Printer Friendly

Politics and headhunting among the Formosan Sejiq: ethnohistorical perspectives.

Headhunting was historically practiced from as far north as highland Formosa, as far west as Assam, and southwards to the Indonesian archipelago (Davison and Sutlive 1991: 153). The last practices of headhunting in the region, which were documented and analyzed by Renato (1980, 1993) and Michelle Rosaldo (1983), ended among the Luzon Ilongot with the 1972 implementation of martial law in the Philippines. Yet headhunting continues to be politically relevant in Southeast Asian and Oceanic communities, not only because headhunting rituals are sometimes celebrated with coconuts or dolls as surrogates for real heads (Cauquelin 2004; George 1993, 1995; Hoskins 2002; Rudolph 2008), but also because ordinary people evoke their headhunting heritage in conversations about topics as diverse as state-community relations or norms of masculinity. In the postcolonial era, in which previously colonized peoples have re-asserted power over their territories, former headhunters now celebrate the deeds of their ancestors, recreate them in art and spectacle, and share memories with anthropologists and tourists. Practices once marked as savage are proudly reclaimed as indigenous identity.

This article is about headhunting on the western Pacific island of Formosa, usually known as Taiwan.' Ongoing studies of Formosan highlanders and the ethnographic corpus of literature dedicated to them are important to Oceanic Studies for many reasons. First of all, native Formosans are the most ancient branch of the Austronesian peoples who peopled the Pacific and Indian Oceans (Bellwood and Tyron 1995). Studies of indigenous Formosa can enrich our knowledge of the aire culturelle that has nourished anthropological thought since the beginning of the discipline. Secondly, the Formosan societies were only relatively recently incorporated into the nation-state. People still recall the words of their elders who had told them stories about non-state forms of political life, anti-colonial resistance, and headhunting. By listening to their narratives, we can learn more about the assertion of political power over their communities, a story all too easily forgotten amidst the much more common Taiwanese tales of democracy and economic miracles. Headhunting belongs to that story, as part of the 'Austronesian complex' of warriors, men's houses, and age classes (Desveaux 1996: 145).

I examine headhunting from the perspective of the Sejiq people, who are often classified as an eastern group of Atayal in the anthropological literature51 place special emphasis on the Sejiq Truku and Seediq Tkedaya groups with whom I conducted 18 months of field research from 2004-2007, and have subsequently visited annually. I was inspired to learn more about headhunting from local men who drew my attention to the fact that their ancestors were formerly headhunters, yet seemed to know little about actual headhunting practices and rituals. These conversations oscillated between masculine pride about warrior/hunter traditions to affirmations that they eventually chose to cease what, in Protestant reinterpretations, was the major sin of their ancestors.

Because it is impossible to observe headhunting rituals that use real heads, I draw upon discussions with Sejiq people, but largely upon Japanese and Chinese language ethnographies, and comparisons with other Oceanic and Southeast Asian peoples in the anthropological literature. Japanese ethnologists and administrators wrote copious volumes during Japanese administration of the island from 1895 to 1945, and the high quality of their work was recognized by their contemporaries (Passin 1947). The early Japanese ethnologists were not anthropologists in today's sense of the profession, as they did less participant observation and did not attempt to build theoretical models. What historian Paul Barclay called 'survey anthropologists,' they were tightly embedded in Japanese colonial practices. Akin to naturalists or surveyors, they compiled volumes of data on local customs and social practices that remain useful as historical documents. (3) What did headhunting mean? Why did the Atayalic and related peoples engage in headhunting? What did the subsequent transformation and abandonment of headhunting entail? What lessons can be drawn from this case for an anthropological understanding of politics?

It is useful to begin with the local term for headhunting, mgaya in Truku dialect. The stem Gaya can be glossed as the 'sacred law.' The prefix m- refers to 'bringing into being' or 'making.' Mgaya is thus 'implementation of the sacred law' (Masaw 1998: 216, Pecoraro 1977: 70, Suzuki 1932: 161). In Sejiq villages, young people and lobbyists for indigenous autonomy usually interpret this in terms of inter-national law, recalling headhunting as an assertion of sovereignty against external invaders. This initial understanding of mgaya is useful, as it helps us envision headhunting as a moral good necessary to a society like the violence of police and soldiers. By affirming their headhunting past in this way, the Sejiq affirm that all men--rather than a state monopoly--held power over the legitimate use of violence on their territory. (4) Mgaya, however, is much more complex than that. As Truku anthropologist Masaw Mowna points out, headhunting was called mgaya, but ordinary warfare without ritual was called muetsipo, meaning 'shooting to kill' (Masaw 1998: 232). (5) Sejiq headhunting was thus part of a social, political and religious complex that cannot be reduced to warfare. I return to the issue of Gaya below, after putting Formosan headhunting into a wider anthropological context.

ANTHROPOLOGICAL INTERPRETATIONS OF HEADHUNTING

Headhunting has long fascinated anthropologists of Oceanic and Southeast Asian societies. Headhunting is not merely primitive warfare, which has always been part of human existence (Keeley 1996). Nor is it simply the taking of body parts as battle souvenirs, as some Allied soldiers in the Pacific are said to have done with Japanese heads. Rather, it is part of a larger ritual and political complex. As defined by Janet Hoskins, headhunting is 'an organized, coherent form of violence in which the severed head is given a specific ritual meaning and the act of headhunting is consecrated and commemorated in some form' (Hoskins 1996: 2). It is a social complex that includes organization (hence politics), violence, and ritual. Without ritual, it is not headhunting, but rather merely a gruesome lbrm of homicide. (6)

Early anthropologists explained headhunting in terms of a spiritual life-force or soul-substance that was concentrated in the skull and could be appropriated ritually by headhunters. In this interpretation, skull racks were reservoirs of power (mana) that could guarantee good crops and protect communities (Needham 1976: 74). This definition of mana as impersonal spiritual power was first proposed by Robert Henry Codrington (1830-1922), but was debated immediately by his contemporaries (e.g. Hocart 1914). Rodney Needham thought that this theory was based on a false causality created by anthropologists thinking in terms of physics and energy. They hoped to explain the efficacy of headhunting by mana, just as they explained radio transmission by radio waves, but this explanation was far from the understanding of actual headhunters (Needham 1976: 82). Theories of mana remain unsatisfactory. Even though heads were perceived to have power to protect crops and community health, mana theories say nothing about who gained power from headhunting or how headhunting expeditions were organized. Other anthropologists focused on the psychology of headhunting.

Robert McKinley (1979), noting that headhunting was continuous across Southeast Asia and Oceania 'prior to the formation of the state,' focuses on headhunting as a ritual to manage 'existential limits of the social world' (McKinley 1979: 95-96). Headhunting is not war waged for territorial reasons, but rather a ritual of community cohesiveness mediating contradictions between life and death, the familiar and the foreign, and human and non-human. He argues that other ethnic groups are perceived as not quite human, as they live across the mountains, a natural realm associated with the spirits, but yet speak. What is important is that headhunters bring the head back to the village, speak to it, give it food offerings, and offer it friendship. The head is central because of the face, which makes it a personal social relationship. Headhunting rituals thus internalize the enemy as a friend and humanize him, which McKinley contrasts to 'our' anonymous modern warfare that 'allow(s) us to forget that our enemies have faces and names' (McKinley 1979: 125-126). Lacking from this approach, however, is attention to who organizes headhunting expeditions and how they convince others to join. McKinley's observation that headhunting is a pre-state practice, although unanalyzed in his article, merits greater attention.

Perhaps the best-known discussions of headhunting are those of Renato and Michelle Rosaldo. Michelle Rosaldo argued that the Ilongot strive for sameness rather than individuality, yet both young and old men find their hearts burdened with insult, envy, pain, grief, shame and other emotions. For young men, the act of headhunting and tossing the head in the air allows them to cast off the shame of childhood and prepare them for marriage. For older men, headhunting allows them to shed the burden of grief (Rosaldo 1983). On October 11,1981, while doing fieldwork, Michelle Rosaldo fell 65 feet down a cliff to her death. Faced with his personal grief and rage, Renato gained an empathetic understanding of the explanation that men hunt heads to vent emotions. This led him to conclude that emotions can animate ritual and other forms of human conduct (Rosaldo 1993).

Rosaldo's psychological approach is also limited. It may explain why individuals participate in headhunting rituals, but says nothing about how the rituals are organized or contribute to political power. Nonetheless, I find three things compelling about their explanations. First of all, the Ilongot have an egalitarian ideology. Secondly, young men and older men experience headhunting differently, with the older men leading the expeditions. There is thus a form of power of older men over younger men; and younger men get access to women through participation in headhunting. Finally, as Renato Rosaldo argued (1980), the meaning and possibility of headhunting depends on the historical and political context.

Some scholars look at headhunting as a form of resistance against state encroachment. On the Indonesian island of Sumba, for example, the Sumbanese engaged in headhunting to resist the incursion of slave traders. It was only after pacification by the Dutch in the 1920s that former enemies transformed enmity and violence into diplomatic rivalry, feasting and alliances (Hoskins 1989). Throughout the region, states have understood the political dimensions of headhunting, suppressing headhunting in order to assert state sovereignty.

Kenneth George takes issue with Rosaldo's interpretation of headhunting as individual mourning. In a study of mappurondo groups of highland Sulawesi, he shows how headhunting was once a form of resistance against powerful coastal states, but was replaced after pacification by headless rituals as an allegory of resistance (George 1995). (7) George's approach is an important contribution, as he discusses the political questions of who hunted heads and whose heads were taken, an approach that highlights collective over individual catharsis. He also shows how the continuing ritualization of headhunting as simulacra of former violence shapes village polity and maintains ideological control of the past (George 1993: 696). As important as anti-state resistance has been in recent centuries, these approaches can explain only part of the phenomenon. They overlook archaeological evidence that headhunting existed long before states; and ethnohistorical proof that headhunting was practiced between groups that were both non-state. It may contain elements of resistance against the state, but is more than that. Most of these approaches, with the exception of George who also sees resistance in new rituals, can scarcely be used to explain why some people seem to accept the substitution of heads with coconuts or other ritual surrogates in present-day situations.

NEW HEADHUNTING RITUALS AND ELITES IN CONTEMPORARY TAIWAN

The Sejiq, like other Formosan indigenous groups (Cauquelin 2004), have used doll heads as surrogates in an autumn 'Ancestor Festival' (Rudolph 2008). Rudolph analyzes these rituals in terms of elite competition for resources, which makes some sense, especially since the government of Taiwan since the 1990s has been eager to fund cultural events that highlight differences between Taiwan and China (Lu 2002). On the surface at least, headhunting rituals at the Ancestor Festival seem to be a paradigmatic example of elite appropriation of government funding. The event is held only if local leaders of non-governmental organizations (NGOs) succeed in applying for government grants, and most villagers take no interest in it. In fact, at one such festival in a Truku (Sejiq) community in 2000, I was struck by the fact that the event was held primarily to attract tourists and that, with the exception of the organizers, Truku participation consisted merely of eating barbecued meat and drinking on the sidelines. Although they seemed to enjoy the feasting, they described the formal event as a failed attempt to attract tourists; and laughed at the organizers, whom they imagine to enrich themselves through such projects. These rituals have been held much less frequently and on a much smaller scale than the mappurondo headhunting rituals observed by George in Bambang, Indonesia (George 1996).

Considering the obvious disconnect between the rituals and village social life, and basing his understanding on interviews with 'ordinary people,' Michael Rudolph argued that the new rituals are largely 'authenticating strategies' of competing local elites using improvised scripts (Rudolph 2008: 116-117). (8) Hsieh Shih-chung (1996) similarly credits the resurgence of 'traditional' performances to the availability of government subsidies to local cultural events. From a social anthropological perspective, Hsieh and Rudolph make an important contribution to our understanding of such rituals by highlighting the new tensions created by the creation of elite groups in previously egalitarian communities. They thus demonstrate how the meaning of headhunting has changed according to the historical context. Without a doubt, the creation of local elites has been one of the most important means by which Sejiq communities and their lands have been integrated into contemporary Taiwanese society.

Lin Ching-Hsiu, in a study of the same Sejiq (Truku) communities, shows how elite groups emerged from land reform and industrialization since the 1950s, as some families were better able than others to take advantage of new economic opportunities. Those who worked for township offices were especially capable of identifying which land could be sold at a great profit, as they had insider knowledge of where industrial parks and other state projects were planned. Lin categorizes the elites into three categories: 1) 'political elites,' loyal to the ruling Chinese Nationalist Party (KMT), composed of local politicians, township employees, teachers, and policemen; 2) 'intellectual elites,' generally opposed to the KMT and affiliated with the Presbyterian Church; and 3) an educated, urban-based elite, generally the offspring of the first two groups (Lin 2011 : 40).

Members of the KMT-loyal elite, through various cultural NGOs, have taken the lead in organizing cultural festivals and have on occasion adopted head-hunting as a prominent ritual theme. Although these activities may appear to be merely political strategies for personal gain, as argued by Rudolph, this can only be partially true. After all, the same leaders could organize rituals with less violent themes and still gain the same economic and political benefits. Instead, having accompanied some of these leaders as they draft grant proposals, I think that they choose to hold headhunting rituals with surrogate heads, much like mappurondo groups in contemporary Indonesia, as 'part of a broader struggle to assert authority and control over their ritual traditions and the past' (George 1996: 15). Aware of the violence and cultural loss that has accompanied political rule by outsiders, they create rituals that emphasize difference between the Sejiq and the politically dominant Han Taiwanese, but recreate a time in which the Sejiq were the holders of political sovereignty with a monopoly on legitimate violence. In discussions with the organizers, I gain the impression that they even take delight in evoking disgust on the part of Han Taiwanese. This aspect of contemporary headhunting ritual, as an assertion of Sejiq agency in the face of state power, remains part of the dynamic even if other members of their own community remain skeptical.

The Presbyterians debate the morality of the rituals, with some pastors identifying it as a sin and others as a legitimate assertion of cultural autonomy. Other Protestant sects, notably the True Jesus Church, have denounced such rituals entirely and discouraged their members from even attending. (9) It is worth noting that contemporary headhunting rituals reveal a tense relationship between local leaders and ordinary people; as different social groups of people now give the rituals different meanings. The changing social and political contexts of the rituals, as well as increasing contradictions between what is said to be done in headhunting and what is actually done in the rituals, end up 'promoting a sense of loss and irony among villagers' (George 1996: 90). In the conversations I had with non-elite villagers, they laugh at the festival organizers, depict them as self-seeking compradors, and evoke an egalitarian past when such activities would have been impossible. Taking their concerns seriously, there is a need to reflect further on the historical meaning of Sejiq egalitarian values and violence, especially as statelessness seems to have been a defining feature of headhunting societies. This requires examination of historical documents, using the perspectives of Sejiq villagers to guide the way.

HEADHUNTING AND NON-STATE SOCIETIES

Village conversations, ethnographies, and colonial documents all lead to the notion that the Sejiq had an 'egalitarian society' until they were incorporated into state forms of governance under Japanese administration. Such societies can be usefully defined as those in which 'political leadership is weak and ranking and stratification among adult males are absent or muted' (Boehm 1993: 227). This is not to say that there was complete equality among all group members. Often men had more power than women, older men had power over younger men, and men with better skills or more experience in such vital activities as hunting and warfare wielded influence in the communities. In addition to power in the household, women exerted considerable authority as shamans and healers. Nonetheless, there were no permanent offices of power, and all men or women could expect to gain in power through age and the accumulation of experience.

Egalitarian societies, as described from ethnographic studies world-wide, are all marked by intentional leveling mechanisms and egalitarian ideologies in which there was a tendency for group members to restrict any accumulation of power by adult males. Men who try to assert too much power could face sanctions ranging from ridicule through disobedience to execution (Boehm 1993: 230). Christopher Boehm was interested in why other higher level primates, as well as humans living in chiefdoms, kingdoms and states, all have marked social dominance hierarchies with authoritative leadership, whereas humans living in autonomous, small-scale societies have managed to create egalitarian societies. He called this social dynamic, common to all egalitarian societies, a 'reverse dominance hierarchy' (Boehm 1993: 228). This dynamic, which by no means presumes that the group aims to reverse a pre-existing dominance hierarchy, is merely the end result of leveling mechanisms by which small groups prevent the emergence of dominant individual leaders or hierarchy. (10) These societies were by necessity of very small scale, as dissidents could exit the community and seek hunting grounds elsewhere. This is the political organization that the Sejiq and other colonized peoples remember with nostalgia--especially in the context of criticizing would-be community leaders who collaborate with the state. When ordinary villagers laugh at NGO leaders for sacrificing doll heads, they follow the logic of a reverse dominance hierarchy.

But how did violence figure in such non-state societies? Bruce Knauft argued that, in 'simple human societies,' decisions were made from casual consensus among men and asymmetries were downplayed or denied. In these societies, violence had little to do with territoriality, property, ritual status, or male leadership concerns, but was rather used as a form of social control or an outlet of male sexual frustration (Knauft 1991: 391, emphasis added). Obviously, there is a contradiction between the idea that all men are equal in egalitarian societies and the assertion that their violence includes social control. For Knauft, such institutions as gerontocracy, ritual eldership, priests, big men, etc., emerge only with increased complexity of hunter-gatherer societies (Knauft 1991: 395-396). An egalitarian ethos did not preclude attempts to gain power, and headhunting was one way for men to gain social ascendancy in egalitarian societies (Davison and Sutlive 1991:157).

Following this logic, we can hypothesize that headhunting, as a ritualized form of homicide rather than a random act of violence or organized warfare, reflected inherent political tension in societies where a reverse dominance hierarchy was challenged by ambitious individuals who sought to monopolize ritual and augment their political power. Yang Shuyuan (2005)illustrates this, in a study of Formosan Bunun headhunting, as a process of oscillation between egalitarianism and nascent social stratification when military leaders and shamans used head-hunting rituals to consolidate power. In Boehm's terms, this is oscillation between a reverse dominance hierarchy and an orthodox dominance hierarchy.

As demonstrated below, headhunting expeditions were fraught with danger, leading to problems of recruitment. From the perspective of younger men, the incentives to join were to assert masculinity or emerge ritually into manhood. This idea was institutionalized in many societies, including the Sejiq (see below), where headhunting was often a prerequisite to marriage (Hoskins 1996: 18, Rosaldo 1983: 146). Headhunting thus met basic needs for reproductive success. From the perspective of older men, headhunting presented an opportunity to organize rituals and emerge with increased political status. In the right circumstances, this could lead to the development of incipient chiefdoms. This is where headhunting as ritual becomes more important than headhunting as violence. In fact, in incipient Oceanic chiefdoms it was only in ritual that the chief could assert his particular status (Barraud 1972: 69). In spite of any individual psychological motivations for participation, the fact that young headhunters made collective vows rendered these rituals political (George 1995: 253). In Southwest Timor, headhunting led indeed to indigenous state formation, as control of ritual violence served as the basis for political power of one clan (McWilliam 1996: 129). Headhunting rituals were thus in many parts of Oceania and Southeast Asia a form of incipient political power that could have ended the reverse dominance hierarchy in the favor of certain individuals--but these rituals ended with colonialism and the introduction of the state. I suggest that similar dynamics happened on Formosa.

REMEMBERING HEADHUNTING: CONTEMPORARY LOCAL EXPLANATIONS

Formosan memories of headhunting cannot be understood without reference to the colonial past. The Austronesian inhabitants of the western plains first encountered state institutions after arrival of the Dutch in 1624 (Andrade 2008); and were gradually subjugated to Qing Dynasty political institutions over the following two and a half centuries (Brown 2004, Shepherd 1993, 1995). Inhabitants of the central mountains and eastern Formosa were first subject to state rule, usually following violent military expeditions, after Japan annexed Formosa in 1895.

The Qing and Japanese states dealt with the presence of mountain headhunting groups in different ways. Although settlers from China took over the western plains, Qing administrators forbade settlers from entering the mountainous territories of the Austronesians. The Japanese, however, were positioning themselves as a new industrial power, and needed to exploit the island's camphor resources for the production of both celluloid and gunpowder. They encountered great resistance from highlanders, especially from Atayalic peoples. Takekoshi Yosaburo thus wrote, 'They mostly live in mountain recesses, are extremely ferocious and attach great importance to head-hunting. This group is more uncivilized than any of the others' (Takekoshi 1907: 219). The challenge facing Japan was to finish annexation of Formosa by pacifying 104,000 'savages,' of whom 20,527 were estimated to be Atayal (Takekoshi 1907: 227). That colonial encounter shaped subsequent experiences and memories of headhunting.

While doing research in Sejiq villages, I found that local people had mixed feelings about headhunting. Christians tended to identify it as a sin, some arguing that their current situation of landlessness, poverty, and social disorder resulted from those past errors. Young men were more likely to be proud of former headhunting customs, sometimes making light of it through humour. While huddling around the fire on winter evenings, men evoked laughter by teasing, 'If you had come here as an outsider a century ago, we would have cut off your head.'

Young men described their ancestors as affirming masculinity through headhunting; whereas now they do so by hunting,joining archery contests, or fighting cocks. When they discuss these popular activities, they refer to headhunting and claim to practice a modern equivalent that demarcates them from the supposedly more effeminate Han Taiwanese. Some showed me Japanese-era photos of the head racks that once adorned their grandfathers' homes. Others took pride in comic books or films depicting headhunters as valiant warriors protecting tribal lands." Ordinary villagers also interpret headhunting as a way of protecting sovereignty. They say that their ancestors always resisted external aggression, even in situations where the Han Taiwanese would capitulate. In the contemporary political climate, some men assert that they are the only ones who would fight off Chinese aggression without compromise. Modern indigenous rights activists also use the Mandarin gloss for headhunting (chucao, [TEXT NOT REPRODUCIBLE IN ASCII.]) for street protests.

Village-based cultural researchers and religious leaders also refer to headhunting customs. I took careful notes on one Presbyterian sermon, which was delivered in Mandarin:
 Our population in the mountains grew so rapidly that there was no
 longer enough game to provide food for everyone. Some people
 decided to leave the mountain and live in the valley. As a
 community, they decided that an equal number of people should live
 on the mountain and in the valley. In order to measure the size of
 each group, since they didn't know how to count back then, the
 people stood together in the two places and shouted loudly. The
 group on the mountain shouted so loudly that all the leaves fell
 from the trees, but the sound from the valley was weak. When the
 mountain people sent some of their own to live in the valley, they
 discovered that they had been deceived. The valley people had
 intentionally shouted in low voices in order to get more people.
 Lying was a violation of Gaya, the sacred ancestral law, so this
 created enmity between the groups. Henceforth, the mountain
 people tattooed their faces and went down into the valley on
 headhunting raids. (12)


This story is recorded in Japanese ethnographies. Kojima Yoshimichi interpreted this myth to explain the enmity between the Atayalic peoples and the people of the plains, whom he understood to be the Hart or pingpuzu ([TEXT NOT REPRODUCIBLE IN ASCII.], plains aborigines) (Kojima 1996 [1915]: 24). In the case of this sermon, the pastor used this myth to denounce the lack of consensus among the three Sejiq dialect groups (Truku, Tkedaya, and Teuda) on the issue of creating an 'autonomous region' around the collective tribal identity of 'Taroko.' He thus used the myth to explain enmity between different groups in the same tribe; and also to declare the righteousness of his own political position.

In both contexts, the myth reveals a great deal about Sejiq perspectives on society. First of all, the basic political unit is the face-to-face community. Second, some groups follow Gaya, whereas others violate Gaya in unforgivable ways. Those who follow Gaya must hunt heads to ensure its perpetuation. Only men who have captured heads could tattoo their faces, which made them attractive to potential wives. Recognizing this social logic, Japanese observers noted that headhunting was a moral good (Mori 2008 [1917]: 314). This moral claim is kept alive whenever Sejiq employ memories of headhunting to affirm masculinity, assert political sovereignty, or advocate political rights. In all cases, some Sejiq emerge as Sejiq balay ('real humans') opposed to Japanese, Chinese, or Taiwanese outsiders who do not follow Gaya and cannot be fully human; and even as opposed to members of their own group accused of immorality. In the past, however, headhunting was a very real practice and needs to be unpacked for its political implications within the context of its historical political organization.

FORMOSAN POLITICAL ORGANIZATION

Wei Hwei-lin established a typology of Formosan societies in terms of political authority. These were: (A) patrilineal, 'quasi-horde' societies, e.g. Atayalic groups and the Yami; (B) composite tribal societies with patricians, e.g. Saisiat, Bunun, Tsou; (C) matrilinear societies with age ranks, e.g. Puyuma, Amis (Pangcah); and (D) dominion-based societies with 'noble' ranks and local defense societies, e.g. Paiwan, Rukai (Wei 1965). Huang Ying-kuei classified the Paiwan, Rukai, Tsou, Amis, and Puyuma as chief (type A) societies; the Bunun, Atayal, and Yami as big-man (type B) societies; and the Saisiat as an intermediate type (Huang 1986). Both of these approaches characterize the Atayal, and by extension the Sejiq, as egalitarian bigman groups.

Atayal and Sejiq communities, located in mountain forests from 500 to 1200 meters above sea level, were very small. Before the Japanese relocated them, Atayalic communities had an average of 50 people, compared to 400-600 for the Paiwan, 700 for the Pangcah, and 800 for the Puyuma (Rudolph 2003: 361). It is worth noting that the Pangcah and Puyuma were coastal societies, and that the Paiwan were located in both coastal and mountainous areas. The Paiwan and Pangcah groups were thus nestled between enemy territory and the sea and built up more complex political structures as their populations grew in limited space. In comparison, Atayalic tendencies to fission and maintain acephelous political organization was facilitated by relatively numerous exit opportunities in the vast forests of north and central Formosa.

Sejiq groups were egalitarian and acephalous, without permanent positions of political power. These communities lived from swidden agriculture, hunting, and gathering. As fertility of the land became exhausted after a few years, they moved in search of new land. These practices, due to conflict over hunting territories, could lead to violence between competing groups. Sejiq communities needed areas with fertile soil that were easily defended, clean, far from the tombs of other groups, and judged appropriate according to divination (Kojima 1996 [1915] :65).

Studies of the Atayal show community life revolving around membership in different groups responsible for work, hunting, ritual, taboo observance, and collective punishment for crimes. These groups had overlapping memberships. One individual could simultaneously belong to a hunting group of several men and to a ritual group of several families. The most important group was the band, or alang. The members of an alang were often brothers, their children and women who moved in from other villages at marriage. This grouping, also called kingal tama ('one father'), could refer to the descendants of the same patriline (Yu 1980 : 97). Depending on necessity and circumstances, an alang was able to incorporate outsiders, and rituals were done to make them kin. Within an alang, they formed nniqan gaga (gaya in two of the three Sejiq dialects), ritual groups of people who eat together, but are not necessarily of the same biological family (Kojima 1996 [1915] : 232, Huang 2000:7). These ritual groups were responsible for the moral behavior of their members and any necessary pig sacrifices. Simultaneously, there was a tendency for small groups to leave and establish settlements elsewhere due to ecological limits and the tensions inherent in a reverse dominance hierarchy. These are both common themes when people recall their histories of migration in the mountains.

Only in the colonial period were communities required to select 'leaders' to manage relations with the state. This person became known as the toumu ([TEXT NOT REPRODUCIBLE IN ASCII.]) or 'chief' (e.g. Masaw 1998, Wang 2006, Wei 1965, Yu 1980). This word, derived from the Hokkien word thau-lang ([TEXT NOT REPRODUCIBLE IN ASCII.]), has its origins in contact between Hart settlers and aboriginal peoples (Masaw 1998 : 34). The chief was usually elected, but became hereditary in some communities. This colonial practice differed from precolonial politics in which men gained influence because of their courage in war and hunting, intelligence, and moral reputation, but in which there were no permanent institutions of political power (Masaw 1998 : 35-36).

Conversations revealed the difficulty of finding a Sejiq gloss for 'leader.' Older men and women are considered ludan, or elders. Like the Big Men of other Pacific Islander societies, ludan were important for conflict resolution, the redistribution of goods, and as orators. Ludan held political influence, but elders remember that all of the older men in the community had exactly the same rights in collective decision making, with no individual holding the right to give orders to others. Even though the ludan had prestige, their only privilege was being invited to every marriage in the alang. They did not form a political class and had no administrative forms of organization (Yu 1980: 98).

Some people suggested the word mdudul. Fernando Pecoraro, a Catholic priest who composed a Taroko (Truku)-French dictionary, defined this word as 'the chief, one who leads, guides, exhorts' (Pecoraro 1977 : 54). The Sejiq Truku of Hualien call these people ngalang kari ('according to his words') to indicate that some people are more reliable to give good advice. Pecoraro also translates tama (father) as chief (Pecoraro 1977 : 293). Most research partners insisted that there is no corresponding word for leader, suggesting that I use kbsulan, which means 'older sibling.' They insisted that this is a reflection of their egalitarian ethos.

In further evidence of a desired reverse dominance hierarchy, ordinary Sejiq often accuse the relatively wealthy of selfishly hoarding money and goods, and insist that Gaya requires sharing. They attack any individual who aspires to wealth or power, including the organizers of modern headhunting rituals, as lohei (thieves) or naqex Inglongan (mean, heartless). This continuing refusal of authority led Japanese and subsequent anthropologists alike to praise Atayalic groups for their political values. As Sejiq anthropologist Masaw Mowna wrote, Sejiq political traditions contained 'the spirit and foundation of democracy' (Masaw 1998: 48). This was the Sejiq law of Gaya.

GAYA AND UTUX: SACRED LAW AND ITS ENFORCEMENT

So far, I have discussed Gaya in three contexts. First, it is the sacred law. Second, the ritual group responsible for the behavior of its members is the nniqan gaya. Third, the sacred law was implemented through mgaya, or headhunting. Taiwanese anthropologists have thus engaged in debates about Gaya and its meaning as a social group, a political institution, or law (e.g. Jian 2002, Masaw 1998, Wang 2003). The Sejiq are no longer members of ritual groups in the same way, although church congregations have replaced them to a certain extent and the pork from pig sacrifices is still distributed to members of clearly defined kinship groups. Some people even told me that anthropologists made up the term nniqan gaya. Some anthropologists (Yu 1979) predict replacement of Gaya with the modern state. What remains is Gaya as a gloss for 'law' and 'culture.' When they asked me if hunting is permitted in Canadian law, for example, research participants used the word 'gaya canada.'

Pecoraro was fascinated by Gaya, which he glossed as 'custom, rule, tradition, taboo, the liturgy of the hunt and headhunting.' He cites mudu Gaya ('to follow Gaya') and smlyeq Gaya ('to violate Gaya'). He added, 'Effectively, GAYA is at the center of the life of the [Sejiq] Taroko, the source, criteria and the judge of their entire personal or social life from birth to death--and after! GAYA is certainly the most sacred reality for the [Sejiq] Taroko' (Pecoraro 1977: 70). (13) Catholic and Protestant missionaries translated the Ten Commandments as the Ten Gaya. Sejiq Christians testify that they accepted Christianity because they perceived it to be similar to Gaya--the major difference being that they no longer engage in headhunting.

Nowadays, the Sejiq refer to Gaya mostly in the context of sexual relations, e.g. premarital sexual relations contravene Gaya. Because marriage is supposed to be a monogamous relationship for life, divorce also violates Gaya. In these circumstances, the Sejiq sacrifice a pig to appease the ancestor spirits (utux). The breach of Gaya, as compared to the Christian concept of individual sin, involves the larger community. The Sejiq previously believed that the utux punish immediately any violation of Gaya by causing the community to suffer from disease, misfortune while hunting, or accidents. The Sejiq also refer to Gaya in terms of property rights. They say that it is a major violation of Gaya to sell land that has been inherited from the ancestors. Hunting practices are also discussed in terms of Gaya (Huang 1996). Although sacrifices continue, the influence of Christianity has led some Sejiq to think of infringements of Gaya as individual sins rather than as collective responsibilities. Anthropologists thus argue that a decline of collective Gaya contributes to alcoholism, prostitution, divorce, and conjugal violence (Yu 1979).

In the past, observance of Gaya was visible in facial tattoos, as men and women who had proven obedience to Gaya through action had the right to tattoo their faces. Like elsewhere throughout Southeast Asia and Oceania (Hoskins 1996: 23, Davison and Sutlive 1991), women proved obedience to Gaya through weaving, whereas men did so by joining headhunting expeditions. As tattooed individuals were considered to be better marriage partners, facial tattoos were useful for reproductive success. After death, only tattooed individuals could cross the rainbow bridge of the ancestors (hakaw utux) and become ancestral spirits (utux ludan). Japanese administrators, aware of the connection between tattooing and headhunting, criminalized tattooing.

Any law needs enforcement, and Gaya is enforced by the utux. The Sejiq make a distinction between ancestral spirits (utux ludan), good spirits (utux malu), and evil spirits (utux naqex) (Jian 2002 : 158, Wang 2006 : 138-148). Christians translate the name of their deity as Utux Baraw ('the highest spirit'). The utux judge between good and evil, punishing wrong doers with the curse of lumuba. This curse is used to explain why people who sell ancestral land, commit sexual crimes, or steal subsequently become sick, get injured by falling in the mountains, or have mysterious automobile accidents (Lin 2010: 108). In the past, the headhunting rituals of mgaya made the presence of utux very real in Sejiq villages.

HUNTING HEADS ON FORMOSA: MOTIVATIONS

Japanese ethnographers and administrators found it difficult to untangle different motivations for headhunting. In 1915, Kojima Yoshimichi summarized research on Atayalic groups for the Governor-General Provisional Committee on the Investigation of Taiwan Old Customs. He emphasized that headhunting was different from war, because getting a head was the only goal. There was no attempt to wipe out the other community or reduce its military strength. Even if property was taken, this was not the primary reason for headhunting. Kojima identified three reasons for headhunting: 1) conflict resolution within the community; 2) retribution for the murder of close kin; 3) the desire of young men to prove their courage. He noted that some people reported headhunting could be done if two men competed for the same woman, or for a community to ward off misfortune, but doubted these reports due to the lack of documented cases (Kojima 1996 [1915]: 258-261).

Torii Ryuzo (1870-1953) entered Formosa's mountain areas in 1896 and, in four extended expeditions before 1900, was the first person to use modern photography to research aboriginal peoples. His apprentice and interpreter Mori Ushinosuke (1877-1926) spent nearly thirty years on the island. He was the most prominent ethnographer doing research during the military conquest of the mountains (1909-1914), working with the Bureau of Aboriginal Affairs. He established the ethnic classification of the mountain peoples that was adopted by the colonial state in 1913. He published many ethnographic studies, most notably the Taiwan aboriginal gazetteer (Mori 2008 [1917]), and gathered most of the aboriginal collection for the Taipei Museum, of which he eventually become director (Tierney 2007: 90). Mori's list of reasons for headhunting were to: 1) gain entry into the male initiation group; 2) establish innocence and resolve legal conflicts; 3) seek revenge for the death of close kin; 4) compete for marriage; 5)get rid of infectious disease and other forms of misfortune; or 6) prove (masculine) courage and gain respect (Mori 2008 [1917]: 315).

In 1932, Suzuki Tadasu of Japan's Bureau of Savage Administration on Formosa published the Gazette of Taiwan Savage Customs with the 'Friends of Administering the Savages.' He dedicated an entire chapter to headhunting, providing a longer list: 1) a prerequisite for male initiation; 2) intra-community conflict resolution; 3) retribution for murder of close kin; 4) courtship rivalry; 5) prevention of disease; and 6) proof of masculine courage (Suzuki 1932:163-165).

In the 1930s, anthropologist Furuno Kiyoto did field research on religion in Formosa. This was part of his Durkheimian research on religious life and practices that he subsequently continued in the Ryukyu Islands and throughout Southeast Asia (Yamashita 2004: 104). Reasons he gave were to: 1) ensure agricultural fertility; 2) alleviate misfortune; 3) drive away infectious disease; 4) demonstrate (masculine) courage; 5) determine a sacred judgment; or 6) take revenge (Furuno 1945: 431).

After the arrival of the Republic of China on the island, the new government took over administration of highlander communities and published similar works. Under the direction of Taiwan provincial governor Hsieh Tung-rain in 1972, a multi-volume Complete Gazetteer of Taiwan Province was published with the 5th volume dedicated to the Atayal, Saisiat, and Bunun tribes. In a description of Atayal headhunting, they listed six reasons for headhunting: to 1) prove masculinity; 2) clear accusations; 3) revenge murder of kin; 4) give emotional vent to resentment; 5) prevent epidemics; or 6) win over a bride (Li 1972: 24).

Masaw Mowna, also known by the Chinese name Liao Shou-chen, was an anthropologist from the Truku subgroup of Sejiq. His analysis of headhunting, based on a review of Japanese ethnographies and interviews with elders, remains the most detailed published study of the practice. He makes a clear distinction between two kinds of headhunting. Mesibo was headhunting against an enemy group for: 1) holding harvest rituals every three to four years; 2) gaining spiritual power from the head; 3) obtaining permission for facial tattooing and acquisition of ritual clothing and jewellery; 4) revenging enemy groups; 5) preventing epidemics; or 6) winning over a bride or venting anger. Ngdeolon was practiced to gain the sacred judgement of the ancestors in the case of a conflict within the group. This was important to keeping peace, as it directed violence toward outside groups and prevented intra-group violence (Masaw 1998:216-218).

The main defect in these exhaustive lists is that they confuse individual motivations and collective activities. First of all, there are individual motivations, such as the need to vent emotions. Young men may have been motivated by the desire to prove their masculinity. They had to participate in a headhunting group in order to join the initiation group and gain the right to receive a facial tattoo. Yet this was probably not an absolute prerequisite to marriage, like Igorot headhunters who reported that 'To take a head is not necessary to marriage, but the young women like it better' (Kolkmar 1974: 230). The reasons related to marriage and masculine courage were thus linked to individual reproductive success, and provided incentives for young men to join the expeditions. There were, however, also disincentives, as headhunting was extremely dangerous and there was no guarantee that the men would return alive. The incentives of tattoos, initiation, and higher chances of access to women were thus ways of encouraging cooperative behavior. I emphasize that older men led the expeditions and augmented their own political power through the cooperation of young men.

There are also collective reasons for headhunting. With the exception of Kojima, who attempted the first explanation of Formosan headhunting, all observers note that headhunting was done to eliminate infectious diseases and other misfortunes from the community. The historical context is surely important here. The Spanish influenza pandemic of 1918-1920 hit Japan severely (Rice 2003), including a 6-8% death rate in the armed forces (Kawana et al. 2007), and had highest mortality among young people. It is likely that a higher mortality rate in Formosan indigenous villages--caused by the infection of highlanders by Japanese soldiers and other military personnel--could have led to a rise in headhunting (see below). (15) Beginning from this period, anthropologists and native informants were certainly highly conscious of the danger of infectious disease and thus more likely to mention this dimension of headhunting. It is important to note that only ritual could make the connection between headhunting and prevention of disease.

The choice of which heads to hunt was surely also important. Mabuchi noted that headhunting generally happened between competing groups who lived along different drainage systems (Mabuchi 1974:186). This was thus related to conflicts about the hunting territories in forests between river systems. Communities were unlikely to attack just any village, as headhunting would certainly contribute to further hostilities. This is probably why headhunting was done to take revenge for the death of close relatives. The issue here is not why they hunted heads, but which heads were hunted. As only the heads of enemy groups were considered legitimate targets, headhunting in this sense was arguably part of a feuding complex like that seen in human societies worldwide (Boehm 1984). Masaw Mowna was even able to produce a table showing patterns of which villages went to war with which groups. Although most villages went after the heads of neighboring Bunun, Kavalan, Amis, etc., groups, there were also longstanding hostilities between the Truku and other Sejiq groups (Masaw 1998: 214). It seems that headhunters went primarily after other Austronesian groups, particularly when there was conflict about hunting territories, rather than after Han settlers in the more remote plains that were less important to their livelihoods. (16)

HEADHUNTING AS CONFLICT RESOLUTION

Sejiq headhunting practices seem unique due to the institution of headhunting to resolve intracommunity conflicts, which I have not found elsewhere in the literature on headhunting. According to Furuno, this is why headhunting is called mgaya (Furuno 1945: 432), or 'implementation of Gaya. (17) In the event of a legal conflict between two community members, as in when Person A accuses Person B of theft yet Person B insists on his or her innocence, they could resort to mgaya. This could also be done to regulate legal issues in sexuality (Furuno 1945: 432). The disputing parties would first consult a respected elder, who would try to find a compromise. If a compromise could not be found, the utux (deceased spirits) could make a judgment by granting a successful headhunt to the innocent party and inflicting damage on the other party. Heads could only be hunted from outside the community; and victoriously bringing back a head would restore one's social standing. (18)

As Kojima pointed out, a number of factors made it difficult to launch headhunting expeditions and limited their frequency. First of all, competing parties could be immediate family members. During the ordeal, they were considered enemies, and not permitted to have any contact. They could not hunt or farm together, and could not even greet one another. Since this was considered unlucky, this practice alone limited the frequency of headhunting. The parties themselves might not even participate in the expedition, especially if the party was a woman, an elderly man, or someone trapped within the boundaries of the Japanese aiyusen ([TEXT NOT REPRODUCIBLE IN ASCII.]). (19) The two parties would have to convince other members of their group of their own righteousness, and convince young men to join them on a headhunting expedition. Considering the dangers involved in headhunting, this process was quite difficult. Finally, they would proceed only if bird oracles indicated that they would succeed (Kojima 1996 [1915]: 259). Clearly, headhunting for conflict resolution was a last resort. These limitations on headhunting, all part of Gaya, curbed fatalities by isolating headhunting incidents in space and time. Headhunting, which was never a form of indiscriminate violence, was proscribed as much as it was enabled by the ritual process.

HEADHUNTING AS RITUAL

The Sejiq headhunting rituals, summarized here largely from the descriptions provided by Furuno and Masaw, are similar to descriptions from throughout Oceania and Southeast Asia (e.g. Barraud 1972, Folkmar 1974, George 1995, Hoskins 1996, Klokke 2004, McKinley 1979, Rosaldo 1980). Headhunting was done at the conclusion of harvest rituals and before ancestor worship rituals. In the event of gaining revenge, expressing anger, or resolving internal conflicts, it happened shortly after the event. In the case of mesibo, headhunting groups were composed of 3 to 5 young men. In the case of conflict resolution, groups consisted of 40 to 50 men. If it were done for revenge, it could involve all men in the alang. There was, however, no compulsion to join, and men could refuse for such reasons as illness, pregnancy in the family, or dreams predicting misfortune. (20) The expedition would be postponed in the event of death in the community, marriage, eclipse of the sun or moon, or mourning. One man emerged as the expedition leader (Masaw 1998:219-220).

During expeditions, family members of the participants observed a number of taboos. They had to extinguish the domestic fire, and begin a new fire. It was considered unlucky if this fire were extinguished. During the expedition, women were not permitted to weave. They could not borrow food or give food in hospitality. In the event of conflict resolution rituals, they could not communicate with members of the opposed group. They could not participate in wedding rituals, talk loudly, or swear. Non-participating men could not hunt or make house repairs. Any sexual violation could result in tragedy for members of the expedition. Family members could not hang a wooden pestle horizontally, or place blankets on a bed. Before leaving on the expedition, members could not have sexual relations, argue with others, or wash their face or clothes (Masaw 1998: 221). During this period, they also abstained from alcohol (Furuno 1945: 434). Headhunting was thus a collective ritual that involved entire families, ritual groups, and even entire alang. It was only accomplished at great cost and inconvenience.

Before launching the expedition, the leader held an oath ceremony over water (Truku, sbalay qsia). He held a branch in his right hand, dipped it in water, and sprinkled group members, saying 'We are now going out to implement Gaya. This water will protect us and help us achieve our goal' (Masaw 1998: 223). The members of the expedition slept together and related their dreams in the morning so that the expedition leader could interpret the dreams. Like the Dayak (Klokke 2004: 165), Iban (Davison and Sutlive 1991) and Igorot (Folkmar 1974: 223), Sejiq headhunters consulted bird oracles. If any man had an inauspicious dream, or if the oracle bird sisil (probably the Grey-cheeked Fulvetta) made agitated sounds or flew across their path, the expedition would be cancelled (Masaw 1998: 224-225). (21) There were thus ample exit opportunities, and no possibilities for the leader to compel men, as these were not armies or militias with the power to punish defectors. The oath, however, made the ritual political, as the reputation of the young men was at stake if they backed out after the oath (George 1995: 249). Once the expedition began, the leader could expect obedience from group members (Furuno 1945: 433).

Upon arrival near the settlement of the targeted group, they stopped to build a hut for a 'ritual to call the enemy spirits.' Facing the members, the leader shook reeds in the direction of the settlement, praying for a successful head hunt and safe return. After shaking the reeds over each person, he planted them in the ground, with the belief that the spirit of the captured head would protect the headhunters. After the ritual, they observed the settlement to determine a strategy (Masaw 1998: 226). They captured heads by attacking homes, smoking people out of their houses, or attacking people walking in the forest (Masaw 1998: 227). As for the weapons, Masaw lists steel knives (adopted from Chinese traders during the Qing Dynasty), wooden bows, and rifles (Masaw 1998: 222). According to Japanese police sources, there was no discrimination based on age or gender, as even heads of women or children could be taken (Suzuki 1932: 171).

After successfully taking a head, the successful warrior placed it in a net over his shoulders and they all ran back to the ritual hut. They washed the skull in a stream and tied the hair in a braid. They tied some hair on the knife that had been used to cut off the head, and put some in talismanic bags that would bring successful hunts and harvests to the holder. Finally, they returned to the village singing victory songs. The person who cut the head received a decorative bell to hang on his clothing while dancing (Masaw 1998: 227-28). Like elsewhere in Southeast Asia (George 1995: 245), all prohibitions and taboos ended as soon as the head entered the village. The successful warrior groups were thus greeted with ecstatic shouting, song, dance, and alcohol.

Headhunting ended in a community-wide ceremony. The head was displayed on a skull rack in front of the house of the person who captured the head. They fed the head rice, as well as sacrificed pigs and chickens. They gave it alcohol to drink, which would mix with its blood, and believed that drinking this mixture with the skull would strengthen their own hunting prowess. All members of the community, old and young, male and female, joined hands and danced around the head. These initial festivities lasted all night. The skull was then taken to the home of the chief, and placed on a community skull rack. A ritual was then held to send the spirit of the deceased away, praying to it saying:

We used a lot of courage to invite you here. Your family must be at ease. Having received so many offerings, you must be very happy. We welcome your family to come and join us here. You are so lucky to have come to our alang, but perhaps you are lonely? We ask your brothers to also come. When we go hunting, please give us lots of prey (Masaw 1998: 230).

This transformation of the head into a friend and brother was common to many 'Indonesian' groups (Furuno 1945: 432, McKinley 1979). The transformation ceremonies lasted for three days of intense drinking and dancing, during which community members gave sacrifices to the skull, asking it to become an ancestor. On the fourth day, the skull began to stink and the expedition leader would declare the end of the festival. In the final ritual, the men hunted and returned to feed it wild meat, entreating it to refrain from anger and protect them. The remnants of the head were made into talismans to be kept at the chief's home, the taboos were lifted, and the dried skull remained on the skull rack (Masaw 1998: 230-231).

The heads became important focal points in the community. As Kondo Katsusaburo, who was adopted into a Truku community in Musha, participated in a headhunting expedition in 1931, and subsequently published a serial article in the Taiwan nichinichi shinpo newspaper ([TEXT NOT REPRODUCIBLE IN ASCII].), described the practice, 'Aborigine men and women, it doesn't matter which, have a passion for heads, and a way of conceptualizing heads, which Japanese people would never understand' (Barclay 2008).

In the event that the expedition ended with the death or injury or a member of the group, they cast away the captured head in the forest. They then tore off their clothing and returned naked to the village at night. A shaman had to hold an exorcism before they could bring the corpse of their deceased compatriot back to the village. If they determined that a death occurred due to the violation of a taboo, the guilty party had to compensate the family of the deceased. All weapons and the talismanic bag used during the expedition were considered useless and discarded (Masaw 1998:231-232). The risk of this happening was certainly high, as the victims fought back on treacherous mountain terrain.

The successful capture of a head had consequences for the whole community. For the person who caught the head, it was the most glorious moment of his life. Any young man who had carried the head for part of the return trip or who gained the permission of the head-hunter could obtain the facial tattoo signaling adulthood. Leading figures in the hunt and subsequent rituals could claim a certain spiritual power from the head. The knife used in cutting the head, some plants carried on the expedition, as well as the remaining millet, alcohol and ashes from the head ceremony could be used in healing rituals performed by women shamans. Parts of the head, especially the hair, were added to the talismanic bag kept in the loft of the leader's house (Suzuki 1932: 176, cf. George 1995: 237). Any man who had participated in the head ceremony could also claim spiritual power that would be useful in future expeditions (Masaw 1998: 228). (22)

These ethnographic descriptions point toward a politics of headhunting. The villagers may have celebrated as a community, and young men collectively gained the right to tattoo their faces, but there were also other power dynamics. It is certainly important that elders were able to increase their own political power by holding the rituals and installing the skull racks in front of their own homes. Shamans, usually women, also claimed spiritual power from the ritual goods produced through headhunting rituals. Yet, there were limits to the power these elders could hold. Younger men may have had incentives to prove their courage, but headhunting expeditions were difficult to mount and dissenters could exercise a right to not participate--if only by claiming to have dreamt about inauspicious omens. If any would-be leader tried to exercise compulsion rather than merely the influence of his oratorical ability, people had the relatively easy option of splitting off and forming their own communities elsewhere in the mountain forests. By the time anthropologists could study these phenomena, however, headhunting was already near an end.

THE END OF HEADHUNTING IN FORMOSA

As part of his job in Savage Administration, Suzuki Tadasu published statistics on headhunting and resistance. In what were known as incidents of 'savage damage' (bangai, [TEXT NOT REPRODUCIBLE IN ASCII.]), Japanese statistics detail deaths and injuries done to military personnel and civilians by the native people. These statistics were disaggregated by military or police occupation, as well as by ethnicity (Japanese or Taiwanese). It is impossible to know if headhunting rituals were part of these deaths, or if they were acts of war. The deaths and injuries inflicted by native people against other native people is probably a closer proxy of headhunting rates. These statistics are probably underreported, especially in the first years of Japanese administration. Furthermore, it is impossible to determine which deaths were part of headhunting rituals and which were other forms of homicide. Nonetheless, the fact that Suzuki chose to present these in a section on headhunting suggests that he thought these statistics were the most reliable proxies of headhunting prevalence. As imperfect as they may be, they are the best indications we have of overall trends from 1895 to 1930.

[FIGURE 1 OMITTED]

The relatively low reported homicide rate from 1895 to 1900 is certainly due to feeble Japanese police presence in the mountains and inability to collect reliable statistics. There are, however, three notable spikes in the homicide rate, followed by a steep decline in the 1920s. The first spike began in 1901 when 22 deaths were reported, and peaked in 1908 with 70 deaths. This probably reflects wider collection of homicide data by Japanese police. The second spike occurred from 1910-1914, when Governor-General Sakuma Samata launched a 'Five Year Plan to Pacify the Savages,' culminating in the three month Taroko Battle of 1914. It is not surprising to see that violence spiked in this period, with 81 documented homicides in 1914. In a wider context of increased violence, young men probably perceived that they had little to lose and wanted to prove their masculinity through both headhunting and anti-colonial resistance. The third spike, which happened after pacification, is quite suggestive. This happened during the 1918-1920 influenza epidemic, as a record 96 homicides were reported in 1920, suggesting that native communities perceived headhunting as a way to ward off disease. Some older men say that the utux had punished them for capitulating to the Japanese, causing great misfortune, and needed to be appeased. Native homicides then dropped off significantly, which Suzuki could report as the success of Japanese police in elimination of headhunting practices.

Throughout the region, externally imposed states took credit for the end of headhunting (Barclay 2003). This was as true of Japanese administrators on Formosa (Suzuki 1932) as it was of the American rule in the Philippines (Keesing and Keesing 1934). It is possible, however, that the Sejiq and other headhunting peoples gave up the practice voluntarily as new technologies, especially with the spread of rifles, escalated the practice into efficient and bloody homicide. (23) At any rate, it is insufficient to explain the end of headhunting as merely a result of its criminalization. We have to understand why aboriginal communities chose to obey the Japanese on this important issue.

A comparison with the end of headhunting in New Georgia, studied by Martin Zelenietz (1979), is instructive here. On New Georgia, headhunting was a ritual needed to launch newly constructed canoes. Although this was obviously different from highland Formosa, the headhunting ritual was nonetheless similar in that it was an initiation rite for men and collecting heads augmented the authority of men who claimed to be chiefs. More than one man claimed power in each village, leading to competition between would-be leaders. In the 1880s, the introduction of the steel axe increased the efficacy of the killing and led for the first time to large-scale organization of head-hunting expeditions with central leaders or strongmen. New Georgians themselves terminated the practice because they perceived that it adversely affected their ability to engage in new trade relations with the colonial powers. They justified the end of the practice by the presence of British rule, and reinforced the decision with conversion to Christianity (Zelenietz 1979: 104).

This explanation, which can only speculatively be applied to the Sejiq case, is consistent with the logic of a reverse dominance hierarchy. Because the moral code of Gaya forbids the accumulation of power in the hands of one individual, the Sejiq tried to avoid conditions where one person monopolized rituals. In the past, they could simply leave and found new settlements. With the imposition of Japanese rule, however, they were forced to settle in villages and thus lost the former exit strategy. Cooperation with the Japanese police to end headhunting, however, would have limited the power that this ritual gave over would-be village leaders. This can best explain why so many ordinary people in the villages took pride in saying, 'in the past we were headhunters.' They take pride not in the fact that they once took heads, but in the fact that they ended this usurpation of power by would-be elites.

Japanese attempts to channel headhunting practices into inter-tribal warfare, however, added a new dimension to the story. In conversations about this past, Sejiq villagers expressed resentment about the fact that headhunting was eventually done only when commanded by the Japanese military to put down other tribes. In those cases, the headhunting expeditions were clearly led by local collaborators who sought to gain material advantages from the Japanese and consolidate local power. Villagers depict those expeditions as some of their most morally reprehensible actions in recent history. The new context, in which headhunting practices were manipulated by a colonial power, meant that headhunting was no longer associated with Gaya.

CODA: MEMORIES OF FORMOSA

In the 'postcolonial era,' both ordinary people and would-be elites have since turned to alternatives for at least some aspects of what their ancestors had previously gained from headhunting. For ordinary men, for whom headhunting had been an expression of masculinity (e.g. cf. Davison and Sutlive 1991), other warrior-like activities took the place of headhunting. During World War II, the Japanese recruited Formosan highlanders for military service. In Kawanakashima, one of my research sites, there were so many volunteers that the Japanese had to select the recruits by lottery (Walis and Yu 2002 : 166). Those who returned alive from the war are commended for their courage and praised for possessing a 'Japanese spirit.' Older Sejiq men today say that they participated in the war because headhunting had ceased and they needed a new way to prove their masculine courage. This is similar to Buaya society in Luzon, where veterans who had killed at least two people gained the status of headhunters (de Raedt 1996: 168). Male prowess was still valorized, but was now subsumed into a modern military ethos. Nowadays, young men claim that hunting for animals has replaced headhunting as an affirmation of masculinity. Of course, since headhunting cannot be reduced to warfare or the chase of prey, this substitution was only partial . (24)

Would-be village elites now turn to modern headhunting rituals, albeit without heads, in their quest for power. The Sejiq men who run community development associations apply for state funding to hold headhunting rituals as part of the autumn 'Ancestor Festival' designed to attract tourists and contribute to cultural renaissance. In grotesque spectacles, they use doll heads to act out the rituals of the past. These development association activities are part of electoral strategies that can get some men elected to new political positions (Simon 2010). These practices are similar to the headless simulacra of past rituals that permit Sulawesi highlanders to maintain ideological control of the past (George 1993). These new leaders now affirm their position in new political structures by hunting heads in the foggy forests of a half-forgotten history.

In Taiwan, this apparent affirmation of headhunting tradition is challenged; showing that the meaning of headhunting depends on social context. Although some pastors support the new rituals, others denounce them as sin. The True Jesus Church forbids its followers from participating in such activities entirely. The Presbyterians are divided on the issue, as some of the pastors are involved in organization of the new rituals and others denounce them. Those who oppose the rites perceive that there is a contradiction between the traditionalist ideas behind such a cultural revival and Christianity. Ordinary people, moreover, laugh hysterically at the spectacles as ridiculous actions of would-be leaders in search of cash and political power. Through these actions, they thus maintain their egalitarian ethos and affirm their own position as the only true followers of Gaya. Sejiq oscillation between egalitarianism and power-seeking strategies thus continues, but in radically different contexts.

ACKNOWLEDGEMENTS

The author would like to thank the Social Science and Humanities Research Council of Canada for funding field research in Taiwan. Paul Barclay, Lin Ching-Hsiu, and Chou Wan-yao, as well as the anonymous reviewers, all made copious and helpful comments to improve this article. Any errors that remain are mine alone. Tigger Wise, herself an accomplished author, shepherded it through the entire process at Oceania. Finally, special thanks are due to members of three Sejiq communities, too numerous to name, who have generously shared their lives and opened their homes to me over the past decade. Mhuway namu balay!

REFERENCES

ANDRADE, T. 2008. How Taiwan Became Chinese: Dutch, Spanish and Han Colonization in the Seventeenth Century. New York: Columbia University Press.

BARCLAY, P. 2001. An Historian among the Anthropologists: The Ino Kanori Revival and the Legacy of Japanese Colonial Ethnography in Taiwan. Japanese Studies 21 (2): 117-136.

2003. 'They have for the Coast Dwellers a Traditional Hatred': Governing Igorots in Northern Luzon and Central Taiwan, 1895-1915, pp. 217-255 in Julian Go and Anne L. Foster (eds.), The American Colonial State in the Philippines: Global Perspectives. Durham: Duke University Press.

2007. Contending Centres of Calculation in Colonial Taiwan: The Rhetorics of Vindication and Privation in Japan's ' Aboriginal Policy." Humanities Research 14 (1): 67-84. (trans.).2008. Serial IIl: Kondo Katsusasboro among Taiwan's Atayal/Sedeq Peoples, 1896 to 1930. The View from Taiwan. (http://michaelturton.blogspot.com/2008/11/serial-iii- kondo-katsusaburo-among.html,last accessed April 20,2011).

BARRAUD, C. 1972. De la Chasse aux tetes a la peche a la bonite. Essai sur la chefferie a Eddystone. L'Homme 12 (1) : 67-104.

BELLWOOD, P., J. FOX, and D. TYRON (ed.). 1995. The Austronesians: Historical and Comparative Perspectives. Canberra, Australia: Australian National University.

BOEHM, C. 1984 Blood Revenge: The Anthropology of Feuding in Montenegro and Other Tribal Societies. Lawrence, KS: University Press of Kansas.

1993. Egalitarian Behavior and Reverse Dominance Hierarchy. Current Anthropology 34 (3): 227-254.

BROWN, M. 2004. Is Taiwan Chinese? The Impact of Culture, Power, and Migration on Changing Identities. Stanford: Stanford University Press.

CAUQUELIN, J. 2004. The Aborigines of Taiwan, Puyuma: From Headhunting to the Modern World. London: Routledge.

COHN, B .S. 1996. Colonialism and its Forms of Knowledge: the British in India. Princeton: Princeton University Press.

DAVIDSON, J.W. 2005 [1903]. The Island of Formosa Past and Present: History, People, Resources, and Commercial Prospects. London: Macmillan & Company (Reprinted by SMC Publishing, Taipei).

DAVISON, J. and V.H. SUTLIVE, Jr. 1991. The Children of Nising: Images of Headhunting and Male Sexuality in lban Ritual and Oral Literature, pp. 153-230 in V.H. SUTLIVE, Jr. (ed.), Male and Female in Borneo: Contributions and Challenges to Gender Studies. Williamsburg, VA: Borneo Research Council.

DE RAEDT, J. 1996. Buaya Headhunting and Its Ritual : Notes from a Headhunting Feast in Northern Luzon, pp. 167-183, in J. HOSKINS (ed.), Headhunting and the Social Imagination in Southeast Asia. Stanford: Stanford University Press.

DESVEAUX, E. 1996. Contro1e de naissances et classes d'age. L'Homme 36 (138) : 143-149.

FOLKMAR, D. 1974. Social Institutions of the Tinglayan lgorot. Philippine Sociological Review 22 (1-4):217-242.

FURUNO, K. [TEXT NOT REPRODUCIBLE IN ASCII.] 1945. Takasagozoku no saigi seikatsu[TEXT NOT REPRODUCIBLE IN ASCII.] [The Ritual Life of Takasagozoku]. Tokyo: Sanseido.

GEORGE, K.M. 1993. Lyric, History and Allegory, or the End of Headhunting Ritual in Upland Sulawesi. American Ethnologist 20 (4): 696-716.

1995. Violence, Solace and Ritual: a Case Study of Island Southeast Asia. Culture, Medicine and Psychiatry. 19: 225-260.

1996. Showing Signs of Violence: The Cuhural Politics of a Twentieth-century Headhunting Ritual. Berkeley: University of California Press.

HOCART, A.M. 1914. Maria. Man 14: 97-101.

HOSKINS, J. (ed.). 1996. Headhunting and the Social Imagination in Southeast Asia. Stanford: Stanford University Press.

HOSKINS, J. 1989. On Losing and Getting a Head: Warfare, Exchange and Alliance in a Changing Sumba, 1888-1988. American Ethnologist 16 (3): 419-440.

1996. Introduction: Headhunting as Practice and as Trope, pp. 1-49, in J. HOSK1NS (ed.), Headhunting and the Social Imagination in Southeast Asia. Stanford: Stanford University Press.

2002. Predatory Voyeurs: Tourists and 'Tribal Violence' in Remote Indonesia. American Ethnologist 29 (4): 797-828.

HSIEH, S.C. [TEXT NOT REPRODUCIBLE IN ASCII.] 1996. 'Chuantong wenhua' de caokong yu guanli: Guojia wenbua tixi xia de Taiwan Yuanzhumin wenhua [TEXT NOT REPRODUCIBLE IN ASCII.] [The Control and Management of 'Traditional Culture': Taiwan Aboriginal Culture under the System of National Culture]. Shanhai Wenhua [TEXT NOT REPRODUCIBLE IN ASCII.] [Mountain and Ocean Culture] 13: 85-101.

HUANG, C.H. [TEXT NOT REPRODUCIBLE IN ASCII.] 1996. Dong Saideke qun de shoulie wenhua [TEXT NOT REPRODUCIBLE IN ASCII.] [The Hunting Culture of the Eastern Sejiq Group]. Minzuxue yanjiusuo ziliao huibian [TEXT NOT REPRODUCIBLE IN ASCII.] [Institute of Ethnology Data Journal] 15. Taipei: Academia Sinica.

HUANG, K.C. [TEXT NOT REPRODUCIBLE IN ASCII.] 2000. Taiyazu de Shehui Goucbeng gaga, niqan yu qalang [TEXT NOT REPRODUCIBLE IN ASCII.]--gaga-niqan [TEXT NOT REPRODUCIBLE IN ASCII.] qalang [Atayal Social Construction: Gaga, Nigan and Qalang]. Tianye Quanshi yu Xin Shengdai Renleixue : Zhongyanyuan Minzusuo, Taida Renleixi, Qingda Renleisuo Yanjiusheng Tianye Peixun Jihua Chengguo Fabiaohui[TEXT NOT REPRODUCIBLE IN ASCII.] [TEXT NOT REPRODUCIBLE IN ASCII.] [Fieldwork interpretation and a New Generation of Anthropology: Proceedings from the Graduate Field Training Project of Academia Sinica Institute of Ethnology, NTU Department of Anthropology, and Tsinghua University Graduate Institute of Anthropology]. Taipei: Academia Sinica Institute of Ethnology.

HUANG, Y.K. [TEXT NOT REPRODUCIBLE IN ASCII.] 1986. Taiwan Tuzhuzu de Liang Zhong Shehui Leixing je qi Yiyi [TEXT NOT REPRODUCIBLE IN ASCII.] [TEXT NOT REPRODUCIBLE IN ASCII.] [Taiwan Aborigines' Two Social Types and their Meanings], pp. 3-43 in HUANG Y.G. (ed.), Taiwan Tuzhu Shehui Wenhua Yanjiu Lunwenji [TEXT NOT REPRODUCIBLE IN ASCII.] [Anthology of Research on Taiwanese Aboriginal Society and Culture]. Taipei: Lianjing Publishing.

HUANG, C.K. [TEXT NOT REPRODUCIBLE IN ASCII.] 2004. URM yu Taiwan shehui yundong : Cong Wu Feng shjian tanqi URM~-, ~ ~r~-~)3 [TEXT NOT REPRODUCIBLE IN ASCII.] [URM and Taiwan's Social Movements: From the Wu Feng Incident]. Tainan: Private publication.

JIAN, H.M. [TEXT NOT REPRODUCIBLE IN ASCII.] 2012. Zuling yu tianzhu : Meixi Tianzhutang chuanjiaoshi chutao [TEXT NOT REPRODUCIBLE IN ASCII.] [TEXT NOT REPRODUCIBLE IN ASCII.] [Ancestral Spirits and God : Exploring the Missionary History of the Meixi Catholic Church]. Taipei : Furen University Press.

KAWANA, A., G.NAKA, Y. FUJIKARA, Y. KATO, Y. MIZUNO, T. KONDO, and K. KUDO. 2007. Spanish Influenza in Japanese Armed Forces, 1818-1920. Emerging Infectious Diseases 13 (4): 590-593.

KEELEY, L.H. 1996. War BeJore Civilization. Oxford: Oxford University Press.

KEESING, F. and M. KEESING. 1934. Taming Philippine Headhunters: a Study of Government and of Cultural Change in Northern Luzon. Stanford: Stanford University Press.

KLOKKE, A.H. (ed.). 2004. Fishing, Hunting and Headhunting in the Former Culture of the Ngaju Dayak in Central Kalimantan. Phillips, ME: Borneo Research Council.

KNAUFT, B. 1991. Violence and Sociality in Human Evolution. Current Anthropology 32 (4): 391-428.

KOJ1MA, Y.[TEXT NOT REPRODUCIBLE IN ASCII.] 1996 [1915]. Fanzu Xiguan Diaoeha Baogao Shu, Di yi juan, Taiyazu[TEXT NOT REPRODUCIBLE IN ASCII.] [TEXT NOT REPRODUCIBLE IN ASCII.] [Report on the Customs of Savage Tribes, Volume 1: Atayal Tribe], translated by the Institute of Ethnology, Academia Sinica. Taipei: Academia Sinica Institute of Ethnology.

KOLKMAR, D. 1974. Social Institutions of the Tinglayan Igorot. Philippine Sociological Review 22 (1-4): 217-242.

LI, R.H. [TEXT NOT REPRODUCIBLE IN ASCII.] (ed.). 1972. Taiwan sheng tong zhi. Juan 8, Tong zhou zhi. [TEXT NOT REPRODUCIBLE IN ASCII.] 8 [Taiwan Provincial Gazetteer, Roll 8, ].Taizhong: Taiwan sheng wen xian wei yuan hui.

LIN, C.H. 2010. Women and Land: Privatization, Gender Relations and Social Change in Truku Society, Taiwan. Ph.D. thesis, Social Anthropology, University of Edinburgh.

2011. The Circulation of Labour and Money: Symbolic Meanings of Monetary Kinship Practices in Contemporary Truku Society, Taiwan. New Proposals: Journal o['Marxism and Interdisciplinary Inquiry 5 (1): 27-41.

LINDHOLM, C. 2008. Culture and Authenticity. London: Blackwell Publishing.

LU, H.Y. 2002. The politics of locality : making a nation of communities in Taiwan. New York: Routledge.

MABUCHI, T. 1974. Ethnology of the Southwestern Pacific: the Ryukyus-Taiwan-lnsular Southeast Asia. Taipei: Chinese Association for Folklore.

MCKINLEY, R. 1979. Human and Proud of It! A Structural Treatment of Headhunting Rites and the Social Definition of Enemies, pp. 92-126 in G. N. APPELL (ed.), Studies in Borneo Soeieites: Social Process and Anthropological Explanation. DeKalb, IL: Northern Illinois University Center for Southeast Asian Studies.

MCWILLIAM, A. 1996. Severed Heads that Germinate the State: History, Politics, and Headhunting in Southwest Timor, pp. 127-166, in J. HOSKINS (ed.), Headhunting and the Social Imagination in Southeast Asia. Stanford: Stanford University Press.

MASAW, M. [TEXT NOT REPRODUCIBLE IN ASCII.] 1998. Taiya Zu de Shehui Zuzhi [TEXT NOT REPRODUCIBLE IN ASCII.] , [Atayal Social Organization]. Hualien : Tzu Chi University Research Center on Aboriginal Health.

MORI, U. [TEXT NOT REPRODUCIBLE IN ASCII.] 2008 [1917]. Taiwan Banzokushi[TEXT NOT REPRODUCIBLE IN ASCII.] [Gazetteer of Taiwan k Savage Tribes]. Taipei: Southern Materials Centre.

MORRIS, A. 2004. Taiwan's History: An Introduction, pp. 3-31 in D.K. JORDAN, A. MORRIS, and M.L. MOSKOWITZ (ed.), The Minor Arts of Daily Life: Popular Culture in Taiwan. Honolulu: University of Hawai'i Press.

NEEDHAM, R. 1976. Skulls and Causality. Matt 11 (1): 71-88.

PASSIN, H. 1947. A Note on Japanese Research in Formosa. American Anthropologist 49 (3): 514-518.

PECORARO, F. 1977. Essai de dictionnaire Taroko-Francais. Paris: EHESS.

RICE, G.W. 2003. Japan and New Zealand in the 1918 Influenza Pandemic: Comparative Perspectives on Official

Responses and Crisis Management, pp. 73-85 in H. PHILLIPS and D. KILLINGRAY (eds.), The Spanish Influenza Pandemic of 1918-19: New Perspectives. London: Routledge.

ROSALDO, M.Z. 1983. The Shame of Headhunters and the Autonomy of Self. Ethos 11 (3): 135-151.

ROSALDO, R. 1980. llongot Headhunting 1883-1974. Stanford: Stanford University Press.

1993. Introduction: Grief and a Headhunters Rage, pp. 1-21 in R. ROSALDO (ed.), Culture and Truth: the Remaking of Social Analysis. Boston: Beacon Press.

RUDOLPH, M. 2003. Taiwans multi-ethnisehe Gesellschaft und die Bewegung der Ureinwohner : Assimilation oder kulturelle Revitalisierung? Hamburg: LIT Verlag.

2008. Ritual Performances as Authenticating Practices: Cultural Representations of Taiwan's Aborigines in Times of Political Change. Hamburg: LIT Verlag.

SHEPHERD, J. 1993. Statecraft and Political Economy on the Taiwan Frontier, 1600-1800. Stanford: Stanford University Press.

1995. Marriage and Mandatory Abortion among the 17th-century Siraya. Arlington, VA: American Ethnological Society.

SIMON, S. 2007. Paths to Autonomy: Aboriginality and the Nation in Taiwan, pp. 221-240 in C. STORM and M. HARRISON (eds), The Margins of Becoming. Identity and Culture in Taiwan. Wiesbaden: Harrassowitz.

2010. Negotiating Power: Elections and the Constitution of Indigenous Taiwan. American Ethnologist 37 (4): 726-740.

STAINTON, M. 1995. Return our Land: Counterhegemonic Presbyterian Aboriginali~ in Taiwan, MA thesis in Social Anthropology, York University, Toronto. [TEXT NOT REPRODUCIBLE IN ASCII.] [Gazetteer of Taiwan Savage Customs].

SUZUKI, T. [TEXT NOT REPRODUCIBLE IN ASCII.] -1932. Taiwan Banjin fuzokushi z~:,,~ .

Taihoku: Riban no tomo Hakkojo.

TAIHOKO TEIKOKU DA1GAKKU, DOZOKU JINRUIGAKU KENKYUSHITSU[TEXT NOT REPRODUCIBLE IN ASCII.] [TEXT NOT REPRODUCIBLE IN ASCII.] [Taihoku Imperial University Research Institute of Folklore and Anthropology]. 1935. Taiwan

Takasagozoku : keito shozoku no kenkyu [TEXT NOT REPRODUCIBLE IN ASCII.] [The Formosan native tribes: A genealogical and classificatory Study]. Taihoku (Taipei): Taihoku Imperial University.

TAKEKOSH1, Y. 1996 [1907]. Japanese Rule in Formosa, translated by George Braithwaite. Taipei: SMC Publishing.

T1ERNEY, R. 2007. Ethnography, Borders and Violence: Reading Between the Lines in Sat6 Haruo's Demon Bird. Japan Forum 19 (1): 89-110.

VAN BREMEN, J. and A. SHIMIZU (ed.). 1999. Anthropology and Colonialism in Asia and Oceania. Richmond, Surrey: Curzon.

WALIS N. and YU G.H. [TEXT NOT REPRODUCIBLE IN ASCII.] 2002. Tuiwan yuanzhuminshi : Taiyauzu shipian [TEXT NOT REPRODUCIBLE IN ASCII.] [TEXT NOT REPRODUCIBLE IN ASCII.] [Taiwanese Indigenous History: Atayal Tribe]. Nantou : Taiwan Wenxian guan [TEXT NOT REPRODUCIBLE IN ASCII.]

WANG, M.H. [TEXT NOT REPRODUCIBLE IN ASCII.] 2003. << Cong gaga de duoyixing kan Taiyazu de shehui xingzhi gaga [TEXT NOT REPRODUCIBLE IN ASCII.] [TEXT NOT REPRODUCIBLE IN ASCII.] , Taiwan renleixuekan [TEXT NOT REPRODUCIBLE IN ASCII.] 1 (1) : 77-104. [TEXT NOT REPRODUCIBLE IN ASCII.] 2006. Taiyazu [TEXT NOT REPRODUCIBLE IN ASCII.] [Atayal Tribe]. Taipei : Sanmin Publishers.

WEI, H.L. [TEXT NOT REPRODUCIBLE IN ASCII.] 1965. Taiwan tuzhu shehui de buhio zuzhi yu quanwei zhidu[TEXT NOT REPRODUCIBLE IN ASCII.] [TEXT NOT REPRODUCIBLE IN ASCII.] [Tribal Organization and Authority Systems of Taiwanese Aboriginal Society ]. Kaogu renleixue kan [TEXT NOT REPRODUCIBLE IN ASCII.] [Journal of Archaeology and Anthropology] 25-26:71-92.

WILDE, C. 2004. From Racing to Rugby: All Work and No Play for Gogodala Men of Western Province, Papua New Guinea. Australian Journal of 'Anthropology 15 (3): 286-302.

YAMASHITA, S. 2004. Constructing Self and Others in Japanese Anthropology: The Case of Micronesia and Southeast Asian Studies, pp. 90-113 in S. YAMASHITA, J. BOSCO and J .S. EADES (eds.), The Making of Anthropology in East and Southeast Asia. New York: Berghahn Books.

YANG, S.Y. [TEXT NOT REPRODUCIBLE IN ASCII.] 2005. Taiwan gaodi de zhengzhi tixi chutan: yi Bunun ren wei li de yanjiu [TEXT NOT REPRODUCIBLE IN ASCII.] [TEXT NOT REPRODUCIBLE IN ASCII.] [Preliminary explorations of Taiwao Highland Political Systems: Research on the Bunun People as an Example]. Taiwan renlei xuekan [TEXT NOT REPRODUCIBLE IN ASCII.] [Taiwan Journal of Anthropology] 3 (1): 185-219.

YU, G.H. [TEXT NOT REPRODUCIBLE IN ASCII.] 1979. Dong Saideke ren de liang xing guanxi [TEXT NOT REPRODUCIBLE IN ASCII.] [Gender Relations of the East Sejiq People]. Zhongyang yanjiuyuan minzuxue yanjiu jikan [TEXT NOT REPRODUCIBLE IN ASCII.] [Bulletin of the Institute of Ethnology, Academia Sinica] 48:31-53.

1980. Taiyazu Dong Saidekequn de buluo zuzhi [TEXT NOT REPRODUCIBLE IN ASCII.] [Tribal Organization of the the Atayal East Sejiq People]. Zhongyang yanjiuyuan minzuxue yanjiu jikan [TEXT NOT REPRODUCIBLE IN ASCII.] [TEXT NOT REPRODUCIBLE IN ASCII.] [Bulletin of the Institute of Ethnology, Academia Sinica] 50:91-110.

ZELENIETZ, M. 1979. The End of Headhunting in New Georgia, pp. 91-108 in M. RODMAN and M. COOPER (eds.), The Pacification of Melanesia. Ann Arbor: University of Michigan Press.

NOTE

(1.) The use of the term Formosa may seem atavistic to some readers, especially due to disputes in contemporary Taiwan about the terms of 'Taiwan' and the 'Republic of China/In order to sidestep such debates, which lay far beyond the scope of this article, but more importantly to avoid conceptual confusion, it is important to select words as carefully as possible. Following conventions in the linguistic anthropology of Austronesia, I thus use Formosa (and the adjectival form Formosan) to refer to the geographical island and its Austronesian inhabitants in the longue duree. Taiwan refers to the contemporary sociological constellation, and Republic of China refers to the state.

(2.) The word dayan in Atayal means 'human being.' In the three eastern dialects, the word for 'human being' is sejiq (Truku), sediq (Teuda), and seediq (Tkedaya). These words for 'human being' were transformed into ethnonyms by Japanese anthropologists Utzukawa, Mabuchi and Miyamoto (Taihoku Teikoku Daigakku 1935). These names are fiercely debated by local activists, who claim identities as indigenous peoples and hope to establish autonomous regions (Simon 2007). For the reader's convenience, I use one ethnonym throughout this article. I use the word Sejiq because I spent most of my time with the Truku and this is their pronunciation.

(3.) For historical perspectives on Japanese colonial anthropology, see Barclay 2001, Barclay 2007, and van Bremen and Shimiju 1999. For a similar perspective on British survey anthropology in India, see Cohn 1996. For Cohn, the main goal of colonial surveys was to create social categories for administrative purposes.

(4.) I think that this explanation is a bit ex post facto, influenced by notions of 'sovereignty' highlighted in contemporary televised debates about 'Taiwan' and the 'Republic of China,' as well as debates about proposed 'indigenous autonomy zones.' Besides, the ethnohistorical data show that much of the headhunting occurred between groups now subsumed under the same 'ethnic' group or indigenous 'nation.' Since the ethnonyms themselves are the product of colonial encounter and statecraft, it would be an anachronism to imagine that anyone ever engaged in warfare or feuding on behalf of a large and imagined 'nation' called Atayal, Sejiq, Truku, Bunun, etc. To the extent that their ancestors were protecting territory, it was in defence of hunting grounds for their patrilocal band (alang).

(5.) The Atayalic groups do not have surnames. Instead, the personal name is followed by the patronym, the personal name of the person's father. Following the instructions of one of my research informants, I thus cite Sejiq scholars by their personal names here and in the bibliography.

(6.) The taking of heads has also been a part of warfare and raiding in Europe, up to the early 20th century in the Balkans. On the rare occasions when Montenegrin warriors took heads, however, the act was intended to insult the victims by preventing a proper burial (Boehm 1984 : 112). As there was no ritual associated with the practice, it does not fit the narrow definition of headhunting used in this article.

(7.) Mappurondo is not the name of an ethnic group; in fact, the people George studied call themselves Todiulunnasalu ('people of the headwaters'). Mappurondo, derived from a Bugis and Mandar word, means 'already in place.' George uses the term to describe autochtonous ritual practices, as opposed to Christian and Muslim practices introduced from the lowlands (George 1996: 25-27). I have seen no signs of such well defined traditionalism in Taiwan.

(8.) I argue that there is no reason to look for "authenticity' here. Staged headhunting rituals, although obviously not authentic in the sense of using real heads or faithfully reproducing the earlier rituals of violence, are nonetheless meaningful to the organizers. As Charles Lindholm argued, the quest for authenticity is 'of great significance to those who are swept up in the modern torrent of change, and who must try to spin a lifeline out of thin air to keep from drowning in the deluge" (Lindholm 2008: 145). George likewise points out that surrogate heads are "equally real, powerfully vested signs of violence' (George 1996: 60).

(9.) In Indonesia, highland communities that convert to Islam or Christianity likewise abandon commemorative headhunting rituals (George 1996: 202).

(10.) Elsewhere, Boehm suggested that the egalitarian warrior society can sometimes be an adaptation to a predatory state, as tribal people fight for local autonomy against an expanding empire. Mountains, with their difficult terrain and marginal economic value, have tended to provide a territorial base for such societies. Montenegrin resistance to the Ottoman Empire is a good example of this type of political adaptation (Boehm 1984: 40-41). Whether or not Formosan highlander societies also fit into this model, resisting Chinese, Japanese and Okinawan expansion for centuries, would likewise make a productive ethnohistorical project.

(11.) This contrasts with the way in which urban-based social movements deplored the Wu Feng myth once taught in school to emphasize the distinction between mountain savagery and Confucian civilization. In this myth, a Confucian scholar sacrificed his own life to convince Tsou people to give up headhunting (see Huang 2004, Rudolph 2008: 49).

(12.) The first time I heard this story, a Presbyterian minister engaged in the movement for Taroko autonomy deplored in church the fact that the so-called Taroko tribe has always been divided into enemy factions; and that they have even waged war between competing villages. In his sermon, he used this story to explain why the Truku supported the establishment of a Taroko Autonomous Region, but the other two sub-groups of Teuda and Tkedaya did not. He thus gave a new interpretation of an old trope going back to colonial days. For him, division is the original sin of the Taroko tribe; and must be overcome by the diligence of the church and its project of autonomy.

(13.) Translation from the French by the author. The use of all capitals for GAYA is from the original.

(14.) Throughout most of Oceania, Christian missions were common accompaniments to the "civilising" goals of the state and probably contributed directly to the cessation of headhunting. On Formosa, however, the Japanese state constructed Shinto shrines in the villages, and actively discouraged conversion to Christianity, which they saw as a potential source of western influence. Only one of the three villages in this study converted to Christianity during the Japanese period, and believers were forced to hold services secretly in a cave in order to avoid arrest. Christianity arrived in the other two villages only after the Japanese left in 1945, long after the end of headhunting. The contribution of Christianity toward ending headhunting was thus rather minimal.

(15.) Thanks to Gao Zhizheng (Gui Giling), Paiwan tribal elder and President of the Indigenous Medical Association, for pointing out the connection to an actual pandemic in private conversation.

(16.) The most documented cases of 'headhunting' are murders of Chinese camphor workers, and occasionally of Japanese, for revenge when the outsiders failed to carry out promises made in exchange for access to camphor trees (e.g. Davidson 2005 [1903]: 416-417). Davidson calculated that 303 attacks occurred and 635 persons were killed in 1898, arguing that 'this death roll is evidence that the Japanese have not yet solved the savage question' (Davidson 2005 [1903]: 428). These acts should be understood as acts of war, not ritual, and thus are not central to this analysis.

(17.) Mgaga in the dialect known to Furuno.

(18.) If we were to compare this to a wider range of behaviors and outcomes worldwide, as Boehm has done, we could see that this practice would, like limitations on feuds, 'prevent escalation of serious individual conflicts to the level of intensive group conflict at close quarters' (Boehm 1984: 227). In this case, the violence was redirected outside of the community, thus increasing the likelihood of its survival.

(19.) This was a fortified line of electric fences, armed soldiers and mines used to encircle indigenous communities and pressure them to surrender to the Japanese.

(20.) Desveaux interpreted this differently, seeing headhunting as a form of birth control because men (in the western Formosan Siraya tribe) often avoided getting their wives pregnant in order that they could participate in the warrior expeditions (Desveaux 1996: 146). This was an extreme case, and led to women giving birth only in their 30s so that men could avoid the supernatural dangers caused by pregnancy (Shepherd 1995: 3).

(21.) In a curious variation, the Dayak consulted a hawk, whereas the Sejiq consulted one of the smallest birds of the jungle.

(22.) The fact that the glory was shared and power gained by any man who participated in the expedition surely limited fatalities. This meant that only one headhunting expedition was needed in a generation in order for all men in a small community to become eligible for facial tattoos.

(23.) In fact, as one of the anonymous reviewers of this article observed, trade in guns and weaponry in other parts of Southeast Asia co-existed with headhunting for centuries and may indeed have intensified its expression. The question is why the Sejiq and other Formosan groups gave up so quickly after getting access to technology that would have made headhunting even more efficient.

(24.) Young Sejiq men also say they hunt and engage in such activities as cock-fighting to prove their masculinity. They evoke the memory of headhunting to explain why they take pride in these activities suppressed by the state, as well as to why they relish job opportunities in the military, the police force, and the fire department. This may be similar to Papua New Guinea, where rugby has replaced headhunting as an expression of masculinity (Wilde 2004). Obviously, none of these activities can replace the complex ritual of headhunting. Making such a claim, however, is a potent statement of pride in both their family histories and contemporary lifestyles.

Scott Simon

University of Ottawa
COPYRIGHT 2012 Blackwell Publishing Limited, a company of John Wiley & Sons, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2012 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Simon, Scott
Publication:Oceania
Article Type:Report
Geographic Code:80OCE
Date:Jul 1, 2012
Words:15045
Previous Article:The silent missiles of Guadalcanal: magic as rhetoric in the Sandline controversy.
Next Article:The process of kinship in the paternal/fraternal house of the Austronesian-speaking Amis of Taiwan.
Topics:

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters