Leo Howe, University of Cambridge
Performance approaches to the interpretation of ritual highlight process, presence, strategy and uniqueness. They are often favourably contrasted to textual approaches said to emphasize meaning, structure and stability. However, this sharp distinction can be maintained only by ignoring the notion of inscription, which is central to text. Inscription is a political process which involves risk, strategy and struggle. Moreover, commonly performance approaches neglect questions of risk in ritual action. Ritual is often seen as suppressing risk, but many rituals are events involving high risk. This article considers various Balinese rituals in which risk is a prominent feature, and suggests ways in which a focus on risk may aid our understanding of ritual.
One reason the metaphor of ‘performance’ has been popular in the analysis of ritual is the concern that textual approaches insufficiently appreciate the importance of a ritual’s occasion. Text is considered inferior to metaphors of dialogue and of performance, the latter advanced to compensate for the apparent failings of the former.1 Metaphors of text tend to construe ritual as an almost automatic acting out of rules based on an underlying cultural logic. This turns ritual action into symbolic action, and analysis becomes the disclosing of the meaning embedded in the symbols. For various reasons performance theorists argue that rituals involve more than this. Officiants and participants always bring their own competencies, reputations and interests to a ceremony, so that the ritual becomes a specific performance rather than an exercise in repetition, and what is important about it is what the personnel make of their opportunities. Consequently, textual metaphors, with their emphasis on meaning, displace the doing, the performative dimension of the action. Ritual action, however, works not only on the level of meaning but also emotionally and socially, the events of any particular occasion being conditioned by many factors other than the prescribed rules.
Despite a growing consensus about the merits of the notion of performance, I think that it pays inadequate attention to elements of risk that inevitably accompany performances. I shall also argue that the criticism of text is problematical. This article, therefore, has three aims: to make a case for the continuing significance of text; to trace some of the relations between text and performance; to comment on the neglected aspect of risk. The ethnography concentrates on various Balinese ritual performances in which risk figures prominently.

Text and performance
My starting point is a recent discussion of text and performance by Schieffelin (1998), whose fine interpretations of Kaluli ritual and shamanism (1985; 1996) have done much to enhance the appeal of the performance metaphor. He argues that text and performance are not the same, and while they ‘may be produced out of one another, this is very different from saying they are reducible to one another’ (1998: 199).Texts share some qualities with performances: they both have a definite sequential pattern, an internal structure, and may be self-referential. Nonetheless, ‘it is precisely the performativity of performance for which there is no analogue in text’ (1998: 198). Performativity, thus, recaptures the elements of uniqueness, strategy, evanescence, presence and becoming in social action that are thought to be lost by a concentration on the meanings in fixed texts. According to Schieffelin, texts are ‘changeless’ whereas ‘performances are assertive, strategic and not fully predictable’, so that performance can never be text, and its strategic properties are destroyed when it is considered as, or reduced to, text. . . . Unlike text, performances are ephemeral. They create their effects and then are gone – leaving their reverberations (fresh insights, reconstituted selves, new statuses, altered realities) behind them. (1998: 198–9)
While there is much here I agree with, Schieffelin’s description of text fails to mention what is perhaps the most significant feature of the text metaphor, inscription. Once inscription is taken into account, the relation between text and performance becomes more subtle and intricate.
By ‘text’ Schieffelin seems to mean something already accomplished, a fixed and enduring entity which has a specific set of meanings; a novel, for example. This static view of text resembles the idea that a ritual is little more than a mechanical playing out of the rules thought to structure it. Clearly this is one way text can be understood, but it is limited. It is text in this sense that informs several of Geertz’s most influential articles (1973a; 1973b), particularly his essay on the Balinese cockfight (1973c). Geertz’s approach there is basically semantic: ritual is primarily communicative because symbols carry meanings, and understanding the efficacy of symbols is mostly a matter of revealing the logic of thought and the motivational force that underpins them. Rituals are ‘texts’ which can be ‘read’ for their symbolic significance. Like the Balinese cockfight, they provide a ‘metasocial commentary’ on experience; they are stories which the Balinese tell themselves about themselves (1973c: 448). These essays have been widely criticized for their essentialism and ahistoricism (Asad 1983;
Crapanzano 1986; Thomas 1989: 25–7; 1994: 93–4;Wikan 1990), criticisms that indicate that Geertz is treating text as a stable and immutable entity. Ironically, it was Geertz himself who provided the most lucid account of the importance of inscription to the text analogy, in his essay ‘Blurred Genres’ (1980). There he compares text to game and drama. Noting that while game and drama are natural analogies for social action, in that people describe their own activity in terms of players, actors and audiences, the text analogy is ‘outlandish’. ‘Prima facie, the suggestion that the activities of spies, lovers, witch doctors, kings, or mental patients are moves or performances is surely a good deal more plausible than the notion that they are sentences’ (1980: 30). However, Geertz maintains that it is the apparent dissonance between text and social action which gives the former its interpretive force. Following Ricoeur, Geertz argues that the key concept of the text metaphor is inscription, or the fixing of meaning. Speech, like all other forms of social action, is evanescent. However, it can be inscribed in some process of recording, and then its meaning, ‘the said, not the saying’ (1980: 31), is more or less permanently available.Thus, while the physical activity of social action disappears as soon as the action is completed, its implications and ramifications can be fixed and retained by being inscribed.
I want to make two comments about this argument, both of which reveal a closer relation between text and performance than is at first apparent. First, in ‘Blurred Genres’, Geertz places more emphasis on the product of inscription than on the process. The translation of the saying into the said produces cultural objects of diverse kinds (myths, cockfights, jokes, widow burning, etc.) which are the texts to be ‘read’ for the meaning inscribed in them. However, this formulation appears to freeze the text and endow it with a fixed meaning. But inscription is a process, and its importance lies in the fact that it is continuous. Therefore, its products are always provisional and always in the process of change as they are inscribed anew. Text is best construed, then, not as a fixed entity with definite meaning, but as a struggle about who can get what inscribed. This introduces ideas of risk, stake, claim, strategy and competition, and it is thus pre-eminently a political process. The appearance of textual stability derives from the fact that inscription involves the elimination of alternatives, or their inscription in inferior forms (such as in the vernacular instead of in a sacred language, or in memory rather than in writing). By excluding alternatives, what is inscribed becomes authoritative, legitimate and even obligatory; but not permanently so, because established texts are always open to subversion and revision.
My second comment refers to Geertz’s emphasis on the recording and fixing of meaning. This contrasts with the stress, in much writing on ritual, on the doing: certain actions have to be performed; it is their enactment which is significant, and questions of meaning may not always be relevant (Lewis 1980). Consequently, inscription should not be confined to issues of meaning, but should be applied more widely. In ritual, what is often important is the apt performance of what is prescribed and, according to Asad (1993: 62), this involves ‘not symbols to be interpreted but abilities to be acquired according to rules that are sanctioned by those in authority’. A concentration on the decoding of meanings loses sight of the performative aspect of even the most textually bound and prescribed ritual activity.
If inscription is as much about doing as it is about meaning, then everything about a performance is liable to be inscribed, not just its purported meaning. As Bloch (1974) pointed out, much ritual activity is highly formalized and in restricted codes, and therefore cannot carry a rich informational content. An important aspect of the roles participants play, of the manipulation of objects, of the incantation of unintelligible words, of the songs and dances, is the enacting and consolidating of specific, often hierarchical, social relationships. While anthropologists may stress the purported meaning, participants care whether the doing is effective. In this sense the inscription is what is done. The action is inscribed in the bodily movements, habits, LEO HOWE 65 memories and experiences of people at the moment it is being carried out. My point here is that inscription is just as applicable to acts, skills, abilities, operations and procedures, as it is to meanings.
Given these considerations, the text analogy is much closer to performance than one might think, and the notion of inscription is the crucial link between them. Both are political processes and both are concerned with the ephemeral nature of social action. It is tempting to think of performance and inscription as two separate processes, the latter being a re-description of the former in some other medium. This works if it is the meaning of a performance which is being fixed in, for example, writing. However, performance and inscription can be the same process (or two aspects of the same process) when what is at issue is not so much meaning, but the action itself, when the acts inscribe themselves.
It has often been pointed out that rituals are unique because they are situated in particular places and times. They are never pure replications of previous performances or anticipations of future ones (Tambiah 1979: 115). Strathern has argued that acts are not replicable but only substitutable.
Re-enactments are always an ‘innovation and improvisation’ and so ‘every act contains its own performative hazard; it risks its effects’ (Strathern 1989: 292–3). In other words, each act refers to similar acts in previous performances, as well as to itself. So, if ritual is a text, it is one re-written every time it is performed. As a performance proceeds it is simultaneously inscribed in the very acts which make up the performance.
One might respond that this perspective misses the point of the performance approach, which is to highlight the specific ways in which actors conduct themselves on a particular occasion. It might be argued that nothing in the text metaphor, or even in the notion of inscription, can provide the theoretical leverage to analyse the unique, creative and personalized activity which make a particular ritual either compelling or routine. However, while a performance is, in some sense, always a unique and fleeting event, two qualifications are needed. Both stem from the fact that the performance approach tends to treat specific rituals in isolation from others, giving the impression that they are unconnected to previous performances. If text tends to accentuate the generic nature of ritual, performance tends to exaggerate its uniqueness. But if what happens at one performance is linked to what has happened at others, then issues of creativity, spontaneity and uniqueness become relative, not absolute.
The first qualification is that a stress on the performative dimension slights the rule-governed nature of much ritual action. It is no coincidence that the performance analogy has achieved its most convincing results when applied to shamanic curing ceremonies, because in these participants have much freedom of action, and the outcome depends on how a shaman performs and how the audience reacts. In many other rituals, however, the precise implementation of rules is crucial to success, and performativity may not be an issue at all.
The second qualification relates to the effects engendered by ritual action. A concentration on a single performance leaves out of account the way that the effects it generates have a bearing on future performances, and how they become part of the context in which future performances are situated 66 LEO HOWE (Bauman & Briggs 1990: 60, 67, 79). If, as Blackburn (1986: 168) puts it, ‘performance . . . is whatever happens to a text in context’, then the stress on the latter risks missing the contribution the former makes to linking successive contexts into broader sequences of action. Any especially memorable performance (successful or otherwise) may become a bench-mark, not only for evaluating subsequent ones, but also for influencing how they are actually conducted. Actors’ claims to be doing something new, better or different, are perforce framed by the ways things have been done in the past and the results these have brought forth. Performances, however improvised they may appear, are never isolated activities; they are always in relation to or against previous performances which act as remembered precedents.

Risk and performance
Regarding risk, shamanic seances are, so to speak, at the extreme end of the scale. Leaders and audience have licence to invent – indeed, this is almost required – so risk is inevitable. (In the next section I argue that risk may also be a feature of rituals in which rules have to be followed precisely if disaster is to be avoided.)
Most rituals are staged to achieve an end, so there is always something at stake in performances. Because the outcome cannot be known in advance, success and failure (however these may be measured: instrumentally, aesthetically, evocatively, morally, etc.) are contingent. Ritual is therefore inherently risky. Rituals are also often dangerous because of the unavoidable contact with powerful and unpredictable forces. Indeed, contact with such forces is usually the point, as ritual frequently is concerned to tame, manage or harness these forces. This has been noted many times but is sometimes lost sight of in approaches to ritual that emphasize the intellectual and cognitive aspects of meaning. Given this, it is surprising that many performance theorists have neglected issues of risk and danger.
The analogies between ritual and theatre as performances, developed by Turner, Schechner and others (see note 1), ignore questions of success, failure and risk. Even in Tambiah’s pioneering work on ritual as performance there is no mention of risk; indeed, he opposes the uncertainty and chanciness of play and games to the predictability of ritual (1979: 117–19). There is also no reference to risk in recent books on ritual by Bell (1992; 1997), de Coppet (1992) and Humphrey and Laidlaw (1994).The only explicit discussions I have found concerning risk in ritual performance is in the work of Schieffelin (1996; 1998), though it is implicit in the work of Atkinson (1989), and both scholars write about shamanic seances.

Atkinson’s analysis of shamanic seances amongst the Wana of central Sulawesi is couched within a performance framework.Wana become ill when vital forces leave the body and wander far off in the wild. A cure is effected by a shaman retrieving these forces. The shaman cannot do this on his own but must use spirit helpers which he has obtained, appropriately enough, by himself travelling in the forest. In a curing performance the shaman goes on a mystical journey with these helpers to visit powerful deities in the sky to ask for the return of the missing vital elements of the patient.These are usually surrendered on condition that the deity is later given a feast by the patient LEO HOWE 67 and his or her family. Such forms of reciprocity are integral to the performance and take many forms. During the shaman’s journey he engages both his spirit helpers and the deity in dialogue. He also addresses those assembled at the seance, but in cryptic language. The spirits and deity ask for many different things and the shaman passes on these requests to the audience, who are expected to produce them immediately. However, because these requests are couched in unusual language, the audience has to work out what they are. If they have problems, the shaman may give further clues. If they still fail to figure out the request or riddle, the shaman’s life is placed in danger. In a sense, to save the patient, the shaman stakes his own life by depending on his audience to solve the cryptic requests. If the audience does not hear the shaman (seances being very noisy), if they cannot decipher the clues, if they get bored and do not heed a request, any of which may be due to the shaman’s incompetent performance, the shaman’s soul becomes pained and may leave his body, in which case he becomes a patient. So the failure of an audience to support the shaman can be fatal. Successful shamans are created by their audiences who construct a social consensus to follow them (cf. Lévi-Strauss 1963: 180).
Schieffelin’s account of the failure of a shaman during a seance among the Kaluli of Papua New Guinea discusses other issues. The seance involved two shamans, an established performer and a younger aspirant.While the latter succeeded and enhanced his reputation, the former failed for a variety of reasons. He was unable to convince the audience that he was being effective. He misjudged the mood of those assembled and failed to adjust his performance in light of the responses of the audience. ‘Consequently his timing was awkward, and he seemed more intent in driving forward his own agenda than synchronizing audience rapport’ (Schieffelin 1996: 78). He lost credibility early on, and once off track the audience never allowed him to get back. As Schieffelin points out, trust is essential for the creation of credible social identities, which then sustain a collective definition of the situation and which in turn engender a successful performance. This makes performance ‘fundamentally interactive and inherently risky’ (1996: 62, 60).
Those present at a shamanic performance may be glossed as passive watchers and listeners, that is, as an audience. But to the extent that they remain nothing more than this the seance is bound to fail. The shaman’s task is to constitute these people into an active presence with which he strives to forge a dynamic relation. The shaman keeps the audience active and interested by providing difficult, confusing and enigmatic commands and riddles that engage the audience. Because life is at stake, both the patient’s and the shaman’s, the risks are high, and much depends, not on the meaning of symbols, but on the performative competence of the shaman: on his ability to attune himself to the mood of those assembled, and to manipulate, cajole and excite them. In such rituals, risk becomes a measure of the importance and value attached to a performance.When a life is at stake a ceremony to effect a cure needs to be challenging, otherwise the life is undervalued.

Extrinsic risk in Balinese ritual
Many other kinds of ritual2 also involve risk, but in different ways. Generally speaking, risks may be either extrinsic or intrinsic to rituals. The former 68 LEO HOWE are risks which accompany the enactment of a ceremony but are not built into the structure of the rite, whereas the latter are integral to the rite itself, part of its very essence. Any particular ritual may involve both kinds of risk.
Anthropologists have probably paid insufficient attention to risk precisely because much ritual action is rule-governed, thus appearing to render ritual free of risk; following the rules produces the desired effects. In this sense, rules insulate the participants from issues of accountability, success and failure. This is a conventional perspective in which ritual is seen as removing or minimizing the risks involved in activities whose outcomes are uncertain (Malinowski 1974). Moreover, if a ritual is just about following rules, then it is a performance only in a trivial sense, since the actors invest little of themselves. However, often the rules are not well known, or there may be disagreement about what they are, or knowing the rules does not necessarily make it easy to implement them.
Incorrect performance, therefore, is one form of extrinsic risk. In Bali, even small ceremonies are not free of this. Balinese devote a large amount of time and a huge range of material resources to ritual. Much of this activity is small scale, with relatively few offerings, conducted within the home (or the rice field), involving only very close kin, and a priest as required. The aim of those organizing the ritual is to have it carried out properly and accurately so that the object of the rite receives the greatest possible benefit: ancestors continue to protect their descendants; long life, health and prosperity are conferred on children; the rice crop is protected from pests and made bounteous; malevolent spirits are placated to prevent them causing mischief; and so forth.
Securing this accuracy is, however, difficult because even small ceremonies are complex. It is crucial that the right amount and kind of offerings are prepared, that these are constructed from materials which are unsoiled by prior usage or theft, that the components of the offering are arranged in a specific way, that the offerings are made whilst one’s thoughts are pure and calm, that the offerings are given ungrudgingly, and that the priest’s prayers and other utterances are correct. Mistakes vitiating the rite are clearly easy to make and may eventuate in supernatural sanctions. In acknowledgement of such risks, rituals often include offerings that express the performer’s sincerity that every attempt has been made to carry them out correctly. They may also include an offering of a bag of money whose sole purpose is to redeem errors and omissions that may have occurred inadvertently. Larger ceremonies involving vast quantities of offerings escalate these risks. If a household is beset with misfortune it is not unusual for a medium to reveal that ancestral spirits have been angered or offended by ritual omissions, inaccuracies or other mistakes. Families are often burdened with extra ritual activity to compensate for past ritual deficiencies.
If all Balinese rituals carry the risk of incorrect performance, other rituals involve additional uncertainties. One which is particularly acute concerns the fact that many ritual performances entail gambling one’s status. In large part rivalry is confined to the major life-crisis ceremonies of tooth-filing, marriage and cremation. Such rituals have a dual aim (Tambiah 1979: 115). One, as mentioned above, is to procure long life and prosperity for the person undergoing the rite, to usher the soul safely into the next world, or to please the LEO HOWE 69 ancestral spirits. This aim exists irrespective of the material level at which the ritual is performed. However, the way that a household decides to celebrate a ceremony inevitably makes a statement about its status, wealth, character or civic responsibility, because others pervasively interpret performances in such terms. A household may assert a ‘caste’ status which others dispute by using ritual items which are the prerogative of a higher group, or it may proclaim growing affluence and importance by inviting many and prominent guests. Conversely, it may limit expenditure on the grounds that ritual inflation is inimical to community harmony.
Almost everything that occurs at a ceremony can take on a competitive edge: the number and importance of guests; the cars in which they arrive; the number of animals slaughtered; the type, quantity and aesthetic quality of offerings; the food served; the excitement the gathering generates. After such ceremonies there is often a discussion about how well it went, and whether it was better or worse than the equivalent ceremony performed by a rival household. The immediate motivation for this competition is the fear Balinese have of being shamed (majengah, a potent combination of envy, anger and embarrassment) at others’ success, and the concomitant drive to do better next time.
Mechanisms exists to moderate this competition, only a few of which can be mentioned. The cultural disapproval of arrogance and boasting, and injunctions to be humble and modest, impose some restraint, as does the fact that ceremonies should ideally be performed at a level commensurate with a household’s caste status or financial ability. Additionally, some village associations have instituted regulations specifying the quantity and quality of the foods to be served, and guidelines on the appropriate numbers of guests to be invited (Warren 1993: 157). Villagers can complain to officials if a household exceeds these restrictions. Moreover, staging a lavish ceremony requires a large amount of unpaid labour, which creates punitive reciprocal labour obligations.
Decisions on exactly how to mount a major ceremony are therefore fraught with risks. A lavish ceremony enhances a family’s prestige and may please the ancestors, but the labour debts incurred generate problems for the future. If the household is too generous, poorer villagers complain. But if the host family is too stingy, helpers are alienated and may disrupt the proceedings.
A performance exceeding conventional practice may also result in supernatural sanctions on the household. Poffenberger and Zurbuchen (1980:
125) recount an instructive story from a village in eastern Bali, where a family conducted an extravagant cremation in an attempt to improve its status in the village. On the night after the burning they were woken by strange sounds, and when they investigated they saw the ‘departed relative, his mouth stuffed with the elaborate offerings, moaning that he was too heavy . . . to enter heaven’. This shows how the risks involved in status competition are transformed into risks concerning the ostensible aim of the ceremony: the punishment for over-ambitious status claims was borne by the deceased.
It is, therefore, extremely difficult to balance the desire to assert status, the requirement to please ancestors and descent group deities, and the need to 70 LEO HOWE maintain harmony with other villagers. The various and sometimes contradictory considerations that must be taken into account when making strategic decisions about how to mount such ceremonies mean that carrying out a ritual can never be a matter of merely following established rules. In that sense, all of these ceremonies are performances that entail both stakes and risks. The risks can be minimized or increased, but they are rarely absent.
A third form of extrinsic risk warrants mention. There are certain kinds of ritual and dramatic performances in Bali which attract witches who try to disrupt the action. Witches are present at ceremonies surrounding death and cremation because they want to devour the body of the dead. Balinese paintings of cremations and other rituals often portray demons and witches at the heart of the action, attracted by the flesh of the dead or the vulnerability of the living (H. Geertz 1995). Similarly, people undergoing tooth-filing are thought to be in danger because they are in a very exposed position. Lying down flat on their backs with their mouths propped open by two sticks of sugar cane, surrounded by many people and in pain, they are relatively defenceless to attacks by witches or sorcerers who may introduce poison into their bodies through the mouth; the priest or smith doing the filing may in reality be a witch. At some tooth-filings I have attended in the village of Corong, in south Bali, host families employ the services of a man credited with sufficient supernatural power to be able to distract the witches, on the basis that witches cannot resist testing their strength against powerful opponents.

Intrinsic risk in Balinese ritual
Large ceremonies, such as temple festivals, major life-crisis rites and so forth, are not single events, but consist of many separate rituals. Several rituals are staged to placate malevolent spirits so that the main event may proceed unhindered. The rationale is that the gods and ancestors called down by the priest into the ritual’s arena are always accompanied by retinues of minor deities and spirits, some of which are highly capricious, selfish and greedy, and they may possess the unwary and the vulnerable. These spirits need to be dealt with appropriately lest they cause mischief.
However, these spirits are not always negative presences, since they can be enlisted to give people extra strength when, for example, they have to lift the very heavy cremation tower which carries the corpse to the burning ground. The rhythmic music, shouting and boisterous activity attract the spirits and encourage them to help. But there is always a fine balance between assistance and hindrance, making their presence problematical and hazardous. Calling down the gods entails the invasion of potentially uncontrollable spirits, and various ritual procedures are need to constrain them.
Other kinds of performance intentionally involve the highly dangerous presence of witches and other powerful forces. One example is the dancedrama performance known as Calonarang whose principal characters are Rangda, the hideous, death-dealing arch-witch of Balinese mythology, and her adversary the Barong, which looks like a Chinese dragon (Belo 1949). The drama is usually put on in times of trouble for the village or as part of the LEO HOWE 71 celebrations at the annual festival of the ‘death’ temple. It is often performed near the graveyard. It begins quietly with much joking and slapstick, but around midnight, when the anticipation and tension in the audience is high, Rangda erupts into the arena causing mayhem and fear. Rangda’s acolytes possess the Barong’s supporters, who try to stab themselves with daggers while the Barong tries to safeguard them. Sometimes Rangda shoots off into the crowd scattering it in all directions, and people are genuinely afraid. The performance ends in a stand-off between Rangda and Barong (Geertz 1973b:
114–17), but there is always the fear that the dangerous magical forces released may get out of control.
A second example of rituals that intentionally involve the risky presence of powerful forces is the performance of specific and highly charged plays in the shadow theatre repertoire, especially those including climactic scenes from the final battle between the Pandawas and the Korawas (as told in the Mahabharata), in which kings and princes meet their deaths. The Calonarang can also be put on as a shadow play and is equally dangerous. These plays are so magically potent that only puppeteers with great inner strength are thought capable of controlling the forces let loose.
What is characteristic of such performances is the production of concentrations of power which are very dangerous because the power is morally neutral. Once conjured up, this power is difficult to contain as it threatens to overflow the boundaries that ritual procedures mark out to control it. I want to elaborate this theme by discussing a type of ritual in which risk is prominent. Shamanic seances are risky because the shaman’s life is placed in danger, but there are also rituals which place social order in jeopardy, and consequently these are high-risk performances. Many rituals in Bali, particularly complex collective ones, have risk built into their very structures. One intention of such rituals is to test the capacity to rule of kings, lords and other leaders. By first releasing forces which have the potential to escape from control, leaders demonstrate their power by bringing them under control. Such rituals are efficacious only if the forces unleashed are potentially dangerous, and they are successful if these forces are ultimately managed. Success is therefore never a foregone conclusion, and failure is always a possibility. (A particularly good example of this kind of ritual is the Swazi Incwala ceremony; Kuper 1947.)
The ritual I describe took place in Klungkung, in east Bali, in 1842. It was a post-cremation ceremony for a king who died in battle around 1809. Its existence is known through a variety of both Balinese and Dutch written sources. Perhaps the most important Balinese text, which describes the action from the point of view of a commoner, has already been brilliantly analysed by Vickers (1991). What makes this ritual stand out is that its centrepiece was the killing of a rhinoceros. This was a gift demanded of the Dutch by the reigning king of Klungkung in return for permission to open a trading post, and the beast was shipped over from Java.
Ceremonies of this magnitude involve the slaughter of large numbers of different animals. Normally a water buffalo would be the largest animal to be sacrificed. However, with the rhinoceros, said to equal three water buffalo in 72 LEO HOWE size (Vickers 1991: 90), the king of Klungkung had an opportunity to assert both his control over the realm and his equality with the Dutch. But the rhinoceros is also a threat because no one knows how to kill it. There is anxiety both about their ability to despatch it effectively, which would reveal their incompetence and weakness relative to the Dutch, and that many humans may die in the attempt. In the text the threat is partly averted by drawing on a complex discourse involving the strength and firmness of the king, and by extension his realm, whose power can defeat the almost invulnerable animal. It is also averted by a notion that talk and trickery can overcome the power of the animal. A method of killing the rhinoceros is found (using a corral and pit into which the animal falls), but the worry remains that this is still a great risk, and it is discussed in terms of betting against different options (Vickers 1991: 99).
Apart from the problem of killing the rhinoceros, there are other sources of risk and threat that need to be surmounted, and Vickers (1991: 95) makes abundantly clear that the king’s position is the stake in the ritual performance. Success in executing all phases of the ritual enhances his position, while failure to control events leads to defeat. The ritual is a gamble and the king must be game to undertake it.
The most significant threat is the production of chaos arising from the random mingling of all the categories of people which made up precolonial Bali. At a major royal and state ritual the king has to ensure that all his subjects and vassals are present, since absence indicates independence and enmity. The town is thus flooded with commoner subjects, lords and nobles of different ranks, and various kinds of priest who officiate. The crowded, lively, energetic and bustling atmosphere this creates is known as ramé and is something most Balinese enjoy. If markets, cockfights, drama performances and rituals are not ramé, it is both a crushing disappointment and a negative comment on the competence and status of those responsible. If one cannot attract large crowds it is because one is not potent enough to ensure their attendance.
However, there is another side to ramé, since it also has connotations of licentiousness, immorality and chaos (Vickers 1991: 94). In crowded situations, quarrels and fights break out, lewd behaviour occurs, theft is likely, and morality is forgotten. Large sections of the text are devoted to descriptions of impending chaos and to the behaviour and state of mind of those in the crowds. The people get dirty and use foul speech, men sexually assault women, and the noise from the shouting, music and rifle firing is like thunder or waves crashing. The text even uses images of war and destruction to represent these scenes, possibly implying that this could be the end result for the realm if the king cannot keep control of events. As Vickers points out, it is essential that the ritual be ‘crowded, full and ramé’ but this inevitably entails ‘violence, theft and devastation’ (1991: 96–7). It is part of the very essence of the ritual to produce this situation, for from it flows the energy which, if harnessed and controlled, brings the ritual closer to a successful conclusion.
The emotions unleashed in the crowd by the frenetic activity also swing ambiguously between positive and negative poles. At one time, people are LEO HOWE 73 happy and pleased as they watch the drama performances, or are calmed by the quiet and orderly priestly rituals. At another, they are fearful and confused when the cannons go off, and then angry when they rush to seize the offerings dedicated to the malevolent spirits but cannot get what they want. The mixing of the castes, of men and women, of lords and commoners, and the taking of opium and drink, creates conditions in which people indulge their vices. But such indulgence is opposite to the restraint of normal sociality and so makes them sad and confused. While the huge gathering in itself is significant, even more important is that the different ranks of Balinese society come into much greater proximity than is normally allowed. Participants of different status almost become equal as commoners forget the forms of separation, distancing and deference ideally due to their superiors. Even noble women get caught up in the push and shove (Vickers 1991: 103). The energy released by the mixing of the king’s subjects at times threatens to overwhelm the power of the king to control it. If the purpose of the ritual is to create an ordered realm and to establish proper relations with the gods and ancestors, this can only be done by first releasing the forces of chaos which threaten to undermine the very basis of that order. I leave the final words to Vickers (1991: 103):
When the king holds the ritual he is bringing on the threat of chaos by crowding everyone together. He provokes a crisis by testing whether he has the resources, in terms of subjects and support from the [priests] and the other lords, to succeed with the ritual. In the case of this [ritual] the gamble is great, because the king has obtained a live rhinoceros, an act of daring which extends the interaction to the foreign forces of the Dutch.That is why the killing of the rhinoceros must be carried out effectively, otherwise the demonstrated inferiority will result in a lack of followers and support. The ritual holder must be, in the words of the poem, purun metoh, ‘game to make a stake’. This ceremony was performed 150 years ago and contained an extraordinary event. But contemporary collective rituals are structured by similar themes. In modern Bali the holding of annual festivals (odalan) in important temples, lasting up to nine days, recreate in a diluted way the scenes of ramé, chaos and happiness prevailing in the 1842 ritual. Not only is the whole village mobilized, but traders, visitors, gamblers, tourists and even the regional Balinese media descend on the village to watch the drama performances, attend the cockfights, sell food, clothes, medicines and toys, and attend the gods that come from other villages. Such events are the highlight of the year and are eagerly awaited. Cockfights can attract several thousand people, some of whom come from Java, and the night-time entertainments act as a magnet for the youth from surrounding villages. While people bring offerings and pray, others are eating, drinking and gambling. At these events boys and girls flirt, jealousies flare, insults are traded and fights break out.While in one sense all this is incidental to the main event of making obeisance to the temple gods, in another it is central, because without the crowd the ceremony would be quiet (sepi) and as such a failure.
Such ceremonies are organized and managed by the village’s foremost priests and the leading village officials. The latter are often direct descendants of the village’s pre-colonial high-caste ruling household. These people are still credited with inherent potency (sakti). They are thought capable 74 LEO HOWE of both attracting the large crowds that make ceremonies enjoyable and successful events, and of controlling them. At temple festivals and other major rituals the presence of these powerful figures is said to inhibit the antics of the crowd. Their charisma and power creates rame, but equally it contains the chaos that potentially the crowd can unleash. In a crowd there is always the possibility of points of power outside the control of the ritual’s managers, but if the latter are really potent these can be minimized. These ceremonies must take place so as to maintain proper relations with the village gods, but they are also daring ventures that test the authority and control of leaders.
The discussion cannot be left here because it is necessary to ask if such rituals ever actually end in failure, the point being that risk may be represented as significant, when in fact it is minimal, in order to convey the illusion that the leader, in controlling the risk, is indeed powerful. I do not doubt that this happens, but it remains the case that important life-crisis rites held by householders are sometimes seriously disrupted by participants who are angry with the hosts. Thus, in Corong village, what was hoped by one family to be a very large and prestigious marriage ceremony turned out a failure when many villagers failed to arrive on the morning of the wedding day to help with the preparations. They stayed away in retaliation against the groom who, having got into a quarrel with a man whose house was regularly used for gambling, asked the police to raid it, and thus became an object of hate. Many cremations are disrupted if villagers are insulted by arrogant and unjuztified status claims or if they disliked how the deceased acted during life. Cremation towers and offerings may be destroyed and the corpse physically abused (Connor 1979). These rituals are failures, the hosts suffer ridicule and ignominy, and the journey of the departed to the other world may be jeopardized. A very spectacular example of how magical forces may run out of control concerned a massive, all-Bali ritual held at the Besakih temple complex in 1963, organized by both the governor of Bali and the Déwa Agung, the direct descendant of the last king of Klungkung. As the ritual got under way, the volcanic mountain on which the temple is situated exploded. Balinese disagreed as to whether the ritual was a success or a failure, since although the temple was relatively unscathed enormous damage was caused to the surrounding countryside and many people lost their lives.

What does risk tell us about ritual? One thing to note is that a focus on risk reinforces the idea that it is the inner intention of the ritual that is important. Rituals, however, may have several aims. The ostensible reason for the killing of the rhinoceros was to usher the soul of the departed king into heaven. At this level success is simply about correct performance. But a second aim complicates the issue. This concerns how the ritual puts into motion certain kinds of forces which must then be channelled and controlled by the ritual’s managers. Failure in this aspect of the ritual may have adverse consequences not only for the prestige of the living king but also for the fate of the dead one. This second aim is therefore integral to the ritual as a whole and not merely an adjunct to it. This has implications for Humphrey and LEO HOWE 75 Laidlaw’s (1994: 94) argument that the intentions of the participants are irrelevant to the identification of an act as a species of ritualized action. In this Balinese ritual it does matter what the living king’s intentions are, for these determine how great a risk he is willing to take, which in turn influences many aspects of what happens during the event.
Highlighting effectiveness rather than meaning concentrates attention on issues of success and failure, and winners and losers. In this context it is worth recalling that Lévi-Strauss considered ritual to be the opposite of games. Games, he said, are disjunctive because, through the contingent events of intention, chance and talent, they transpose an initial symmetry of the players into a distinction between winners and losers at the end. Ritual, by contrast, conjoins because even though asymmetry is postulated in advance, it ends by ‘making all the participants pass to the winning side’ (1972: 32). However, this interpretation does not accord with the unpredictability of some rituals, in which risk renders ritual actions and events contingent and which, like games, may finish with winners and losers. The distinction between ritual and game is not nearly as sharp as Lévi-Strauss conceives it.
In view of this, risk prompts a new metaphor for ritual. Ritual has been analysed as communication, drama and performance, but it is surely important to see some ritual action as test, trial, examination and contest. Although the performance metaphor is enlightening, what it presently lacks, at least in some of the ways it has been used, is the sense that, because something is at stake, it is a gamble, and that the ritual’s managers must dare to conduct it. Such a gamble is only meaningful if one may lose or win something of value; the more important the value, the larger the bet, and the more significant the victory or failure.
Speaking of ritual as a test in which core cultural values are at stake recalls Appadurai’s discussion of tournaments of value (1986: 21). If a state ritual is a test of the king’s power to control the forces unleashed by the performance of the ceremony, it is also a contest, implicit or otherwise, with other kings and leaders, because the king’s failure to control ritual events indicates that he does not possess the power to be a king. This may entail his subjects transferring their allegiances to another centre of power, a process often invoked by Balinese to account for the rise and fall of leaders of diverse kinds. Appadurai limits tournaments of value to ‘staged’ and dramatic forms of economic competition such as the kula and the art auction. Harrison has generalized the idea by arguing that what is really important is that ‘the competition is for the control of prestige values’ (1992: 226). In Balinese royal rituals there are many, varied prestige values: rank, ritual prerogatives, positions of authority, the control of magical weapons and heirlooms, and privileged access to divinity.While in part the right to these depends on the power credited to the king by virtue of legitimate descent, they also accrue to those who can accumulate this power in other ways. Such power is finite and scarce; as it concentrates in one site it flows away from others. Once in possession of such power it must continually be renewed and demonstrated (by ritual action, warfare, and the protection of clients) or it slips away. Royal rituals, village temple festivals and even the major life-crisis rites of householders are forms of competition in which these prestige values are re-distributed.

According to Geertz (1973c: 418), the Balinese cockfight is comparable to warfare and political contest. Since collective rituals are also forms of political contest, sometimes described using images of warfare, violence and destruction, then Balinese collective rituals can be compared to cockfights. In Balinese cockfights the larger the centre bet the less it is about money and the more it is about status; in Baudrillard’s terms the cockfight is about the ‘production and exchange of sign values’ (1981: 117).The centre bet is always even money and the two birds in the fight are matched as closely as possible.The outcome is therefore highly unpredictable and the bet entails high risk. As Geertz (1973c: 430) shows, in high-stake fights the underdog won as often as the favourite. Who, then, is the opponent in the ritual? In one sense an implicit opponent is any rival king, village leader or householder whom one wants to outsmart. But in another sense the opponent is oneself, or one’s own identity and status as leader. In the collective ritual the king pits himself against the forces he himself releases, recalling Strathern’s point that the performance of any act ‘risks its [own] effects’. Now the king is constrained to perform the ceremony at the highest level possible, otherwise he is cowardly and mean, and thus not fit to be a king. On analogy with the cockfight, as the stakes in the ritual grow larger, in terms of the resources mobilized by the king, the opposing forces which these unleash also progressively increase to a level at which they are on a par with the power of the king to contain them. In this sense such a ritual is an even money bet.

In preparing this article for publication I am particularly indebted to Stephen Hugh-Jones for providing constant encouragement and incisive comment. The paper was presented at a senior seminar in Cambridge, and I take the opportunity to thank the participants, especially Malcolm Ruel, for some very helpful suggestions.
1Performance theory finds its inspiration in diverse sources: Goffman’s (1959) ideas about the ‘presentation of self ’ in everyday life, Austin’s (1962) notion of the illocutionary or performative aspects of language, analogies connected to theatre, and so forth. Consequently, performance appears to mean rather different things to different people. Turner’s (1974) view of performance as social drama has themes in common with some of those in Tambiah’s (1979) work, but Turner’s later writings (1990) on the similarities between ritual and theatre, carried on and developed by Schechner (1994), seems less central.Whereas Bell (1992: 43) argues that a prominent feature of performance theory is that ritual communicates, Schieffelin (1998) argues that it is precisely against this perspective that performance theory makes its stand. If there is a common theme it is that performance does not merely refer to or talk about existing situations, but ‘does something in the world’ (Laderman & Roseman 1996: 16). It is this version of performance that I find most helpful and persuasive.

2 I am not interested in producing a general theory of the relation between ritual and risk, because there are so many different forms of ritual and so many different things accomplished through ritual action. My intention is to point out some of the ways risk may be an important element in some rituals, particularly the more complex and collective ones, and how a recognition of this may have specific implications for the analysis of these particular rituals.

No comments: