political solidarity is grounded in large part in the ability of communities and organizations to inspire commitments and penalize deviancy. As social anthropologists, in particular, have long been aware (Douglas 1970), the range of organizations capable of doing so is quite varied. We grossly oversimplify the origins of social power and the grounds for allegiance, then, when we automatically assume that their real basis is class or market position.

For the peasantry in the Tengger highlands, differences of religion and party affiliation were rightly viewed as reflecting far-reaching disagreements on what form social life should take and the nature of power and privilege. Religion and party, in other words, were indices of two very different ways of life, each with a slightly different culture, and each with its own mechanisms for inspiring and enforcing social allegiance. In a small way, one might say, it was here as it is in warfare between nation states: each side had its own class structure and internal political hierarchy, and the contest united people from different classes under a broader social banner.

To see this fact as evidence of a “false consciousness” on the part of the poor (or wealthy), or, even more misleadingly, as evidence of the absence of class influences, is to misconstrue class, substituting a unitary model for the more complex way in which class usually works. It would imply that the political influence of class is somehow more real when it operates independently of other social cleavages (which it never does) and results in conflict groups neatly polarized along lines of property ownership.

As has been widely noted, Marx believed that class would evolve along just such lines in the industrializing West, clearing the ground of religion, ethnicity, and other “archaic” status allegiances. With such traditionalist loyalties extinguished, politics would become a simple contest between two groups, the owners of the means of production and their propertyless laborers (Giddens 1973, 30; Parkin 1979, 5). Within the framework of such an idealized history, it was easy to believe that religion and ethnicity were spent forces, incapable of inspiring commitment or enforcing a political will. Similarly, to identify the forces that ultimately underlay social evolution, it seemed to make sense that one need only look to relations of production. Herein lay the hidden key to political and economic life alike.

Like the history of Europe since Marx’s time, however, politics in Pasuruan shows that class is neither a unitary force nor the sole ground for social allegiance. This point is essential to the noneconomistic understanding of class. Class can work in conjunction with other social forces to create segmentary conflict groups, linking people of different classes but common race, religion, ethnicity, or whatever, against rival organizations also incorporating people of diverse class backgrounds. As Max Weber (1968, 926-39) emphasized, in other words, status groups identified in terms of religion, ethnicity, nation, or life-style can engender interests that bridge objective class groupings and have a moral and material integrity of their own. Although their identity appears grounded in noneconomic criteria, status groups of this sort invariably become important players in the contest for power and resources. We should not be fooled into thinking, therefore, that “economic” conflict is synonymous with class struggle. Competition for control of productive and cultural resources can be cast along more complex lines.

Our modern egalitarian values aside, then, there is nothing illusory about arrangements that link people of diverse class backgrounds in common cause. Nor, by extension, is it in any sense irrational or fetishistic for one to give one’s life for one’s nation, one’s religious community, or some other social grouping irreducible to narrow economic advantage. Modern Java, and modern history generally, remind us all too often that men and women sacrifice themselves for more varied interests.

To say, then, that politics in pre-1965 Pasuruan was influenced by religion and regionalism, then, is not to imply that class in its circumstantial sense (see ch. I) had little political influence. In fact, the contrary is the case. Class played a strong role in the political contest, but its impact was mediated through other social organizations in such a way that it sometimes worked to create segmentary alliances between classes, rather than neatly polarizing them in the way Marx envisioned in Europe.

In lowland Pasuruan, for example, the influence of land scarcity, rice shortages, and inflation worked to restrict the political options of the poor, forcing them into alliance with their NU patrons. Even working within an economistic framework, one can see that it was certainly in the best interests of a peasant in an NU-dominated lowland community to ignore communist appeals, even if they made moral sense and he or she was not a particularly pious Muslim. To join the PKI in such a village would have been tantamount to social suicide. Hence this actor’s self-interest did not coincide with his interest as defined in a dichotomized class model, pitting owners of the means of production against laborers. A wider balance of forces than class has to be examined to determine actors’ “economic” interests.

Clearly, in this instance, for example, the strength of rival parties would have to be factored into any assessment of the costs and benefits of different political options. So, too, would the influence of religious organizations and the state. Given this complex equation, the fact that an individual might choose to ignore PKI appeals is not necessarily evidence of false consciousness. It shows instead that more than relations of production must be considered when assessing an actor’s social position and the structural interests it entails.8

It is important to note, finally, that both the aliran and productivist model of Javanese politics fail to grasp the importance of modern Indonesia’s most critical political institution: the armed forces. Here is an organization grounded in neither cultural primordialism nor control of the means of production alone. Yet from 1957 on, it played a growing role in the nation’s political and economic life. Since 1965 the armed forces have been the dominant force in the New Order polity. Their influence again underscores the important fact that all sorts of organizations can inspire and enforce allegiance, and that power does not flow from relations of production alone. The same point was made a half century ago by Weber (1968, 926; see also Giddens 1973, 43) in his analysis of the organizational distinctiveness of the state and the plural bases of social power. Modern political history only demonstrates the continuing relevance of this truth.

Here in modern Indonesia the state’s intervention in 1965 marked the end of the aliran and the beginning of a new era in state-society relations. Its actions imposed new limits on the question of what Indonesia was to be. More than any other single institution, in fact, the New Order state created the foundation for the changes sweeping today’s Java, transforming rural production, and challenging received ideas of community and authority.

Notes

1. Until well into this century, however, many Islamic communities continued to maintain shrines dedicated to the spirits of village founding ancestors. THese often looked similar to Javanist dhanyang, with a large tree, an open grassy space, and a small shrine marker. Village origin myths stressed that the ancestors revered at these shrines were humans, however, rather than, as is often the case with Javanist shrines, guardian spirits of a nonhuman sort. In general, in addition, the primaryindices of Javanism – prayer celebration by dhukun ritual specialists rather than a Muslim elder, food offerings for earth spirits, holy water, and drinking and dancing – were also less common.

2. Clifford Geertz’s (1960, 128) comments on Javanist social organization (or the lack therof) in the early 1950’s apply equally well to that in Pauruan at this time: the Javanist community had “nothing… which could even in the remotest sense be called a church or a religious organization…[T]here is only a set of separate households geared into one another like so many windowless monads, their harmony preordained by their common adherence to a single tradition.”

3. At the time of its founding under the Japanese, Masyumi grouped Muslim tradi-tionalists and modernists in a single organization. Angry at the declining influence of traditionalists in the party, Nahdatul Ulama withdrew from it in 1952, leaving Masyumi to modernist Muslims (see Samson 1978, 199).

4. In East Java as a whole, the overall percentages for the four major parties were: PNI, 22.8; PKI, 23.2; NU, 34.1; and Masyumi 11.2 percent. In Central Java, the proportions were: PNI, 33.5; PKI, 25.8; NU, 19.6; and Masyumi, 10.0 percent. See Feith 1957, 78. NU won first place in the vote in fourteen of East Java’s twenty-nine regencies, and four of Central Java’s thirty-two (alfian 1971).

5. In the midslope community where the PNI land scam had been most massive, however, the PNI leader forced thirty recipients of land to return it to him after the violence of lat 1965. His technique was similar to that used in 1964, but now it was all the more effective because of the fear squatters had of being accused of being communists. Approaching each family privately, he offered them a small house plot, explaining that if they did not give up the remainder of their landholding, he would denounce tham as communists. On receiving land, he then sold it to outside investors or gave it to regency officials. Eventually the PNI leader was replaced by a Golkar village chief. This mand had actually supported the squatters and was one of the most decent village officials I ever met. He was unable to reverse the land deals, however, since several of the people involoved still had influential government ties. The corrupt village chief died a lonely man, openly despised by the villagers, who, once he was out of poewr, called him a thief to his face.

6. There is perhaps no more revealing document on the uncertain alliance of military leaders, party organizations, civil politicians, police, and religious groups involved in the anti-PKI purge than the anonymous eyewitness intelligence report on the violence in East Java written in November 1965; the picture that emerges from this confirms that in East Java the NU formed “the vanguard of the movement to crush the PKI and its mass organizations” (Anonymous 1986, 142). But it also indicates that there was hesitancy about the killings even within NU’s ranks. This is a useful reminder of the fact that NU has often operated more as a loose alliance of religious leaders and their civil representatives than as a centralized party. Similarly, while the report shows that the military and the police took the initiative in coordinating attacks on the PKI, their efforts were at first handicapped by nationalist and leftist supporters in their own ranks. Elsewhere, the document shows, military and police units provided only follow-up support for bloodshed that was initiated by nongovernmental organizations. The document also confirms that during the first days of violence there was considerable regional variation in the scale and organization of anti-PKI activity. As in Pasuruan, this variegated organization took on a more coordinated, disciplined form only afte the military settled scores in its own ranks.

7. It is easy to discern the influence of Parsonian models of national development, with their emphasis on the importance of normative consensus as teh basis for society, in Geertz’s rather optimistic assessment of aliran. In fairness to Clifford Geertz, however, it is important to note that, unlike with Parsons, this emphasis was as much an expression of hope for Indonesia as it was a belief that such a consensus could ever be achieved; the vision was tempered by a realistic fear that things might go wrong. In one of his last works before 1965-66, for example, Geertz spoke sadly of the aliran’s “failure to make party politics work” (1965a, 150). Rather than demesticating political passions, the aliran seem only to have heightened them: “With each tremor at the national level local equilibrium was disturbed and all the hard-earned agreements, arrangements, and understandings were dislodged.”

8. Wallerstein (1979,172) makes a similar point, noting that status groups are not just based on shared identity, but often function to allow people to compete for resources. Having correctly recognized that status groups can play such practical roles, however, Wallerstein goes on to draw the mistaken conclusion that, in the end, status groups must be “blurred collective representations of class” (1979, 181). This sounds vaguely Marxist, since it keeps class in the foreground. But in fact it distorts Marx’s own views on class, which always begin from relations of production. For Wallerstein, it would seem, any social group that competes for resources is a class. Besides confusing the concept of class, this argument fails to do justice to his own African examples. The point is that there are social organizations other than (lateral) class that can compete for scarce economic resources—political parties, ethnic groups, religious communities, and so on. Competition for resources can be based on a more varied array of groupings than class, and this fact requires us to attend to all solidary groups in a society, not just those specified by relations of production.