Balinese communities at the time of the early royal edicts appear to have been organised around a council of seniority-ranked elders, who were paired and grouped into ceremonial moieties. This pattern of precedence-based village organisation is still a defining characteristic of contemporary Bali Aga communities (Reuter 2002b). The edicts refer to the localised communities represented by these elders as banua (Old Balinese) or wanua (Old Javanese), rather than desa. Most likely the meaning of the term banua in Bali at that time referred to something more like a ‘village’ than a ‘domain’. It is possible that the newer term for village, desa, was introduced later in order to distinguish between a single village community and territory (desa) and a very old settlement (banua), which once had a very large ‘territory’ on which a whole cluster of newer hamlets were established as the population grew, naturally or by immigration. When these newer settlements became independent villages, the original village lost direct control over the land that fell within their respective ‘village territories’ but retained a special status as ritual overlords in the domain as a whole. In historical perspective, therefore, the term banua in Bali has consistently carried the meaning of ‘a unit of land or territory shared by a community’, while the size and complexity of the unit of land and community thereby designated appears to have changed over time in the direction of greater complexity. This trend is evident in the contemporary situation.
Bali’s central highlands and northern coast today are home to about 100 Bali Aga villages. These villages form at least 13 distinct banua or ‘ritual domains’ of varying size, clustered around regionally important temples (pura). These ritual alliance clusters are depicted in Figure 1.
Each domain is a bounded space in that all of the villages within certain boundaries normally partake in the one encompassing ceremonial order. While domains are named after and are socially oriented to a centre, the boundaries, while not emphasised very much in ritual, are quite clear. This is because the participating villages have definite boundaries nowadays, and probably always did. The royal inscriptions of classical Bali already mention streams and ridges as village boundaries. In any case, banua networks, figuratively speaking, do not have any gaps or outliers.
It is a common feature of most of the larger domains, however, that they incorporate smaller subsidiary domains, each with its own regional temple of lesser status. Such complex layered networks are nevertheless conceived as single banua with a common origin rather than overlapping, historically unrelated domains. There are a few cases where two distinct domains do share part of the same network of villages. However, this usually means that one of the two was formerly a sub-domain of the other, has attempted to break away from it to become an independent domain in its own right, and has had only partial success in doing so. [4]
From the perspective of local participants, the term banua designates a bounded ‘spiritual territory’. More specifically, a banua is conceived as a large and clearly defined area of agricultural land whose fertility, along with the prosperity (rahayu) of its human occupants, depends on the ancestral deities enshrined at the domain’s paramount temple or pura banua. Likewise, the smaller divisions of land within a domain are under the protection of the lesser and more local ancestor deities of ‘village temples’ or pura desa. These lesser village deities are often identified in narratives of origin as the children or children-in-law of the domain’s paramount founder deity. The congregation of a pura banua is thus a network of otherwise autonomous villages, just as the congregation of a pura desa is a group of otherwise autonomous households (umah, ‘house’, or kuren, ‘hearth’). This raises the question of why these independent villages would find it desirable to continue to recognise the precedence status of the origin village.
A banua is a self-evident fact to its participants; a paramount natural (sekala) and supernatural (niskala) reality with a social character. The force of this social reality is vested in the agency of post-human or non-human agents—the local ancestral and non-local Hindu deities. The relationship with ancestral and other deities is enacted dramatically in a ritual setting. A ritual domain is thus experienced in ritual as a given and is described officially as non-negotiable. Simultaneously, from an ethnographic perspective, the spiritual territory or domain of a temple is a socially constructed reality, a man-made and negotiated boundary and a set of internal relations inscribed on the land by the narrative and ritual performances of a temple’s human congregation.
I would like to reconcile these two perspectives. Participants, speaking from a position of individual lived experience, are quite correct in their perception of the banua as a non-human facticity. First of all, the land itself is part of a self-evident material reality that predates human occupation, although land is also transformed by human beings for the purpose of agriculture or house-building; land is thereafter classified, respectively, as tanah ladang or tanah pakarangan. It is also self-evident that nearly all land is visibly occupied and claimed by someone—agricultural land by means of regular cultivation and the construction of garden huts and fences, and house land by means of regular habitation. Given that particular people will argue a claim to the right to occupy particular pieces of farm and house land, this land cannot easily be taken away from them. The banua and its subsidiary desa therefore are not only a material given, but also a ‘supernatural’ or rather ‘post-human’ facticity, in the sense that contemporary society is encountered as a reality that was created by now-deceased and invisible others in the past, namely, by one’s own historical predecessors or ‘ancestors’.
From the point of view of an interactionist social theory, this still may not seem like an entirely satisfying explanation of the status quo in a banua. Although a particular land tenure arrangement, once it has been established, may constitute a self-evident social fact, the more important questions are how such arrangements were established in the first place, and by what means they are maintained or changed. Indeed, local participants are immensely interested in the same questions, though their intuitive answers may be phrased in an unfamiliar idiom.
The Bali Aga are always keen to discuss how relationships between particular people and places came to be. Each domain has its origin narratives, and so does every village within it. Often enough, these narratives exist in different and competing versions. Ritual performances aimed at maintaining good relations with the ‘supernatural’, which are regular, frequent and elaborate in the highlands of Bali, serve to commemorate this origin history. As are the narratives that inform them, these rituals are subject to variation and can also become a means of contesting origin-based status claims. For the Bali Aga, the answer to the question of how a village or domain came to be in the hands of particular people is, ‘history’, and as for the means by which this claim to the land is maintained or challenged, the answer is ‘commemorating [or forgetting] particular ideas about history in a ritual context’. If I am to convey accurately what a banua is or can be, my two principle tasks therefore are to do justice to the regulated practices by which the Bali Aga conceive of the present as a self-evident product of the past, while at the same time acknowledging that the present is also the not-so-self-evident product of a strategic remembering or reconstruction of the past.
Let me begin by looking at regulated practice. Usually, when it is not in (com-) motion, a banua ‘rests’, so to speak, on a shared conviction among participants that the founders of all the villages in the domain originated from a single source village, namely, the village where the domain’s paramount temple is now located. Origin narratives retell how these founders left the source village, cleared land somewhere nearby to establish gardens and hamlets, and how these hamlets eventually became independent villages with their own local temples. In these temples, the branch village founders are now enshrined to be venerated by their many descendants. These deified founders still own the land because death is not the end of life, of rights and obligations, but simply the beginning of a gradual transformation from a human to a post-human state of being.
There are different degrees of post-humanisation or sacredness on a temporal scale of precedence. This means that the ancestral founders of a branch village, even though they are local deities in their own right, still have ritual obligations towards their own ancestors, who are enshrined in the origin temples of the domain’s source village. (Their obligations in fact mirror those of members of a small kin group around a branch house temple towards the older, larger temple of their clan or local clan segment.) In practical reality, of course, the ritual obligations from junior to senior ancestors have to be met by the living descendants of the junior ancestors in the branch villages. As a consequence of this remembering of ancestral relations, the pura desa of the source village is transformed into a regional pura banua, which is thereafter maintained collectively from regular contributions (peturunan) paid by every household in every one of the branch villages.
Figure 2 depicts the logical pattern of these relationships in schematic form. The diagram is based loosely on the internal organisation of the very large banua of the regional temple Pura Pucak Penulisan, which has four major sub-domains centred on the temples of the ancient villages of Bantang, Selulung, Kintamani and Sukawana (Reuter 1998). Each of these secondary ritual centres has its own following of subsidiary villages. What the figure does not show and as we shall see later, the sub-domain of Kintamani, for example, contains yet another, smaller
sub-domain consisting of a cluster of villages around the small regional temple Pura Tebenan. In short, the logic of the status relationship that gives an origin village ritual precedence over its branch villages can be applied recursively wherever people recognise an extended sequence of historical migrations and settlement foundation events. The length of such a historical sequence of migrations may seem unlimited theoretically, but in ritual practice connections of precedence are restricted within the more narrow confines of a socially recognised history. My research has shown that the depth of ancestral origin and migration connections in a domain usually amounts to only two layers (origin and branch), and rarely exceeds three layers, as depicted in Figure 2 (see Reuter 2002a). Four layers are very rare, and I have no example of a five-tiered domain.[5]The pattern shown in Figure 2 may look much like a network of ancestor temples, and in some ways it is. However, what we are dealing with here is not a simple kinship system nor a complex one, but a Lévi-Straussian house society in a ritualised modality (see Reuter 2002b). The organising principle is a sequence of inhabited places represented symbolically by sacred ancestor houses or ‘temples’. These sacred houses are ranked in order of their foundation. In short, the contemporary order of ritual obligations maps and retraces a path of origin through a socio-historically inscribed landscape.
The emphasis in this system of social relations lies on a series of place names within a landscape rather than on personal names in a genealogy. Villagers in the highlands are unable to recount a genealogy of named predecessors from themselves all the way back to either the clan or village founders, let alone the domain founders. People’s genealogical memory usually spans no more than four generations, and rights to an inheritance as well as duties to ritually post-humanise recently deceased predecessors are serious issues only within that range. Beyond four generations, what serves as a means of connecting people within kin groups are once again links between sacred places, rather than named ancestors. The connections in this case are between older and newer ancestral ‘origin temples’ or sanggah kamulan, the private temples of groups of agnates whose individual houses are usually located within a single compound. The older temples in a sequence of source and branch sanggah become sanggah gede (great sanggah) and the oldest sanggah within a clan is elevated eventually to the status of a so-called pura kawitan (‘temple of the source, beginning, or trunk’).[6] One reason for this reliance on place markers and forgetfulness of personal ancestor names was that despite their agnatic ideology and patrilocal residence, members of Balinese kin groups in fact often chose to incorporate male outsiders through uxurilocal marriages (nyentana) and strategic adoptions.
This openness of place-based categories is even more evident at a village or domain level. Villages may afford a special status to members of the houses of the founding emigrants who originated from an older source village. In most villages, however, there are also newcomer houses founded by more recent immigrants from elsewhere. At a domain level, flexibility about ancestry is manifest in the fact that sometimes entire villages are said to have been founded by immigrants who were not at all related to the source of the domain by descent. Their relationship to the source was established on the basis of their use of domain land, which incurred a ritual debt that had to be paid in accordance with the customs of the domain. In this way, newcomer settlements were culturally and ritually incorporated to such an extent that they soon became difficult to distinguish from genuine branch villages that were founded by emigrants from the source village (Reuter 1999).
On whatever level of organisation they may operate, models of human relations that are based on ‘place’ rather than ‘name’ have the advantage of greater inclusiveness, of being able to accommodate a diversity of people within its folds. This is all very well, but encounters with newcomers also need to be regulated. People will tend to protect their land and other resources from new claimants unless they have a good reason for sharing. This problem returns me to my second task, that is, to analyse processes of memorising and reconstructing the past which serve as a strategy for negotiating and contesting the distribution of land and other resources.
I will approach this analytical task first at the village level. The most important aspect of history to be either actively remembered or conveniently forgotten in this context is that certain people are the agnatic descendents of the village founders, or literally, ‘the people of the trunk’ (wedan), while others are not. This distinction is almost universally necessary given that village founders were nearly always joined by unrelated immigrants later on, by ‘newcomers’ or, literally, ‘people of the leaf’ (pendonan). In an agrarian society such newcomers required land and as long as vacant land was still available it was freely given. Indeed, newcomers were rather welcome because together with a share of land, they were also given a share of the fixed cost of celebrating the festivals of the village temples in honour of the founding ancestors and spiritual owner-caretakers of that land. In most highland villages, though not in all, land was thus given to the newcomers outright rather than being leased to them as dependent clients or sharecroppers.
In most cases, the botanic metaphor-based distinction of trunk and leaf people thus has limited significance when it comes to controlling access to land. While members of the villages founder group tended to have larger landholdings overall, newcomers had to be given villages land or sufficient other land to be able to make a living. The twofold social classification of wedan and pendonan is still very important, however, due to a distinction between two types of land, namely ‘village land’ (tanah ayahan desa) and ‘owned land’ (tanah milik). [7] Only the men who farm classificatory village land are represented in the community’s seniority-ranked council of elders and priest-leaders, while others are not. The council (ulu apad) consists of the male heads of eligible households and meets on new and full moons in the village longhouse. Divided into ceremonial moieties, the men sit in order of precedence (depending on when they became married householders) on the massive side beams of the longhouse floor, with the ranks decreasing from the tip (and uphill) to the trunk (downhill) end of the beam (compare with the nili ele councils of Seram, Boulan-Smit, this volume). [8]
Belonging to the people of the village trunk or an early, now fully integrated immigrant group that has been admitted into the village council thus signifies access to political and ritual authority. In short, material resources, in the form of land, were given away for two reasons: to share the material costs of ritual with others and to acquire symbolic resources in relation to these others. These symbolic resources take the form of ritual and political authority over a group of client households which are economically independent.
Contemporary banua have much less relevance for land usage rights than desa. A ritual domain is not at all definable as a landowning corporation, whereas a village can be, or at least formerly could be, classified as such. Prior to Indonesian independence, village land was often owned by residents collectively and in many cases individual plots of land were rotated continuously from recently defunct households of older, retired or widowed couples to the households of newly married couples, irrespective of any relationship by descent between them. After following a shift from subsistence agriculture to cash crops such as coffee, which require long-term labour investments, this rotation system was abandoned in many villages in favour of a de facto inheritance system, whereby the special character of ‘communal land’ became more and more nominal (see also Waelty 1993). In other villages all land has been classified as ‘privately owned’ for as long as anyone can remember, and is passed on by inheritance within kin groups. [9] In both types of villages, the introduction of new legal codes by the Dutch and Indonesian governments has encouraged a shift to more modern notions of landownership, especially from the 1970s onwards. Nevertheless, there is no such thing as individual landownership free of ritual obligations in highland Bali.
There is evidence that the banua was still relevant for maintaining land claims until the 1930s and the desa has remained important until today. Occupants of classificatory village land can be dispossessed if they fail to meet their ritual responsibilities to the desa and, theoretically at least, to the banua of which it is a part. Individual ownership of so-called ‘private land’ (tanah milik pribadi) is also contingent on meeting ritual obligations towards the village (desa) or the hamlet (banjar), or both, as well as social obligations towards one’s extended kin group (at least for as long as the older generation can delay a formal transfer of land certificates to their heirs). Even the non-resident owners of land used for commercial purposes are subject to a ritual levy for village and local domain temples.
This evidence suggests a historical process whereby land as a material resource always remained under the control of relatively small units of social organisation. In addition, as I have argued earlier, there has been a longstanding principle of allowing or even encouraging the alienation of material resources for the sake of gaining symbolic resources, namely in the course of accommodating immigrants and newcomers into a house, village or domain. Indeed, even the more regular mode of banua expansion, through a process of emigration, is posited quite explicitly on a granting of material autonomy to branch villages while the source village retains a privileged access to symbolic resources. In short, the apparent tendency towards a gradual devolution of material resource control to smaller, subsidiary social units is in fact the result of the growth of originally small units of ritual organisation, such as single villages, into larger units, which are concerned increasingly with the distribution and control of symbolic resources rather than land usage rights or other material resources.
The ritual process of a gradual post-humanisation or deification of ancestral founders follows a similar logical pattern, in that gains in spiritual status are accompanied by the loss of material bodies, in successive layers of decreasing density (from kasar to halus). One possible explanation for this historical process of ritual alliance-building might be that alternative coercive means, which would have allowed for a similar expansion of the social order by an accumulation of material power, were never available to any group within Bali Aga society, which has been politically marginal as a whole for centuries.
Within a village, to control symbolic resources means to have ritual status as well as political authority. The people of the trunk who man the village council are not only the ritual leaders, they make and amend local customary regulations, settle disputes, impose sanctions and take political decisions on behalf of the community as a whole. In a domain, by contrast, the source village does not wield any such power in relation to branch villages outside the context of the annual ritual of the domain temple and perhaps in a number of other ritual matters. There is a political link between desa and banua, however, insofar as any claim to village founder status and associated rights is a local historical claim embedded in a wider domain history. The banua as a whole may be said to have a political dimension as well, but the orientation of this politics is toward participants’ relationships with the outside world rather than an internal politics of land rights. Ritual domains have served as a powerful means for organising Bali Aga responses to a wider world beyond the highlands, a world that, I believe, would have long swallowed them up if it had not been for their intricate ritual networks of regional integration and cooperation. [10]
I would now like to retell the origin narrative of a particular banua in order to illustrate that local participants themselves also reflect on this shift in significance in historically expanding place-based social categories, as they are transformed from material into symbolic resource pools. The example I have chosen is a narrative that describes how the village of Tebenan is transformed into the depopulated ritual centre of a sub-domain.
Once a group of subjects approached the king of Bali at his residence near the ancient regional temple of Penulisan to ask leave for establishing a new settlement ‘downhill’ or teben. He granted their request for land, but also warned that disaster was sure to befall them if their village council were to grow beyond a membership of 33 household heads. The new village was called Tebenan to commemorate its origin from upstream Penulisan. Its rice fields were fertile, people’s lives were leisurely, and soon the population grew until the village council had a membership of 200 heads of households. It then happened that a ritual was planned in which a deer was to be sacrificed, offered to the gods, and divided among the council members. But much of the forest had been cleared and deer were hard to find. On the day of the ritual, the hunting party finally returned with a small animal. Once the victim had been sacrificed and divided, the meat portions obtained were fewer than the number of eligible members. Arguments began, a fight broke out, even brothers killed one another, and the survivors were scattered in the eight directions of the compass as leaves blown by a wind. They founded eight new villages: Ulian, Gunung Bau, Bunutin, Langahan, Pausan, Bukhi, Bayung Cerik and Manikliyu, closest to the old Tebenan. The land was plentiful, but their lives were short, their labour hard and fruitless. Then the senior priest-leader (kubayan) of Manikliyu had a dream in which he learned that their lot would improve if only they remembered their common origin and resumed the cult of the village temple of the abandoned Tebenan. He enticed people to search for the foundations of the forgotten temple in the forest, and they uncovered nine [i.e., eight around a centre] sacred xylophone ‘leaves’ (don selunding), indicating that they had found the precise site of the old temple. The former village temple was then rebuilt to become a sub-regional temple. Ever since, the banua of eight villages has gathered there for an annual joint ceremony to honour their deified ancestors and to ask for a blessing upon their crops. Until today there may not be more than 33 elders in the ‘village council’ of Manikliyu. Before these elders take their ritual meal after a new moon meeting, they speak these words: ‘King’s men, insiders, elders of the heart, shoulder, elbow ... [and so forth in order of rank]. Come together, there are more food portions than people!’
The present origin narrative confirms Desa Manikliyu’s role as a ritual leader among the eight villages of this sub-domain. The origin point of the sub-domain is identified as Pura Tebenan. The reference to the higher authority of the ancient kings of Penulisan, as the ultimate and now deified overlords of the land, reflects the fact that all villages in this sub-banua also participate within the larger ritual order around Pura Penulisan.
When their usage in specific social contexts is taken into consideration it becomes evident that the origin narratives of a domain cannot be interpreted as disembodied texts. A structural analysis might reveal their intrinsic logic and the pattern of relations between their conceptual constituents, but even these elements have a contingent or non-intrinsic meaning. As Paul Ricoeur (1981: 217) has argued, narratives of origin bear reference to a cultural world per se but also to particular situations within that world.
Bali Aga origin narratives are rarely written down, they are not recited formally in ritual contexts, nor are they told in a ritual language marked by distinct formal characteristics. Most often origin narratives are retold spontaneously among a group of priests, elders and ordinary participants during the quieter moments of a festival at a banua temple. Their conversation can be described as a ‘discussion’ insofar as there is much room for polite disagreement about the narrative’s content and interpretation. Therefore each narrative must be seen not as a monolithic text but as a tentative consensus reached by particular people at a particular time and place.
At the same time, such narratives also refer to a shared material world and a shared way of life, and thus contain non-ostensive references that bear meaning beyond any of the specific contexts in which the narrative might be retold and discussed. I agree with Ricoeur that origin narratives are not only the focus of a conflict of interpretations but may also depict important existential conflicts or ‘boundary situations’ in the world, which are of general relevance. A compromise approach towards the interpretation of myths of origin may thus be: to hear them, as the spoken discourse of specific authors with specific intentions, and simultaneously to read them, as the textual testimony of a collective world of shared experience and knowledge.
The tale of Pura Tebenan, if it is read rather than heard, illustrates how the social unity of a domain is contingent on the symbolic labour of recollecting and re-enacting a common origin. The path of origin is marked first by a lateral expansion, by the growth of a new branch domain out of the unity of the Penulisan domain. Then a second expansion is caused by the violent rupture of the village unity represented by what is now the Tebenan sub-domain temple. Processes of lateral expansion or sudden fission establish relationships between parts and a whole embodied as a sacred origin point. With each progressive differentiation, the degree of ‘social unity’ is diminished in the more immediate sense because the people involved no longer share the same, finite material resources of a single village. The idea of a social unity, however, is reintroduced simultaneously at a broader, more symbolic level by bringing several villages together in the ritual context of the origin temple and its ceremonial order.
The idea that only a finite number of people can derive a reasonable share from a single sacrifice may be taken as a metaphor, suggesting that in any given place there is a natural limit for sustainable solidarity on a material plane. This limit is often violated. The narrative hints at a golden mean, to be achieved by avoiding a practical condition of scarcity that arises from the togetherness of too many people in one place, and an equally undesirable condition of social isolation, which would arise if one were to define the small world of a single village as the only relevant sphere of social interaction. The problem of negotiating this balance is the archetypal ‘boundary situation’ in the world, which this origin narrative addresses. The proposed solution is to allocate material and symbolic concerns to separately conceived levels of social interaction. Thus, the narrative can be read as a culture-specific commentary on a general dilemma of social cooperation and competition in agrarian communities with finite land and expanding populations. It also illustrates some of the historical processes involved in the evolution of larger, regional social institutions, which in turn may have facilitated the emergence of small states (negara) in South-East Asia and in other parts of the Austronesian world.