It is unclear what, if any forms of socio-spatial organisation larger than desa existed in this part of Bali before or independent of the multiplication and expansion of puri in the 18th century. On the one hand it is reasonable to expect that ‘ritual domains’ along the lines of the Bali Aga banua described by Reuter (this volume) might have existed, but on the other, the evidence of local oral history suggests that the settlement of much of the area coincided with rather than predated puri expansion.

The anthropological record is also somewhat ambiguous on this subject. While the Dutch scholarly colonial orthodoxy that ‘the village forms a closed, self-contained unit’ (Goris 1984: 79) has long since fallen from favour, subsequent writers continued to take for granted the village as the natural unit of analysis.^[12] This focus has been at the expense of recognising modes of organisation beyond and between villages. The existence of such modes of organisation is, however, evident in the literature. This evidence includes:

Despite such widespread evidence, the implications have not been pursued systematically with the exceptions of Lansing’s (1991) work on water temples, Schulte-Nordholt’s (1988a, 1991a,1991b) on pre-colonial state temples, and Reuter’s (1998, 2002a) more recent work on Bali Aga banua. None of these refer to the ordinary villages and temples of south Bali. Recent ethnographic evidence, however, indicates traces of banua-like forms, especially the further one moves uphill from the puri centres of Ubud, Tegallalang and Payangan, towards the more unequivocally Bali Aga areas documented by Reuter. ^[13]

For example, Desa Sebatu, according to local tradition never subject to puri control, is the centre of a network of some five desa linked by reciprocal ritual ties. They are not, however, referred to as a named collective entity. The ritual cycle in these desa is, as in mountain banua, tied to the old lunar calendar (sasih) rather than the Hindu-Javanese one (wuku). On the other hand, they utilise the services of Brahman high priests (pendeta) in some of their rituals, which is evidence of influence from the Majapahit lowlands.

In nearby Pujung (Talepud), Leo Howe (1980: 13-27) reports a similar blend of lowland and highland customs, as well as local oral traditions including a version of the Rsi Markandeya story, which link Talepud to nearby desa. Unfortunately, he gives little detail of contemporary practices of trans-desa organisation. According to my inquiries in Pujung in 1996, however, it is the centre of a group of nine ritually linked desa but these are not referred to as a banua. Pendeta do not officiate at temple rituals here and the ritual cycle is tied neither to the lunar nor wuku calendars, but to the local cycle of the traditional rice crop. The form and seating arrangements of desa meetings likewise appear to be a fusion of elements characteristic of mountain and lowland forms. Local opinion, however, sees it more as the transplantation of the forms of their village of origin in Karangasem (East Bali), forms that are themselves more consistent with those Reuter characterises as Bali Aga.

Immediately downhill of Pujung and Sebatu are a number of small desa (e.g, Kebon, Tangkup, Cebok), which have no bale agung of their own but share that of the older village from which they originated (e.g, Kedisan)—a mode of relationship consistent with processes of linkage in both the mountains and East Bali (Reuter 2002: 38-41; Stuart-Fox 2002: 46-51). At roughly the same elevation and a couple of ridge/valley systems to the west, near Payangan, is yet another group of eight desa, linked to a shared temple, known as Pura Banua, in Desa Bukian. Unlike the category of Bali Aga temples of the same name (pura banua), this Pura Banua is not understood as the centre of a ritual domain so much as a regional temple with a unique history. According to local oral narratives, its origin lies not in a ritual alliance but a defensive one, at the time of the Payangan wars of 1843, with the temple being a place of assembly in times of crisis. However, the term and metaphor of banua was chosen, which suggests familiarity with the concept, and, as Reuter (2002: 80) notes, this area is one of the few puri-dominated areas with strong ritual links to Bali Aga temples. ^[14]

The evidence of all of these examples consists merely of traces of various kinds, and there is no evidence of systematic organisation or a sense of collective identity as in the mountain banua. David Stuart-Fox (2002: 46-51), writing of similar but different groupings of villages further east in Bali, reminds us, however, that no matter how ancient and timeless they may appear, all these groupings of villages are the result of concrete historical processes. I would suggest furthermore that it is to these processes that we need look if we are to understand their contemporary forms; a point to which we will return later.

If we look further downhill, where the political and ritual dominance of puri increases there is progressively less evidence of such forms or of ‘ritual domains’ (banua) other than pre-colonial ‘kingdoms’ (negara). This would suggest that banua are either a form distinctive to the mountain regions for some reason, or that they have been eliminated or obscured in the areas subject to Majapahit puri domination. There remains, however, the evidence listed above, even in relatively downhill areas, of elements of inter-desa organisation, articulated through links between temples. Are they negara or banua or something else? The remainder of this paper considers this question by summarising and examining my own ethnographic evidence of a more substantial network of linkages in the upper Wos Valley.