PART ONE
THE KODI CONSTRUCTION OF THE PAST

1
The Imported Past
Foreign Sources of Power

Brought across the wide seas, carried over the wide oceans
To fall at our feet and be grasped by our hands
The stalk of foreign banana [papaya] now sits at our ancestral hearth
The sweet gourd from overseas is offered to our own forefathers
                From a Kodi song about imported heirloom valuables

Kodi is the kind of society that, fifteen years ago, when anthropology had a somewhat different orientation, might have been described as "isolated from history." Sumba remained in the backwaters of the Dutch colonial empire for several hundred years. Sporadically involved in the sandalwood trade of the sixteenth and seventeenth centuries, and later in the export of horses and slaves in the eighteenth and nineteenth centuries, the peoples of this island were never integrated into the Indic kingdoms or mercantile sultanates that lay to the west of them. Until the beginning of the twentieth century, rather, they lived in relative autonomy, divided into many small feuding domains and without a centralized polity or a single indigenous ruler. In the west, each domain had its own language and its own stories of the origins of time and of the calendar, the acquisition of fire, and the development of agriculture.

And yet indigenous social institutions, particularly those that express ideas of hierarchical difference, are conceived of and legitimated in the terms of external political powers. The tension between ideas of local origins and imported authority in fact lies at the very heart of Sumbanese society, inscribing all Kodi social forms in historical space and time. The most important heirloom objects stored· in ancestral houses come largely from other islands, and even certain objects that could have been locally made (such as a wooden trough or drum) are traced to a faraway kingdom in the west. Present symbols of local political orders are given a foreign history; their power was "imported" it is argued, and realized anew on the shores of Sumba.

In trying to account for this tension analytically, we are caught between silences in the documentary record and a confusing series of contradictory

claims in the oral record. Written sources on precolonial Sumba are extremely scanty, with specific material on Kodi almost nonexistent. As the domain farthest from the administrative centers of Waingapu and Waikabubak, Kodi was the only region not visited by Dutch missionary writers such as D. K. Wielenga, Ten Kate, and Albert C. Kruyt, who traveled extensively throughout the island in the early twentieth century. Louis Onvlee, the missionary-linguist who spent over thirty years on Sumba, visited Kodi only once, for a few days in 1932, and confessed that Kodi was the only Sumbanese language he could not understand (personal comm.). Until the 1950s, when Kodi was visited by F.A.E. Van Wouden and Rodney Needham, there were no published materials specifically concerned with the region.

The history of events in Kodi thus remains largely obscure, at least before the beginning of this century. But we can glimpse certain larger processes, through which ideas of foreign authority were taken up and incorporated into notions of indigenous origins. The imitation and assimilation of incompletely understood foreign powers was sometimes accompanied by a disavowal of their real origins, with foreign objects redefined as "indigenous" and made the basis of local claims to rule.

My treatment of themes in the island's history will focus on the interactions between the Sumbanese and four foreign places, each of which had a significant impact on both the events themselves and the perceptions of the nature of power and its deployment. These are "Java" (more a mythical construct than a reference to the actual island), Bima (on the island of Sumbawa), Ende (on the island of Flores), and Batavia (the colonial capital of the Netherlands East Indies).

Origins from "Java"

The Sumbanese say that their ancestors migrated from the west, crossing over a "stone bridge" (lendu watu ) that once connected the island to lands to the west. There are many local versions of the shared mythic tradition of this migration. All of them agree that the first ancestors arrived at Sasar, a treacherous cape along the north shore whose name, in Malay, means "lost" or "off track." In some accounts, the two ancestors had a wooden ship that smashed on the coral reefs and left them stranded on the island. In other accounts only one of them, Umbu Walu Mandoko, traveled to the island by ship, and it was his descendants who settled along the banks of various streams. The other ancestor, Umbu Walu Sasar, often identified as Umbu Walu Mandoko's older brother, came down directly from the sky, riding on his horse; his descendants settled the drier areas

A Kodi woman displays her high rank by wearing imported wealth: ivory
bracelets, heirloom ceramic beads, and a bark-embroidered betel pouch.
1980. Photograph by the author.

inland (Couvreur 1917, 209). Both men were originally said to have come "from Java," and to have been driven from their homeland by mysterious circumstances (Wielenga 1916-18, 21:3; Kruyt 1922, 471). Warfare, quarrels between the two siblings, and disputes over their marriages are sometimes cited as reasons for their decision to leave.

These immigrants met an indigenous population, which they gradually displaced by means of a division of powers, a conquest, or the dying out of the original inhabitants. Early Dutch writers generally agreed with Kruyt's interpretation that an "originally democratic society" once existed all over the island. In the east, that society was replaced when a few noble families took power into their own hands, dislodging the traditional religious leaders who guarded the relics of their ancestors (Kruyt 1922, 467). In the west, the ceremonial leadership of a calendrical priest like the Rato Nale ("Priest of the Sea Worms") remained unchallenged, though warfare and political struggles between competing clans raged on. Wielenga noted that many Sumbanese domains are divided into two parts, one older and one younger; he explained this peculiarity by the fact that "one part represented the original inhabitants, who had ownership of the land, and one part were later-comers, who received, with the permission of their elders, a share of the land to live on, or else seized the land without permission through force. In both cases, the older part possessed spiritual powers" (cited in Kruyt 1922, 468). A final possibility was that the original inhabitants might have been almost completely wiped out, with few traces of their presence remaining.

In West Sumba, most domains are traditionally divided into two parts—for example, Lamboya and Patialla, Laura Marada and Laura Letena, Lauli Deta and Lauli Wawa, Kodi Bokol and Kodi Bangedo. It is not clear whether these in fact represent "older" and "younger" parts or are simply geographical divisions that were settled by different clans, with no disparity of rank or genealogical status implied. The names themselves distinguish "pastures" (marada ) and "mountains" (letena ), "highlands" (deta ) and "lowlands" (wawa ), and "larger" (bokol ) and "smaller" (bangedo ; lit., "founded by Umbu Ngedo") divisions. The division between "politics" and "religion" was less clearly established in the west, since there the title given only to a priest in the east (ratu ) could be used for a chief as well (Kruyt 1922, 469).

All West Sumbanese trace their descent from the ancestors who landed at Sasar, with the exception of the few hundred descendants of an earlier population, called the "Lombo" people in Laura and the "Karendi" in Balaghar. The indigenous people, who are believed to assume the form of wild animals or carrion-eating witches, are recognized as the original owners of the land. Before the migration from the west, they were said to have had no knowledge of fire or agriculture, living as hunters and gatherers of wild forest foods.

The original migration is recounted today in the paired couplets of ritual speech (panggecango ), a verse form that marks the most important

texts in the oral tradition. An invocation of the ancestral journey runs like this:

Much of the passage cannot be translated, since it consists of a series of place names and names of ancestors, which blend into one another. The division of Kodi Bangedo is said to have been settled by the descendants of a Lord Ngedo (umbu being an East Sumbanese title for a nobleman, and Ngedo remaining a common name in the area), but little else is known of such an ancestor. The other place names are recited in a sequence that suggests an itinerary from Kodi into the highlands (Rara, Ede, Manola), across various eastern boundaries (Karendi, Bukambero, Weyewa), down to the south coast (Rua, Wanokaka, Lauli, Lawonda), and then, with a final detour to East Sumba (Kambera, Kanata, Melolo, Kabata), returning to the northern coast (Laura, Tana Righu). These verses must be recited at rites bearing on the distant past, when the souls of ancestors of the whole island are invoked· The "strange, foreign land" of tana dawa, tana

[1] Lombo means "end" or "point" in Kodi, but this reference of an ancestral migration "as the day breaks over Lombo" is often interpreted as referring to an origin from the island of Lombok, which lies three islands west of Sumba (after Flores and Sumbawa).

ndimya is usually glossed as Java and Bima (one of two sultanates on the island of Sumbawa), though it may also include Lombok, Bali, or Flores. A similar series of couplets recited in East Sumba mentions Malacca, Singapore, Makassar, Ende, Manggarai, Roti (Enda), Ndau, and Savu (Kapita 1976b, 13).

The "Java" that is evoked as the origin of the Sumbanese should not necessarily be taken in its literal sense. As one of the earliest chroniclers noted, "The Sumbanese call everyone who comes from overseas a foreigner (tau jawa ), so the category includes Europeans, Arabs, Chinese, Javanese and inhabitants of other islands in the archipelago" (Couvreur 1917, 213). The great lord who ruled over the foreign kingdom from which the ancestors came is ambiguously rendered as Rato Ndimya, Rato Dawa, and his kingdom is not so much Java or Bima but any distant land to the west. Sumbanese oral tradition contains many narratives about heroes who traveled to this distant land; the presence of Javanese krises, heirloom porcelain jars, Indian patola cloths, and other imported finery suggests that relations with these distant states took the form of trade and perhaps tribute.

Written records offer only a few scattered references to substantiate the existence of a tie between Java and these outer islands. In the Nagarakrtagama (conventionally dated about 1365), Sumba is named as a subject of the Majapahit Empire (1294-1478). A fleet of Gajah Mada sailed to Dompo, Sumbawa, in 1357, and apparently laid some claim to the island, though there may never have been any physical landing on Sumba (de Roo 1906, 185). It is not known how much contact there actually was with Java at that time, and no Hindu or Buddhist remains have been found. The main resource that attracted traders since the seventh century was sandalwood, whose fragrant bark was used to make incense, fans, and clothes chests and was much sought after by Chinese merchants. In 1522, Magellan's chronicler Pigafetta mentioned sailing past "Cendana" or the "Sandalwood Island"—which, later maps indicate, must have been Sumba (de Roo 1906, 187).

The sandalwood trade attracted other European powers to the Timor archipelago, prompting competition for control of the waters. In 1566, Portuguese traders settled on the neighboring island of Solor and built a fortress to protect Christian converts from Moslem sea raiders. The fortress was attacked and captured by Dutch forces in 1613, which sparked an intense rivalry (Fox 1977, 63). In 1636, a Dutch ship was wrecked on "the unknown island of Sandalwood," and some of the men were left behind, but no expedition was sent to look for them (de Roo 1906, 188). The Portuguese of Larantuka built a small fort at Tidas, on the southern

coast of West Sumba in Wanokaka, but its dates and use remain mysterious. In 1726, the government in Batavia recognized the presence of the Portuguese on the island; the fort itself, however, was not spotted until 1902 (de Roo 1906, 188).

Local legends often associate the ancestral migration to Sumba with the fall of the Hindu-Buddhist kingdom of Majapahit. Mythologies of aristocratic origins are extremely widespread in Indonesia, and it can hardly be true that all the peoples of the outer islands were descended from exiled Javanese princes. Nevertheless, the legend of Majapahit does have a long history on the island. Coifs, writing in 1880, reported: "Every evening a light can be seen in the direction of Monboro: the natives there say that it is on the tombs from the Modjopahit period which are there" (cited in Needham 1987, 21). No actual tombs with Javanese inscriptions were found at that time, though, nor have any been discovered since. In Anakalang, Alfred Buhler was told that the ancestors descended from the sky to settle first at Majapahit, later migrating from there through Bima and Flores until they eventually reached Sumba (1951, 57).

Present-day speculations about the origins of Sumbanese ancestors often include "the Majapahit" as a category of the prestigious past. In Kapunduk, East Sumba, I was told in 1988 that the noble families of the eastern part of the island had immigrated from Majapahit, and this was why they still had so much gold. In Kodi, some people speculated that Pokilo and Mangilo, the brothers who founded the ceremonial system, were from Majapahit, but the members of other clans were not. When a film crew from Java filmed scenes of horse battles on Sumba using local riders as extras, there was additional speculation that the ancestors of the Sumbanese must have been from Majapahit, because they also were horsemen. Such conjectures, however, should most appropriately be interpreted as reimaginings of the past, and not oral tradition. As Geertz (1973, 398) has noted, Majapahit assumes the mythical status of illo tempore in much of Indonesia: a glorified time of origins that may not correspond to any actual historical time or place.

The Sultan of Bima

The sandalwood trade was administered by the sultan of Bima, whose claims to control the island appear in the records of the Dutch East India Company. In 1663, a merchant named Van Heijst reported that the sultan of Bima was having troubles delivering Sumbanese sandalwood because of the complicated political situation at the time. The sultan said that the Sumbanese had been his subjects "from antiquity" but "now it seemed

that they wanted to become independent." He then asked for assistance from the Company to crush the rebels so that the trade could continue. The Dutch were willing to help only on the condition that they be paid with the profits of the sandalwood, but they would not promise to build a fortress on the island. The planned military expedition never took place, however: the sultan had to travel to Makassar, and when he returned the Company ships were deployed elsewhere (de Roo 1906, 189).

In 1675, further correspondence from the Company to the sultan affirmed that "from ancient times these lands have belonged to the King of Bima and his viceroy Turilia Gampo, and the Company has no plans to interfere with this dominion, but simply wants its contract to be fulfilled." The sultan said that Portuguese Christians from Larantuka had been fomenting unrest among the local population, forcing him to wage war against insurrections on Sumba and in his other possessions, even though he realized that the Dutch and the Portuguese were supposed to be at peace. The authorities in Batavia answered that they would have nothing to do with such internal strife: "The ruler of Bima can do as he likes in the areas that he controls, but the Company can give him no help; however, the authorities recognize that he must wage war on Sumba, because he has enemies there" (de Roo 1906, 190-91).

It is not clear how the situation was resolved. In 1726, a Dutch merchant named Engelbert suggested there may have been an alliance between the ruler of Melolo, East Sumba, and the "Black Portuguese"—Portuguese-speaking mestizo Christians, later known as the Topasses (de Roo 1906, 193; Fox 1977, 63). The documents indicate that there was some strain in Sumba's tributary relationship with Bima, which was accepted in some regions but resisted in others. Each sultan claimed a much larger territory than he was really able to control, because he wanted to direct trade and impose taxes. Thus, the claim that Sumba belonged to Bima "from antiquity" was more rhetorical than real, even though it did reflect a long-established trading pattern. Yet the Dutch policy of using the coastal sultanates as intermediaries for dealing with distant islands had the effect of strengthening the sultan's hand, since he retained control of the export market. In 1775, Tekenborgh wrote that several domains along Sumba's northern coast had close relations with Bima, which included marriage alliances and gifts of "people" (i.e., slaves) supplied when needed to the Bimanese court (de Roo 1906, 228). The people of Memboro were most intensely involved in trade with the sultanate, and as a result appeared "more civilized" than peoples of the rest of the island.

Accounts collected from Sumbanese informants indicate that Bima and Java were associated with prestigious trade relations, titles, and imported

objects, but people did not see themselves as the subjects of a foreign power. At the beginning of the twentieth century, Wielenga heard that strangers from tana ndima had settled in Kodi and Laura and intermarried with local people (Wielenga 1916-18, 20:139). In 1920, Kruyt was told that the people of Laura still remembered paying tribute to the sultan of Bima. The eastern part of the island, his informants said, had been dominated by Bima much earlier, but the "more democratic" domains of the west had had only sporadic contact. The "more refined" and "articulate" people that he met along the northern coast were supposed by him to have developed these superior skills as a result of a long involvement with Bima (Kruyt 1922, 472).

As in the somewhat mythologized depiction of Java and Majapahit, Sumbanese accounts represent their relations with outside powers in terms that suggest relative equality, not subjugation. Indeed, a profound ambivalence is expressed in stories about both European traders and the Bimanese, as in this one, collected by Kruyt in Weyewa in 1920:

In the old days, the Sumbanese were friends with the "white foreigners" and the Bimanese. They crossed the stone bridge which went from Sasar to the other side. Both groups once came to visit a harvest feast, and followed local custom by engaging in a calf-kicking contest. The foreigners won the first time, and many Sumbanese suffered broken legs. When they moved to boxing, however, the Sumbanese defeated their guests, making blood stream down their faces. Finally, the Sumbanese and their guests started shouting insults at each other, and the foreigners left. A little later, the Sumbanese found an eel. They sent a messenger on horseback to invite the others, but he took so long that the Sumbanese went ahead and ate it up. When the guests finally came, the eel was gone. The Bimanese were furious and became violent. In the end, peace was reestablished, and they agreed to share a meal. The Sumbanese gave them meat that they would not eat (perhaps the Bimanese were already Moslem, and the meat was pork). They became angry again and went home. As they returned, they told their hosts, "We'll get our revenge!" The revenge came in the form of a smallpox epidemic, which killed many people.
(Kruyt 1922, 471)

The story cannot be interpreted literally, since it collapses a mythical time when Sumba was still connected to the other islands of the Lesser Sundas with an event of recent history, the smallpox epidemic of the late nineteenth century. Relations with these foreign groups are cast in the idiom

of contests, which later lead to quarrels and the angry departure of the guests. The Europeans did not return, and although the Bimanese came back, they were unable to share a meal. The final explanation interprets the epidemic introduced from other islands as a punishment for violations of the host-guest relationship: because of cultural differences, exchanges of food were not possible, and a reciprocal relationship could not be maintained.

In a somewhat similar vein, the East Sumbanese nobleman Oembu Hina Kapita has written an antiquarian folk history that contains an account of the fifteenth-century Bimanese viceroy Turelia Nggampo, who came to Sumba to establish a power base there:

This power cannot be compared with that of Dutch colonialism or the Japanese occupation, but was only a recognition of the superior power of the Sang Aji Ruma Mawa Ndapa, since the local rulers of Sumba maintained their own authority. The sovereignty of the Great Raja of Java and the Raja of Bima was not visible, but remained always in the hearts and memories of the people of Sumba, and became the stuff of myths and legends about the hanggula ratu jawa, hanganji ratu ndima , "the crown of the Javanese ruler, the hajji title of the Bima ruler." . . . These titles were given to local nobles, as synonyms that also preserved a difference in sense. The hanggula was the one who had been in power but was no longer active, while the hanganji was the one who was still ruling and still active. The prince who was no longer active was also given the title karaingu , coming from the word karaeng in the language of the Bugis or Makassarese.
(Kapita 1976b, 17)

It is interesting that Kapita here transposes the contrast between the two foreign sources of power to a differentiation of indigenous types of rulers: the distant but all-encompassing power of Java is presented as the senior, passive party and opposed to the closer and more immediately effective power of Bima. He also brings in a title taken from the rulers of the South Sulawesi kingdom of Goa, which was also used by the related rulers of Sumbawa (Andaya 1981, 164). The combination reveals the complex lineages of foreign powers who were evoked by Sumbanese nobles in claiming an authority legitimated by "great lords" who lived overseas.^[2]

[2] Kapita's folk etymologies can be compared to historical evidence on the derivation of these titles. Sengaji was a title used in the Moluccas in the seventeenth and eighteenth centuries to refer to "a village head who had almost the same rank as a raja" (Knaap 1987, xx). Ellen (1986) gives it as a title in the sultanate of Ternate. In Manggarai, once the tributary of Bima, sengaji refers to the "Highest Being" (G. Forth 1981, 444).

The people of Memboro claim the closest ties with Bima; indeed, a section of Manua Kalada, the ritual center of the domain, is still called Nggaulu Ndima ("the Bima enclosure"). The houses of Bimanese mercenaries who once served the raja of Memboro were located there; also, a number of Sumbanese were said to have migrated to Bima, where they guarded a great gun called Kambeku at the mouth of the Bima river (Needham 1987, 22). Kapita reports an encounter between foreigners from Bima and an indigenous wild spirit which has resonances with Kodi oral traditions: In the village of Sangu Mata, the Bimanese started to excavate a channel when they met a spotted snake with a human head. Terrified, they abandoned their undertaking and returned home (Kapita n.d., 10). The snake resembles Pala Kawata, a spotted python-man who is said, in Kodi, to have defended the island against intruders and even accompanied the culture hero Lendu overseas in the search for life-renewing powers, which resulted in his return with the sea worms (see chapter 3, text #1).

Both "Java" and "Bima" are often cited as the ultimate origins of titles, objects, and finery. Traditional political authority was legitimated by reference to gifts received from foreign powers. Javanese rajas made gifts of fine silk Indian textiles—the patola cloths—to subsidiary local rulers, including those on Sumba. Although few of the cloths survive today, they have been extensively copied on local ikat textiles. In the eastern part of the island, they shaped decorative motifs once restricted to the nobility (Adams 1969), and in Kodi they were the model of the man's loincloth and funeral shroud (hanggi wola remba ). Traditional rulers were called "lords of the silk headpiece and the patola cloth" (ratu hunda rangga, ratu ruu patola ), indicating that ownership of the textiles symbolized claims to office. In East Sumba, sumptuary rules made the double-dyed, rust-and-indigo cloths a noble prerogative. The more diverse textiles of the west expressed claims to rank through complexity of design, the icon of gold ear pendants, and the use of deep indigo backgrounds. Wearing an ivory-handled sword or dancing with a gold pendant strung around the neck likewise served as a statement that one had noble ancestors. These particular usages must be placed in the wider context of the archipelago in colonial and precolonial times.

Moslem Mercenaries: A Predatory Expansion over the Seas

The coastal sultanates of Sulawesi, Sumbawa, and Flores figured importantly in trade with the eastern islands from an early period and became intensely involved in local politics in the. eighteenth and nineteenth centuries. Many heirloom objects in Sumbanese houses are traced to contacts

with dawa ronda (literally, "foreigners in cotton sarungs"), a category referring to all the Moslem seafaring peoples who came to the island. The Sumbanese construction of the past is heavily vested in these "history objects" traded from the west, but before we can understand their involvement in local events we must examine the cultural heritage they represented.

Given the importance of military conquest to many of these sultanates, it is hardly surprising that a great many of their sacred objects were weapons, most famously the Javanese kris, swords, spears, and even cannon. Some of these were captured from the enemy, others were acquired through trade, marriage alliances, or the miraculous "discovery" of an unusual and thus apparently spiritually potent object (Andaya 1975, 120).

In his study Conceptions of State and Kingship in Southeast Asia , Robert Heine-Geldern describes the importance of regalia in Southeast Asian political life, including the idea that certain objects, such as the royal sword of Cambodia, had their own magical force. This attitude reaches its most developed expression in the Malay Peninsula and Indonesia, where it culminates in "the curious conception prevalent among the Bugis and Makassarese of Celebes, according to which it is really the regalia which reign, the prince merely governing the state only in their name" (Heine-Geldern 1956, 10). The regalia were conceived as immortal and immobile, defining the center of power and geographical space, while the ruler was mortal and mobile, serving, says Shelley Errington (1989, 129), as "a kind of mouthpiece" in contrast to the "stable silence" of the object. Only after his death did the ruler acquire the full sanctity of his objects, for then his own clothing, teeth, and personal effects could join the store of valuables that formed the heirloom treasure of the kingdom.

In the oldest Bugis kingdom, Luwu, royal objects served as placeholders for the titles and responsibilities the ruler could bestow on his subjects (Errington 1989, 124). Titles that included ritual obligations were attached to sacred objects and were given for the recipient's lifetime (Errington 1989, 200). Each family, noble or not, had its own collection of "leavings from the ancestors," which were passed on through the generations and cherished as talismans of the power of the past.

Errington's interpretation emphasizes the stability of local inherited objects both as representations of power and as assertions that descent and "white blood" qualified a ruler to claim authority through sacred objects (1989, 125). Writing about the southern Buginese state of Bone, Andaya (1975, 120) presents an image of much greater political turmoil and change, in which the transfer of sacred objects (gauking ) could legitimate usurpation or even conquest: "The ruler could be deposed at any time, but the gauking and the rest of the regalia would continue to be accorded the

highest veneration in the community. A ruler without the arajang (regalia) had no authority to rule whatsoever, whereas the arajang retained its power by virtue of being considered the representative of the gods on earth."

In 1666-69, the Company fought a war against the kingdom of Goa in South Sulawesi, then one of the most powerful and extensive empires in the history of the archipelago. Just three years earlier, Goa had conquered Bima, and a huge migration of Makassarese to Sumbawa began. A thousand men in twenty-eight ships arrived in 1664, followed a short while later by eight more (de Roo 1906, 243). Fearing the political campaigns of the Buginese prince and Dutch ally Arung Palakka, the refugees fled to the south. Some of them settled in Sumbawa and Flores, intermarrying with coastal Moslems in Bima and Ende, while others roamed the seas from one kingdom to another disrupting normal processes of trade and government (Andaya 1981, 217-18). In 1675, a large community of Makassar refugees was reported to have formed under a Daeng Mamanga at Ende on Flores, and roving bands of Makassar, Bugis, and Mandar refugees started to move farther eastward (Andaya 1981, 163-64). Invited by Arung Palakka, by then the conqueror of Goa, to return home, they refused, suspecting they would be enslaved by the victors (Andaya 1981, 217). Their presence had a great impact on the Sunda seas:

These refugees constituted an unstable element within the area. They cast their lots with one or another factions within a particular kingdom, thereby creating unnatural or transitory governments which survived at the pleasure of the refugees. Such arrangements bred resentment in the local populace and the eventual expulsion of the refugees. They were then cast adrift once again seeking a home and an ally and making every ambitious leader in a kingdom vulnerable to the attractions of such a powerful group of armed warriors.
(Andaya 1981, 217)

Sumba at the time was a tempting target: since the early fifteenth century, traders from South Sulawesi had visited the island, and by the early 1600s there was a steady market on Sumbawa for products obtained on Sumba.

The character of trade in the Sunda Seas had changed significantly by that time. Although Sumba continued to be called the "Sandalwood Island" by European mapmakers of the sixteenth and seventeenth centuries, its supply of the white fragrant wood quickly dwindled, and most trade interest shifted to Timor (Fox 1977, 61). In the eighteenth and nineteenth centuries, the island's main exports became living things: horses, buffalo,

and human slaves. The new "commodities" involved foreign traders much more intensely in local politics and colored perceptions of outside powers with a new dimension of terror.

The Makassar-Endehnese appear in colonial documents relating to Sumba in complaints about "pirates" who raided Dutch ships and competed with them for control of trade in the Sunda Sea. Company officials soon realized that their own chances of exploiting Sumba for slaves depended on the exclusion of these rivals from the area. De Roo (1906, 195n.2) presents the Dutch perspective clearly:

The Makassarese power in those times made the waters of the Timor archipelago unsafe not only because they committed piracy and captured slaves, but also because they sold imported goods more cheaply than the Company and paid higher prices for local products, as well as selling gunpowder, lead, and rifles to the enemies of the Company. . . . These Makassarese over the years caused great pain and suffering to the Company. Now and then, when their actions got completely out of hand, military expeditions were mounted against them, which would temper their outrageousness for a while but were not enough to stop it completely.

Sumba, in fact, was to remain an intermittent battleground for the two competing forces for the next 150 years.

In 1750, Van den Burg concluded an oral contract with Sumbanese rulers along the northern coast, binding them to trade exclusively with the Dutch East India Company, and not with the Portuguese or the Makassarese. Presents of weapons, beads, gold, and a flag and staff were distributed to those who consented, from ten different regions of middle and East Sumba (de Roo 1906, 196; Kapita 1976b, 21). In 1755, a second written contract was signed with the raja of Mangili, East Sumba, in Kupang, and the Dutch sent an official named Beynon to investigate conditions on the island. He described it as very large, sparsely populated, and disrupted by constant regional warfare; the export of sandalwood, cotton, slaves, and livestock, he concluded, would not be safe until the whole region was pacified (de Roo 1906, 204-5; Kapita 1976b, 21).

The year 1753 saw an intensification of Makassar slave trading activities, and a large raid was conducted in 1758, with hundreds of people captured in the eastern areas of Melolo and Kanatang (de Roo 1906, 196n.2; Needham 1983, 38). The refugees also began to work as mercenaries for local rulers: they supplied forces to the raja of Lewa in central Sumba, for example, for an attack on his enemies in Melolo, accepting fifty-five slaves in payment (Needham 1983, 20). By 1775, fleets of thirty

to forty Makassarese praus came into Sumbanese ports each year; the island had become their "general rendez-vous or nesting place" (de Roo 1906, 227; Needham 1983, 21).

The holdings of the Company passed into the hands of the Netherlands East Indies government in 1800. In 1820, the Dutch ship Pamanoekan , under the command of J. Batiest, was on its way from Java to Makassar and became stranded on Sumba's western coast, in the region of Lamboya Patialla. Taken prisoner by the local people, Batiest and his men stayed on the island for many months, until an Endehnese ship passed by and took them to Makassar (de Roo 1906, 240). The captain's description of his time on the island, recorded by J. D. Kruseman, trade commissioner of the Timor area, provides the earliest glimpse of daily life in West Sumba and the shifting political situation. The Sumbanese that he encountered drew a very sharp distinction between outsiders—people from beyond their own island or domain—and insiders—those who shared their feasts. The members of the Pamanoekan crew seem to have experienced both states, one when they first arrived, the second after they had spent some time in a single village. In the words of Batiest:

Concerning honesty and faithfulness, the people are prone to great extremes. Men who not only robbed the survivors of the Pamanoekan of everything they could take, and even cut their clothes off their bodies and led them as slaves to the mountains, would not steal even so much as a piece of firewood from a neighbor, even if he needed it very badly. At and during harvest time, the homes are often empty for months, and although nothing is hidden, nothing is missed. . . . If they can steal something from a foreigner, they will boast about their skill and delight in their accomplishment; but once the same foreigners have become united to them as members of the clan, sharing a meal of goat meat sacrificed to the deity, then they will have nothing more to fear, and they and their belongings will become just as safe as the life and goods of a native. The people are in fact very friendly, so much so that their goodwill seems almost a contradiction of their fierce appearance. They would not kill a foreigner or enemy lightly, and would never strike an ally or friend.
(Kruseman 1836, 70-71)

Batiest reported that West Sumba had "a society without kings or chiefs" led by elders of the hereditary nobility who gained their followers through skill in speaking and daring in warfare. Slaves were kept in noble households, but they "ate from the same dish as their masters" and worked and rested beside them. He described his hosts as brave, generous to a fault,

and very impressive orators, who ruled through persuasion at large, consensus-based tribal councils and did not coerce others to accept their authority (Kruseman 1836, 72-74).

The hereditary nobility was recognized primarily in council meetings, where their speech was granted the most authority and their eloquence could be demonstrated: "Such meetings always take a long time, even when the case at hand is relatively clear, because a speaker must always be answered by someone else who interprets his words, and the people believe that no important issue should be resolved too quickly. Large gatherings also give them a chance to display oratorical abilities, and this is a society in which rhetorical skill is a source of great pride" (Kruseman 1836, 74). Celebrations were held for the rice harvest, coconut harvest, and alliances with neighboring districts, but there is no mention of the swarming of the sea worms or the pasola jousting (see chapter 5). A lunar calendar is suggested, however, by references to the "fasting month" of October, in which sacrifices are dedicated to a spirit who protects ancestral graves (Kruseman 1836, 82).

Domestic life was characterized as "a truly rare example of morality and chastity among Orientals as is found nowadays only in the most remote corners of the world" (Kruseman 1836, 70).^[3] People were industrious and kind, if a bit cowardly by Batiest's standards. Although he found Sumbanese warriors strong and cruel in their appearance, their wars seemed "almost child's play" involving much strutting and shaking of spears with relatively little bloodshed (Kruseman 1836, 74). Each side carried swords and spears in a battle formation, but stayed at such a distance that their weapons rarely struck anyone. When someone was injured, the battle was immediately stopped and sacrifices were made to compensate those who had suffered losses.

Relations of traditional hostility between domains did affect travel and communication, however, even if they rarely resulted in the conquest of territory. Heads were taken in wars waged against neighboring domains, in a pattern of continuing enmity for which there was no historical explanation:

The Sumbanese do not know much about their own history. The current generation is even ignorant of the cause of the wars between

[3] Social customs may have changed since Batiest's time. In the 1970s, the domains of Lamboya and Wanokaka had one of the world's highest reported rates of gonorrhea, which the local doctor attributed to frequent changing of sex partners before marriage (Mitchell 1982a, 12). However, it seems most likely that Batiest's remarks concern the absence of prostitution or homosexuality, a state that continues to the present.

Laboya and Manukaka [Wanokaka]. But there is such enmity between these domains that whenever someone crosses the borders set by the ancestors, armed or unarmed, man, woman, or child, they must be captured and are put to death immediately. . . . The only exception is made on the occasion of death feasts, where relatives in the enemy territory may be invited with a white banner to join in mourning for a shared ancestor.
(Kruseman 1836, 75-76)

Respect for the dead and for the spirits of the deceased was the primary idea behind all of the traditional feasts, which were led by the elders of the clan in the center of the village. The sequence of events was much the same as is still found on Sumba: guests were greeted with offerings of betel nut, entertained with singing, dancing, and oratory, and fed large platefuls of rice and shares of the sacrificed pigs, goats, and water buffalo. The description of elaborate shared ritual celebrations prompted this early observer to a nostalgic evocation of the vanished world of antiquity:

One can hardly imagine the feelings a civilized European has when he sees all this! One sees a mixture of manners, customs, and habits which bring to mind so many past eras and make one go back to the time of our own ancestors, when men lived in a state of natural happiness, such as the era of the Romans, who celebrated in their camps, hanging their weapons in the trees as they tired of victory, removing all memory of war and destruction from their minds while they gave themselves over to the innocent pleasures of a sacrificial celebration under Italy's warm beautiful sky.
(Kruseman 1836, 81)

The account appears idyllic, especially when contrasted with reports from East Sumba in only a slightly later period—the late nineteenth century, after 1860, when slavery was forbidden in the Netherlands East Indies and the Dutch forces tried to regain control of the Sunda seas. At the time of Batiest's visit, however, isolated western districts like Kodi and Lamboya, many days away from the centers of trade, were to remain sheltered from the most intense raiding for only a few more years.

In 1843, an Arab horse trader, Sharif Abdulrahman, founded the port town of Waingapu at the best natural habor on the north coast of Sumba. He was "an extremely enterprising but sinister character" (Needham 1983, 24), well connected to both the Dutch resident Gronovius and the Endehnese. Authorized to develop the export of horses from Sumba to Java,

Flores, and Sumbawa, he was soon also involved in the burgeoning slave trade (de Roo 1906, 248).

Endehnese communities rose up along the northern coast and soon were deeply involved in local politics: in 1860 Etto, the crown prince of Ende and married to the daughter of the Sumbanese chief at Patawang (de Roo 1906, 245), sent ten ships and five to six hundred men to help the raja of Kapunduk wage war against his enemies in the interior. The Dutch resident attacked and sank the ships at Kapunduk, then signed an agreement with four local rulers who said they wanted to be freed from the oppression and molestation of the Endehnese. Sharif, though married to Etto's sister, advised the Dutch to expel the Endehnese in order to stop the slave trade, perhaps because he feared them as trade competitors and a challenge to his own supremacy (de Roo 1906, 266; Needham 1983, 28).

In 1861, the export of slaves to Lombok and Sumbawa was reported to be dying out because it was no longer profitable, but the Endehnese were now plundering the interior, burning villages and capturing people to sell them to Sumbanese rulers on the coast. "The spread of the Endehnese plague not only outside Sumba but now into the interior has become even worse than it was before" lamented a Dutch report on the slave trade (de Roo 1906, 245). Because the Dutch were stronger on sea than on land, this new development threatened their already unstable control of the island's politics.

The Makassar-Endehnese have been presented as the villains of the island's history, "the scourge of Sumba" who caused tremendous suffering wherever they went. Needham (1983, 39, 49) shares the interpretation of many Dutch writers when he says:

It does not call for great imaginative powers to conceive how the Endehnese domination of Sumba would have proceeded if it had not been forestalled by the Dutch intervention, or what would have been the condition of the island if the Endehnese had wreaked their will without restraint or limit. . . . The history of Sumba presents the example of a land formerly ravaged by the slave trade but eventually liberated from the terror of forcible transportation to distant countries.

It is useful to remember that the Dutch also participated in the slave trade until 1860 and were not sharply distinguished from their rivals by local populations, who suspected all outsiders of coming as robbers and marauders (penyamun ). The leverage exercised by these outside groups depended on warfare, slavery, and island instability, which created conditions

ideal for ambitious local leaders to form alliances with the Endehnese, Dutch authorities, or Arab horse traders to further their own political goals.

External and Internal Slavery on Sumba

Slavery seems to have had a rather different meaning in the closed, indigenous context described by Batiest in Lamboya of the early 1800s relative to the "open market" in which human beings were traded as commodities along the northern coast at the end of the nineteenth century. Samuel Roos, the first Dutch controller sent to the island in 1862, wrote that slavery was an indigenous institution "so deeply ingrained in the Sumbanese character that it would be hard to bring it to an abrupt stop" (1872, 11). He noted that a ruler's power depended on his control of slaves, who provided a fixed pool of labor for the cultivation of wet-rice fields and whose status was marked by ritual and legal subordination. A very large number of people were slaves: Gregory Forth (1981, 462) found that nearly 38 percent of the population of Rindi was of slave descent, and of them over 90 percent had been attached to the noble clan. Village heads from Kapunduk and Lewa estimated in 1988 that fully 75 percent of their populations was descended from slaves.

Slavery in the eastern part of the island resembled the "closed systems" of other parts of upland Southeast Asia (Reid 1983, 161-63). Slaves were inherited, connected to noble houses, and identified with those houses' paternalistic power. The nobility themselves were called the "mother moon and father sun" (inya wula ama lado )—those with dominion over the area—and their hereditary servants were the "feet of the sun, the feet of the moon" (wisi wula, wisi lado ). This form of slavery implied an obligation on the part of the master to assist his subjects, by finding them wives and homes, paying their bridewealth and (in the colonial period) their taxes, and providing clothing and food as "a kind of poor relief" (Versluys 1941) when times were hard. Hereditary slaves (ata pa helu or ata memango ) were never sold, sacrificed, or used for hard labor, and they could be transferred from one house to another only if they accompanied a noblewoman as the "bearers of her sirih pouch." Referred to as "children in the house" these servants were often invested with important ritual duties. At funerals and important ceremonies, hereditary slaves were dressed in gold, fine textiles, and ivory, and they paraded the finest ornaments of the house.

Quite different was the fate of war captives, outsiders who were taken prisoner and could be used for ritual sacrifices. Called the "feet of wild

pigs, paddy gathered on horseback" (wisi wari ruta, pare pa mandara ) (Versluys 1941), they were treated as casually acquired plunder or booty. Reports from precolonial times say war captives were killed at the funerals of important nobles in East Sumba (Kruyt 1922, 540). In West Sumba, some domains were linked in gruesome exchanges of sacrificial victims for ritual purposes. In Anakalang, a Weyewa girl was purchased and sacrificed so her skin could be used to cover a sacred drum; and in Lauli, a Wanukaka captive was strangled as an offering to a python spirit (Kruyt 1922, 540-43). In Weyewa, captives were sacrificed whenever a sacred house was rebuilt (Kuipers 1990, 20-21). In Kodi, I photographed the skulls of sacrifice victims buried under the pillars of the headhunting house (Urea Katoda) in Ndelo (see p. 313) and was told about the sacrifice of young slave girls to cover the "drum with human skin" (Hoskins 1988a).

The demand for victims for ritual sacrifices, however, could never have been as great as that for live captives in the late nineteenth century. Certainly, the development of an export trade in human beings changed the nature of raiding and regional warfare profoundly. In oral histories, my Kodi informants recalled the 1880s as a time of escalating violence and attacks between one domain and another. Headhunting, a ritualized form of traditional enmity between domains, was accompanied by a new greed for captives as sellable property. A nobleman traveling to the coast to sell some of his own captives risked ambush and decapitation himself, as in the locally famous case of Rato Malo, whose head was stored in the Kodi village of Ratenggaro for thirty years before his son negotiated for its return (Hoskins 1989a). Other cases I heard of concerned raids on Weyewa and Tana Rio, from which prisoners could be transported to Wai Kalo to be sold to the Endehnese. Headhunters who traveled along the south coast, to Gaura and Lamboya, were more isolated from the trade and were locked in a cycle of vengeance killings that seems to have been linked to patterns of political achievement within the society (Hoskins forthcoming [1]).

The intensification of slave raids and the export of human captives in the second half of the nineteenth century arose because of the mercantile rivalries of Endehnese traders, Dutch colonial officers, and corrupt middlemen like the Arab Sharif. While they did not invent the sale of human beings in the area, they certainly seem to have developed it to an extent unimaginable in the precolonial context. The impact of slave raids was felt most heavily along the northern coast, which Resident Gronovius claimed in 1855 had been almost totally depopulated by slave raiders (cited in Fox 1977). In the more fertile interior and southern coastal areas, the impact was more one of destabilizing local politics, since a few indigenous warlords

obtained monopolies on gunpowder and firearms, which they used to raid others.

The peoples of West Sumba, organized into ceremonial confederations but not under the rule of a single noble lord, were more successful than those of East Sumba at defending their autonomy. Although the Endehnese traveled throughout the interior, they were never able to establish permanent bases in the west or make alliances with important local rulers. One Endehnese soldier told the Dutch controller A. L. Couvreur that the bodies of people of West Sumba "could not be pierced" and that special magical preparations protected them, produced by their "secluded priests" (ratu sepi ) (Couvreur 1917, 213, 215).

Moslem mercenaries were both admired and feared for their control of a superior technology of war and sea travel; consequently, some uneasy alliances were formed despite cultural differences. Two ancestral villages in Kodi, Manu Longge and Wei Hyombo, were founded by Moslems and contain Islamic burial stones in their centers. The descendants of these early ancestors later "converted to paganism" by eating pork with their fellows at marapu feasts. A great many other villages have a house called the Uma Dawa, or "Foreigner's House" where people from other islands lived and intermarried with local families, eventually becoming officially adopted. One Florinese visitor, a man remembered only as Rato Daing,^[4] became the brother-in-law of the first Kodi raja, Loghe Kanduyo, and the father of the second, Ndera Wulla. He sailed away from the island before his son was born and never returned, but left behind a legacy of alliances with foreigners that his son was to repeat.

At the start of the twentieth century, when the Dutch tried to dissolve the Endehnese communities of the northern ports of Waingapu, Memboro, and Wai Kalo, a number of Endehnese moved into Kodi, establishing the village of Pero (Needham 1968). Now four generations old, the community of six hundred Moslems is distinguished by its architecture (Bugisstyle wooden bungalows instead of tall thatch towers over a bamboo frame), Islamic cemetery, and involvement in sailing and trade. Most men make a living from fishing and livestock sales, while women weave sarungs with commercial dyes, not the traditonal Kodi indigo. A few wealthier merchants own motor-powered boats that can travel as far as Ende. There is a small mosque, and in the 1980s eight of the most prominent family

[4] The name is in fact a coupling of two different titles. Rato is the Kodi title given, usually posthumously, to leaders in feasting or warfare, while Daeng is a title for middle-level Bugis or Makassarese nobles, which has the literal meaning of "older sibling" (Errington 1989, 197).

heads had made the pilgrimage to Mecca. Like their predecessors, present-day dawa ronda are suspected of transporting contraband and violating government regulations on dynamite fishing, but they have been able to negotiate compromises with local authorities that allow them to continue in these activities as long as they are not overtly disruptive.

Batavia and the Dutch Colonial Project

Dutch traders and colonial officers remained a distant presence for the Sumbanese until the beginning of the twentieth century. Like the early rulers of Java and Bima, they claimed dominion over the island but for a long time made no moves to impose state control. When the colonial army finally arrived to enforce rules against slave trading, regional warfare, and the plundering of foreign ships, they were greeted as "the foreign mother and stranger father" (inya dawa, bapa ndimya ), whose paternalistic power was conflated with that of earlier kingdoms that had intermittently legitimated local rulers in return for tribute. The Dutch came bearing prestigious gifts, gold and silver staffs of office that were conferred on prominent Sumbanese, to create the offices of raja or bestuurder (ruler) and raja kecil or onderbestuurder (subruler).

The goals of Dutch colonial policy for Sumba were not articulated until the late nineteenth century because the island was judged to be of little economic value, remaining "scantly regarded and neglected" (Fox 1977, 164). In the provincial capital of Kupang, the first person to pay much attention to the island was Resident Gronovius, who wanted to expand the horse trade and bring Dutch planters to settle the land, giving them land on credit to cultivate coffee, sugar, cotton, pepper, and tobacco. Although the project was never realized, his suggestion that Christian Indonesians from other islands be encouraged to migrate to Sumba was prophetic:

To the idea of colonizing the Sandalwood Island, I still remain devoted. My stay there and my travels through the island convince me that such an undertaking would be crowned with success. There would be great blessings in this for the development, civilizing and protecting of a dumb but good-natured population, who are now the prey of usurious traders, pirates and insignificant but vexatious rajas. I would hope that if the Government agreed to such an undertaking, a colony of Rotinese and Savunese would be transferred to Sumba.
(Gronovius, cited in Fox 1977, 164)

Christian Savunese began to settle on Sumba in the 1800s, though not initially because of Dutch policy. A royal marriage alliance between rulers on the two islands was the reason for the first colony, established at Kadumbu on the northeast coast in 1848 (Wijngaarten 1893). Only a few Rotinese ever came, but later in the nineteenth century a considerable number of Savunese mercenaries were brought by the Dutch to help "pacify" the island and control wars between local rulers.

Dutch policy at the turn of the century was based on an assessment of the "character" of different populations on the islands and their usefulness for carrying out the colonial project. The Savunese had developed a reputation for bravery and skill in battle and so were recruited into the army. It also seems the Christian converts on Savu may have been pressured to leave the island and seek their fortunes elsewhere (Fox 1977, 172). Although armed and often uncooperative, they appeared to the Dutch more controllable than the Moslem Endehnese. But the Savunese settled only in the region of Melolo, where they maintained a bounded, endogamous community, while the Endehnese continued their predatory expansion throughout the western parts of the island, cementing military alliances with local rulers by intermarriage.

The Sumbanese of this time were often depicted as naive victims of the two invading forces. The Dutch resident Humme in 1876 described the Sumbanese as "timid and cowardly . . . never having left his island, [he] considers any foreigner a dangerous wild animal from which he quickly takes flight" (cited in Fox 1977, 171). Unfamiliar with firearms, the local population was quickly terrified into leaving the territories conquered by the Savunese and was unable to take them back. Soon, however, Sumbanese were purchasing firearms from both the Endehnese and Savunese, as well as hiring foreigners as auxiliary troops in wars between domains. The Sumbanese rulers were described as living "mainly from warfare which they conduct in an inhuman fashion," intent on capturing slaves for export (Koloniaal Verslag 1877, 37).

The Dutch policy of encouraging Savunese migration after 1890 should be interpreted as part of the wider Islampolitiek , which aimed to produce a buffer between Dutch-dominated areas and areas where Islam had diminished Dutch influence (Bigalke 1984; Kipp 1990). The Dutch feared the political power of Islam and did not trust any of the Moslem seafaring peoples in the region. They therefore allied themselves with the Savunese, defending their presence on the island with military power. In 1875, the ruler of Batakepedu tried to drive both the Dutch and the Savunese from his territory. After an attempted negotiation, the Dutch gave the Savunese ruler a gunboat to transport armed men and weapons; they crushed the

Sumbanese rebels, forcing them to accept the presence of foreigners along the coasts (Fox 1977, 172).

The Savunese settlement in Melolo carved out an ecological niche on the parched northern coastline by tapping the sap of the Iontar palm. Associated with Christianity and education, the "Savunese foreigners" (dawa haghu ) traveled to the western part of the island as schoolteachers and village evangelists, bearing a religious message in the form of the Malay Bible. Local Kodi perceptions were that the supernatural arsenal of the Savunese included black magic and witchcraft (marango ). Because of these suspicions, to this day few Sumbanese will eat the small brown patties of lontar sugar produced by the Savunese. Their link to an alien faith practiced by white people also supposedly gave them access to magical procedures that made the skin invulnerable to bullets, caused abortions and miscarriages in one's enemies, and involved sacrifices to the spirit of wealth.

Mistrust, however, is combined with a recognition of a shared cultural heritage and a long history of contact between the two islands. Of all the foreigners, the Savunese were and still are the closest to the Sumbanese. Savu is also often referred to as the "younger brother" of Sumba, for it is said that after the ancestral migration across the stone bridge at Sasar, one junior member of the party continued on to Savu. It should be noted that the younger brother in these narratives is usually more clever and more enterprising than his seniors. While Sumbanese commentators acknowledge the industry and ambition of the Savunese, they often disparage their relative deficit in honesty and loyalty. Haghu, the local designation for Savu, is a common name given to Kodi children; it is also the name of the hero of a series of entertaining tales about a crafty younger brother who makes his own fortune, rising from abject orphanhood to great wealth.

In Kodi oral tradition, the arts of metalworking and indigo dyeing were brought to the island from Savu as part of a complex of occult techniques passed down through the generations. A number of Kodi ancestral villages contain a house named Uma Haghu ("Savunese House"), where Savunese ancestors are recognized and metalworking and indigo dyeing are practiced. The spear used in divination is addressed as mone haghu , or "Savunese man"; it is told to "cut through" to the source of trouble and root out the reasons for ancestral displeasure (Hoskins 1988a). The secret combination of dyes used to make the darkest form of indigo is a heritage from a Savunese woman, whose "blue arts" included knowledge of herbalism, contraception, and infertility cures (Hoskins 1988b).

As they gained a greater knowledge of Sumbanese society, Dutch visitors began to speculate that the power of the wealthy warlords was coun-

terbalanced by that of an indigenous spiritual authority, the ratu . Several hundred years of more intensive involvement in the related polities of Flores, Timor, and Roti had prepared them to encounter the recurrent social phenomenon of dual leadership and established a pattern for incorporating it into local administration:

A common feature of many of the political systems of the Timor area is dual sovereignty—a division between a person endowed with spiritual authority and one or more persons who exercise political power on behalf of this spiritual authority. . . . In the signing of treaties, therefore, it was often the executive figures of these various territories who obtained the recognition of the Dutch as rulers and legitimate representatives of their states. Not infrequently, these recognized rulers did not have the authority to command the recognition of inhabitants of their own territories, who either opposed them or recognized some higher traditional figure. Local legends to this day abound with stories about this kind of confusion over legitimate rule.
(Fox 1977, 68)

Reading the accounts of the earliest administrators, Roos and Couvreur, in fact, we often see them struggling with local categories and trying to understand how they could be used for the purposes of colonial administration.

Both Roos and Couvreur were concerned to determine the relationship between military power and genealogical precedence as bases for leadership, and each provides a somewhat different view of the raja as either the head of a descent group or a master of force. Roos (1872, 8-9) argues that descent is an important criterion for social and ritual status, but real leadership must be achieved through conquest:

The rajas should be considered not so much as kings but as the most important person in the domain, where the aristocrats and wealthy warriors remain the main actors, as long as they can remain in power through the control of slaves. The Raja of Tabundung, the head of the so-called royal line from which the most important rajas are descended, is poor. He lacks the power to enrich himself at the expense of others; he is not feared. But the rajas of Batakapedu and Kawangu (Sudu), who rule through robbery and murder, are feared and obeyed, and there are many others like them.

Couvreur (1917, 215), by contrast, argues that the power of the conquering rajas was balanced by a different kind of hidden, supernatural power

exercised by the ratu , saying colonial officials had to be especially cautious in dealing with this figure:

The ratu . . . performs ritual offerings that can bring calamity to the community, so it is necessary to treat him as a friend. . . . In everyday life, he is completely ordinary, a poor little man whose appearance gives no clues of his importance. . . . But the ratu can, if he wants, act as the medium of mystical power in its supreme form, exercising a great influence, greater even than [that exercised by] the chief of the domain; hence, if the chief seems personally weak, or the times are troubled, one should keep a close eye on the ratu's actions. During these periods, colonial authorities on Sumba have discovered that it was not the weak chief of the domain but another person, who never showed himself and was always in the background, who had the greatest influence. That was correct; the authority had simply stumbled onto the ratu's existence, since the ratu in West Sumba is always isolated or in seclusion.

The power of the "hidden ratu " was mysteriously associated with the power of the indigenous population to resist outside domination. The conventional interpretation, presented by the missionary Wielenga and endorsed by Couvreur in his colonial policies, was that there had been a diarchic division of powers between the "priest" (ratu ) and "noble" (maramba ):

The ratu was originally the sovereign. He was the authentic ruler, but then the spiritual and worldly functions became split. The ratu was obligated to live in mystery and isolation, and thus lost his worldly power. He delegated it . . . to another related family or descent line, but in the same clan. . . . So the marapu of the ratu mangu tana are found in a subordinate house in Middle and East Sumba, but in the west they are not subordinate.
(Couvreur 1917, 217)

In the end, though, it is difficult to reconcile this version with the diversity of ritual divisions found in Sumba today, where a clear division between "spiritual" and "worldly" functions is not so clear.

The role of the ratu was, even in these early accounts, concerned not so much with "spiritual" affairs as with agriculture and the calendar. His office expresses the concerns of the whole population to avert plagues, provide rain, and assure the success of the rice crop. He was opposed to the warrior, the master of force, who was not purely "worldly" in his orientation, since he used a vast array of magical weapons, potions, pray-

ers, and charms to enhance his skill in battle. What distinguished them was the fact that the ratu's power was based on concerns shared by everyone, having to do with fertility of the land and crops, while the warrior had the differentiating attributes of a conqueror, who stands out from his companions by his skills in battle.

An important difference between the eastern and western halves of the island, which operated with fundamentally different political institutions, came into play here as well. In Kodi, an independent priesthood maintained a certain autonomy alongside the various military leaders. Because the smooth operation of the complex calendrical cycle was essential to the well-being of everyone, the Rato Nale retained the highest ceremonial rank in the domain—if only by means of a relatively fragile and diffuse form of ritual authority. In the eastern domains where the Dutch had had greater contact with local rulers, this independent priesthood had largely been incorporated into the political power of the nobleman. In Umalulu, the ratu retained only a few shreds of their former autonomy (Kapita 1976b); in Kapunduk, they were assimilated to slaves (Adams 1974); and in Rindi, the priestly function had almost completely merged with other duties of the nobility (G. Forth 1981).

The program to bring "civilization" (beschaving in Dutch) to the island had to begin, therefore, with a change in its leaders, who had to become convinced of the necessity to submit to a central administration and cooperate with it in efforts to control the local population. As Couvreur (1917, 219) states, "Once we know who and what these leaders are, we understand that we can only rule with and through them. The rajas, noble chiefs, and heads of descent groups (kabisus ) must have our full attention, and also the ratu . The first three to be used in governmental administration and information, the ratu only as a source of information, since he can play no direct role in government." The Dutch decision that the authority of the ratu was not "governmental" despite its great importance, was crucial to future interactions with figures such as the Rato Nale, who controlled the center of the Kodi polity through control of the calendar.

As the first civil administrator to take effective control of the western part of the island, Couvreur implemented a colonial policy that shifted the meaning of diarchic terms and brought them more closely into line with Western notions of the division of church and state, religion and politics. As the origin narratives of chapter 3 show, the precolonial division had a quite different foundation, for the ratu was the guardian of important objects and the passive center of the cycle of time. In Kodi, the polarity was established between the passive ritual authority of the rato marapu , the priest whose seclusion within the village protected the crops and the

rhythm of the seasons, and the rato katoda , the war leader who used his magical powers to raid neighboring peoples and appropriate their vitality for his own people. The competition for life and political power was played out between domains through the taking of heads and the capture of prisoners. The power of the Sea Worm Priest was centrifugal: it spiraled outward from him, the unmoving center that held together the round of seasons and agricultural activities. The power of the headhunting leader, by contrast, was centripetal: it turned inward to the center, bringing the trophy heads back into the domain and placing their fertile, vital energies at the disposition of the victors.

In making contact with the peoples of West Sumba, the Dutch tried to displace the value of the ratu and the "founding objects" which represented the indigenous polity, by offering their own alternative: the gold staff of the office of raja. In doing so, they were implicitly playing upon a contradiction that existed within the ritual objects controlled by the ratu —for many of them were imported, appropriated objects, taken from an acquisitive, conquering power, then used to contain the threat of external domination and reabsorb its charismatic power into the center of the polity.

Distant Wealth and Its Distribution

The relations of the various foreign populations on the island to the sources of wealth and power are evident in a Kodi version of the "money tree" story: In the "foreign land" (tana dawa ), a tree grows that bears fruit of gold, silver, and copper. Since the Dutch were the first to gather this fruit, they got the gold pieces. The Moslem sailors were the second, gathering many silver rijkdollars . The Savunese were the third, collecting copper coins which they ate in the form of lontar sugar. When the Sumbanese came, the tree had already been stripped bare, and they were told to travel back to their own island to "work the land and weed the grass" (dari cana, batu rumba ), because the riches of those who do not labor had already been plundered by others.^[5]

Before Dutch control of Sumba became a reality in the early twentieth century, foreign powers were seen primarily as sources of wealth. Trade was the paradigmatic form of the precolonial encounter, and when efforts were made to establish political alliances through signed contracts with

[5] Witkamp (1912, 486) gives another version of this story, presented as the speculation of his East Sumbanese host about the source of money for the Dutch and others. He does not, however, mention the Endehnese and Savunese in this context.

local rulers, the local perception was that they were bound to a new trading agreement, not to a sovereign political state. The legitimation of foreign powers was sought by means of gold and silver objects, some conferred directly by the Dutch on local people, some stolen from shipwrecks.

A story I heard about a sword with a golden handle illustrates this theme. Sometime near the end of the nineteenth century a ship anchored off the bay at Rada Kapal to trade food for metal and cloth. A man from the hilltop village of Ratenngaro spotted one of the sailors carrying the sword and decided to tempt him by sending his beautiful daughter bare-breasted to greet the newcomer. She offered him her betel bag, and the sailor immediately grasped at her breasts. Then her father and brothers grabbed the man and took him prisoner. "According to our custom, he has committed a great offense and must pay a fine" they insisted. The Endehnese ship captain agreed and forced the man to give up his sword. It was called katopo huhu ana , the "sword of the daughter's breasts" and remained as an heirloom in the village for many years. At the time of my fieldwork, it had been pawned to the former raja to pay a gambling debt. "If I still had the sword in my possession, I could have been a raja too" the family of its former owner insisted. "The glimmer of the gold was as bright as that of the raja's staff."

Whether its acquisition is legitimate or illegitimate, accomplished by delegation or by conquest, an important foreign object conveys power and represents an ability to command, or at least to exert a certain mystical influence over events. In the following chapters, we will see how objects become attached to narratives, to ritual offices, and to social action, through a process in which the object becomes an actor in history and history itself becomes embedded in the object.

2
The Local Origins of Time
The Day, Month, and Seasons

What, then, is time? If no one asks me, 1 know. If someone asks me to explain, I cannot tell him.
St. Augustine

The past, in many of its most important aspects, has been imported to Sumba. The ancestors came originally from a faraway land to the west and had to cross the sea to come to the island where they made their home. Valuables which they brought at that time, or traveled to other places to acquire later, are the markers of wider historical processes that bind the whole region together. The past is also concerned with the creation of hierarchical difference, and with the acquisition of objects that distinguish one person from another.

Time, however, is considered a local invention. The ancestors arrived a long time ago, in an age of origins (e nawu ) before people knew the experience of death, or loss, or temporal variation. According to the narrative traditions, it was only after they came to the island that the divisions of time came into being and time units were perceived and named. Perception of the passage of time and of mortality are universals that transcend hierarchical differences and reveal what is shared by all human beings.

"Time" in the English sense cannot, however, be given a precise translation in Kodi. Kodi has a host of words to designate English ideas such as "one time" (heimihyango ), the "present time" (henene ), the "time of day" (piti a lodo ), and the "time of year" (piri a ndoyo ), but none of these has the all-encompassing scope of the English word time . This is not to say that time is not a category of Kodi experience, or not something that can be the subject of reflection and speculation. Rather, this fact only underscores the fact that in English we collapse a number of separate concepts to make our own category of "time," whereas in many other cultures these notions may be kept more rigorously separate. This sepa-

ration does not give these cultures a sense of "timelessness"; instead it allows them to distinguish different aspects of time more carefully than we do ourselves.

The experience of time takes two major forms, which could be said to be universals in human perception: sequence and duration (Goody 1968, 31). In Kodi, these are labeled with the terms katadi and mandeinya . Sequence designates the order in which acts are performed and can be approximated by the English notions of "steps" or "stages" along a continuum. The experience of duration refers to the relative span of events and the intervals between them and is indicated in English with phrases such as a greater or lesser "length of time." The English terms are spatialized, reflecting a need for images of physical objects in space (a ladder, a pathway, a length of string or cloth). In contrast, the Kodi terms are more abstract. Physical "lengths" are designated with another term, maloyo , so the term for duration, mandeinya , has a uniquely temporal reference. Similarly, although sequences often unfold across space, the term katadi does not presuppose a spatial referent; images of steps, ladders, and pathways (lete ) may be used in metaphoric speech, but for simply expressing temporal progression they are unnecessary.

The distinction between sequence and duration is especially interesting because it separates discrete actions that can be repeated and arranged in a particular order, on the one hand, from continuous and indivisible ones, on the other. Sequence assumes the possibility of repetition; duration implies that some processes may be irreversible. Leach (1961) argues that the union of these two logically distinct notions in a single category of time is due to a Western religious heritage: by equating the irreversibility of time with the possibility of its repetition, we deny the commonsense truth of our own mortality.

Kodi narrative traditions dealing with origins Of the day, month, and year reflect this problem. They present time, in the sense of duration, the experience of time slipping away from us, as coming into being with the genesis of oppositions such as night and day, the waxing and waning of the moon, and the wet and dry seasons. The narratives argue, with keen phenomenological insight, that our awareness of time passing appears only with the knowledge of our mortality. When we come to realize that our days are numbered, then we seek actively to number them, and systems of time reckoning come into use. Sequence, the possibility of ordering events and counting them, becomes relevant once our own duration becomes short-lived.

Time is concretized in notions of the day, month, and seasons by peoples all over the world, since these reflect natural oscillations that are part of

all human experience. Artificial intervals such as the week, the epoch, the historical age, or the period, by contrast, are not universals, and in places like Kodi they are in fact relatively recent conceptual imports. Now that many Kodi speakers are fluent in Indonesian, they use words like masa (period) or jaman (era) that have a long and complex history in many Malay cultures and religious traditions (McKinley 1979). These terms have no local equivalents, however, and generally refer exclusively to the period since the Dutch colonial conquest, when "history" broke into local time. The stories told about the local origins of time differences reflect on the "natural" categories that are part of all time experiences and how these diverge from "historicized" categories that follow a particular path to reach the island.

When the Kodi say that time did not begin until their ancestors arrived on Sumba, they give a primacy to their own social experience and endow it with cosmological significance. Aside from such narratives, ultimate origins or eschatological issues receive little formal attention. The past, in most of its aspects, is treated relatively realistically in Kodi narratives. Only through their accounts of how time started do Kodi narrators place themselves, at least briefly, at the origin of a world order.

It has often been noted that a local view of the world is, primarily and most importantly, a view of time. The perception of how time passes is also a perception of life and what it has to offer. As a preamble to our wider discussions, then, let us turn to Kodi stories about the origin of time.^[1]

The Day

The Origins of Night and Day

Once, very long ago, the world existed as an undivided continuum. It was neither dark nor light, but always murky. People lived for a while, and as their bodies wore out they would shed their old skins and rejuvenate. They went through cycles of youth and old age, but they did not

[1] All narratives have been translated by the author from a recorded text, usually with the assistance of Kodi commentators who transcribed parts of the text and helped with the rendering of ritual speech. Some repetition has been edited out, but 1 have made no substantive changes in the telling. Passages recited in ritual speech have been printed in a format that highlights the use of parallel verse (and also its special status as the most authoritative and "unchanging" section of the story). All narrators were older men, with minimal schooling, who "owned" the story concerned. (See the discussion in chapter 3 of narratives that are "held" or "grasped" by a specific owner.)

The first rice and corn harvested from each garden is cut by a woman and
placed in a special basket to be given to the priests who control the os-
cillation of wet and dry seasons. 1984. Photograph by the author.

die. The sun was close to the earth, the sky was not far away, and people could easily go up to the heavens.

One day Mbora Poka built a house that was taller than the houses of any his fellows, and because the sun was so close, the thatch on his house tower caught fire. "Couldn't you move the sun a little farther

away, at least part of the time, so that our houses will be safe?" he asked the Creator.

"Yes, I can create a division of night and day, so the sun will be absent half the time. But with that division come other divisions. People will no longer grow old and then rejuvenate. Their own lives will come to an end, and they will die."

Not long after that, Mbora Poka noticed that his own body was not returning to its youthful form, but starting to grow old and brittle. Gradually his limbs began to rot and disintegrate, and he became the first victim of leprosy [mboghi mopiro , "decay while still alive"]. Finally he sat down under a banyan tree near the river and heard the cry of a blackbird [kula ]. This signaled that his time had come, and he became the first Sumbanese person to die, and also the first ancestor.

Since that day, all other persons have also grown old and died. At their funerals, we repeat his name, saying we must all come to the same end:

This was also the beginning of the reckoning of time, when people began to count the months and number the years [baghe a wulla, ghipo a ndoyo ], because the period that they spent on earth was now limited. The living and the dead became divided, and people began to bear children as the replacements of those gone before, the breath that is born again [dadi cou pa helu, hungato pa dadi ]. The living tried to make bargains with their ancestors: they promised to give feasts and sacrifices in order to extend the time given them on earth, and so exchanges began with the invisible world.²
Narrated by Temba Palaka, Wei Walla, Bukambero

[2] This story is clearly related to the Weyewa myth about Mbyora Paku presented in Kuipers 1990, 36, but it differs from it in several important ways. In the Weyewa variant, Mbyora Paku does not build a house that catches fire but simply asks for the Creator to let the Sun and Moon "live and die again" because the earth is too hot. His actions cause not his own death but that of a respected elder, and he does not suffer first the deterioration of leprosy and then the knowledge of his own demise in the blackbird's song. In some other variants, the earth's heat continues until it is shot out of the sky by Umbu Tindu, in a story with obvious relations to

the Kodi account of the origins of the phases of the moon (Kuipers 1990, 54). Another related story about the origins of differentiated time from a shattered unity appears in McKinnon 1991, 38-62.

The last lines announce several important themes in Kodi attitudes toward time: timekeeping is a calculation that emerged only as a form of struggle against mortality; exchange is the key manifestation of this struggle on the ritual stage; and the creation of descendants is the only effective way to assure one's continuation after death. The finiteness of human life provided the first measure of other finite entities, like days, months, and years. Progressions, intervals, and cycles are given meaning through narrative, which makes the relation between parts intelligible. The sense of loss and incompleteness that resulted from the division of the living and the dead becomes the prime motivation for the performance of rituals.

Another aspect of this story is its highlighting of the disruptive influence of individual striving. Mbora Poka's attempt to build the tallest house was a form of hubris that eventually dislodged the self-renewing continuity of human life. His competitive urge to stand out from his fellows brought about the curse of human mortality, a curse only partly mitigated by the possibility of continuing life in the form of future generations. Even today, it is recognized that tall house towers attract lightning and risk the destruction of whole villages by fire; they are nonetheless built, material proof of the leadership of particular individuals who dare to defy the gods' authority to control the heavens.

The origin stories provide a way to understand both the uniqueness of distinct and irreversible events and the order and structures by which they can be apprehended. The stories must be understood in a creative tension with the commonsense vocabulary used to talk about time, where these abstract concepts are applied to the contexts of daily life and realized in human experience. Ritual time operates on somewhat different principles from the mundane time of routine activities. Before we turn to deliberate divisions and manipulations of Kodi temporality, it is necessary to examine the terms of ordinary speech that characterize daily life.

Time and tense are not always marked in Kodi expressions, and certain conversations are carried on with some vagueness about actual sequences of events. If precision is desired, it can be accomplished by careful use of modifiers such as "already" (mengeka ) and "not yet" (nja pango ), and by a particle, kya (or with a direct object, nikya ), which indicates a completed action. But one can also choose to speak in a "timeless world" to a much greater degree than is possible in English. As a general rule, it is easy to

specify the exact temporal location of events occurring within a few days of present time, but increasingly difficult to specify either their location or their duration as they become more distant.

Although the notions of sequence and duration are not spatialized, an overlap of spatial and directional terms in the vocabulary does order "before" and "after" as "in front of" and "behind." Ancient times are not referred to as "far away from the present" or "distant" but as having "come in front of" (la ma ulu ) specific events in a particular sequence. Generations that follow in time come "behind" (ha muri ) and follow the preceeding one in space. Metaphors referring to the future may use spatial segments when a specific pathway or journey is indicated. Thus, a man may say, "I haven't yet come to that point" (nja ku duki pyango ), in referring, for example, to a sequence of rituals that he is planning to sponsor, but one cannot speak in more general terms of the "times that lie ahead of us." In fact, it is extremely difficult to discuss the future in any abstract terms. The Indonesian phrase waktu yang datang , itself a rather ungainly way of designating the times to come, cannot be rendered into Kodi at all. The future, in the end, can only be suggested by speaking about the world of great-grandchildren (nuhi ), who are categorically merged with their great-grandparents and thus only uncertainly oriented in the future.^[3]

Spatial metaphors come into play in the set of terms that refer to what we might call "process"—that is, change over time, which is conceptualized and represented as movement in space. The "development" of a particular set of plans and their realization are described as their "walk" (halakona ) on a journey that goes from a former state to a future one. The oriented and irreversible aspects of time, thus, are projected onto space, but those which are not given direction by human motivation are expressed in a language that is at once more neutral and more abstract.

Markers Of The Day

The Kodi word for "day" lodo , as in many other Austronesian languages, also means "sun" and refers to the period between sunrise and sunset. There are also terms for "yesterday" (wei myalo ), "last night" (wei hyudo ), "the night before last" (wei hyudo ihya ), "tomorrow" (meraho ), "the day after tomorrow" (haromo ), and even the day after that (haromo ihya ). The twenty-four-hour period is referred to as "a single day and night"

[3] This would seem to be another difference with the Rindi vocabulary for time in Eastern Sumba, which G. Forth (1983, 47) describes as "a series of spatial segments articulated by a number of points of transition."

(hawu lodo, hawu hudo ), but when periods of several days are counted they are generally counted as nights.^[4]

Ritual intervals, such as the period set aside to plan a feast or the reburial of a body in a stone grave, are always counted in years, which are metaphorically referred to as "nights." Thus, if a feast is planned in three years' time, it is stated that it will happen "after two nights" (menge du hudo ). Sometimes informants explicitly linked the concept of "day" to the dry season (mara tana ) and that of "night" to the rainy season (righuto ), making clear the assumption that all major rituals occur after the harvest and usually toward the end of the dry season.

A great many expressions are used to distinguish stages of the day and night. These are particulary densely concentrated at points of transition—the ritually efficacious early hours of the dawn, and the darking time of dusk. A series of terms reported by informants described the light spreading over the land: Na mangahaka , "the horizon becomes light"; delakoka a tana , "the lay of the land becomes perceptible"; ice ura manu , "the lines on the hands are visible"; na hundaka a lodo , "the sun appears"; and mbara kapahudo , "the morning is close by." The term for the morning hours, kapahudo , means literally the "edge of the night" which extends until the chill of the early dawn has passed. From 7:00 to 8:00 A.M. , the sun's rays begin to bring heat as well as light, and people come out of their houses to "warm themselves in the sun" (pa dirungoka a lodo ), wrapped in sarungs and mantles. By eight they have "finished sunning" (menge pa dirungoka ) and begin to do their errands.

The position of the sun in the sky is pointed to with the index finger extended to refer to periods of the day. Both the sun and the moon are perceived to rise in an arc, then to "sit" (londo ) briefly at the zenith before they begin their descent. The rising sun is called the lodo koko , and the setting sun the lodo malo . At about ten in the morning the sun grows "boughs" of heat that pull it to the top position (na karangga dinjaroko a lodo ). The moment the sun "sits" is dinjaro lodo , "high noon" or the "center of the day"; this is also the time that the heat from the sun is said to "sting" with special intensity (mbati a lodo ). The climax of ritual combats, such as the mounted battle of the pasola or traditional boxing, is timed to come at midday, under the "biting" sun. This is when "the sun steps on the shadows" (pandali ngingyo lodo ) so they appear small. Later

[4] This agrees with Nilsson's (1920, 300) contention that primitives count days by nights, although G. Forth (1983, 48) insists that this is not the case in Rindi, Eastern Sumba. However, he does not cite data about ritual counting of intervals of ceremonies, and does concede that nights are usually counted rather than days when Rindi people state how long a lengthy journey should take.

in the afternoon the shadows grow longer, and the sun lowers its "feet" (witti ) as rays reaching down to the earth and "pushing the shadows" (na tularongo ngingyona ) along the ground.

When the sun moves lower in the sky it is said to "look down into the ocean" (na tangera a wei myahi ) with slanted rays in preparation for its final descent. Near the horizon, the sun begins to "sit askew" in the sky (na tangera hoka a lodo ), and its "feet" dip into the water (ndallu ndikya ha wittina a lodo ). Sometimes it is described as "pausing over the waves" (mangga ela panu mbanu ) or dyeing the waters red with its evening glow (na raraka a wei myahi ). It sets halfway (tama hapapa ), and then entirely (tama hambolo ).

Evening begins with a period of twilight, when the sun has set but the darkness is not yet complete (nja hudongo pango ). This period is the time of stealing and ill deeds—hence the expression "they used the twilight darkness, the obscurity of the dawn" (na waingo hudo ndoko, lodo ndango ) to refer to deception and subterfuge. It is opposed to the "full night" (hudo mbolo ), when ritual singing begins (usually about 10:00 P.M. ), which is said to coincide with the first cock's crow (hakuku handoko ). The first few hours after darkness has fallen are called "those in which sleep is still light" (manduru pa lete konggo a hudo ), a time when people can be easily disturbed. These are followed by the deeper, darker hours (na kapandu pohi myalo ), when sleep is heavier. Midnight is the "center of the darkness" (taloro hudo ) and is marked by the second cock's crow, but it is not as symbolically charged as the early hours of the dawn (mari myeraho ), when the morning star Venus is about to make its appearance. The approach of the star is described in spatial terms. At about 2:00 A.M. , it is "still far away" (marou pango a motoroma rara ); by 3:00 A.M. "it is approaching" (na tukeka ), and the third cock's crow is heard; by 4:00 A.M. "it is close" (na maranda ), and about half an hour later "it appears" (na hundaka ).

This moment marks the key transition from night into day. It is when signs are expected from the invisible spirits that they have heard the prayers addressed to them. Such signs may take the form of a flash of light in the sky (a falling star), a large crashing noise, or a shaking and rumbling near the bamboo "spirit ladder" (pahere ), erected in front of a house where rites are held. In Kodi prayers, the moment is designated as

After the morning star has risen, the next stages of the dawn are marked by the cries of the kadoko bird, who sings once "to tease the young men" (kahele mone ), arousing them to defend their women and children); a second time to "tease the older men" (kahele malupu ); and a third time to mark the real dawn of the next day (delakoka ). The rooster must then crow four times (hakuku hambatango ) to confirm the song of the kadoko bird and stir the other domestic animals from their sleep.

The Time Of Daily Social Activities

In addition to a vocabulary of descriptive markers that focuses on the transition between day and night, the Kodinese have "chicken time" and "buffalo time;' which are further delineated with references to routine tasks in caring for domestic animals. Dusk, for example, is called "when the chickens go up to their perches" (ba na detango a manu ), and the period just before dawn is "when they come down" (mburu a manu ). Midmorning is when "the buffalo go out" (pa loho a karimbyo ) of the corral to pasture, and midafternoon is when they return (pa tama karimbyo ). Such markers are as common in everyday speech as is reference to light or darkness.

Although many other daily activities occur with similar regularity, they do not function as time markers. Fetching water and building a fire are a part of every morning, as is pounding paddy for the morning meal, but these are not used to designate temporal sequences. It is the coordination of human activities with the more or less independent movements of animals that is used to generate a sequence of time markers. The importance of domestic animals, especially buffalo, as markers of social and biographical time for longer sequences is further explored in chapter 7.

Evans-Pritchard (1940, 101), noting that the Nuer have as many points of temporal reference between 4:00 and 6:30 A.M. as for the entire rest of the day, claimed that these moments are more finely partitioned because they are important in directing economic and domestic activities. In Kodi, the density of temporal reference at these points seems related to ritual as well as economic concerns. The early stages of dawn are the time of

symbolically significant transitions—when messages sung and spoken by human orators reach their destination in the upperworld, when dream conversations with an ancestor may occur, when visions of medicines or magical phrases are transmitted. Kodi village dwellers are early risers, and they often witness sacrifices to the highest marapu performed just after the appearance of the morning star, waiting for the light of the sun's first rays to interpret the liver or entrails. Divinations performed at this hour are the most significant; when asked why, Kodi priests told me that our consciousness is most receptive in the early hours when darkness fades into daylight.

Numbered Days And Named Intervals

Traditionally, no intermediate unit of time between the day and the month was specified. Upon the introduction of markets during the Dutch colonial administration, "weeks" began to be been counted in cycles of market days (lodo paranggango ) or, for churchgoers, Sundays. The days of the week are not named but instead are counted—"the first day" (lodo ihya , Monday), "the second day" and so on. Government regulations often shift the days the market is held, sometimes scheduling two different days (say, Wednesdays and Saturdays) for these gatherings; moreover, the frequency of and interval between markets are not necessarily constant. Thus, when someone speaks of an event as having happened "three market days ago," it could mean three weeks or only a week and a half ago. Markets are held in four different locations in Kodi (Bukambero, Kory, Bondo Kodi, and Waiha), so the counting of market days varies regionally. Appointments and time references are made in terms of market cycles, and the name of the market's location is specified if there is any doubt about which one is meant.

Cycles of Sundays refer to the custom of "entering the house of the seventh day" a Kodi term for the Christian church. An analogy between the Christian proscription on Sunday work and the Kodi ritual silence for the period of the "bitter months" has created a more common reference, with Sunday referred to as the "bitter day" (lodo padu ) and churchgoing as "entering the bitter house" (tama uma padu ) (Hoskins 1987c). The term is not meant to be derogatory of Christian teachings, but only to stress the local perception that Christianity, too, involves restrictions and that a separate demarcation of sacred time obtains in the Christian religion. Since not only churches but also schools and government offices observe Sunday as a day of rest, the cycle of "bitter days" is becoming increasingly common as a term of temporal reckoning.

The Month

The moon is the most important unit in Kodi time reckoning. Its variability is explained in another story of an initial, enduring entity that was broken apart as a result of human action:

THE ORIGINS OF THE PHASES OF THE MOON

Two brothers in the village of Toda, Myondo Manda and Pyati Lando, were famous hunters. They traveled to Ngahu, a high cliff in Bangedo, and from there they threw their spears into the heavens, trying to see whose spear flew highest. One of the spears caught on the edge of the moon and a piece of it broke off and fell to earth. They caught it in one of their hunting nets [kareco londo laka ] and carried it back home to show the others.

Rato Pokilo and Rato Mangilo, in Tossi, saw the moon glowing in their net and said, "How beautiful! What is it?"

"It's the moon," said Myondo Manda and Pyati Lando; "we struck it with our spears and captured it in this net."

"But you cannot leave it here," said Mangilo and Pokilo. "If this piece of the moon is not returned to the sky, it will always be dark, and we won't be able to sing or dance in the moonlight or travel home from long journeys. Give it to us and we will return it."

"What do we get for giving you the moon?" asked Myondo and Pyati. They were promised that each year they would receive the first fruits of the harvest. Everyone who made a garden in the territory of Bali Hangali would have to honor the people of Toda with a small sacrifice of a chicken or dog and a gift of a knife and a spear. This showed the control that Toda once had over the moon, which they surrendered to their older brothers in Tossi.

Mangilo and Pokilo took the moon down to the seashore and released it into the sea at sunset. It rose as just a small sliver at first, but gradually grew fuller and fuller, until it became very full and red, then began to diminish again. From that point on, the moon was always variable, waxing and waning, "dying" and coming back to life. The net that had contained the moon [kareco londo laka ] was taken back to Tossi and stored beside the great gold breastplate that belonged to Mangilo. The home of the treasures was called

The people of Tossi watch the moon, while the people of Toda use the moonlight to sing and ask the spirits for better harvests and more descendants.

Narrated by Rangga Pinja, Toda

In this story, the theme of mortality and loss is applied to a heavenly body as well as a human one. Wounded by human attack, the moon now suffers the same processes of aging and death that people do. It differs from people, however, in being capable of regeneration. The regenerative power of the moon is also for the first time connected to objects: the net used to capture the moon, whose guardians are now in the ceremonial center of Tossi. The idea that some persons are privileged controllers of time is introduced here.

The narrative suggests other associations if we consider it in the context of related mythological traditions. The shattering of the moon and its fall to earth are common themes in other parts of the island, where they are associated with ties between the moon and precious metals. The Sumbanese often say that the sun is fashioned from gold and the moon from silver. Although gold and silver are found on the earth, they are said to be deposited there when a star falls from the sky (G. Forth 1981, 441). A creation story from Lewa in central Sumba tells how two suns and moons were forged by the founding ancestors from stone cliffs on the fifth level of the earth. At first both were hung in the sky, but since it was too hot, one sun was removed and set aside. The two moons got into a quarrel over a woman, and one of them struck his counterpart with a dibble. He in turn was stabbed with a knife. The stabbed moon fell to the earth and died, while his companion still bears the scars of the battle at his home in the sky (Kapita 1976b, 219, 230-31).

The life and death of the moon are associated with the life and death of human beings in everyday life as well. For example, it is widely believed that a person should die at the same lunar phase in which he or she was born—that is, that the cycle must be completed. Similarly, it is considered inauspicious to plant crops when the moon is at its fullest and appears "red" (taru rara —that is, just as it begins its decline) or when it is not visible and "dead" (mate wulla ) (see also G. Forth 1981, 208). The net that captured the moon is treated as one of the most precious heirlooms of Tossi because the moon itself is seen as a wealth object, but one that

had to be returned to the sky. Because the people of Tossi helped to return the moon to its proper place, they established a special relationship with it that in turn gives them a privileged position in the calendrical rites. The story about knocking down the moon is also about the exchange of wealth and the potential violence involved in dislodging a wealth object from its original home. (This exchange theme is further developed in the story of the origins of the wet and dry seasons in a trade of fire for water; see below.)

Methods Of Moon Counting

The phases of the moon are described and distinguished by both laymen and ritual specialists, though the degree of specification varies widely. Gregory Forth (1983, 54) recorded a list of twenty-nine named phases from one informant in Rindi, Eastern Sumba, but notes that for most people "the total seems to be incidental to the classification; for some Rindi denied knowledge of the number of nights, or days, in a month, while others claimed there were thirty" (1983, 53). Many Kodi informants provided terms for the full moon (wulla taru ) and for five or six phases of waxing and waning, but usually they (like the people of Rindi) enumerated the rest simply as "several nights when it gets darker" (piri hyudo na kapandu taruhinikya ).

Dogs are believed to be able to see the new moon before it has become visible to human beings. In the calendrical rites, invisible spirit dogs (bangga marapu ) patrol the village of Tossi and the beaches where the sea worms will swarm, and may bite anyone who offends the taboos of the yearly festivities. The first phase of each new moon is when "it is seen by dogs" (pa ice bangga ), and the death of the old moon (mati wyulla ) is signaled by dogs barking. The new moon, moreover, is associated with new plans and new projects and the budding of fruits and garden crops, as expressed in this couplet referring to an intention that has not yet surfaced in open speech:

Kodinese share the feelings of many other agricultural peoples that it is better to plant crops when the moon is waxing and still "young."

As the moon travels on its path, it has several "resting places" that mark its journey. At the first quarter, it hovers so it can be seen "looking up from the village gate" (tangara binye ). Three days before it reaches its fullest point, it is dangerous and inauspicious (wulla nja ndaha ); during

this period, wild creatures of the land and sea are particularly active, and people should not venture out to hunt or fish. As the moon becomes almost completely full, it becomes auspicious again. The wulla mburu manu , or "moon that comes down with chickens" marks the beginning of a good time for feasts and night dancing, which continues as the full moon becomes "red" (taru rara ) and stays in the sky until sunrise.

The full moon signals a period of intensified nocturnal activity and amorous conquests. In ritual couplets, illegitimate children are called "children of the light of the full moon and the fleeting darkness of the evening" (ana kandeghu wulla taru, kapandu pohi malo ), and the erotic charge of this phase of the moon often features in haunting love ballads. After the full moon "sits" at its zenith (londoka a wulla ), seven dark nights are counted to determine the day the sea worms will swarm for the nale festivities. This event marks the beginning of the period of license. The moon then becomes "old" and "slender" (wulla malupu, wulla malaka ) and begins to amble uncertainly toward its death. The aging moon rests briefly on the high plateaus to the east (mangga la panuna ), then shifts into an uncertain existence where its visiblity is open to debate (wulla pa palumungo , the "contested moon") before disappearing.

The Seasons

There are only two "seasons" in the Kodi year: the dry season (mara tana ) and the rainy one (righuto ). The division is presented in traditional narratives as stemming from another incident in which human actions played a role, provoking an alternation between periods of abundant rainfall and others of drought. In this story, the theme of exchange comes most prominantly to the fore, and is negotiated by a wily cultural hero.

THE ORIGINS OF THE WET AND DRY SEASONS

When the first ancestors first came to Kodi, they were still connected to the upperworld of the sky and its inhabitants. Relations were not good, however, for the Kodi people wanted to plant gardens, but their crops were constantly being eaten by a huge boar. The boar belonged to Rato Byokokor, the giant gatekeeper of the flood gates of the heavens. He lived in the sky beside a dam that controlled rainfall, along with his brother Manjalur, and they were called

One day, the boar ran into the garden of Lete Watu of Wei Kahumbu.^[5] Lete Watu was not afraid of anyone, so he threw his spear at the boar and struck it. The wounded boar ran home, scrambling up a series of stone steps [lete watu ] that led to his master's home in the sky.

Lete Watu followed him to the upperworld and found the bloody boar dying in front of Rato Bokokoro's house. "This pig was destroying my gardens" he said. "I have a right to take his meat."

"This was my favorite boar" said Rato Bokokoro. "We will eat you instead, since you have killed him." His words were unclear, because Rato Bokokoro had a harelip and could not talk distinctly.

The two were about to fight, but Lete Watu was too clever to allow that to happen: "Don't eat us, elder brother. We will bring you something else to eat, as compensation for your loss. Then we can be at peace, since your boar will no longer be destroying our gardens."

He brought one of his own buffalo and killed him so they could share a meal together and make peace. They decided to divide the meat from both animals equally.

As soon as the two shares were distributed, the people in Rato Bokokoro's party immediately began to chew at the bloody carcass with their jagged teeth, while those who came with Lete Watu set about cooking their share in a large pot. In a short time, the smell of the roasting meat began to travel over to Rato Bokokoro's house.

"What do you do to make that meat smell so good?" the harelipped giant asked.

"We cook it with fire" answered Lete Watu, and handed him a piece of the cooked pork. Rato Bokokoro noticed that not only was the meat superior in taste, but it was also much easier to chew.

"How do you make this fire?" he asked. "You must give me the secret. We are always hungry here because it takes us so long to chew the meat."

Lete Watu said, "I will give you the secret to relieve your hunger if you will help us to relieve our thirst. On land, we can make fire whenever we want, but our crops are parched and dying because there is not enough rain. You are the gatekeeper of the flooding waters from the south [mandoko ], so if you can give us the tools to make rain, we will give you the tools to make fire."

They exchanged the objects needed to start a fire and start the rainy season. Lete Watu gave Rato Bokokoro a kohe , a wooden tool whose two sticks could be rubbed together to make sparks, and showed him how to use it. Rato Bokokoro gave Lete Watu his "toys": a gourd filled with water from a sacred source [tabelo wei hyari ] and, to Ra Hupu, his com-

[5] Lete Watu is a familiar trickster figure who appears in a number of Kodi myths. His Weyewa counterpart, Yanda Mette, has a somewhat similar adventure, as reported in Kuipers 1990, 14-15.

panion, a bamboo pipe with more of this water, along with lightning stones. "This is the water that will feed your crops," he said, "and here are the lightning stones that will punish any infractions of the rules of planting and harvesting." These objects can still be found in Wei Kahumbu and Bondo Kodi.

Lete Watu told Rato Bokokoro to be careful in storing the fire tools, because any stray sparks could destroy the flammable thatch of Sumbanese houses. Rato Bokokoro told Ra Hupu and Lete Watu to use the power that they had over rainfall wisely, and not to ask for rainfall more than once a year, so that when it came there would be enough to nourish the crops until they were fully grown. In this way, it became the task of the rainkeepers to maintain the seasons in the proper balance.

"These tools are yours" he said, "but whenever you pray for rain you will have to say my name and my brother's name, because we are the owners of the rain waters." Then Lete Watu and his party went home to earth.

A short time afterward, Rato Bokokoro's village caught fire because they did not use the tools properly. "You have cheated me, Lete Watu!" He called. "What you have destroyed by fire can also be destroyed by water!"

And he sent a huge flood, which washed over all of the first Kodi villages and destroyed their homes.

Because of Rato Bokokoro's anger, he does not always keep his part of the bargain. Sometimes the rains are insufficient or unpredictable, not coming in the months that they were requested. Sometimes the rivers flood and the young crops are destroyed, other times the ground is parched and the dust is so thick nothing can be planted until very late. A house named after Rato Bokokoro has been built in Tossi, and it faces the House of the Sea Worms, which controls the calendar. But its master does not always heed the pleas of those who make offerings to him.

Narrated by Gheda Kaka, Wai Kahumbu

In each of the stories about the origins of temporal divisions, human action destroys an initial plenitude that can be restored only partially and elliptically. Mbora Poka's ambitions lead to his own death, but he gains the power to help succeeding generations. Myondo and Pyati return the moon to the sky, but now its cycle is variable and inconstant. Lete Watu receives the tools to ask for rain, but he antagonizes the guardian of the heavenly waters. The destruction of an original unity is replaced by a more complex and confusing concatenation of particulars.

The beginnings of separation and opposition are also the beginnings of organized exchange, through ritual prayer and obligatory offerings. In these stories, the origins of time are linked to the beginning of mortality

and to the consequences of human ambition: Mbora Poka's unusually tall house, Pyati Lando's skill with a spear, Lete Watu's ability to trick Rato Bokokoro with his words. An original unity is violated by the disruptive power of an individual who wants to stand out. In its place, an oscillation is established—the cycle of day and night, of wet and dry seasons. This in turn may also be disrupted by efforts to gain control over the seasons and the heavenly bodies. Compromise and accommodation are the hallmarks of exchange, which is given the paradigmatic form in these early narratives of a negotiation between a clever, forceful invader of someone else's space and the earlier, superior, but more passive power.

In the first story, Mbora Poka dies as a result of this ambitions, but he thereby becomes the first ancestor. In the second, the two brothers from Toda achieve an important role in harvest ceremonies through a temporary disruption of the flow of time. The theft of the moon interferes with the division of day and night, requiring the leaders of Tossi to intervene; the wounded moon is returned to the sky, and its hunters secure increased ritual importance by acknowledging the temporal authority of their seniors. Finally, the story of Lete Watu's victory over Rato Bokokoro describes the dangers of human guile. Although he tricks the sky giant out of valuable tools, what he receives in return is ambivalently charged: division is presented as the necessary precondition of exchange. There is a certain familiarity to the contrast between the cultured, smooth-speaking Lete Watu, who possesses the art of fire-making, and the more primitive, animal-like Rato Bokokoro, who controls the natural power of rain-making. Unlike in the famous lowland South American myths of fire stolen from the jaguar (Lévi-Strauss 1969b; T. Turner 1985), the cultural art is not taken from the creatures of nature but traded against another art, one dependent not on cunning but on access to heavenly waters.

I interpret this opposition as a version of the diarchic divide: Lete Watu, representative of the human intruders, takes on the role of the younger brother and addresses Rato Bokokoro, lord of the skies, as his elder brother. It is a gesture of respect and deference and at the same time potentially derisive, since it is often asserted in Eastern Indonesian diarchies that "the older brother is stupid."^[6]

Lete Watu's use of the kin term creates a framework of ranked sibling-ship, in which cannibalism can be avoided and a mutually beneficial exchange relationship can be established. Rato Bokokoro is, if not fully human, at

[6] For other cases, see Traube 1986, 51-65, on the Mambai of Timor; Schulte Nordholt 1971, 262-74, on the Atoni; and Lewis 1988, 17-19, on the Tana Ai of Flores (where the opposition between older and younger siblings is transposed to female and male, with a similar symbolic valence).

least potentially so, and he is won over by Lete Watu's offer. But he is too careless to handle fire properly, so eventually he loses his house and his sympathy with the "younger brother."

Rato Bokokoro is one of the original inhabitants of the island, a race of primitive hunters and herbalists from whom present-day witches (tou marongo ) are supposed to be descended. Although they had detailed knowledge of medicines and the healing properties of roots and barks, these people did not know how to grow gardens. Their ability to open and close the gates that dam the flow of waters from the sky was for them a trivial skill used only to avoid getting wet when returning from a gathering trip to the forest. For an agriculturalist, by contrast, it is the supreme power, the determiner of feast or famine, and not to be used casually. Rato Bokokoro is stupid to surrender his rain-making toys to someone else who values them much more highly than he does himself. Yet despite his blunder, he retains a hold on power. His "toys" must be ritually manipulated in seasonal prayers, where his name is repeated in invocations to coax a bit more rain to fall.

The story ends with the uncertainty of reciprocity from an antagonized exchange partner. Because of Rato Bokokoro's anger, the stone steps that once connected the heavens and the earth have been withdrawn, and the stone bridge that once connected Sumba to Bima and Java to the west has been washed away. Separation and exchange have replaced unity and sharing.

In a universe of diverse origins and disparate peoples, the problem is no longer how to divide basic resources, but how to reallocate them once they have already been distributed. The shifts and changes that occur over time are henceforth irreversible, and there is no unbroken cycle returning to the origins. Detailed and elaborate narratives associated with the founding of the calendar continue the theme that valuable ritual objects (such as the trough to catch the sea worms) are initially seen as mere toys, undervalued by their original owners and only later recognized as important, once they have been put in the proper place.

The Beginnings of Time and the Construction of the Past

The division of wet and dry seasons defines the year as a seasonal cycle, but in this story they have not yet been elaborated into the twelve named "moons" of the Kodi calendar. The wet and dry seasons exist as simple opposites, the "repetition of repeated reversal" that Leach (1961, 126) believed to be characteristic of the "primitive" conception of time.

At the end of these three stories we come to a conceptual period in which "time" as an experience of discontinuities could be said to exist, but it is still incompletely "cultural" in that it is not marked by the collective festivities and memories of the calendar. This fact was expressed to me in the field by a storyteller who said that after the exchange of fire and water and the division of the moon, "there were months, but they were not yet Kodi months, and there were years, but they were not yet Kodi years." That is, they had not been classified within the tradition handed down by the ancestors and were not part of a ritual sequence (katadi marapu ). I interpret the storyteller's statement to refer to the fact that "the past" as a cultural category refers not simply to time itself as an abstract entity, but to a specially constructed understanding of time in which the experiences of previous generations are put into a meaningful framework.

The cultural variability of the past has been the subject of much recent debate, which also concerns the notion that particular cultures may perceive duration in very different ways. The debate has often centered on whether there is a universal form of time perception, with duration as its basis, or whether all time concepts are culturally relative. Durkheim ([1912] 1965), Whorf (1964), and Geertz (1966) have been presented as the defenders of the relativist position, holding that all such concepts are socially determined, while Bloch, Turton and Ruggles (1978), and Appadurai (1981) suggest instead that "all speakers must at some level apprehend time in the same way" (Bloch 1977, 283) in order to communicate with each other and be capable of eventually changing the prevailing conceptual order.

The debate has started out on the wrong foot, however, by failing to distinguish between "time" as a perceptual experience with certain universal features, on the one hand, and "the past" as a culturally constructed image of a particular society projected backward onto previous generations, on the other. Certainly, at one level, the people of Kodi perceive days, months, and annual seasons in much the same way as peoples all over the world. Because of their closeness to the equator and the climatic conditions they experience as gardeners and stock-keepers, they have not developed a quadripartite notion of fall, winter, spring, and fall. The binary opposition of wet and dry seasons they use is commonsensical for any resident of an equitorial zone. The vocabulary just elaborated for speaking of time in relation to diurnal cycles and social activities uses a number of different time scales, each of which classifies time in relation to specific purposes (agriculture, animal tending, ritual stages).

In everyday ways of talking about time, events are correlated with each other through proximity and by reference to the life cycle of specific individuals. Great cultural importance is placed on sequence, since the

order of birth, of building houses, and of acquiring valuables determines rank in a ceremonial context. Attention to sequence produces a memory of precedence, but not one of chronology: the relative position of an event (as before or after another event) is remembered, but it is not situated in reference to an absolute time scale.

The particular characteristics of Kodi notions of time, in other words, do not lie in a fundamentally different concept of duration, or in the absence of duration in favor of repetition, but instead in the plurality of temporal scales and their relationships. Rather than being subsumed in a single sequential chronology, these different ways of expressing the passage of time coexist for the numerous special purposes that they serve. The portrait that we have drawn of Kodi time before the creation of the calendar, hence, is of a diverse but not necessarily confusing series of time concepts, applicable in different contexts.

The calendar itself, as we shall see, provides a loose integration of the notions of day, month, and season and is the first step toward the synchronization of different cycles and processes in a shared chronology. As I use the term, a chronology refers loosely to an independently calibrated gauge of events over time through which various time scales can be brought into meaningful relationship with one another. I do not restrict the term to actual dating systems; it is, rather, a more encompassing frame of reference, one that, in Kodi, is articulated in traditional narratives.

The coexistence of a number of varying temporal frames is characteristic of the intellectual systems of many non-Western societies. Robin Horton (1967, 176-77), for example, describes the time systems of many African groups as equally particularistic, piecemeal combinations of loosely connected segments:

In traditional Africa, methods of time-reckoning vary greatly from culture to culture. Within each culture, again, we find a plurality of time-scales used in different contexts. Thus there may be a major scale which locates events either before, during or after the time of founding of the major institutions of the community; another scale which located events by correlating them with the life-times of deceased ancestors; yet another which locates events by correlating them with the phases of the seasonal cycle; and yet another which uses phases of the daily cycle.

The distinction between special-purpose time scales and a more encompassing chronology is similar to the distinction between special-purpose primitive money and our own multipurpose money: a common currency

is not needed until it is necessary to generalize across contexts that under normal circumstances are assumed to be separate (Maltz 1968). A multiplicity of time scales is also found in the West, of course, in that we can speak of events occurring during the Kennedy administration, at the end of the summer, and during someone's adolescence; the important difference here is that these expressions can also be rendered in dates, while those of many other societies cannot.

Kodi "time," therefore, is plural and not necessarily consistent, and builds on notions of duration and sequence, both of which have correlates in the natural world (the movement of celestial bodies, the changing of the seasons, the waxing and waning of the moon) and in human biological experience (the processes of aging, death, and reproduction). These notions of time are then brought together in the ritual calendar, to which we turn next. From the story of the creation of the calendar we start to see the construction of a "past," or rather several "pasts," which reveal Kodi understandings of their own heritage and its imprint on present-day life.

3
The Past in Narrative
The Creation of the Calendar

Apart from the practical value of the calendar it is used in order to supply the framework of a narrative account of the year. Whenever one of the old men is asked about the moons he does not give a sober account, but will proceed at once to recite a story in which he gives the successive names in a more or less detailed and flowery description of what takes place in each. An intelligent old native will make it clear that not everyone knows the names of the moons, that only those having to do with gardens and knowing how the year runs are acquainted with them.
Bronislaw Malinowski,
"Lunar and Seasonal Calendars in the Trobriands"

Malinowski's description of the Trobriand response to questions about the calendar in a sense replays my own field experience. When I first began my investigations and traveled to speak to a couple of respected older men, I was encouraged to focus on the calendar, "the Kodi months and the. Kodi year" which might give an overarching structure to the vast project of recording Kodi tradition. "If you don't know where to start" an old man advised me in an early conversation, "start with the sea worms. That is where we start ourselves."

The enumeration of month names was also an occasion for enumerating an annual round of social activities, the times of feasting and the times of famine, and also, I was to discover, the division of ceremonial tasks among ancestral villages and the location of objects that represent the historical process of acquiring these rights. The creation of the calendar at a consensus meeting of the ancestors provides the narrative frame that orients, to a certain extent, all Kodi storytelling about the past. All accounts of how crucial social institutions came into being are in some way related to the structure of the sea worm festivities and the divisions of the year. Knowledge of the calendar is focused on a few important ritual centers, invested in named religious officers (the Rato Nale, or "Priest of the Sea Worms"), and seen as something of a "native science."

Having said this much, however, it should be added that there is no "master narrative" of the origins of Kodi society or the calendar. The

stories that exist are "owned" by certain houses and "attached" to objects within them, and told by their descendants in full recognition of their partial and partisan content. An idea nevertheless persists that, scattered through the memories of a great many different storytellers, there is a common, encompassing story line that—although contested and debated in many of its details—is both inscribed on present-day social institutions and part of a shared, narrativized past.

This chapter represents my attempt to reconstruct that shared story line, through the presentation and analysis of seven different texts. Each one, collected from a descendant of an ancestor involved in the story, was considered a relatively "legitimate" version of a portion of the past, but each was also the subject of other interpretations and counterclaims, which I discuss alongside it.

Kodi thinking about the past is expressed not only in narratives, but also in exchange valuables, the architecture of villages and gardens, the names of places, and movement through the landscape. A feature of the narrativized past, however, is its malleability—and this feature is also its undoing, for stories are given less credence than many other forms of evidence because of local recognition of their almost infinite variability. It is important, therefore, to begin with considerations of genres of narrative and their relation to knowledge about the past. Once we have situated narrative in the context of other forms of "traces" or "signs" left by the ancestors, we can better understand how these particular stories are fixed in time and place, and why debates still rage about them.

A sense of temporal positioning, of how human acts have been ordered in time, is clearly crucial to problems of legitimacy and historiography in Kodi. The origin narratives discussed in this chapter convey an ambivalent attitude toward the locus of authority, often defined by contrast to the locus of action in a diarchic division of agency. In the colonial period, this ambiguity came into play in reshaping authority, moving it away from the ritual centers and into a newly defined "political" sphere, separated from the officers of the calendar. Later transformations of exchange transactions and the "new order" penetration of the state into local affairs disturbed indigenous ways of talking about the past, leading to a new authoritative discourse in "history" (sejarah )—a progressive, oriented vision of a discontinuous past, informed by the rhetoric of a new national future.

None of these transformations has brought about a complete change: the different forms of historical consciousness developed, respectively, in precolonial times, the Dutch period, and the years since independence still coexist in dialogue with each other. By studying their complex interrela-

tionships we can analyze the processes of formation and interpretation that lie behind them.

Genres of Stories About the Past

Knowledge of the yearly cycle and the control of time is a key strategic resource of the local polity. Calendrical knowledge is concentrated in the most prestigious genre of storytelling, the recitation of narratives that are "attached" to persons, objects, or locations. These narratives usually concern origins, and knowledge of an object's origins can carry power over that object. No ancestral relic can be addressed without pronouncing its special couplet names; yet it is both dangerous and inappropriate to pronounce these names without an understanding of their sources and significance. The classification of a narrative as "fixed" or "grasped" by a particular person, therefore, implies certain standards of authentication and legitimation that do not apply to other genres.

Kodi storytellers recognize four categories of narratives,^[1] which are differentiated by their relations to space and time: "stories from long ago" ngara kedoko e nawu ; "stories performed with songs,' ngara kedoko lodo ; "stories that are held or attached" ngara kedoko pa ketengo ; and "the voices of the ancestors" li marapu . "Stories from long ago" which make up the first group, happen in a distant, bygone era that is roughly similar to the "once upon a time" frame of Western folktales. They refer to an age when trees and animals could still talk, people did not yet die, and the Kodi did not yet live at their present village sites. They include animal fables of a vaguely philosophical and speculative character (Needham 1957b, 1960), accounts of how people acquired shared human characteristics, and the stories about the origins of days, months, and seasons that we have

[1] These four genres can be compared to Adams's (1970, 1979) enumeration of three similar categories in East Sumba: harvest myths told in the fields, clan myths told on the porches of clan houses, and chanted texts in ceremonial language performed inside the clan house for the invisible audience. Her list marks a similar progression from unmarked, ordinary language to more fixed, invariant speech but does not differentiate between the "stories from long ago" and the "stories performed with songs." Christine Forth's (1982) dissertation on Rindi oral narrative concentrates on nine orphan (ana lalu ) stories that I classify among the "stories performed with songs;' though her texts are much briefer than the Kodi ones. Kuipers (1990), writing about Weyewa ritual speech, notes a similar progression toward invariance and greater textual authority in the movement from divinations to misfortune rites to feasts of fulfillment (woleko ), but his discussion focuses only on genres of couplet speech and does not include vernacular narratives.

"Stories that are attached" can be narrated only by their rightful owners,
who usually prefer to speak on the wide verandas of their ancestral houses,
where they can request permission from the dead to tell each tale. 1985.
Photograph by the author.

already surveyed. They do not refer to social divisions between people, houses, or territories, and can be told by anyone in an informal context.

"Stories performed with songs" in the second group, are elaborate performances that alternate vivid scenes of dialogue and action with commentaries sung or chanted in ritual couplets. Their recitation requires a specialized, and usually paid, storyteller, who speaks for an entire night. Moreover, because the stories are sacred, they must be preceded by the sacrifice of a chicken to the ancestors to ask permission. They feature standardized heroic protagonists: the male figure is called Ndelo in the greatest number of stories, though occasionally Rangga or Haghu; the female figure is almost always Kahi or Leba. The story typically begins with the birth of the protagonist and usually concerns orphanhood, abandonment, and a struggle to recover a lost position of privilege. It is punctuated by long songs of lament, in which the hero's or heroine's trials and tribulations are recast in poetic couplets. It ends with a happy reso-

lution, usually a marriage or a triumphant feast of celebration. Although these stories may "belong" to a specific village, they can be told by a number of talented storytellers, and are a form both of entertainment and of edification.

The third genre comprises stories that are "attached" to persons, places, or things; often, therefore, they are called in Indonesian the "property" (milik ) or "legacy" (warisan ) of a particular descent line. They are highly sacred, but need not be told in an elaborate style, and at times are in fact quite truncated in their presentation. Although they contain short passages in couplets, they are narrated in the vernacular and are more "realistic," even "historical" in their content than stories in the first two groups: that is, they take place in actual, named locations, concern the ancestors of the storyteller, and have a minimum of supernatural elements. They may be jealously guarded, so that not everyone is allowed to listen in when a performance occurs. These stories document how people and objects came to their present homes and contain historical "evidence" in the form of references to named valuables, places, and persons.

The fourth genre, the li marapu —"voices of the ancestors"—is a chanted, relatively invariant series of couplets that is performed only when actually invoking ancestral spirits. Recited almost exclusively beside the sacred pillar in an ancestral house, these narratives refer elliptically to the events of the "attached stories" but do not elaborate on them. Their purpose is not to explain the past, but simply to evoke it, using a chain of place names, ancestor names, and the names of sacred valuables to "travel back" to the marapu and summon them as an invisible audience for a ritual event.

The genres are distinguished by the conditions of performance, their "truth frame" and the community they address as an audience. While the first two groups are considered part of the heritage passed down through the generations, it is expected that a narrator may change or at least embellish their story lines to achieve the proper effect. In contrast, the "attached narratives" are supposedly invariant, and a narrator must assert that he is doing his best to repeat the tale as he heard it from his father or grandfather. The li marapu , as ritual invocations, are so rigidly structured that a mistake in pronouncing the couplets could necessitate the paying of a fine or, in an extreme case, result in physical danger to the speaker: if someone speaks wrongly in an important ancestral house he risks illness and even death, sanctions from the listening ancestors who decide that their voices have not been properly rendered.

Each genre has characteristics of what Bakhtin (1981, 84) has called the chronotope : "In the literary artistic chronotope, spatial and temporal in-

dicators are fused into one carefully thought-out, concrete whole. Time, as it were, thickens, takes on flesh, becomes artistically visible; likewise, space becomes charged and responsive to the movements of time, plot and history." The chronotope defines generic distinctions as different relations to time: in the first category of narratives, the relation is reflective and generalizing; in the second, time is used to edify by means of dramatic tension, comedy, and song; in the third, an authoritative but partisan vision of a sequence of legitimating events is provided; whereas the fourth is characterized by a definitive and invariant ritual text.

The "stories from long ago" and the "stories told with songs" recall the romance genre in many ways, taking place in an abstract adventure-time where no significant changes occur. At the end of the story, the initial equilibrium that had been destroyed by chance is restored, and everything returns to its own place. For some storytellers a psychological development occurs, with the hero or heroine becoming more knowledgeable as a result of these various experiences, but it is not characteristic of the genre as a whole. The songs are repeated, sung once at each stage of the protagonist's travails; their very repetition reinforces the idea of an unchanging time frame in which the individual simply moves through a series of stages. The space of the "stories with songs" is also abstracted: although specific villages are named, the same story can be told by different storytellers in different locations. Thus, the narrative itself is not fixed to particular locations but simply uses them as a frame for the action.

The "stories that are held or attached," by contrast, occur in concrete space, with great importance given to named locations and features of the landscape. The time that they contain is a time of sequences and orders, but not yet a "chronology"—in the sense of a time situated in relation to an external time scale, such as the passage of years. Nonetheless, such a chronology could be said to be emergent in the form of the stories themselves, since they are about the origin of a single external time scale: the ceremonial system that synchronizes events to provide a unifying cycle for social life in the domain.

The li marapu are conceived not as narratives about the past but as the actual voices of the ancestors—that is, the past speaks directly in the present. The spirits play a ventriloquist's role, "placing" the words into the mouth of the speaker and requiring that they be pronounced correctly. As he chants or recites these couplets, the priest is said to be a simple mouthpiece of the spirits ("the lips told to speak, the mouth told to pronounce" wiwi canggu tene, ghoba tanggu naggulo ; see Hoskins 1988b). There is no loss of consciousness on the part of the speaker, who does not become possessed. He remains responsible for his words, and this respon-

sibility is immediately checked by the ancestor for whom he speaks, who can punish any errors that misrepresent him or other traditions.^[2]

The Politics of Narrative Collection

The methodological problems that I encountered in trying to piece together this collection of narratives illustrate cultural attitudes toward the ownership and guardianship of the past. Members of the community who were interested in my project warned me that I would be told many lies and that I should insist on seeing the objects to which each narrative was attached in order to validate its legitimacy. They also told me to listen only to the "masters of the house" (mori uma ) when collecting its narrative traditions and to pay no attention to the "talk on the veranda" among those less directly involved. Some of my taped narratives were given to me in confidence, with a request that they not be shared with rival groups. Most storytellers, however, were willing to let their words be widely known, because they were proud of their claims and wanted others to recognize them.

My informants disagreed about whether this project should be a simple compilation of diverse stories told in different villages or should result in the distillation of an authoritative "history of origins." I soon became convinced that no single "true version" existed, but the project was haunted by the possibility that such might have existed in the not too distant past. Although aware of discrepancies, my Kodi informants liked to imagine that a unifying canonical text could be pieced together "if only a few more old men were alive today, those who really knew." With this aim in mind, they applied themselves with particular rigor to the criticism of the texts collected.

The future use of these materials was also of great interest. One of my assistants, a former schoolteacher named Guru Katupu, proposed that at the end of my stay I leave a notebook of the complete texts, in Kodi and Indonesian, in the district office where it could serve as a reference work

[2] The frequency of these punishments was much debated. Ritual speakers are Often older men, and when one of them dies abruptly after an important ceremony speculation inevitably arises about whether he might have "said something wrong" in the course of his oratory. A number of both speakers and singers, moreover, are blind. One such singer, Ra Kambura, told me that his loss of sight allowed him to concentrate on listening to the spirits and thus protected him from error. Another, Rangga Pinja, who lost his sight some four years after I met him, interpreted his blindness as a punishment not for mistakes he had made in speaking but because one of his sons had been convicted as a thief and blemished the purity of his descent line.

for generations to come. The suggestion was vetoed by his mother's brother, Maru Daku, who noted that there were people who might want to destroy such a record, much as they had burned local government archives in the past. He proposed instead that printed versions of individual narratives be given to those who had provided them to me, and that these be stored in their own ancestral houses, "so those spirits could watch over them and protect them separately." The younger man here was affirming a belief in a master narrative and a synthesis of Kodi traditions, while his senior was cautioning him against the possible consequences of forcing a consensus that did not exist. I decided on a compromise: the documents I left in the district office listed the names of the spirits worshipped by all the people of Kodi, in the shared ritual traditions of the harvest and rites of affliction, but it did not include partisan versions of the acquisition and distribution of sacred objects.

Real historical time and space can never be completely assimilated into expressive forms. Each of these four narrative genres offers a lens through which we can view a Kodi construction of the past, but none is absolutely definitive: even the "attached stories" which carry the greatest authority, are challenged and contested. The Kodinese see history as materialized in the house and its possessions. Narratives about ancestral exploits are themselves placed among the "possessions" of a house and are permanently attached to it, subject only to the variations of memory in its descendants. Remembering becomes a sacred duty, and ancestors can punish those who remember them falsely or neglect what they have said. This is why I was told to listen only to people who were as permanently attached to these houses as the narratives were: if one of them were to deceive me, he would suffer the consequences, because these stories are "heavy" enough to crush offenders who play games with them.

Presentation of Texts

For the purposes of exposition, I present seven narratives, each identified as the product of a specific narrator and clan. They are arranged in a sequence that corresponds roughly to the one in which "the events" are said to have unfolded, but there is no occasion in Kodi when all of these narratives would be told together. The central narrative—the one most widely known, though in abbreviated form—is text #5, concerning the founding of Tossi by Mangilo and Pokilo, the Sumbanese equivalent of Romulus and Remus, two orphaned boys who stumble into possession of the sea worms and the most important valuables in the domain. The other

texts all refer to this one in some way, usually by including Mangilo and Pokilo themselves or by "anticipating" or "repeating" some of their actions.

I have chosen the fullest and most lively texts to present as a first version, but I follow these with a discussion of alternate versions, to convey a sense of both the common ground and areas of disagreement. Leach's early (1954b) reminder that mythic discourse provides "more a language of argument than a chorus of agreement" can be explored in terms of the controversies surrounding these texts.

TEXT #1: THE COMING OF THE SEA WORMS TO KODI

Lendu Myamba was the first Sumbanese to bring the knowledge of certain things from across the seas. He came to the island along with many others in a migration from the west, crossing over the stone bridge [kataku lendu watu ] at Cape Sasar and entering from the north coast. After a few years, the first settlers began to experience hardships. Their garden crops were not large enough, and there was nothing to eat during the hunger season. Lendu set off to look for new sources of food [mandata ]. He was accompanied by Pala Kawata, his mother's brother, who took the form of a giant python.

They traveled across the western seas to the splendid kingdom of Rato Ndimya, a foreign lord who had tremendous wealth. Lendu hoped to secure powerful valuables and magical knowledge from him, but first he had to perform a variety of trials. When he arrived at the palace of Rato Ndimya, he was treated as an honored guest, and a large water buffalo was slaughtered in his honor. The meat was not skinned or separated from the bones but chopped up and cooked all together. Then he was told to eat the huge pile of flesh, skin, fur, and bones so that his host would not be offended. Realizing that he could not do it alone, Lendu prayed to his mother's brother Pala Kawata, who told him to ask to be served by the palace gates. When the huge plates of meat were brought to him, the python slithered up behind, opened his mouth, and swallowed it all whole.

In the second trial, Lendu had to play buke : a post as slender as a hair was set up fifty meters away, and Lendu was told to hit it with a small sharpened bamboo dart [karaki ]. He again called on his mother's brother, who gave him a piece of sticky resin to put on the dart. When he set the dart flying straight, the resin made it head for the hair-thin post and stick there, so that there was no doubt that it had struck fast.

The third trial was the board game kule , played with two rows of eight holes and twenty-four dedap seeds. The goal was to arrange eight in a row, and even though the number of seeds was not great enough,

Lendu managed to win with the help of Pala Kawata, who magically created new seeds from his mouth.

Once he had passed all of these trials, Rato Ndimya announced that he had been searching for a son-in-law to marry his beautiful daughter Nyale, and Lendu had proved himself worthy. "But what" he asked, "did you come looking for on this long voyage?"

"I was searching for eternal life" answered Lendu. "I have grown tired of hard garden work in this dry land."

"Alas, there is no eternal life" said Rato Ndimya, "but I will give you a gift of returning life, to renew the land with fresh waters. I will give you the sea worms [nale ], spirits from the deepest ocean who will bring you fertility and the birth of new generations. Each year, if you receive the sea worms well and they are abundant, your rice harvests will be good and your descendants Will be plentiful."

With the help of Pala Kawata, Lendu produced a magnificent bride-wealth payment of gold valuables, which he presented to the foreign lord. Rato Ndimya's daughter was told to prepare to return with him as his bride. She was dressed in all of her finery and sat beside him as he was given a series of farewell gifts.

The most important gift Lendu received was the karaba rica , a small trough to hold the worms and preserve them with ginger and spices. The worms themselves were not inside the trough, but he was told that they would come to meet him on a specified day of the year when they swarmed along the western beaches. They would continue to come to the island as long as they were greeted with a ritual combat on horseback called the pasola . He also took with him the megapode bird [wondo ; a long-legged forest fowl that builds large mounds of mud in which to lay its eggs] and several wild tubers, caterpillars, and honey. These were all wild foods that could be eaten in times of famine while waiting for the harvest of garden crops.

As Lendu's ship prepared to return home, Rato Ndimya's daughter refused to come with him. She threw herself into the ocean instead, her body parts breaking up into many tiny small pieces, which would wash up along the beaches of Kodi in February—red pieces from her rosy, betel-stained lips, blue and black pieces from her long flowing hair, golden pieces from her smooth skin. If he found the right spot in Kodi, he could release the sea worms from his trough and call the other worms to swarm and reconstitute the lost body of his sweetheart in the sea.

Lendu and Pala Kawata returned to Sumba, following along the same pathway as the first ancestors:

They followed the northern coast to the region of Bukambero, where the megapode bird was released. She immediately began to gather up bits of mud to make a huge nest to lay her eggs. The nest was as tall as a high-towered house, so the bird was called Wondo Cabeka ["the great house builder"]. Her eggs were incubated in the mud nest until they hatched into a man and a woman, who became the first inhabitants of an area called Ngundu Ngora Tana, Pyapo Ndara Lewa ["the cape of land stretching out, the round cheeks of tall horses"]. They took the wild foods brought from overseas and settled there.

Lendu observed the techniques that the megapode bird used to construct her house and taught the other Sumbanese to build their own homes in the same fashion—standing on wooden piles with tall thatch towers, so that their finest heirlooms could be stored in the tower. He himself build a large house with a great tower in the interior of the district of Bukambero.

Narrated by Temba Palako, Pati Merapati, Bukambero

This narrative marks the beginning of a sequence of events. The descendants of Lendu and the children of the megapode bird are the present inhabitants of Bukambero, numbering some four thousand; their home is at the northern edge of the domain of Kodi, and although most them are fluent in Kodi, at home they speak their language (paneghe bukambero ), which is most closely related to Laura. Because Bukambero is a marginal territory, however, this narrative is a potentially controversial part of the corpus of the "Kodi past." As the first architect of Kodi villages, Lendu had an important role to play in defining the region's cultural identity, but he is not acknowledged as a founder of Kodi. In Tossi, people play down his importance, stressing the fact that Lendu received the sea worms overseas but (as we shall see) was not able to keep them for himself.

Lendu's courtship of the daughter of Rato Ndimya, who later metamorphoses into Inya Nale, the female spirit of the sea worms, recalls the Javanese tradition of Nyai Lara Kidul, Goddess of the South Seas. She became the special protectress of the House of Mataram after a similar series of events:

According to Mataram tradition, she was a princess of Pajajaran who had been driven from the court when she refused a marriage

arranged by her father. He laid upon her a curse: she was made queen of the spirits with her place beneath the waters of the Indian Ocean, and would only become a normal woman again on the Day of Judgement. . . . After Senapati [the perhaps mythical founder of Mataram] had spent three days with her in her underwater palace, the Goddess promised him the support of her spirit army.
(Ricklefs 1981, 38)

Like the Javanese princess, Inya Nale is the female consort of the "ruler" of Kodi and can be summoned to appear only by those who possess certain sacred objects (green cloth in Java, the sea worm trough in Kodi). Her unwillingness to marry Lendu is sometimes interpreted as showing that he was not an appropriate leader for all of Kodi, thus requiring the transfer of the worms to others.

While the people of Bukambero are acknowledged to be the custodians of an ancient and in many ways powerful form of knowledge, other people in Kodi are profoundly suspicious of them and of this version of the origin of the worms. Lendu is recognized to have had the worms in his possession at a certain point, but versions told in Mete and Tossi say that it was in fact Temba and Raghe who brought the worms from overseas. Another form of secret knowledge that Lendu is widely believed to have brought to the island is not mentioned in this myth: the knowledge of secret poisons (pawunu ), which find their way into food served to guests in Bukambero and which work, like witchcraft, to consume the internal organs of the victim and transfer all his strength and vitality to the witch who has afflicted him. All the people of Bukambero are suspected of being involved in witchcraft, though most deny that the occult arts are practiced by more than a tiny number of the indigenous inhabitants. The descendants of two matriclans (Walla Kyula, Walla Ngedo) and the patrician of Wei Wyalla Pati Merapati are particularly notorious for their knowledge of poisons, herbs, and wild foods.

In the narrative, Lendu receives nale in disguised form: the gift of the daughter is in fact the gift of the worms. Lendu believes that he is receiving a human bride, but in fact he receives only the shattered pieces of her body—the sea worms which will bring him the renewal of life that he seeks. The trickster is thus, in one sense, tricked, since he does not get what he bargained for and his most precious gift cannot be used in his own homeland but must be passed on.

The text introduces a symbolic trope that we will encounter in other narratives as well: the idea of an original location which in some way proved unworthy. This trope is useful because it is open to multiple

interpretations, and the tension between different versions is often not resolved. Thus, from the perspective of this storyteller, the narrative establishes the ritual priority of Bukambero, since the worms were brought by Lendu from overseas. From the perspective of other storytellers whose narratives follow, however, it establishes only that Bukambero was the first of a series of different locations for the worms on Sumba, but not their definitive home. In denying the consequences of the story told in Bukambero, the people of Mete and Tossi incorporate many similar elements (especially the idea of a series of games and trials) into accounts of subsequent events.

TEXT # 2: THE JOURNEY OF THE SEA WORMS AND THE FOUNDING OF TOSSI

Since Lendu lived in the interior and could not release the sea worms near his home, he traveled out to the coastal region, passing through the deep forests of Honde Ryara. There he met two small boys, Mangilo and Pokilo, who were living all alone in a "Monkey Shack" that served as an outpost for hunting monkeys and wild pigs. Mangilo and Pokilo saw the worms and thought they were a plaything—a bauble made of cotton, a skein of colored strings [maghana lelu, mangguna hario ]. They wanted to play with them, but Lendu was not sure that he should surrender them. He invited them to come stay with him in Bukambero.

A little while later, Temba and Raghe came by. Temba and Raghe had migrated from islands to the west, sailing to the northern promontory of Sasar, where their wooden boat was smashed to pieces on the stone bridge that then linked Sumba and Sumbawa. Their father was Tana Mete ["Black Land"], their mother was Ndabi, and their descendants were also "black" [i.e., members of the village of Mete]. They traveled to the island of Sumba with many other companions, but they were the first to reach the western tip of Kodi. They sailed around the island to the east, past the southern districts of Anakalang, Wanokaka, Lamboya, and Gaura, and eventually to Kodi. They stopped in Balaghar and set up a stone, called the Temba Raghi stone. Near the western tip of the island they stopped and planted a garden at Kule Ndako, the "Wandering Board Game"^[3] named after a game they played on their trip.

One day they wandered farther inland and discovered there were other people living in the area, who did not have gardens or cook their food. They were hunters, skilled in herbal medicines, and eaters of raw foods, who knew many poisons and occult secrets. They lived with the

[3] Kule is a game played by inserting dedap seeds into a piece of wood in particular formations. It is traditionally associated with the sea worm festivities, and it is taboo to play it in the period preceding the arrival of the worms.

wild animals and wild spirits of the region [marapu la kandaghu ], and some of them were said to be witches and eaters of human flesh. Temba and Raghe made a peace pact with one of them, an old woman who lived in a hut made of bitter creepers [warico lolo kapadu ]. She had no fresh water, only brackish water, and no cured tobacco or dried areca nut, only fresh leaves and fruits. But she lived in a fertile valley and was willing to let them live beside her if they promised not to harm her or her husband. Temba and Raghe cut off bits of their fingernails and hair and scraped a bit of flesh from their tongues to be put into a bamboo flask as proof of their pact. They settled at the upper end of the village, in what became Mete Deta, and the old woman and her descendants settled in Mete Wawa. But the village was still empty and lonely.

Temba and Raghe invited Mangilo and Pokilo to join them in settling the new territory of Kodi. But Lendu and his brother, Atu Awa, were reluctant to part with them. Temba and Raghe played a trick on the boys to persuade them to move to the coast: they filled their drinking gourds with coconut water and offered them to the boys.

"Where does this delicious water come from?" asked Mangilo.

"It comes from the coast, where every day you can drink this sweet liquid," Temba and Raghe said. They did not tell them how parched the region could be in the dry season, or that fresh water was not available for much of the year. Each day, they fetched new coconuts to feed to the boys so that they would agree to join them in their new home.

Temba and Raghe wanted to adopt the boys, but their guardians in Bukambero did not agree. "If you take them away like that, you will be stealing our younger brothers, and that will be the beginning of an unending enmity, like that of cat and mouse. Our people will come to take heads from your people, and your people will come to take heads from ours."

"No, we do not want that, so we will give you something to compensate you for the food and betel nut that you have given to raise them." Lendu Myamba was given a huge gold breastplate, the marangga bali byapo ["breastplate of both sides of the river" the couplet name for the Kodi territory], to secure his blessing. That breastplate is still to be found in Wai Walla, Bukambero, as proof of the fact that the masters of the Kodi calendar spent their youth there. It also assures the people of both regions that they are not strangers to each other, and cannot take heads on the warpath.

Lendu had to provide a countergift for the gold valuable that he had received, and he also wanted to provide the boys with something that would bring them good fortune in their new home. He decided to give them the "plaything" they admired so much, but it came with ritual requirements.

"My home is too far from the seashore," he said, "so I will give these

worms to you so you can play with them like cotton baubles, amuse yourselves with string games. But there are also responsibilities. Once you release the worms into the sea, they will come back each year on a specific day, and you must count the years and measure the months to determine the proper day. They must be greeted with a mounted battle, the pasola , so you must find a village site where the horses can run without hurting their feet. You must teach the others to observe the taboos that come with welcoming these special creatures, and then you will receive good harvests and many descendants."

Mangilo and Pokilo took the worms close to the sea shore, but as the boys played, the worms got washed away in the river water.

"Where are the sea worms?" Temba asked.

"They have disappeared," the boys answered.

"Then go look for them, all over Kodi if necessary!" said Temba.

They searched all along the coast, and found that the worms had been caught by the roots and trunk of a banyan tree at Kawango Wulla [the "Moon's Banyan," near Tossi]. The worms were rescued and taken out to the sea at Kawoto, near the western tip of Sumba at Cape Karosso. But the land there was too rocky for horses to run without hurting their feet. Then they took them to the midpoint of the coastline, Halete, in the center of the line of ancestral villages called Pola Kodi, "the trunk of Kodi." There, the worms swarmed in great numbers and washed up on a sandy beach where they could be easily collected. A large flat grass field lay beside the beach, perfectly situated to receive the thousands of horses and riders of the pasola . The site was renamed Kapambolo Nale Hari, Karangga Rica Marapu ["Platform for Sacred Sea Worms, Beam of Spirit Pule Wood" the present site of the pasola combat]. This was the best site for the sea worm woman [Nyale] to be reborn.

When the trough was placed at this site, in the shade of the Kawango Wulla tree, the next day they found on the beach a great porcelain urn that had washed up from the sea. The urn was filled with cool, fresh water, which Mangilo and Pokilo drank, surprised that it was not salty after being tossed for so long by the waves. "This will be our village" they said, "and here will be the home of the sea trough and the urn that cannot be lifted." It was clear that after this long journey, the sea worms would remain at this site, which is where Pokilo and Mangilo built their own village, Tossi.

Narrated by Guru Kedu, Mete, Pola Kodi

This narrative uses the trope of a journey to establish both a series of areas that have some residual rights to participate in sea worm rituals and the primacy of the two oldest villages in Kodi, Mete and Tossi. Since it was collected from a descendant of Temba and Raghe, it places them in the

spotlight, stressing the events that made them the "first to claim the land, and first to build a village" (kapunge tana, tandai parona ). However, as in the story of Lendu from Bukambero, this narrative finishes with the sea worms being passed on to another owner. In some versions told in Tossi and Wai Kahumbu, the trip overseas to obtain nale is carried out not by Lendu but by Lete Watu, the trickster who exchanged fire for water in the narrative about the origin of the seasons (see chapter 2). Lete Watu was the younger brother of Mangilo, who later split off to found the village of Wai Kahumbu.^[4]

The long search for the proper place to hold the horse jousting combines practical considerations with political ones: the form of pasola is taken as a heritage from the overseas kingdom of Rato Ndimya, linking the promise of fertility that comes from the worms' swarming to a ritual condition that these visitors be welcomed with a proper spectacle. Lendu, Temba, and Raghe surrender their earlier rights in order to place the worms in a new location that is not merely close to the sea but at the very center of the line of ancestral villages along the western coasts.

The relationship of Mete and Tossi in this narrative is that between an adoptive father and his young sons. It contrasts sharply with the one we see in text #5, where Mangilo and Pokilo turn on their earlier benefactors and usurp the site of the "unopened land and round stones" (mboka tana, mbola watu ) where the ancestors first settled. But let us first turn to two other stories of change and disruption in the times of the earliest settlements.

TEXT #3: RA HUPU AND THE DISRUPTION OF THE REGION

After several years, the region grew more inhabited, but it became unsafe. Settlers along the western beaches began to notice that their pigs and chickens were disappearing mysteriously. The raids escalated, and soon not only small animals but also horses, buffalo, and even people were perishing. The trouble was caused by Ra Hupu, the younger brother of Mangilo.

When Ra Hupu was just a boy, he began to throw small nets of cotton string in front of the house. He threw the nets at grasshoppers, as a game. But when he caught grasshoppers, poor people began to die here and there. Their souls were caught in the webbing of the magical nets.

[4] Lete Watu strongly resembles the Weyewa figure of Yanda Mette (Kuipers 1990) and is also associated with two brothers, Anda Mangu Langu and Mete Mangu Dulu, who were, in some accounts, the first Sumbanese to arrive on the island. Some say, using a folk etymology derived from Indonesian Malay, that Anda, the elder brother, went first into the tides, while Mete "first waited" (Ind. mangu dulu ), then followed him. Others say that Langu and Dulu were the names of the wives who accompanied these two brothers to the new land.

His spirit was fierce, it was not good, so he could not control this magical power. When he grew up, he bought thread of many colors—black, white, yellow, and red. He wove nets from this thread, and smelted heavy weights for the nets from gold.

This time, when he went down to the seashore to cast his nets, they were powerful enough to endanger not just poor people but important ones. He would cast them in the sea and catch the bearded manduli fish and the mangata fish with a straight tail [couplet name for the nobles and wealthy men who became his prey]. Ra Hupu grew into a huge mar with a forelock of yellow hair who could breathe flames of fire. When others objected, he went on a rampage to punish them, sending out lightning bolts or fierce winds, which blew so furiously that all their tender young crops were destroyed.

The people of Kodi decided to meet in Tossi to establish a system of order that could control his rampages. "We must send you off to cool your head. You will go into exile in the next valley, where you will live at the corner of the river and the edge of the tides," said Rato Mangilo.

The meeting was held in the Uma Batango (Council House) in Tossi, presided over by Temba and Raghe, known as Temba who established the villages and Raghe who owns the land [Temba tandai parona, Raghe kapunge tana ], since they were the founders of the region who established the first rights to land. All of the lands in Kodi were divided between the ancestral clans and houses, and a single mother-father village had to be chosen to keep order and oversee the annual cycle of the months.

At this meeting, Ra Hupu was banished to the other side of the river, to "soak his head and cool his liver" in the fresh waters that flowed down from the highlands. He was told to

With these words he was exiled to the other side of the embankment [Bali Hangali], where he and his descendants would live on their own territory, separate from the original villages of the "trunk of Kodi" [Pola Kodi] but still acknowledging their ritual preeminence. He was forbidden to steal from his fellow Kodinese, but his fierceness and fire were given a new focus in the skull tree erected in the center of his village, where the heads of enemy highlanders could be hung. On the warpath, his magical weapons and control of the winds could once again be used—but in Kodi they would not be unleashed for as long as the Sea Worm Priest [Rato Nale] remained confined during the month awaiting the swarming of the worms.

When he established his residence at Mba Ronggo, Ra Hupu moved into the territory of Pala Kawata, the python who had helped Lendu overseas. Pala Kawata came originally from the highlands, but in certain years he would visit the drier regions of Kodi, bringing with him abundant rains and unusual fertility. Pala Kawata took a fancy to Ra Hupu, so he adopted him as his nephew and moved his own residence farther upriver to the gates of the river and the source of the swamps [binye loko, mata rende ]. The descendants of Ra Hupu have preserved a special ritual relationship with Pala Kawata, so that if they need more rainfall they can ask for it by shaking the trees at Mba Ronggo in the early hours of the dawn, or by making offerings of betel and chicken feathers at the sites associated with their ancestral names.

Narrated by Ngila Pati, Bondo Kodi, Mbali Hangali

This text serves largely to justify the land rights of the people of Bondo Kodi, who established rights to all the territories that border on the river. This right is expressed socially by the fact that they must receive the heads of all sacrificed buffalo and pigs throughout this region. It also comments, obliquely, on the problematic trope of first owners versus later owners that we noted in the first two narratives, since in this text Temba and Raghe preside over the consensus meeting, but it is held in the Council House in Tossi.

Ra Hupu, like Lendu, is an adventurer who is allied by fictive kinship with the python. While all sides agree that Ra Hupu was powerful and disruptive, his descendants stress the ways in which his magical powers could be put to the service of others—in obtaining rainfall or enemy heads, for instance. Detractors emphasize the losses suffered by his fellow villagers, which led to his exile. Because Ra Hupu and his descendants still control lightning and rainfall, members of other villages must go to them and present a sacrificial animal to request that rites be performed to bring rain or control lightning bolts.

The text views the central consensus agreement that established the ceremonial system from the periphery, the outermost edge of Greater Kodi. It also presents a part of the rituals of agriculture, whose more complete version is found in the next narrative.

TEXT # 4: THE CREATION OF RICE AND OF THE BITTER AND BLAND MONTHS

The early settlers of Kodi were having a hard time making a living. While the soil was rich and fertile, rainfall was erratic, and long periods of hunger plagued the first generations who opened up gardens there. A diet of root crops, corn, vegetables, and beans was not enough to fill their bellies or to give them the sense that they had eaten at all.

Pala Kawata, the giant python who lived in the highlands, decided to make the supreme sacrifice to feed the others. His own daughter, Mbiri Koni, was nearing marriageable age, but she found none of her suitors to her liking. One day he came home and told her to dress up in all of her best finery. She put ivory bracelets on her wrists, a string of colored ceramic beads above them at the forearm, a gold hamoli pendant around her neck, and wrapped herself in fine indigo textiles dyed with intricate patterns. She presented herself to her father, glowing with girlish beauty, and he led her off into the center of his garden plot to meet her new "husband."

When they arrived at the platform built to hold the rice seeds, he took out his small harvesting knife and killed her. Then he cut her body into small pieces and buried it throughout the garden. He rinsed the blood from his loincloth, wiped his knife clean, and concealed it in the folds of his waistband. He went home to his wife, without a word about what had happened.

After their daughter did not return for two or three days, her mother grew very worried. "Perhaps someone has carried her off without our permission" she said. "Perhaps she has eloped."

"No, my wife," Pala Kawata answered. "That will never happen. Your daughter may not be visible now, but she is with you in the gardens, and she will show her face after four days."

On the fourth day, the mother went out into the gardens to call her daughter. "Ooooo, Mbiri Koni, where are you?"

"Here I am, Mother," came the daughter's voice, but she was nowhere to be found. Finally, the mother went to the seed platform and called again. "Ooooo, Mbiri Koni, where are you?"

"Here I am, Mother," answered the daughter's voice, "don't you recognize me? The mother turned in confusion to her husband, who told her: "Your daughter's body has returned to us in the small green sprouts that you see just breaking out of the ground at your feet. Her ivory

bracelets have returned in the white tubers at the side of the garden, her ceramic beads in the beans, her gold pendant in the corn ears beside the rice. She has been transformed into the garden crops to save us all from hunger."

Her mother was so distraught she refused to eat any of the new crop. She became the old woman who guards the upright stone at the entrance to each garden plot [waricoyo ela watu kareke ] and stayed in the garden with her daughter. Each year, when the rice seeds are planted, we say that Mbiri Koni has died and her mother goes into mourning for her.

We, the Sea Worm Priests, are also in mourning—we follow the rules of kabukuto kalalu , we sit in silence and brood. We wait throughout the bitter months, and do not allow anyone to beat the gongs or disturb the young crop. Then, when the rice crop is tall and golden and full [pa ihi ; lit., "pregnant"], we hold the sacrifices to allow the harvest to happen. Once the sacrifices are done, the month is bland [kaba ] and the crops can be eaten.

Narrated by Ra Ndengi, Rato Nale, Bukubani

Different versions of this story play on the changing identity of the "knife" and the person who wields it. In one, the girl meets a handsome young man, Rato Malogho, who is in fact a field rat. She marries him and becomes pregnant by him, but when he tries to take her home to his village under the ground, she cannot fit through the hole. He kills her and drags her body in piece by piece. Once she is completely underground, she remains there for four days before being reincarnated in the form of the rice crop. In another, Mbiri Koni is sacrificed by her brothers and sisters who migrated to the island with her, but her erstwhile suitor is so saddened that he chooses to become a field rat in order to join her underground.

In all these versions, the fate of Mbiri Koni clearly parallels the fate of Inya Nale, the foreign beauty whose body was transformed into the sea worms. Each is a bride whose body is dismembered and transformed to create a resource needed by the whole region. The "gift of a woman" in marriage is thwarted, but a symbolic equivalent—the source of fertility and vitality—is substituted.

In the following pivotal text, we learn of the consensus agreement that leads to the regulating of the harvest by the priests of the sea worms.

TEXT #5: THE KODI MONTHS AND THE DIVISION OF RITUAL TASKS

When the leaders of Kodi gathered to divide up the lands and the various ritual tasks, they did so by using the divinatory powers of objects. Four ritual objects were brought to this meeting to make the selection: the

kule game with eight holes, the buke stakes and darts, the kadiyo top made of buffalo horn, and the kalayo discus. These were all divination tools which the first ancestors brought with them from Sasar.

Representatives of each of the ancestral villages played each other in the traditional children's games:

Mangilo and Pokilo defeated all the others, easily smashing their opponents' tops and outdoing them in contests of skill and strategy. This established their right to serve as the "mother-father" figures of the region.

There was then a division of power between the two brothers based on birth order. Mangilo, the older one, was told:

Given the most sacred of the ancestral heirlooms, Mangilo retired into relative seclusion. He served as the guardian of the calendar and the yearly seasonal cycle, but took no active part in enforcing the rules that he established.

Through his calculations, the times of planting and harvesting were controlled and agricultural activities were coordinated throughout the

region. No one could hold feasts once the preparations of the gardens had begun, and none of the tender young plants could be picked before they reached maturity. Rato Mangilo received the lightning stones used to direct lightning bolts at offenders of these prohibitions, and promised to remain in strict confinement in the months preceding the coming of the sea worms.

Rato Mangilo then named the twelve months of the Kodi calendar [see table 1]:

The first month of preparations before the sea worms came was called [1] Nale Kiyo, and it was during this month that he performed sacrifices to make the tree crops of pinang and coconut ready for the harvest [kaba wei kapoke ; lit., "bland waters on the sprouts," since water from the sacred urn was poured on them]. The month of the worms' arrival was called [2] Nale Bokolo, and it was when the pasola combat was staged. It was preceded by offerings of betel nut on the tombs of one's ancestors [hengapungo ] and followed by the sacrifice of a chicken in each lineage house [tunu manu nale ]. Then came [3] Nale Wallu, for the "remaining" [wallu ] worms of the second swarming. [4] Bali Mboka, the "return of small buds," marked the beginning of the harvest season, followed by [5] Katota Lalo, the "small red flowers" of a common bush, and [6] Katoto Bokolo, the "large red flowers." Feasting on a large scale was not supposed to begin until [7] Rena Kiyo, the preparations for celebrations, then [8] Rena Bokolo, the larger celebrations. Feasting could continue through the period called [9] Padu Lamboya, but had to stop once the priests had performed the sacrifices necessary to start the Kodi "bitter months," or [10] Padu Kodi. Silence had to be observed throughout [11] Habu, the "nesting" [habu ] season for birds, and [12] Mangata, the month when the white flowers of the mangata bush were visible.

His younger brother Pokilo received a fast-traveling horse [ndara halato ] and was given the task of patrolling the area to guard against theft or trespassing onto another man's gardens. He was told:

His was the task of roaming through dangerous border territories, mediating between litigious parties, and leading war parties, if necessary,

(Table continued on next page)

(Table continued from previous page)

against invading groups. He was the "master of force" whose powers counterbalanced the elder "master of time."

Then they divided up into separate villages. Each man of importance was told to establish his own village so that they would have enough villages to intermarry. The villages took the name of the site they occupied, and the tree which served as their altar. The lara marapu , or "path of the ancestors" recited in each village, began with Temba and Raghe, the founders of the land, and then went on to name those who dug the first pillar holes and put together the foundation stones at each village site. The indigenous spirits of the earth [tagheghe mori lyodo, wondo mori pyada , "forest fowl lord of the region, megapode lord of the area"] also received invocations and sacrifices, as did the deputy hamlet deity [inya mangu tana, bapa mangu loko ] in the gardens.

There were twenty-four villages formed from the "trunk of Kodi" (Pola Kodi), and seventeen on the other side of the embankment (Mbali Hangali). All of these villages come together now to sacrifice chickens to the sea worms and offer the first fruits of the harvest to Toda.

Narrated by Ra Holo, Rato Nale, Tossi

This version of the central narrative of the division of powers and the establishment of the calendar was obtained from a 1980 interview with both the Rato Nale of Tossi and a retired Kodi minister, Pendita Ndoda, who was descended from the Sea Worm House and helped with translations and transcriptions. It is the closest thing to a "charter" in the Malinowskian (1954) sense for the Kodi ceremonial order as it exists today. Parts of this narrative may be strongly contested, however, and there are many different interpretations of its relevance to present action.

The most widely known and quoted part of this narrative is the division of power between Rato Mangilo and Rato Pokilo. These couplets provided the first model of a diarchic divide, which was then repeated, in subtly different ways, in a number of other areas. The Kodi assertion that it provided an "unchanging template" is at least partly supported by the fact that the only previous recording of the narrative, by Onvlee and Kapita in 1932, cites the couplets in almost exactly the same form as I heard them in 1980, even though the wider political situation was quite different at that time. A consensus that Tossi had the role of ceremonial leader because of its control of the calendar remains, along with ideas that Bondo Kodi has access to rain magic and Toda receives the first fruits of the harvest. All other ancestral villages are the ari ana , or "younger siblings and children" of these founding villages and trace their ancestry to their founding figures.

Beyond this core template, however, interpretations of the events that

occurred at this distant meeting vary widely. The first alternate interpretation that I will develop emerges from the background to Onvlee and Kapita's visit to Kodi in 1932, and illustrates what was at stake in manipulating versions of the ancestral agreement. The Kodi transcripts of texts recorded by Onvlee and Kapita include a version of this story, narrated by Haghe Tyena of Mete and Rehi Kyaka Ndari of Tossi, that repeats the peace pact between Kodi and Bukambero established with the transfer of the breastplate and adds that the first gardens were made at Kule Ndako. These tellers also said that Temba and Raghe brought sacred stones with them from their homeland to the west, which they used to consecrate the land and make it "bland" (pakabaya ) enough to cultivate (Onvlee and Kapita 1932).

The interview was dominated by Raja Ndera Wulla, the district ruler at the time, who summoned the two storytellers to give their account, but "insisted on speaking first while all others remained silent" saying that he had invited the others "to confirm his accounts, in case there were suspicions that he had made misleading statements" (Onvlee 1973, 57). The primacy of Tossi had to be stressed in his presence, and ambiguities about its position played down. The raja finished the interview with an invocation of his ancestor, Rato Pokilo, and an assertion that "I was given the land of Kodi from the beginning of the earth and stones, and if I now command, it did not begin when I first held the Raja's staff, but because Pokilo and his descendants have always been those who commanded from the time of our ancestors and forefathers" (Onvlee 1973, 59).

In a later recollection of the 1932 visit, Onvlee reports that Haghe Tena came to see him privately in Waikabubak and "presented a story in which the relationships were rather differently represented" (Onvlee 1973, 59). Mangilo and Pokilo received the sea worms and their ritual offices as a gift from Temba and Raghe, who sought them out as companions in order to defend themselves against enemy attack. Mete was the oldest village and the real "elder brother," whose authority was based on the "unopened land and the round stones" (mboka tana, mbolo watu ), while Mangilo and Pokilo were just boys who received these important ritual offices without knowing what to do with them. Haghe Tena described them as "locusts scampering on cassava leaves, bats balancing on banana leaves" (papa enggena kabala rou katete, pa pandeta panighe rou kalogho ), meaning, as Onvlee (1973, 60) says, "they did not have themselves to thank for the good place that they currently occupied."

Reinterpreting this account with the benefit of hindsight, I would classify it within the same trope as the Bukambero story. We have, on the one hand, a group that maintains its prior rights to the worms as a source

of traditional authority, and on the other, a group that stresses the significance of the transfer while implying that the first site was unworthy. During my own fieldwork, tensions between Tossi and Mete were muted, presumably in part because Tossi was no longer the center of district government. But I did record a narrative about how the people of Tossi tricked the people of Mete into abandoning their original village site and moving a short distance to the north. The source of this story, however, was not Mete, but another faction within Tossi: the storyteller was himself a descendant of Rato Mangilo and told the tale as an example of the cleverness of his ancestors, who used their superior ritual position to appropriate the most sacred village site in the area.

TEXT #6: HOW TOSSI AND METE CHANGED PLACES

The place where Tossi is now located once belonged to Mete. It is the oldest village site in Kodi and was founded by Temba, who built the first villages, and Raghe, who owned the land. Tossi was originally located to the north of them, and it was founded by Mangilo and Pokilo.

One day, as he took his horse to the pasola , Rato Mangilo rode past Mete and began to think: "Here we are, the guardians of the most sacred objects in the land, but we do not have the best village site. It would be much better if we were right at the edge of the pasola field, so as soon as we untied our horses we would be ready to ride out to receive the sea worms." He thought and thought, but found no solution. So he told Rato Pokilo to call Lete Watu.

Rato Mangilo told him: "I want you to figure out a way for us to change places with Mete, so they won't be closer than us to the field."

Lete Watu answered: "If you want to exchange villages, first go to a feast and tell them not to give any meat or rice to the people from Mete. When they ask why they haven't received their share, tell them that it has been a long time since they held a feast. Goad them into agreeing to feed the others, because once they hold a feast in their own village, we can outwit them and take over their village site."

Embarrassed by what they heard at the feast, Temba and Raghe of Mete planned a two-day feast [woleko ]. They invited people from Kodi, Bangedo, Bukambero, and Karendi. Lete Watu built a platform outside the village gates to receive them, so they wouldn't enter the village. They all received food and betel at the gates and then settled down to sleep. Lete Watu called over four or five people and told them to take the swords of the Karendi and Bukambero people and stab all the pigs, horses, and buffalo. After killing the animals, they put the swords back in their scabbards and did not wipe off the blood.

The next day, the guests were horrified to see that all the animals were already dead and rotting, and Temba and Raghe were deeply embarrassed. They called out, "Stop the dancing, stop the slaughter. What has happened here? Who killed all these animals?" Their bodies were so numerous they couldn't be counted. Everyone gathered together and members of each region were asked. "No, we didn't do it," they all answered. They called Lete Watu to the village gates, and he suggested that they look at people's swords to see if they bore traces of blood. They looked at the swords of Kodi, Bangedo, and Balaghar and found no blood. They looked at the swords from Karendi and Bukambero, and they were full of blood. The people of Karendi and Bukambero were ashamed and ran away to settle at the farthest ends of the domain.

Then Lete Watu said, "This feast was a failure. There must be a curse on this village site:

The village site is no good. If there is anyone else who would agree to come here, I would advise you to trade with them."

At this point, Rato Mangilo said: "Perhaps we are the ones who should help you out, since we are the mother-father village, the mother of the water jar and the father of the moon net. If you wish, I promise to drag the funeral stone that we have prepared for Temba until it stands by the former site of Tossi. And we will cut down the great tree of Tossi that stands by our house and replant its sprouts at our own new village site down here."

So they agreed, and the two villages changed their sites, with Temba moving back, away from the pasola field, and Tossi moving down. They sacrificed a buffalo and pig to establish themselves in the new location and banish the curse that afflicted the site. Mangilo brought the sacred water jar and the moon net and made his new home the place

Narrated by Ra Katupu, Tossi, Pola Kodi

This narrative accounts for two things: the estrangement and isolation of two peoples who intermarried with the indigenous peoples of the area, the people of Bukambero and Karendi; and the displacement of Mete at the oldest village site. In both cases, it celebrates the triumph of crafty newcomers over earlier inhabitants.

The narrative told in Tossi about the acquisition of the sea worms does not refer to a change in village sites. The location of the Kawango Wulla, or "Moon's Banyan" is assumed to be unchanging. It now stands near the present village gates, at the entrance to the pasola field, about two hundred feet from the beach. If the waves once washed up at its roots and deposited sea worms there, marking the spot on which the village was to be constructed, one wonders why Mangilo and Pokilo first chose to build their houses farther inland. Ra Katupu, the narrator of this version, reconciled it with the story told by the Rato Nale by saying that "Mangilo and Pokilo were young boys when they first came, and still ignorant. Once they grew up, they saw that Temba and Raghe had a better location, and so they decided they wanted it."

Contradictory assertions about the centralization of Tossi's power, said to be unmoving yet at the same time continually tested and reaffirmed, emerge in the events of the early twentieth century (see chapter 4). Before the colonial period, such assertions cannot be directly associated with remembered history, but political processes contesting Tossi's position through often conflicting interpretations have left traces in narratives such as this one. The emphasis on trickery and guile is much less evident in the next narrative, which details the agreement to extend the calendrical system beyond the Greater Kodi valley to the distant region of Balaghar.

TEXT #7: THE CREATION OF A BRANCH OF
THE SEA WORM FESTIVITIES IN BALAGHAR

Many years after the first settlement of Kodi, the population filled the territories of Kabihu Pola Kodi and Kabihu Mbali Hangali, so people moved across the river into the valley of Bangedo and even farther across an estuary into Balaghar. Since they were still descended from the original founders, they still had to travel back to Tossi and Bukubani for the calendrical rites.

One day, Pyunggero, from the village of Wainjolo Wawa in Balaghar, asked for a meeting with the elders of Tossi. Rato Mangilo and Rato

Pokilo, now old men, attended, as well as Temba and Raghe, Ra Hupu, and many others. Pyunggero stood to make a speech;

He was asking the officials of Tossi and Bukubani to delegate their powers to people in Balaghar, who would serve as lower-ranking ritual officers but would be able to preside over their own ceremonial calendar. If they agreed, there would be a division of territories expressed as a separation of the food served at feasts, which would henceforth be "meat cut up separately, rice heaped in a new plate" [ropini kabiyo, hanggani ngagha ].

Rato Mangilo and the elders of the main ritual villages in Pola Kodi considered this request, but said that it would be necessary to test the leaders of each of the villagers in Balaghar, as they had been tested themselves, to see who was capable of controlling the calendar:

The method of divination that was required for this occasion, however, was not the usual interrogation with a spear or reading of animal entrails, but once again the playing of children's games, which combine elements of skill and chance to determine the best candidates.

So a competition was held, first with the kule board and the buke darts. The representatives from Balaghar played against the officeholders in Tossi. Rato Mangilo played kule against Lere Ura, from Waingyali, the oldest village in Balaghar. If Rato Mangilo had won, it would have been a sign that the spirits of the heirloom objects concerned—the holy branch and sacred bough, the ancient water jar and the net of valuables—were not in agreement and could not be moved. But Lere Ura won, taking all of Rato Mangilo's seeds. Pyunggero played buke against Rato Pokilo, and defeated him.

The people of Tossi called for a rematch, this time with the tops and discus [kadiyo, kalayo ]. The new series was played by Lete Watu, from the village of Wainjoko, and Rato Jadi, from Kaha Malagho. Once again, the challengers won—but they won a victory that was divided between two players, rather than a simple conquest on all fronts. So the sacred prerogatives of Rato Mangilo to control the annual calendar and the agricultural rites were divided between two villages in Balaghar—Wein-gyali for the nale offerings, and Kaha Malagho for the padu offerings. Likewise, the powers of Rato Pokilo to control the borders and control the rains were divided between Wainjolo Wawa and Wainjoko.

Lere Ura from Waingyali received a small trough to hold the sea worms and a trap to scoop them up, just like the ones that Lendu had brought from overseas and given in turn to Mangilo. Rato Jadi of Kaha Malagho received a twig from the sacred kapok tree (Wei Marongo Ra-ra), which stood by a source in Tossi, where offerings were placed to begin the ritual silence of padu . Rato Pyunggero received a fast-traveling horse [ndara halato ], of the same descent line as the one used by Rato Pokilo, which he could use to patrol the region and enforce land boundaries. Lete Watu received a bamboo tube used to ask for rain [onggolo ura ]. All of these objects were brought back to Balaghar, and feasts were held to consecrate them in their new home.

Narrated by Maru Daku, Wainjolo Wawa, Balaghar

This text legitimates the transfer of the sea worms to a new region by "summarizing" events from many of the earlier narratives, especially text #5, and placing them in a new context. Perhaps because of its skill in

borrowing elements from the familiar stories of the founding of Kodi, this narrative provoked the suspicion of various local people who listened to it on tape. Their ambivalence may reflect an uneasiness about the legitimacy of the Balaghar rites. Despite the acknowledgment that the source of their calendrical ritual was the older "trunk of Kodi" many people in Greater Kodi dismiss the smaller nale rites held in Balaghar as insignificant, asserting that nothing of importance was transferred.

The villages of Balaghar were the most remote in all Kodi for most of this century, since no road reached them until 1988. As a result, this most recently settled region remained a bastion of tradition, where less than a fifth of the population had converted to Christianity. The calendrical rites of Balaghar, though they are derivative in terms of the mythical mandate, are performed with very full participation of the population and a passionate involvement, which contrasts with the more desultory performances in Tossi, now a village of many converts and extensive contacts with schools and government agencies. It is precisely because of the continuing importance of ancestral rites in Balaghar that many people suspected this detailed narrative was an "invention of tradition"—a bit too finely wrought to be genuine.

The narrative was collected from one of my finest informants, Maru Daku, who was among the first Kodi converts to Christianity in the 1920s but then returned to traditional worship late in his life (Hoskins 1985). Although he was recognized as one of the best ritual speakers in the region and a skilled compiler of ancestral lore, he also provoked mistrust and criticism. As someone who had been a practitioner in both systems, the Christian and the pagan, he was said to have been punished by his ancestral spirits and the Christian God for "worshipping two sets of gods"; this charge diminished his credibility in reciting the most sacred stories. His gift for rhetorical flourishes was seen as "adding too many spices to the stew" and thus concealing the basic flavor and consistency of their original form.

Since each time a story is narrated it takes a slightly different form, Maru Daku's critics may simply have been accusing him of being too effective at his task of providing coherence. "He has been studying our own stories,' they said, "and so he knew what to put in his account to make it sound good." Maru Daku answered his critics with the physical evidence referred to in the narrative: the sea worm trough in Weingyali, the sacred banyan in Kaha Malagho, the bamboo tube in Weinjoko. A narrative attached to an object can be proved true if the object is efficacious. Few other challenges were mounted.

Problems of Legitimacy, Authenticity, and Hierarchy

Consideration of these seven texts opens up numerous questions about the bases of ceremonial authority in narrative. Almost all of the texts contain, in some form, a trope embracing a tension between an earlier owner of a sacred object or site and a later one. The earlier owner can claim primacy, the later can try to show that he has achieved control. All of these narratives speak of how Kodinese ancestors have gradually gained a certain mastery of their surroundings through a social consensus about the proper use of objects with magical properties.

This transfer of power involves one of two processes: delegation by an older authority to a younger executor or usurpation through trickery. The first stresses genealogical priority and ascribed position, the second individual action and achieved rank. Often, the difference between the two is a matter of interpretation: Thus, the descendants of Temba and Raghe say that they gave the sea worms to Tossi out of generosity and a desire to help two orphaned boys. The descendants of Tossi, by contrast, say that they deserved the gift because of their superior skills in games, which allowed them to trick their benefactor into giving them even the site of his own village. The people of Bukambero, for their part, say that they originally acquired the sea worms but knew they should pass them on to people who lived closer to the beach. And the people of the coastal villages maintain that they were always the rightful owners, so their use of false accusations to drive out the earlier inhabitants was justified by hierarchical privilege.

The differences between the two modes reflects the variety of interacting oppositions that constitute hierarchy. The locus of authority is opposed to the locus of action, and the production of asymmetries of power emerges as "not a principle but an outcome, the result of the application of several principles" (Fox 1989, 52). The "order of precedence" produced by the calendar thus comprises heterogeneous components, which include genealogy, spatial location, personal qualities of skill or bravery, and fortuitous circumstances. Power is conceived not as a single entity but as something that is immanent in a plurality of existing social relations (Foucault 1978).

Power is transferred in these narratives by means of three specific devices: descent (reckoned through persons, places, or objects), game playing, and offerings to obtain the fertility of the land. Descent is established either by the birth of the ancestor within a particular line or by the breaking apart of a whole object into parts that can be moved. Breaking off a sapling from the great tree at Tossi and planting it in Balaghar, for

instance, establishes a plant line of descent, which substitutes for the absence of a human line of succession. The part is always less than the whole, however, and the transfer is thus of a diminished and subservient power. When descent is the criterion of transfer, an accompanying idea of devolution, of a lessening of status as one moves away from the source, is implied.

Such is not the case with things acquired through games and contests. Playing at tops and discus, the kule board and the buke darts, is a very serious business. In the ritual formulas still pronounced today, priests say they are only children who "hold the top's string, grasp the discus's net." They evoke children's games to protest their innocence and inexperience in relation to the ancestors, but also to remind their listeners of the primordial contest in which their ancestors emerged victorious. The theme of "playing" at ritual is related to the idea that the process of acquiring ritually charged objects is also part of the human process of learning and maturation. Mangilo and Pokilo, the ancestors of present-day calendrical priests, appear first as small children, tempted by the sea worms as playthings but not yet responsible enough to control them. They must first find the proper site for their village, then prove themselves through games. Once they have achieved the maturity to serve as the masters of time and of force, they agree to pass on some of their ritual prerogatives to junior deputies—-but only in a divided and diminished form.

As the proper locations are sought and the proper intervals established, the need for rain and new crops arises. In several cases, the "proof" that one's ancestors had legitimate control of a certain territory is established by a ritual demonstration that rain can be obtained to nourish the land. The art of rainmaking is hardly a trivial one in a land as parched and prone to drought as Kodi. Rain is associated with disruptive, rebellious ancestral figures: Ra Hupu, Lete Watu, Rato Bokokoro—a thieving fisherman, a playful trickster, an easily duped guardian.

The power of the rainmaker is at the opposite pole from the power of the lawgiver, in a relationship similar to the elder brother/younger brother polarity that binds Mangilo and Pokilo. It provides the basis for the ritual cooperation of older source villages—the locus of authority—and younger peripheral villages—the locus of action. Rainmaking is represented as a violent assault on the heavenly kingdom, which explains why in mythic accounts it is associated with warfare and headhunting. The control of seasonal rains, thunder, and lightning is invested in stable, unmoving ritual authorities, whereas the disruption of these powers is a necessary infusion of energy associated with the young, mobile manipulators.

Time within the mythic narratives is constructed in relation to temporal

sequences in the annual cycle of ceremonies. It is a common theme throughout Southeast Asia that a period of mourning is ended with a celebration of the renewal of life. Both Nyale and Mbiri Koni are female sacrificial victims whose bodies are transformed into food. Since the harvest of the sea worms anticipates the harvest of rice, Nyale's reincarnation precedes Mbiri Kohl's; indeed, the spectacle of many-colored sea worms actually depicts the abundance of the harvest in iconic form.

Although the past is given great value, simple priority in time is not enough to assure hierarchical supremacy. An analysis of these narratives shows that no one single principle is dominant; rather, several principles must be used in combination. The "order of precedence" that finally emerges bears the traces of past conflicts and negotiations, which are not totally obscured in an idealized collective past.

Different Tellings of the Tales: Historicizing These Narratives

In my opening remarks, I noted that the narratives are composed of a mixture of prose and poetic couplets. Most of the action happens in the prose sections, as well as all of the conniving, scheming, and trickery. The couplets in verse represent primordial statements, "contracts" that supposedly have been passed down, unaltered, from one generation to the next. Examining the records we have of the tellings of these narratives, we find this claim to be justified to a certain extent, but it also becomes clear that the continuity of the literal form of a traditional couplet is quite separate from the continuity of its interpretations.

Since there are no occasions on which all of these narratives are brought together to be performed, compared, or heard, their general consistency is impressive. Despite minor differences, members of all villages agree that Tossi was the ceremonial center and that the most powerful objects were stored in certain ancestral houses, though they may have followed a long and circuitous route to reach them. I suspect that the reasons for the generally high degree of consistency lie in the fact that the Kodinese are very interested in their own history and enjoy discussing and comparing stories.

Conclusions about the current political position that are drawn from these stories, however, vary widely, for the narratives are themselves involved in historical processes and reshaped by them. It is, for example, interesting to compare the views of Kodi's past as recorded by Onvlee and Kapita in 1932, by Van Wouden in 1951-52, and by myself in 1980. Three of the seven narratives in this chapter (nos. 2, 3, and 5) were also recorded,

in clearly recognizable form, by Onvlee and Kapita; they deal with the settlement of the region, the exile of Ra Hupu, and the division of tasks in Tossi. The couplet sections that detail the division of tasks between Rato Mangilo and Rato Pokilo appear almost verbatim in their transcriptions· Although the 1932 texts are somewhat more disjointed than the ones I was given—briefer, choppier, less detailed—this may be because I told my informants I would try to construct a longer, continuous political history from the stories they shared with me. In soliciting the narratives I specifically asked about relations to other stories I had heard, requesting that the narrator make these explicit. My texts were recorded on tape, which allowed the narrator to speak at his usual rhythm without having to pause while words were transcribed. In practice, however, narrators often paused after speaking to repeat the verse passages, checking that I got them right because of their greater importance.

Van Wouden's account of his stay in Kodi does not include texts, but it has many references to the political order and ancestral precedents.^[5] He was told that Mahemba, Pawungo, and Balaghar, the three kabihu of Kodi Bangedo, were all equal and "each of them separately performs the nale ritual," but in Kodi Bokol the villages of Mbali Hangali were superior to those of Pola Kodi. Bukambero had a controversial status. Some people described its inhabitants as "unimportant people who retreated into the hills;' while others said it was "a venerable kabihu from Sasar" (Van Wouden [1956] 1977, 5).

Within Mbali Hangali, Van Wouden describes a division between two halves, also called kabihu , of which one, Bukubani, is associated with "religion" and the other, Barada, is associated with adat , or "custom": "The ancestral house of the first section is called the office for religious affairs, that of the second is the office for customary law." He continues: "The meaning of religion for Bukubani is clear· In the first place it refers to the celebration of the nale ritual. The most important functions of the adat are the determination of the calendar and the ordering of the seasons. Furthermore, Barada was seen as the meeting place of all Kodi Bokol. . . It may be noted that the sections of the two halves of Kodi Bokol intermarry" (Van Wouden [1956] 1977, 6). The passage is obviously a version of the division of power between Rato Mangilo and Rato Pokilo, and reflects the fact that within Tossi there is a division between Tossi

[5] In my discussion of Van Wouden's article, I have adopted his spellings of Kodi names to be the same as my own to avoid confusion. For the record, he wrote Bukambero as Buka Mbero, Bukubani as Buka Bani. I have followed the spellings currently used by government officials and the literate inhabitants of these areas, although it is reasonable to assume that the names were once binomials.

Bukubani (descendants of Rato Mangilo) and Tossi Barada (descendants of Rato Pokilo).

When I discussed these passages with Kodi informants in the 1980s, however, they were unanimous in denying (1) that nale rites had ever been conducted in Pawungo and Mahemba and (2) that the division of Bukubani and Barada existed outside of Tossi itself. The supposed "superiority" of Mbali Hangali could be interpreted as the superiority of Tossi as a ceremonial center, but it did not imply inequality between the peoples of the other villages of Pola Kodi and Mbali Hangali. Bukubani and Barada are the two "halves" of Tossi, representing the descendants of two founding brothers whose progeny can now intermarry. While they are descent groups, not kabihu , they do represent a complicated ritual division of powers (see chapter 4 for further discussion of later developments). These errors on Van Wouden's part were linked, I believe, to the political situation in Kodi in 1951-52 and possibly reflected misunderstandings of the facts as they were presented by local informants.

Van Wouden had great skills as an ethnographer and insight into Eastern Indonesian social formations, but he came as the guest of the current district administrator, Hermanus Rangga Horo, and stayed in his house. Horo, who continued in his position after serving as the third raja during the brief Dutch return to power in the late 1940s, came from the region of Bangedo. During the period of his administration, a small pasola had been performed in the grasslands between his ancestral village of Rangga Baki and Ratenggaro. The performance was a display of his power and influence, but it was not linked to offerings at a separate nale house or to any independent ritual control of time or the calendar.

Despite Van Wouden's position as a guest and the fact that he stayed only two months, his article is extremely valuable. It is not surprising, however, that he accepted an interpretation of the Kodi polity that, consistent with the aspirations of his host, emphasized the ceremonial independence of Horo's homeland Bangedo and did not acknowledge the priority of Tossi or its links to a large number of external, even foreign sources of power—the distant lands of Java and Bima, the earlier populations of Bukambero and Karenali, the competitive structure of neighboring regions. Van Wouden's ([1956] 1977, 19) conclusions about the structure of the Kodi polity can be quoted to illustrate a compelling, if inaccurate, vision of a totalizing polity:

Kodi Bokol can certainly be considered a community that we could call a "tribe." It has a single nale house. There are clearly defined oppositions and an explicit division of ritual tasks between its con-

stituent parts. Both these oppositions and ritual divisions are the expression of a total unity. Virtually all marriages are contracted within the community. If the sea were to swallow all of western Sumba except for Kodi Bokol, it would leave a completely self-sustaining unit in terms of social and religious life. Only the ritual contests with Bangedo would be missing.

In fact, in the 1950s, Kodi Bokol already had a long history of marriage alliances with Endehnese visitors, borrowed heirloom objects from overseas, and wars and confederations with neighboring domains. The nale cult was itself divided between two ritual houses, in a tense balance of powers that. will be the subject of the next chapter. Even in his misunderstanding of the rather confusing assemblage of stories about the past that were presented to him in 1951-52, Van Wouden pulled out a theoretical insight of considerable force. He noted that loose ends remained for the area of Bangedo, which broke up this totalizing vision into competing claims. Turning away from his earlier view that each Eastern Indonesian society could be viewed as a "closed society with a fixed number of groups" expressing "a specific type of social structure characterized by double descent" he came to see that regional variations were "complex and encompassing historical and sociological questions" (Van Wouden [1956] 1977, 148) whose dynamics were inscribed in a wider network of relations of insiders and outsiders local traditions and foreign influences.

4
The Past in Objects

The Colonial Encounter

Vaygu'a [shell valuables] are not indifferent things; they are more than mere coins. All of them, or at least the most valuable and the most coveted, have a name, a personality, a past, and even a legend attached to them, to such an extent that people may be named after them.
Marcel Mauss, The Gift

Objects are not what they were made to be but what they have become.
Nicholas Thomas, Entangled Objects

Local knowledge of the past is organized not only in narrative, but also in the visual and tactile traces left by past events: heirloom objects, features of the landscape, the special relationship established with a particular animal or location. This chapter begins with the "archive of the past" that is found in the Kodi world of objects, examining first the principles by which that world is constituted and then how certain objects have changed in their relations over time. The origin narratives presented in the last chapter all deal with the period when Kodi society was coming into being, the period "of former times and earlier peoples" (la mandei la maulu ), which is nevertheless perceived as homogenous with the present and described "historically" rather than "mythically."^[1] In examining the role of objects to represent the past, we move into the realm of remembered time and documented historical events, for at issue now is the shifting diarchic balance established between an object identified as "indigenous" and one brought by Dutch conquerors. This chapter probes two related questions: first, what is the historiographic function represented by objects, and second, how is their picture of the past subject to reinterpretation? Resolving these issues also engages the political tensions of diarchy,

[1] Following Valeri (1990, 164), I distinguish between a mythical past where events happen that are no longer possible in the present (as in the Kodi stories "of long ago" e nawu ) and a historical past that can be exemplary for the present but that can also be questioned and challenged. The mythical past has a transcendent and unquestioned character that keeps it "walled off from all subsequent times" (Bakhtin 1981, 15); thus it is not a direct force in shaping patterns of present action.

or dual rule, which is variously described as involving two different figures—passive and active, female and male, indigenous and external, "spiritual" and "worldly."

The two questions are related in their resolution, as important objects are endowed with features of gender and agency that place them in complex and changing dualistic contrasts.

History Objects and the Reification of the Past

A number of the most important ritual objects are described as "history objects" (Ind. barang bersejarah ), a term that corresponds to the Kodi expression "the traces of the hands, prints of the feet" (oro limya, oro witti ), which sees these things as the physical marks left by the ancestors. Although many have a utilitarian function—a water container, a weapon, an item of clothing, or an ornament—their most important role is to mark a particular historical moment. They are used didactically, as "evidence" of the past and a reminder of what has been lost, giving a permanent, external form to contingent events and preserving the memory of a promise, a covenant, or an alliance.

These objects are material signs of the past that exist not only as expressions of history but also as objects in history. They can even help to make history by "choosing" their proper location and exerting a mystical force on their human guardians to assure that they end up there. Some "history objects" fit the wider category of heirlooms or the more narrow one of regalia: they legitimate the claims of whoever may come to own them and as repositories of magical power are sometimes believed to affect the processes that they represent.

Most of the "history objects" I studied on Sumba were imported items that came to the island through trade. They included porcelain ceramic urns and plates, gold jewelry, swords, and gongs, which were stored high in the lofts of traditional houses, removed from circulation in exchange. Through a ritual dedication they became the "possessions of the ancestors" (tanggu marapu ), attaining a status at which they were equated with persons. At each important ceremony they had to be addressed in prayer with respectful kin terms and fed with rice and animal sacrifices. If they were sold, mistreated, or lost, they could curse their former masters and exact supernatural revenge. Power objectified in a concrete object preserves an impression of stability even when the object comes into the possession of a rival; thus it can legitimate usurpation while maintaining a fiction of continuity.

A few, however, were objects of local manufacture, including a spindle

that once pierced through the heavens (kinje maniki ), a drum covered with human skin (bendu kalulla toyo ), and a fiddle and flute (dungga mono pyoghi ). Stones used to summon lightning or earthquakes were found in the local area, but their power was revealed to an ancestor by a special vision. Some villages kept not an object but an animal, a horse or dog given a ritual name who served as the "placeholder" for a ritual office. Others had stories of exchanges with crocodiles or pythons, coded as marriage alliances, which entailed a totemistic prohibition on eating the animal's flesh.

When I began to ask the people of Kodi about their past, the first stage of my investigation was not to establish a line of dates and periods, but to make a map and a catalog. The map showed the location of each ancestral village, and the catalog listed the heirloom objects or totemic animals located there. A summary of this catalog (table 2) notes the location of the forty most important named possessions in Kodi. Although many other important objects are also kept as reminders of the past in Kodi villages, for this catalog I limited myself to objects given ritual names, invoked along with the ancestors, and "fed" at sacrifices. I included only objects or spirit-animals that were addressed as ancestors, anthropomorphized to become part of the invisible community that "listens in" at each ceremonial occasion.

These forty possessions are distributed among thirty-one ancestral villages, out of the total of sixty-six villages found in Kodi. Thirty-five villages (many of them small and only recently founded) had no significant heirlooms and worshipped "using only the names of their ancestral founders." Almost all of these villages were "attached" to larger and older villages with a store of heirloom objects, drawing on the power and prestige concentrated in the "mother village" (bei parona ).

A Kodi interpretation of the catalog is that it provides "evidence" (Ind. bukti ) of the veracity of the narratives presented in chapter 3: Tossi's ritual preeminence, for example, is "proved" by the fact that it has the largest concentration of sacred objects. The redistribution of other objects to outlying villages, with separate posts for harvest offerings (in Toda, Bondo Tamiyo, and Kaha Malagho), metalworking tools (Wei Yengo and Wei Hyombo), and the skull tree (Ndelo, Bondo Kodi, Kere Tana, Bongu, Lewata, Ratenngaro, Parona Baroro, and Kaha Katoda) all display the narrative pattern of delegation from the source to an executor. The bush knife that belongs to the village of Watu Pakadu and its associates deserves a special explanation. The story is told that Landa Deta, the founder of this village, was a famous healer who was treating a pregnant woman at the time of the first division of ritual tasks and territories. Since he missed

Among the sacred objects stored in the ancestral village of Toda is an heirloom
spindle that once connected the heavens to the earth by a line of cotton thread.
1988. Photograph by Laura Whitney.

the demarcation of traditional boundaries, he was allowed to make his gardens in all of the tabooed areas, being given a special knife that "cuts without waiting and chops without hesitating, since it does not recognize the crown of sacred land or respect the forelock of forbidden ground."

The "evidence" of objects, however, sometimes contradicts the narratives. Why are the urn and plate entrusted to Rato Mangilo in Tossi now found in Bukubani? How has the skull tree given to Bondo Kodi spread to six other villages? How have metalworking and rain magic traveled into other regions ? The answers to these questions are usually offered in other narratives, which tell, for instance, how the nephew of a famed headhunter in Bongu or Kaha Katoda achieved so much renown that he was allowed to cut off a sapling from the earlier skull tree and plant it in his own village. In some cases, though, a more complex accommodation has occurred, in which supposedly "unmovable" objects have been moved. The most important of these is the great porcelain urn "discovered" by Mangilo and Pokilo near the site of Tossi. When I pursued the question of the urn's location, I found that it had been moved only relatively recently, as a result of events during the colonial encounter.

The story of the urn reveals a present division between the functions of objects as traces of ancestral heritage and their new involvement in a more disruptive, discontinuous "history." The conceptual framework for this discussion therefore begins before the colonial encounter itself, in the constitution of an "imported past" in the form of ritual treasures brought from faraway places.

The Location and Transfer of Objects in the Past

In the traditional ceremonial system, the authority of objects came from their priority in time, expressed by an arrangement in space. The oldest objects were stored in the ceremonial house at the head of each village square, with younger ones in ranked positions at the base and sides. A consciousness of the past was produced and transmitted by visually encoding notions of precedence, sequence, and order; in this way, residues of ancestral migrations became guideposts for future generations. The location of these objects is probably the single most authoritative form of historical evidence on Sumba: land disputes can be resolved by the revelation of an heirloom sword, political struggles can focus on the proper name for a plate or jar, and ritual offices are defined not by the persons who hold them but by the objects that are manipulated.

A. L. Couvreur, the Dutch district officer (controleur ) who commanded during the period in which Kodi was effectively brought under colonial rule, wrote that the relics of the marapu included "ornaments, bits of gold, and sometimes other terrible rubbish as well" noting that a colleague of his once found a disheveled and filthy copy of the Koran in a sacred basket stored in an ancestral cult house (Couvreur 1917, 207). While he considered the Koran an anomalous ritual object for a community of pagan ancestor worshippers, I argue that it is in some way paradigmatic. These "history objects" represent an effort to appropriate an incompletely understood external power—one often coming from Hindu or Moslem states—into a very different symbolic world. It was the power of the Koran that Sumbanese villagers fetishized, but not the content of Islam.^[2] Dutch of-

[2] The category of "fetish" has not been used in Indonesian ethnography, though the magical importance of objects has often been noted and the tradition of family heirlooms, or pusaka , is widespread throughout the archipelago. Fetishism was a classic problem of nineteenth-century anthropology, but in this century only art historians, psychoanalysts, and Marxist economists have continued to use the term. If I venture, tentatively, to join this motley crew, it is in order to propose an analytic alternative to a term that has suffered from pejorative connotations. Pietz (1985, 14) notes, "The discourse of the fetish has always been a critical discourse about the false objective values of a culture from which the speaker is personally

distanced:" In other words, they have fetishes, we know better. But it can also be used to refer to the attribution of mystical powers to objects in a historical context of shifting political realities. The genesis of fetishized relations in the colonial encounter expresses different perceptions of value between the colonizers and the colonized and a problem of translation that is, in turn, presented as a displaced agency or consciousness from persons onto things. Fetishized objects occupy a historically particular cross-cultural space. They are found at the intersection of ideas of descent from an indigenous founder, on the one hand, and power from an external force, on the other—an intersection that in Kodi can be represented by tension between the ancient urn and the new gold staff.

ficials also had their own version of fetishism: they deliberately introduced the gold staff of office as a substitute for indigenous heirlooms, using the prestige of imported objects to reweight and reorder indigenous political systems under colonial rule.

Couvreur (1917, 215) acknowledged the power of traditional objects when he wrote that the person most likely to cause trouble for a colonial officer was not the chief but the secluded priest: "Although he stays in the background, he exercises a great mystical influence" and his guardianship of land and heirlooms can have important political consequences. The passage probably refers to the Rato Nale of Tossi and his guardianship of the urn, since it was importantly involved in the largest and most sustained armed resistance to the presence of the Dutch on the island.

The Urn as a Ritual Object

The urn involved in these events is made of glazed porcelain, probably produced in South China during the Ming period and one of a large number of ancient high-fired ceramics traded widely throughout Southeast Asia from as early as the eleventh century into the nineteenth (O'Connor 1983, 402). It is decorated with blue dragons (perceived as "pythons" locally) and contains sacred water, used to heal wounds in warfare or ritual jousting, and ancient coins, pieces of gold offered by people who have come seeking its blessings. The urn is classified as a female object within the rules of the Kodi ritual division: in the apportionment of powers between the two brothers, its hollow, curved shape made it appropriate to store in the "mother house" of Rato Mangilo, while the roaming horse of Rato Pokilo had to be male, expressing as it did qualities of virility and mobility. The elder brother was told to behave like a woman, remaining confined inside the house for four months before the sea worms came, "brooding" over the new year "as a hen broods over her eggs, as a sow calls to her young" showing nurturing, maternal concern for the young rice crop. The stallion ridden by his brother followed the same food taboos as the priest, avoiding corn and tubers and eating only rice.

The hard, shiny surface of imported porcelain is so different from locally produced earthenware that many people in Southeast Asia trace their origins to a cosmogonic act: the miraculous excretions of the Ceramese divinity Hainuwele (Jensen and Niggemeyer 1939), a gift from the heavens in northern Luzon (Cole 1912), a formation from the clay used to make the sun and moon in Borneo (Bock 1881, 198). Assessment of the age of such jars using objective criteria—form, glaze, decoration, and physical properties of the body—is a developed tradition in many isolated tribal communities, each having its own rigorous standards of connoisseurship (O'Connor 1983, 405).

On Sumba, most ceramic pieces are perceived, realistically, as imported objects, though no one could trace their provenance back to the kilns in South China, Vietnam, or Thailand where they were probably manufactured. In this case, although the origin narrative says the sea worms came from the court of the sultan of Bima, the urn was not traded for but discovered . It appeared magically on the island and "chose" Mangilo and Pokilo to serve as the priest and defender of the calendrical system. The urn washed up on the sand at Tossi to mark the location of the sea worms and the site where the rituals of the Kodi new year could be carried out. The historical ties that brought trade goods like the urn to the shores of Sumba are thus displaced from the imported object itself to the sea worms. This mystification of the urn as an "indigenous" object is related to its distinct appearance and to earlier ritual uses of jars on the island.

The physical properties most admired in the urn were its hardness, the smoothness and whiteness of its surface, and the 'ring of its body when struck. Almost all objects of local manufacture on Sumba were only semidurable: even the finest ikat textiles would eventually tear, even the firmest mats would disintegrate, and none of the local ceramic ware was anywhere near as hard and fine. The imported urn was valued because it was an irreplaceable object, the only one of its series. Its "voice" was compared to the deep sonorities of imported gongs,^[3] its smoothness to the surface of a tortoise shell, its roundness to the seeds of the dedap tree.

[3] There are many other reports of imported porcelain jars that talk. Cole (1912, 12), for instance, tells the story of a jar called Magsawi by the Tinguian of Abra that could speak softly in very low tones; it also went on long journeys by itself and was married to a female jar owned by the Tinguian of. Ilocos Norte. The famous jar of the sultan of Brunei was said to have howled dolefully the night before his first wife died and to have heralded misfortunes of all sorts by emitting strange tones. Its cries may have been the result of air draughts thrown into resonant reverberations by the peculiar form of the mouth of the jar. Usually covered with brocade, it was uncovered only when it was consulted, so it spoke exclusively on solemn occasions (Cole 1912, 11). Among the Datu of Borneo,

villagers brought gifts to the sultan and received water from the sacred jar to sprinkle on their fields and thereby assure plentiful harvests. The "voice" of heirloom plates among the Hua Ulu of Seram is an important characteristic of their value, used in calculating bridewealth (Valeri 1980).

Archeological research has shown that long before the import of high-fired ceramics, urn burials were made on Sumba, dating back as far as the first millennium B.C.^[4] Contemporary peoples have no memory of these rites; today they bury their dead in large stone megaliths or earthen graves. But ideas of the sacredness of pots and their relation to the after-world persist in the sacrificial offering placed in the urn to "ask for blessings" and assure the safe voyage of the soul. The urn was said to have descended directly from the upperworld, the land of the dead, and to have assumed its position as the symbolic anchor of the Kodi year.

The narrative tradition that the urn presented itself to its proper owners established a paradigm for attitudes toward ritual heirlooms found in many other parts of Southeast Asia. Imported ceramics are not "mere handy crockery"; rather, they are marked by "a potency that makes them active agents" (O'Connor 1983, 403). Kodi ritualists recognized this fact when they placed the urn in the house of Rato Mangilo, at the head of Tossi, naming it "the urn that cannot be moved" and laying an heirloom porcelain plate on top to close off the sacred opening, called it the "plate that cannot be lifted."

The Urn Meets the Staff

In 1909, when the Dutch took effective control of the island of Sumba, they asked the people of Kodi to choose a leader to serve as an administrator and representative to the colonial powers. This leader would be presented with the gold staff that created the colonial office of raja, the native administrator of each "self-governing region." A meeting of elders from all the ancestral villages was held in Tossi, in the Council House established by Rato Pokilo, and they agreed to choose this leader from the home of the urn, the most sacred object in the domain. Hence Rato Loghe Kanduyo,

[4] An urn field at Melolo in East Sumba was excavated by van Heekeren in the 1920s and 1930s. Many round-based urns were found with unburnt secondary burials of broken bones, as well as adzes, stone bracelets, shell and stone beads, and smaller votive pots. The smaller pots were highly polished and had elegant long-necked flasks, some depicting human or animal figures with incised geometric designs filled in with white paint. Polished pots of the same type are found today in Lamboya and Memboro, but not in Kodi, where ceramic work is much cruder. Van Heekeren now (1972) dates the urn field to the Southeast Asian Neolithic (about B.C. 1000), but in his first publication of the results (1956) he estimated it as early Metal Age (B.C. 500—A.D. 1000). Heine-Geldern (1945, 148) favors the Metal Age date.

a famous orator and warrior descended from Rato Pokilo, was selected to be the first raja, serving as a spokesman and mediator vis-à-vis outside forces.

His selection was legitimated by an argument involving the genders of the urn and staff and the idea of their complementarity. The urn, a hollow cavity with generous curves, was a female object containing fertile, life-giving water. The staff was classified as a male object, because it was long, firm, and could be used as a weapon; it was therefore included in the category of spears, swords, and metal goods presented as the male gifts of bridewealth. Because the holders of both objects were members of the village of Tossi, the two objects could be properly balanced, along with the two kinds of powers. The male staff served as the "husband" of the female urn. This symbolic marriage linked the two offices, making the representative of the Dutch colonial administration into the junior partner in a division of powers, since the holder of the staff (called the toko , the native term for raja, from the Indonesian tongkat ) owed ritual deference to his senior, the priest of the sea worms.

Contested Claims: The Staff Separates from the Urn

The gold staff that Rato Loghe Kanduyo received was stored in the Council House in Tossi and, as local people tell the story, wasted no time in causing trouble. As an officer of the Netherlands East Indies government, the new raja was supposed to explain its civilizing mission to his people as well as provide labor for bridge- and road-building projects to improve communications. Two years later, in 1911, rumors circulated that Dutch soldiers had enslaved noblemen to work on these projects, insulted the raja, and raped a local woman. An armed rebellion began with the ambush and killing of four soldiers by headhunters from outlying villages, who fled into the forest with the guns captured from the soldiers. The Dutch forces retaliated by burning the raja's village of Tossi, after which they took refuge with his rivals in another river valley.

To avenge the burning of his village, the raja gave permission for military attacks to continue. He took the traditional symbols of governmental power—the urn and plate—off into the bush and hid them. Yet he did bring the Dutch staff of office with him when he rode, in a procession bringing gold and livestock, to meet the Dutch commander and negotiate a peace payment. Instead of talking to him, however, the commander pulled him from his horse, bound him under the house, and made him march to a distant prison, where he soon died. This brutal punishment of the first native ruler united almost all the population in opposition to the Dutch presence. Three years of fighting followed, with rebel forces hiding

in the interior and attacking the colonial army periodically. The Dutch, deciding to starve out their enemies, then forced everyone to move to coastal villages and leave their gardens behind. Finally, pressured by famine and hardship, the rebel forces surrendered and were sent into exile.

This first sequence of events shows the consequences of differences between local understandings of governmental power and those of the Dutch colonial forces. The Dutch expected to find rulers who could command the population and thus concentrated power and wealth in the man chosen as raja. By contrast, Kodi perceived the raja, at least initially, as a mediator and spokesman, who would speak for them in negotiations with outside forces but who had no authority to act without a meeting of elders from each village.

The creation of a Kodi polity, and after that of a regional resistance movement, was one of the unintended consequences of the colonial encounter. Because no single ruler preceded them, the Dutch had to establish the legitimacy of the first Kodi raja themselves. Rato Loghe Kanduyo died in the process of his transformation into a raja, and his nephew Ndera Wulla made the first claims to a new and different form of political power.

Stoking The Flames of Controversy

Once order had been restored, the captured staff of office was given to Rato Loghe's nephew and adopted son, Ndera Wulla. He had taken the urn and plate to Bukubani, the garden settlement farthest from the Dutch fortress, and stored them with descendants of Rato Mangilo who had planted coconuts there. Washed in coconut water to cleanse them after their journey, they were placed beside the lightning stones which were Bukubani's most important heirlooms.

Tossi was rebuilt during the next few years as Ndera Wulla prepared to assume political control of the area. On the day he was officially installed in office as raja, he decided to place the gold staff in his own house and bring the urn and plate to join them. A procession was formed from Bukubani to the Council House, where he lived, to bear these sacred objects and place them under his guardianship. Many people criticized the move, because the urn and plate had never been stored anywhere except in the Sea Worm House, and they speculated that his uncle's own problems had occurred because the staff did not get along well with the urn. They said the urn, which was supposed to be unmovable, had already been moved too many times.

In 1927 and in 1936, the village was burned again, first by a powerful Javanese trader, who resented efforts to control his activities, and then by rivals of the raja within the village itself. In the second fire, flames spread

quickly to all of the houses, and the urn and plate were damaged by the blaze. Charred and cracked, the two old porcelain dishes were taken to the bush to be repaired. Special sacrifices had to be performed to "mend their bones and lift up their souls" after this incident, and repeated prayers were offered to the ancestors Rato Mangilo and Rato Pokilo pleading for unity among their descendants.

At the end of this rite, a special meeting was held to discuss the location of the urn. The raja's rivals, who came from the division of Tossi Bukubani, argued they should have been the ones to hold the office of raja. They particularly resented the fact that the urn and plate had been moved from the Sea Worm House into the raja's house, since he had no hereditary right to them. "This is why our village keeps burning" they said; "the ancestors will not protect us if we do not keep things in the proper places." As a compromise, it was decided that the urn would not go back into the raja's house, but neither would it be placed in the Sea Worm House of Tossi, where it had been scarred by fire three times already. Instead it would be entrusted to the outlying village of Bukubani. There, safely removed from the ritual center, descendants of Rato Mangilo would look after it and offer it sacrifices of rice and meat. Tossi was not surrendering its authority over the objects, but only delegating it to the house of another priest. Subsequently, a new Sea Worm House was built and ritually dedicated in Bukubani, specifically to store these objects. Later, once the political struggles had cooled, the urn could be reclaimed—but probably not while the office of raja remained in Tossi.

The annual festival of the sea worms brought all the people of Kodi back to their ancestral villages to sacrifice chickens to the souls of the dead, collect the sea worms, and participate in a huge ritualized jousting combat on horseback. It was still performed in Tossi, as were the calculations that determined the start of months for planting and harvesting and the four-month ritual silence.

In the late 1930s, these rites were led by a priestess, Mbiri Koni, who bore the name of the rice spirit. She had been chosen to succeed her husband at a divination that occurred during the period when Tossi was deeply divided politically. Because the diviner read her name in the entrails of sacrificial animals, a reading confirmed by spear divination, her anomalous selection was explained as the "will of the ancestors" who this time, exceptionally, favored a high priestess over any of the male descendants of the house. We should note, however, that the role of priest, though symbolically female, became literally so only after the raja developed a strong concern to depoliticize the functions of that office. Once the supreme ritual leader in the domain, the Sea Worm Priest became, in the

1930s, little more than a religious functionary who performed standard rites in the raja's village. Raja Ndera wanted a Tossi priest to continue to lead the region in these ceremonies, but he did not want the ritual leader to emerge as a political rival. Thus, there may have been a strategic advantage to this choice of an older woman, widowed and past childbearing age, as the new priestess of the Kodi year.

Consistent with the constraints placed on the office, Mbiri Koni was a rigorous traditionalist who refused to have anything to do with foreign authorities, withdrawing into seclusion whenever Dutch visitors came to the village. She understood that to retain her authority she could not display it openly, and to continue as a guardian of tradition she could not speak to foreigners. The selection of a female priestess thus intensified and further polarized a diarchic division that had already been shifting in a certain direction. It brought the division between the functions of the urn and staff more into line with Western notions of a division of church and state, emphasizing spiritual power at the cost of secular influence.

Separated Powers: The Japanese Occupation and Independence

Conflicts did not end with the relocation of the urn, but they came to be defined in different terms. The next few decades were ones of great turmoil and suffering. Eight thousand Japanese troops arrived on Sumba in 1942, whereupon they forcibly inducted local people to work on airstrips for a planned invasion of Australia. They required that local rulers supply them with large amounts of food and decimated local herds to feed the occupying forces. A number of people went hungry, most had no clothing, and few seemed to accept the Japanese message that they had been sent to liberate them from the oppressive institutions of Dutch colonialism.

During the Japanese occupation, the cycle of calendrical rites and government functions in Tossi was interrupted by the deaths of both the high priestess and the raja. In some local interpretations, the deaths were linked, as the guardians of the land and government both succumbed to their inability to relieve the suffering of their people. Raja Ndera Wulla was replaced by his administrative assistant, H. R. Horo, who came from Rangga Baki, across the river. The Japanese forces administered Indonesia through the Dutch-established hierarchy of native leaders. Their decision to appoint Horo followed the rules set by the colonial powers that preceded them, but as a consequence it moved the office of raja away from the "mother-father" village of Tossi and gave it to the rival territory of Bangedo.

In 1945, the independence of Indonesia was declared, and five years of

fighting followed to keep Dutch forces from returning to power. Sumba remained protected from most of the violence of that struggle. With no military opposition at all, the new Kodi raja, now presiding over a Council of Rajas in the port town of Waingapu, simply ordered the raising of the new Indonesian flag to replace the Dutch one. In the first elections after independence was realized in 1950, Tossi recovered its position of political leadership when Martin Calei, a descendant of Rato Mangilo, was elected to the new parliament and later nominated to assume the position of regent (bupati ), administrator of the whole island of Sumba.

In 1958, Tossi was burned again. In the same year, Martin Calei, having lost the regency election by a single vote, fell seriously ill and died, as did his mother and father in rapid succession. The series of catastrophes was interpreted by Raja Horo as punishment sent to his Tossi rivals by their own ancestors, who did not approve of the mixing of ritual power with careers in local government. "This is what was needed" he told me, "to teach them that some people wear the high feathered headdress and the black plumes and can be warriors and government leaders. The others, the members of female houses, are the ones who control the seasons, the time of planting and harvesting. They must stay confined to protect the whole region and should not try to wander off to the capital city. The holder of the staff cannot be the same person as the holder of the urn."

This formulation of the diarchic divide was the one used in 1949, when Raja Horo received a new gold staff from Dutch officials, during their brief return to the island before independence was realized. He stressed the legitimating power of objects at a time when the prerogatives of Dutch-appointed rulers were being increasingly called into question. After independence, traditional rulers were left in ceremonial positions and respected, but virtually all power was taken from them: a new nationalist rhetoric asserted that because all of Indonesia was once a "village democracy," past hierarchies and inequalities were the creation of a "feudal" colonial system and should be abolished.

The early 1960s were a time of great disorder and turmoil throughout the country, expressed in Kodi by the burning of seven important ancestral villages—not only Tossi, but also Rangga Baki, home of Raja Horo, and five others that contained significant wealth and heirloom objects. Some of the suspected arsonists were described as "communists"; yet they, too, were engulfed in the flood of violence that spread over the country in 1965, after generals crushed a reported coup. At this point it is unclear just where local politics stopped and national politics began. Anyone who seemed to wield power or privilege was the object of assaults, with quick

retaliation a foregone conclusion. The result was to paralyze not only the national political system but also the local ceremonial cycle.

The office of Rato Nale in Tossi remained vacant for much of the sixties as villagers tried to summon the resources needed to perform a ceremony calling back the soul of each of the destroyed houses, which would permit them to be rebuilt. No one dared to assume the politically dangerous position of high priest lest further arson and attacks be the result. For almost a decade, therefore, Ra Ndengi alone performed the calendrical rites, "holding the year" in place from his secluded home in Bukubani. The lapse in ritual performance created a sense of social fragmentation and a loss of hierarchy. People no longer gathered for large-scale festivities, feasting and jousting to welcome the sea worms, but simply prayed in their isolated garden huts.

In 1972, a divination was held in the newly rebuilt Tossi, and Ra Holo, the grandson of Mbiri Koni, was selected to become the new head priest. He was reluctant to take the post because of his youth and ignorance, however, protesting with a verse that argued he could not perform the rites properly:

He was finally persuaded that his authority would be buttressed by Ra Ndengi, who acted as his senior in the ritual installation. The trials and humiliations of the urn itself (and, by extension, the ancestors who acquired it) were cited in words he reported they used to convince him to accept:

The spirits of all the ancestors were seen as angrily demanding that sacrifices again be offered to them so the prestige and importance of Tossi would be restored. Ra Holo was anointed with water from the sacred urn, but allowed it to remain at its present location in Bukubani. He and the leaders of Tossi agreed that since the urn was the senior object, its ritual authority was greater than the genealogical priority Tossi had over Bukubani. As long as they guarded the urn, the people of Bukubani would be the "older brothers" and Tossi the "younger brothers" in the public context of calendar rites.

Since Indonesian independence the raja's gold staff has no longer been used and is kept as a family heirloom by his descendants. The "immovable urn" has found a new home, one now, after sixty years, recognized as its legitimate location. Once again reassuringly immobile, the urn can anchor the ritual polity and exert some of its mystical power to hold together the year and the people as expressed in its mythical mandate.

Interpreting and Criticizing Sources on the Kodi Past

At this point, I pause to reconsider the role objects play in constituting "evidence" in addition to and sometimes in contrast to the "evidence" of narratives and ritual action. The competing claims made in origin narratives often concern the identity of the ancestor who first acquired a certain valuable, since the valuable's present location is well established. Differences concerning the story of the sea worm trough on its journey to Tossi are all of this kind. A second kind of competing claim concerns why an object found its way to a particular village, and on this question a greater variety of interpretive strategies are possible. Some argue that the transfer

was legitimate, some that it was the result of a theft, some that eventually the object will "return home." Others say that if it has stopped moving, this in itself is evidence of the superiority of its new location.

Competing versions of the story of the urn express differences of opinion on who was really an agent in historical change, and whether the potential for agency invested in objects had changed over time. The people of Bukubani said, "We have always been the older brothers," but this was denied by those in Tossi, who reminded them that Bukubani was once only a small garden settlement where the descendants of Rato Mangilo grew rice and corn. Before a ritual house was constructed for the urn it was not even consecrated as a site for ancestral offerings, much less the calculation of the beginning and end of the ceremonial year. But if Tossi's leader argued, "We are the only mothers and fathers" members of other villages objected that they were exaggerating the contingent event of receiving the Dutch staff of office and neglecting the disputes that led to its loss.

The disagreements concerned interpretation more than the sequence of events. Debates centered on the issue of whether the urn could still play a role in determining its proper home. In the distant past, the great urn was said to have washed up on the shores of Kodi and found its way into the hands of Mangilo and Pokilo almost as if it moved of its own volition. Some people assert that such movement was possible only in the past, and now when objects are shifted it is only because of human actions. Others, however, still read a divinatory significance into the trials and tribulations of objects like the supposedly immovable urn. It was "not happy" that Tossi had become so divisive, that quarrels about who should hold the raja's staff had disturbed the unity of the center of calendrical ritual for the whole region. The fact that the urn was moved when these conflicts escalated, so that it reposed more peacefully in Bukubani, was no accident. "The urn knows," they would say, "where it is supposed to stay."

A sacred object can become reanimated if its authority is violated in a particularly extreme way. Events that occurred in 1986, when a smaller ceramic urn in the village of Waindimu was stolen, were interpreted as supporting this position. The urn was taken to remove the gold coins placed there by people seeking blessings, and when it suddenly disappeared Moslem thieves from a nearby fishing village were suspected.^[5] The wealthiest Moslem merchant owned a small motorboat, which left port a few

[5] Most people agreed that "no Kodinese could have done it" because of their respect for ancestral taboos. Even converted Christians offered testimony of the power of sacred objects, which they said did not violate the rules of conversion because it was an inherent power, not mediated by ancestral spirits.

days after the disappearance, apparently carrying the urn to sell to traders on Flores. A huge storm came up, the boat was smashed to pieces, and the urn drifted back with the waves to the same area where it was originally discovered. Emptied of its contents and cracked on one side, it was recovered and carried into the village, where sacrifices were performed to call back the soul dislodged by this act of violence. The urn's guardian said that the soul of the urn had "taken its own revenge": the thieves were financially ruined, and shortly afterward the merchant went mad.

The lessons taken from these events were these: a sacred object in the right place brings good fortune on its owners and thus legitimates their ownership, whereas the loss of an object is a sign of moral weakness or deception. Possession here is more than nine-tenths of the law; it is itself a form of mystical justification.

The smaller urn was stolen while I was compiling the competing versions of the past of its "elder sibling" in Tossi. The people of Bukubani argued that since the urn now sits soundly in the Sea Worm House of their village, it should have always been there. The past, in their view, showed only a record of error that is better eclipsed by memory. When I first interviewed them, they made no mention of the historical events I have detailed, saying simply that the urn had been entrusted to Rato Mangilo and that his descendants continued to be its guardians. Their silence on the issue seems to conform to the classic anthropological view that unwritten history is largely fictive history, with myths recited as charters to justify present institutions (L. Bohannan 1952; Malinowski 1954). But a closer look at the situation reveals that this apparent presentism is illusory: they were fully aware of the difference between the urn's past location and its present one, but interpreted the fact that it remained in Bukubani as vindicating their own position. If the urn did not belong there, it would have left.

From the people of Tossi I heard two arguments. The descendants of Rato Mangilo in Tossi Bukubani said that the power of objects to determine their own destinies held true only for the primordial period of the ancestors, and now it is human beings who are the actors on the stage of history. Objects serve only as historical evidence for certain attributes, such as rights to an office, but the mythical, magical past is finished. In this view, explanations that attribute meaning to an object's location can only be retrospective. "The true home of the urn was established in the distant past" they said "and nothing can change that."

A second group, descended from Rato Pokilo in Tossi Barada, claimed that sacred objects continue to exert a mystical influence on the behavior of their owners (or their violators) to legitimate processes of innovation.

In some ways, this is the more historical view, because it focuses on the changing relation between an object and a particular temporal context, saying the location is not determined absolutely but must always be viewed relative to shifting circumstances. Their version of this argument, however, gives a different spin to the idea of magical choice. The gold staff of office, in their perspective, proved to be the more powerful than the urn during the colonial period, so the loss of the urn to Bukubani became necessary to preserve a transformed diarchic division of powers, in which Tossi emerged triumphant. In other words, although control of ritual time was the supreme office according to the cultural values of the past, in the socioeconomic world of the Indonesian new order, access to government offices is of much greater significance. The emergence of new cultural forms always involves the incorporation of the outside by the inside, in which the signs and substances of foreign powers are taken and made part of new local systems. While the office of raja is now denigrated as a relic of feudal colonialism, the active, secular orientation given by the staff has enabled other people from Tossi to hold administrative offices and to seek education overseas.

The differing views of the power of objects are evoked as interpretive strategies, not as mutually exclusive representations. Each side might choose in a different context, for example, to argue that the movement of some other object had no ancestral sanction; the fact that the urn itself is now "settled" in its new home would have no bearing on that case. With regard to the urn, the ultimate test is the well-being of the people of Bukubani over time: if they continue to prosper and enjoy good health, the favor they have found with the ancestors who brought the urn and with the spirit dwelling in the urn itself is proved. But if they are struck by illness or misfortune, the location of the urn can be called into question again in a traditional divination.

The attitude of "spiritual empiricism" invoked here is similar to that contained in the phrase "History will be the final judge": that is, the past receives its fullest legitimation in the events that follow it. Because precedents are evaluated by their effectiveness under new conditions, current events provide a retrospective justification for past claims.

The System of Objects and Knowledge of the Past

I began by addressing two different questions, one historiographic, the other ethnographic. The first had to do with the forms taken by a historical knowledge invested more in objects than in persons—a "great things" approach to ordering the past, instead of one based on "great men" or

"great women." The second involved the conundrum of dual sovereignty, or diarchy, where powers were shared between two persons or two objects on uncertain terms.

In answering the first question, I argue that the focus on objects as markers of past traditions creates an impression of stability , which seems to represent enduring offices and relationships as less open to variation than a person-centered genealogical model. At the same time, this stability can be illusory, for the offices and relationships do not in fact remain unchanged, even though the objects that represent them maintain a reassuringly ancient appearance. The traces of the past left in objects can be manipulated, as can other forms of historical evidence, but they remain, from a Kodi point of view, the "material that needs to be explained."

Objects belong to history as a heritage of the past, an uninterrupted process that reveals the continuity of culture over time. This contrasts with the discontinuous history of recent years, called by the Indonesian term sejarah , in which a new ideology of process is associated with the ephemeral importance of persons. In written histories, individual heroes are introduced as the protagonists of a novel form of narrative, set on the stage of irreversible historical changes. Thus Wona Kaka, the military leader of the rebellion against the Dutch forces in 1911-13, is cited as a "historical figure" while people like Ra Holo and Ra Ndengi (as well as their predecessors in the office of Rato Nale) are seen as placeholders in roles whose real meaning is defined by ritual objects.

The authority of the past depends on the existence of cultural standards of validation—a native historicity of some sort. A repertory of rules for appropriate action, such as those involved in the political constitution of diarchy, can be revised and adjusted by reweighting certain elements to respond to new historical conditions.

The traditional anthropological view of dual sovereignty presumes a static separation of powers, in which persons or objects interact to reproduce an established cultural pattern. It is unrealistic, however, to believe that diarchy in the past presented an unchanging template of complete consensus. Rather than portraying the "ragged Forces of History shattering the crystal Patterns of Culture" (Geertz 1990, 326), I maintain that the diarchic pattern constantly shifts to accommodate the importing of foreign models of rule—which often come, through trade routes, in the form of foreign objects. Ideas of hierarchy and wider social orders (whether those of Moslem sultans, Dutch administrators, or Indonesian nationalists) are incorporated into indigenous traditions and change their values. The relation between the urn and staff as tokens of office appears as a dynamic process, in which unities dissolve into oppositions, with a single opposition

reproduced in a new context or else reunited with its counterpart at the opposite pole. The once supreme inner, spiritual source has been marginalized and made subservient to an outer, active executor, showing how the pervasive duality of power in Eastern Indonesia is always and intrinsically contestable, unstable, and politically constructed. The opposition articulated in metaphors of gender or seniority must repress internal ambiguities in order to stress certain differences at the expense of others. An initial pairing of male and female objects, for example, is now expressed more commonly in terms of elder sources and younger executors, and the cultural value of the distant past competes against the attractions of a new social and political order.

The value of the urn was an object of contention because an externally introduced trade item came to be used as a symbol of indigenous rule. The urn stood for the immutable and undeniable power of the great courts to the west of Sumba, yet it was used to legitimate the much more restricted powers of the indigenous timekeeper, the priest of the calendar. While the priest was never a ruler, he was the supreme ritual official of the area and exercised hierarchical control over the social organization of time.

The power of the foreign state was both recognized and denied in the process of offering sacrifices to the urn. A splendid imported object was made the center of the indigenous polity, described as the immovable anchor that held the calendar in place. The fresh water originally found in the urn after many days in the sea was replaced with the water from local sources. Once it had been stored in this sacred vessel, it became a medium of blessings and ritual cooling. The Chinese dragon, a powerful mythological animal for the urn's creators in South China or mainland Southeast Asia, was transformed for these islanders into a python, the giver of rainfall who sacrificed his own daughter so that the people might have rice. The original meaning of the iconography was lost on the way from China to Sumba, but its reinterpretation reflected the historical conditions that brought it to the island. The imagery of fertility and power was invested with a specific sense of distant sultanates and indirect rule, an aura of remote authority and diminished capacity for action.

Objects in Movement and Objects in Place

This drama of the disruption and reabsorption of colonial power through the rearrangement of sacred objects prompts an evaluation of the role of objects in Kodi exchange and the meaning of an object's movement. When the names "urn that cannot be moved" and "plate that cannot be lifted" were given to these objects, they assumed a distinctive place within the

world of Kodi valuables. They became objects that stay in place , and as such were sharply distinguished from the objects that move almost incessantly along traditional exchange paths.

Most wealth objects, even those of great value, are supposed to circulate in transactións between affines. Decisions about when to keep them and when to give them away are motivated by thoughts of possible gain or loss, but not by the threat of supernatural sanctions presented here. The gold ear pendants (mamoli ) used in bridewealth transactions move incessantly, up to nine or ten times a year.

The goods exchanged at marriages are opposed by gender: wife-takers give "male valuables"—swords, spears, gold pendants, and livestock (horses and buffalo)—while wife-givers present a countergift of "female valuables"—men's cloths and women's sarungs, pigs, cooked food. In the context of marriage exchanges, male goods are durable, forming the metal skeleton of the patriline, while female ones are perishable—food that must be consumed, textiles that must be rewoven each generation. The meeting of male and female in the world of objects parallels the marriage of man and wife, since the man's group will acquire descendants from it and an identity that will endure through the generations, while the women's group will "give life" but achieve no continuity from daughters who marry out.

The urn and plate are paradoxical because they are both female valuables and durable ones. Their inalienability places them at the top of a hierarchy of valuables, mobilized in the more encompassing struggle to achieve a lasting influence over time and escape the risks of loss involved in the exchange game. For stability and immortality are not finally achieved until death, when a man is placed inside a stone grave and his finest wealth objects are sealed up with him. Death removes a person from the risks of exchange and begins the process of his transformation into an ancestor, a marapu , whose grave will be decorated with carvings of the gold jewelry, gongs, and buffalo horns he owned in life. To play the exchange game is to struggle against time, trying to achieve immobility and permanence through the acquisition of wealth that can be transformed into the stable value of a splendid tomb.

The object which is "turned into an ancestor" (bali marapu ) by a ritual consecration that makes it inalienable is consciously used as a "vehicle for bringing past time into the present" (Weiner 1985, 210) and making history visible. It acquires a peculiar temporal identity, becoming an anachronism, lifted "out of time" to sit above it. Like the pen used to sign the Bill of Rights or a hat worn at a famous battle, this object must be taken out of history in order to represent it. The treasure kept in each lineage

cult house is a miniature museum, segregating particular objects to protect them from being caught up again in the flow of historical contingencies. But the ancestral patrimony is, at times, disturbed, pulled back into history by irregular events—theft, breakage, burning. On these occasions, the assault on the object is a deliberate assault on the history it represents, as well as an effort to undermine precedent and dismember memory.

The ways in which objects are embedded in history and the politics of representing the past have only recently come under scrutiny (Parmentier 1985, 1987; Thomas 1991). Throughout the Pacific, exchange valuables have become complexly entangled in local politics and colonial expansion. The contingent meanings that become attached to these things must first be analyzed within the framework of an indigenous system of objects and their relations before broader conclusions about reciprocity and regional interactions can be made.

5
The Past in Action
The Rites of the Kodi Year

Truly, carnival is the denizen of a place which is no place, and a time which is no time, even when that place is a city's main plazas, and that time can be found on an ecclesiastical calendar. . . . What we are seeing is society in its subjunctive mood . . . its mood of feeling, willing and desiring, its mood of fantasizing, its playful mood.
Victor Turner, The Anthropology of Performance

Calendrical rites have been described as attempts to abolish time, to annul the past and to return to a primordial era of cosmogonic regeneration. For Mircea Eliade (1954, 85), they represent "archaic man's refusal to accept himself as a historical being . . . to grant value to memory and hence to the unusual events that . . . constitute duration." In the ceremonies of the New Year, he sees a revolt against the irreversibility of history and a claim that "past" and "present" can coexist, the ancestors once again being the contemporaries of the living. The nale festivities held to welcome the sea worms in Kodi display many of the features of the "rites to regenerate time" that he describes: they follow the narrative sequence of privations, chaos, then a restoration of order; they involve ceremonial combats between opposing sides, a dissolution of the barriers between the living and the dead, a "carnival" atmosphere, and a great feast of commensality.

But Eliade is wrong in saying that rites of reversal do away with the past. The point of all these actions is not to destroy the past but to revitalize it, to reflect upon it and reenact it imaginatively, recognizing a continuity of tradition in cultural heritage. The roles that people assume in the ceremonial context are those of ancestral personae, whom they personify on a stage where the present "plays" at recreating the past. The actors are conscious of shifting roles, and the ritual frame allows them to travel back in time while still remembering their own personal, contingent relations.

In moments of liminal separation, they display an ironic distance from social convention—singing bawdy songs, crossing lances with their seniors on the playing field, defying the formal etiquette of most ceremonial events. But they do so in full awareness that this temporary "time out" will soon come to an end and they will reconstitute ordered social relations with a final sacrifice and shared meal.

The rites of the Kodi year involve an oscillation between constraint and excess, strict discipline and Rabelaisian rejoicing, which helps to define the polarities of both social life and the year. The nale festivities and pasola battle provide a unifying, centering focus for Kodi collective life, defining a shared order and ritual discipline for the whole domain. But they also offer an occasion for contests, both physical and verbal, that reveal underlying tensions—between juniors and seniors, women and men, members of peripheral villages and the ceremonial officers of the source.

The outer frame of the calendar is set by prohibitions. As outlined in mythic narratives, the autochthones are excluded, displaced by the descendants of Rato Mangilo and Rato Pokilo, who take over the center. The rites of the year bring people together, assembling them into a totality that observes a single annual rhythm of activities and acknowledges a single source of fertility in the nale spirit that comes from the sea. At the same time, the songs and ritual combat are made up of many voices and many riders, playfully celebrating the physicality of their encounters, exchanging courtship banter and blows in the mounted combat. The restrictions of ritual prohibitions give way to the risky pleasures of a game, and regulation ends in revelry.

The creation of the year is enacted at four ritual performances: (1) the start of the ritual silence at the "roasting of the bitter chicken"; (2) the playful period of courtship songs and games played along the beaches; (3) the sea worm collecting^[1] and the crescendo of pasola jousting; and (4) a feast of chicken and rice to honor the dead souls in each village. Over the last century, the form of these enactments has changed, as has the significance of the historical reimagining they attempt to achieve. In moving from narrative to action, we see how ideas of ancestral precedence have also served as the basis for innovations and changes in ritual sequence.

[1] Sea worms collected on Kodi beaches are marine annelids of the Eunicid family (Leodice viridis ), commonly known as the palolo in much of the Pacific. Rituals to greet the sea worms are reported from the neighboring Indonesian islands of Savu and Lombok (Fox 1979a) and more distant Pacific islands such as the Trobriands (Malinowski 1927), Fiji, and Samoa. The significance of the sea worms for the indigenous calendar is explained in chapter 12.

The "Bitter Chicken" of the Four-Month Ritual Silence

The ceremony to begin the ritual silence of four "bitter months" (wulla padu ) is a simple one, conducted first in the Sea Worm House in Tossi, then in Bukubani, and later in Hangga Koki in Bangedo and Kaha Malagho in Balaghar. In 1980, I witnessed the rite in Tossi, as raw rice was scattered to dedicate a chicken from each house in the village, with these words:

The spirits were assured that the human community would follow the

Before the pasola battle can begin, the Rato Nale, or "Priest of the Year"
(here, Ra Holo), must circle the playing field on a consecrated horse. 1980.
Photograph by the author.

calendrical prohibitions, and in return they were asked to provide an abundant crop for the next harvest.

Small offerings of a bit of chicken meat and cooked rice were placed in eight different locations in the house, drawing different domains of social life together meaningfully and relating them to the cycle of the seasons. The first share went to the right front corner of the house and the base of the main pillar, conduits to the Creator and the Great Spirit of prosperity (marapu bokolo ). Then other offerings were made to the Elder Spirit (marapu matuyo ) at the rock and tree altar in the center of the ancestral village, to the spirits of the dead along the front veranda, to the spirit of hunting and of domestic animals in the rear corners of the house, to the place where wealth objects are stored near the hearth, and to the guardian spirit of the gateway. In this way, the prohibition was spread to different domains of social life, and ancestors, affines, animals, and men were bound together to concentrate their energies on producing a successful harvest.

The prohibitions of the bitter months forbid noisemaking and a wide

variety of activities that could endanger the growth of young rice plants. These include children's games, large animal sacrifice, and the manufacture of white substances that might seem to compete with the production of full, white grains of rice. The prohibitions were summarized by the Rato Nale thus:

The complex list of prohibitions defines precisely those activities that, as nale festivities approach, will be permitted and even joyously engaged in. The origin narratives describe the sea worms as having been won through a series of contests, so the repetition of these contests (the top and discus, jousting, and beachcombing) serves to regenerate the vitality and fertility of the year. By means of temporal regulation, apparently frivolous children's games become a means for recreating the social order.

The period of the bitter months is a time of intense agricultural activity, when fields are prepared and the new crops of corn and rice are planted.

Most people live in their garden huts, usually located several hours or even days from their ancestral villages. They devote all day to preparing their gardens, with none of the distractions of ritual gatherings.

By late December or early January the crops have been planted, and the phase of preparation for the New Year starts with the moon called Nale Kiyo. The priest of the sea worms, Rato Nale, begins to "brood" adopting a pose of immobility and spiritual concentration, which is also ritually enjoined on the female mourner at funerals (kabuku kalalu ). He "sits like a hen on her eggs, like a sow sheltering her young" (bei myanu na kabuku tongo taluna, bei wyawi na karekongo anana ). By remaining totally inactive, he controls the forces that activate the seasons, keeping the calendar in place and maintaining time as it has been culturally constructed. Symbolically female during this period, he provides a model of self-discipline and self-control that rules by example rather than by decree.^[2]

The whole region would be in danger if the Sea Worm Priest were to violate any of these restrictions. Should he leave his house, fierce winds would tear apart the tender young rice plants. Should he eat any of the taboo foods, lightning bolts would strike the fields. Should he fail to control the calendar and coordinate feasting within a seasonal pattern, a tidal wave would engulf the whole coastline and destroy the ancestral villages. As the master of time, the priest of the sea worms represents the collectivity, both the land and the people, and should he fail in his duties, all would suffer.

From Taboos to License

As the young rice sprouts grow stronger, however, and as the time of harvest approaches, a lessening of these constraints begins, expressed in rites to make the early-ripening crops "bland" or "sweet" (kaba ) and harvestable for the festivities to come. The first marker of the relaxation of taboos is a small ritual named for the "sweet waters of the young sprouts" of the areca and Coconut palms (kaba wei kapoke ). A chicken is sacrificed with a prayer by the Rato Nale requesting that the people be allowed to gather their stores of areca and coconut to receive guests. Fresh branches full of young areca nuts will be placed as offerings on the tombs

[2] The rule is that he may not "cross over" into the women's part of the house, and his wife, in turn, must prepare his meals in a special area where no corn is allowed, and she cannot join him, as she usually would, to sleep in the right front bedroom. Except for this period, the Rato Nale has a normal family life. Ra Holo, at the time of my research, had a wife and five children, while Ra Ndengi had three wives and eight children.

of the recently dead on the morning that the sea worms arrive, and the coconuts will be processed into the thick cream used to cook chickens an rice for the festive occasion.

The full moon of Nale Kiyo, approximately three weeks before the predicted sea worm swarming, is when young people begin kawoking , wandering the beaches at night singing to the spirit of rice, calling he back from overseas to return to Sumba and provide a bountiful harvest During this period it is taboo to fish or spear octopus along the seashore so the beaches are deserted during the day but full of bustling, giggling nocturnal activity. Children bring out the tops, discus, and board game that were forbidden during the bitter months and scamper in the sand playing games of hide and seek. Young boys meet to practice patukengo , a form of traditional boxing, in which teams hold on to a single rope an swing their fists at each other to force the other side to let go. Groups o girls gather to watch them, and sing the bawdy kawoking songs: short compositions in light verse (lawitti ), which comment ironically on the upcoming pasola and its ethic of male bravery. They are answered by choruses of boys, who take up the same themes but turn them around t mock female resistance. The wordplay focuses on ordinary objects (the spindle, the loom, the sword, the trough), which are anthropomorphized and made to speak for their owners in comic fashion. The obscenity of the everyday offers a veiled commentary on male-female relations, represented in the homespun idioms of weaving, hunting, fishing, and scavenging along the beach.

Among the most common are two women's songs, which purport to treat the appropriately feminine domain of weaving but in fact comment on male propensities to seek other wives:

The ending of both these ditties, dogho damu , has something of the sense of "you silly little thing" and is used in instructing children. Thus, on one level the song instructs young girls on the proper technique for arranging threads on the loom, but its allusion is to men who try to satisfy two or three women at the same time and are not up to it.

The mockery is flirtatious at the same time that it is critical. During this period of nighttime revels young people are allowed the freedom to initiate liaisons, which they consummate, discreetly, in the darkness of the dunes. But although the songs play a role in courtship, they also serve to warn young girls against agreeing too readily to the demands of their suitors. A whole series of songs carry explicit messages to deflate male strutting and posturing, asserting that young men are not always as irresistible as they seem to believe:

The fondness of men for young girls, newly fertile, is mocked in the last verse, which asserts that they would do better to turn their attentions to the older women who might be more receptive. Implicit warnings against polygamy and age differences are found in the assertion that an older and important husband "cannot keep giving fresh betel" to satisfy his wife, especially if he has other wives to "feed."

Three other songs dwell on the heroic image of the pasola rider as he departs for the field of combat, dressed in a high-peaked bark headcloth and carrying a tall saber with a horsehair fringe. His costume celebrates his virility, as do the jangling bells around his horse's neck, but the songs

suggest that on his return from the combat these objects will look much less impressive:

These attacks on male virility present its splendor as only momentary and its weapons as weak.

Although the majority of the songs are sung by women, groups of young men can compose their own replies, such as the following verses:

The image in the first verse of the unattractive but sexually voracious woman finishes with a description of the trysting in the sand that often ends the evening.

The coy girl who refuses all her suitors until it is too late is the subject of another set of verses:

The girls reply to these verses with short ditties describing unsuccessful attempts to catch them and seduce them, using the idiom of the pasola battle and the male activites of hunting and fishing:

Shrimp and pigs are symbolically female animals, yet here they escape the traps and weapons of the hunter/fisher; even the horse, a man's closest companion, slips from his grasp and leaves the "hero" without a mount. Njelo Kaka, the name given to the "hero" in this ditty, is the protagonist of a series of myths of male achievement, but in the context of women's bawdy songs he is degraded and humiliated—closed off from the impenetrable facade of the woman's body and made into a grotesque parody of the dismembered organ.

The courtship of young boys and girls is supposed ultimately to summon images of agricultural fertility and abundance, reflecting the idea that the sea worms bring back the lost soul of the rice crop and capture it for the harvest. The link is made most explicit in one more serious song, which echoes the prayers of the Rato Nale:

The nale and pasola celebrations link human and plant reproduction, and provide a propitious time to begin an affair, which will come to fruition in the period of harvesting the mature crops of rice and corn.

At the end of the evening, the tone turns serious again, and a haunting ballad is sung to summon the distant nale and place the teasing and enticing courtship songs in perspective:

All of the songs are, in a sense, addressed to the spirits of the sea: the Mother of the Sea Worms, who swarm for these festivities, and her consort, the ipu fish, which swarms in March or April along the same beaches.

The revelries of kawoking continue for about ten days, until the moon waxes and the nights become dark. Starting on the darkest night, the Rato Nale then begins his official count. On the seventh day after the new moon appears, the sea worms should swarm.

The Sea Worm Swarming and Pasola Jousting

On the morning of the swarming, hundreds of people go down to the beaches with traditional woven dippers, ceramic jars, and plastic buckets. In the early hours of dawn, the tiny, multihued worms swim to the surface of the tides and begin to discard their genitalia after reproducing. The sea fills with colorful, pastalike strings, some only a few inches, others several feet long. Men, women, and children eagerly scoop the soft, wriggling bodies up, collecting them in a great number of containers. By the time the sun has risen to its midday position, the worms are largely gone, as they dissolve quickly in the harsh light of day. By then, though, the kitchens in the ancestral villages are full of nale , which are flavored with ginger and garlic and prepared as a condiment, a pungent mixture vaguely resembling anchovy paste. Prized as a delicacy, a rare and wondrous food from the sea, the worms will be eaten in this way along with the chicken and rice offered to the ancestors after the most heated battle of the pasola .

The spectacle of a mounted battle with spears, mandated in Kodi mythical narratives for the entertainment of these unusual supernatural visitors, begins with the massing of thousands of riders in full ceremonial garb. Men and boys arrive wearing several lengths of red, orange, and

yellow cloth tied around their heads and under the chin, supporting tall peaks reinforced with bark cloth or cardboard inserts to stand erect. Draped in mantles of indigo textiles and wrapped in loincloths (nowadays worn over shorts), they grasp their mounts with bare legs, stirrupless in the traditional style. Their horses are decorated with bells hanging from their bridles, cloth banners tied to their manes and tails, and sometimes even added tufts of horsehair at the neck and foot. No saddles are used, but colorful back pads are secured by a cinch around the chest. Members of the core villages of Pola Kodi—"the trunk of Kodi"—are pitted against the more recently founded villages of the other side of the river and the embankment (Bali Hangali, Bangedo). The two sides gather to face each other on the field, awaiting the arrival of the Rato Nale to start the event.

Although the Rato Nale does not participate in the jousting, no one can ride out until he has circled the field. He rides the Ndara Nale, a special horse raised from the herd in Tossi, which follows the same prohibitions as its master: it is not allowed to leave the village for the period of confinement, cannot eat corn or tubers, can never be bought or sold. The taboos observed by the Sea Worm Priest and his mount are supposed to protect the other horses and riders on the field, preventing serious injury or death. Although the shedding of some blood is part of the logic of the event, water taken from the heirloom urn is said to cure wounds of anyone who has not violated the calendrical taboos.

Participants in the pasola represent both ancestral personae who confront each other across the terrain of an early ritual division and ambitious individuals who want to impress others with their horsemanship. The contest is a forum for the display of personal feats of bravery and skill, but it produces champions rather than winners. While great riders are noticed and applauded by the crowds, no score is kept, and there are no final victors. The emphasis is on a display of vitality and fierceness—which of course is open to multiple interpretations. Not surprisingly, given the contentious nature of Kodi life, both sides usually return with claims that they have bested their opponents.

The combat warms up gradually. For the first half hour of the combat the event is more a parade than a contest as riders circle the field in long, easy, loping strides that allow the jingling bells on their horses' halters to ring out, while those watching can observe them and identify individuals by their colorful dress. The best riders do not enter the field until later, leaving the terrain open to younger riders who let their horses run loose before the spectators. Once the field has been trampled, the more experienced horsemen begin to challenge their rivals with verbal taunts: "Where

are you, the counterpart with bent knees, the partner with parted hair?" they will shout, using the couplet that refers to opposing sides in warfare or alliance—papa ndende kundo, nggaba horo longge . "Come to meet me in the wide pastures" they cry; "come to clash spears on the grassy field."

In the nineteenth century, the combat was conducted with real spears, and blood was quick to flow, but since the period of Dutch colonial control only blunt bamboo lances have been permitted. The goal of the jousting has thus shifted primarily to trying to unseat one's opponent, making him fall from his horse or causing the horse to trip. Groups of five or six riders set off in a formation, charging toward the other side and then circling back, with the famous champions in the lead and younger apprentice riders bringing up the rear. Short bursts of intense activity alternate with more leisurely circling of the field, followed by a sudden and concentrated attack.

The first blunt lances rarely strike their mark. It is only the most experienced riders, daring to get close to their opponents, who strike the head or flanks. Falls occur most often in the middle of the battle, when the rider is in great danger of being trampled by passing horses. Broken ribs, blindness, and neck injuries are frequent, occurring almost every year. Deaths from blows to the throat, eyes, or skull have occasionally occurred, but are rare, and are always explained as the result of a taboo violation.^[3] The dangers experienced by the riders are a way of testing their adherence to traditional norms and values, which ensure them protection, not to mention a forum for competitive display.

Three combats are held, but only the middle one is of great ritual significance. The day before the worms swarm, a small "training session" (dongu ndara ) is held in late afternoon in front of the village of Bondo Kawongo. This is said to "soften the horse's feet" and prepare the mount for the later combat. The middle combat, held the next morning at the westernmost field by the village complex known as the "trunk of Kodi" is the most intense, climaxing in the generation of ritual heat that coincides with the sun's ascent to its highest point in the sky. The combat continues for three or more hours, and is called to a halt when the Sea Worm Priest circles the field on his horse. People then return to their ancestral villages to roast a chicken to their ancestors and make offerings. In late afternoon,

[3] In 1980, a man was killed at the pasola in Wanokaka, and it was explained that he had defied the taboo against fishing along the seashore in the month before the sea worms swarmed. His wife reported tearfully that he had set out to spear an octopus, claiming that he was strong enough to resist the ancestral law. When he insisted on taking part in the ritual combat afterward, he was struck at the base of the breastbone, and the lance opened up his windpipe.

they gather near Tossi for the last and final battle, which helps to "cool down" the excitement of the morning and dissipate remaining aggressive feelings.

The pasola has been interpreted by some contemporaries as a substitute or exercise for war, remembering the not-too-distant period at the beginning of this century when the island was torn by tribal warfare and mounted parties of warriors took the heads of highland peoples. Nowadays it expresses an aggression that is more properly interpreted as reflecting the tension within the region between hierarchical ritual order and egalitarian competitive exchange. On the pasola field, a certain kind of supremacy, one based on history and precedent, is challenged by another, and in many ways more pervasive, mode: that of ostentatious achievement. Young men ride to make a name for themselves by impressing girls and prospective in-laws, and older ones display their ability to command, marked by the following they have among their juniors and the fine possessions that they have acquired.

The finest riders at the pasola often come from Tossi, the "mother-father village" of the whole region; but men from more recently founded villages on the other side of the river that have emerged as Tossi's rivals in a wider political context are also coming to dominate the proceedings. In the 1970s and 1980s, the informal leader of Tossi's horsemen was Tari Nggoko, a descendant of the first Kodi raja. Rangga Baki and Ratenggaro, the home of the third Kodi raja and a famous center of headhunting rites, are now the main challengers from Bangedo.

A champion, or "rooster" (maghailio ), is someone who has not been unseated in several years of riding across the field but who has caused many others to fall or bleed at his hands. The greatest champions may hold this title for ten or twenty years. None of them can hold it forever, however. The glory of the pasola rider is ephemeral. "So far" a current champion told me, "I have dodged all their lances and none of them has struck me down. But when my time comes, I will know that I have grown old, and cede my place to the younger men." The pasola is an arena in which daring younger men can challenge and even defeat their elders, defying the usual deference due to those senior in rank and social position.

Today, close police supervision is needed to keep the ritualized violence from erupting into actual battles or skirmishes on the sidelines. It is forbidden to wear bush knives onto the pasola field, to throw stones, or to charge directly across the opponent's lines, although angry riders do at times give chase. Hundreds of frightened horses and spectators retreat hastily if a rider loses control of his mount and breaks out of formation.

The pasola resembles other ritual combats found throughout Indonesia,

Local people walk down to the western beaches in the early hours of the
dawn on a date determined by the Rato Nale to collect the sea worms and
bring them back for a sacrifical meal shared with the ancestors. 1988.
Photograph by Laura Whitney.

such as Balinese cock fighting (Geertz 1973), Savunese rock-throwing battles (Fox 1979a), Minang buffalo duels, and other animal combats involving elephants, tigers, bulls, or oxen (Reid 1988, 1987). R. E. Downs argues that all of these confrontations began with the dualistic opposition of two groups within the society, who engage in "a ritual struggle representing the two halves of the universe" (Downs 1955, 59). Many present performances are interpreted by him as having lost this original meaning: "Often they would seem to have degenerated into mere sporting events indulged in at the time of general festivities, without any consciousness on the part of the participants of their religious significance" (Downs 1955, 55). As for the pasola , the struggle of opposing principles is still clearly tied to the constitution of the ceremonial system as presented in traditional narratives, though the particular form of the combat was probably imported from Java.

Consistent with the Kodi idea that the pasola was introduced from a

splendid kingdom to the west, the Malay epic Hikayat Hang Tuah mentions jousting as part of the entertainment at the Javanese court of Majapahit. The first detailed description of the weekly tournaments (senenan ) held in the squares near the royal citadel comes from a Dutch account of the events in Tuban, and is summarized by Reid (1988, 187):

About four in the afternoon the younger braves of the court would converge on the square after parading through the city on their magnificently attired horses. There they would engage in a series of charges and manoeuvres, one generally pursuing the other down the length of the field, with the aim of knocking each other off their horses with blunted spears. In reality, this happened seldom, and most attention was paid to the horsemanship displayed in the constant wheeling and turning on the square. The king was always present for these occasions and, at least in Mataram, took part in the jousting.

The similarities with present-day pasola are very strong, and the Kodi combat shares the aspect of serving as a "metaphor of war in which young aristocrats displayed their qualities" (Reid 1988, 188). Before 1600, the tournaments seem to have been quite bloody, as provisions are mentioned for the families of victims (Ma Huan, 1433; cited in Reid 1988, 187). The "knights" of Java were described by Tome Pires in 1515 as proud horsemen who often provoked deadly combat on an individual basis: "The noblemen are much in the habit of challenging each other to duels, and they kill each other over their quarrels, and this is the custom in the country. Some of them kill themselves on horseback, and some on foot, according to what they have decided" (cited in Reid 1988, 187).

Jousting tournaments declined in Java as warfare became less frequent, and they were gradually replaced by animal combats. The coupling of human and animal combat suggests that the blood of one has often been used to represent the blood of the other. Ritual cock fighting on Savu is held in conjunction with ceremonies to Welcome sea worms, and Savunese explanations of the combat closely parallel the pasola : "It is a struggle between classes of spirits represented by men, the spirits of the land the spirits of the sea; and it is a substitute for warfare between clans and villages" (Fox 1979a, 165).

Sumba, in a way, is similar to the societies of classical Greece and Sparta, where athletic champions were often political and military leaders as well. This seems to have changed somewhat with the suppression of interregional warfare in the 1920s, as well as the institution of present government policies controlling feasting. In effect, an informal and competitive

political system is increasingly being replaced by an externally appointed administrative hierarchy in which horsemanship is no longer an important factor. Thus, the pasola is being promoted now as a folkloric celebration and a way of entertaining visiting dignitaries; it is no longer a religious rite and the basis of the Kodi polity.

Roasting a Chicken to the Dead: Sacrifices in the Ancestral Villages

At the end of the second pasola battle, people return to their ancestral villages to prepare a meal. A somewhat chaotic scene follows, as hundreds of chickens are sacrificed to those who have most recently died, whose graves were scattered with betel nut in the morning. On this occasion, all the various members of a house unite with their lineage brothers on an equal footing, without the hierarchical divisions of the feasting season. Instead of competing to see who can bring the buffalo with the longest horns or the largest pig, participants at the nale feast all come with a single chicken, which they share with their ancestors. The markers of rank and seniority are conspicuously absent as everyone shares cheerfully and generously in the reserves of rice set aside for the feast, flavoring their meat with small tasty servings of the sea worms themselves, spooned on the side of the plate as a relish.

The nale feast is the only one in which everyone can be included because no embarrassment is involved if one cannot bring an adequate contribution. It gathers together all the members of a house and can, as a result, be an occasion for the telling of mythic narratives and the singing of songs forbidden for so long. It can also be a time to discuss new projects for the dry season: the erection of a new stone grave, the building of an ancestral house, the giving of a feast. The marapu must be informed of these projects through sacrifices, and may also be told of changes in residence or group composition, recent marriages, or a preference for male or female descendants. At this gathering, "no one is afraid to come with empty hands, and no one must go home with an empty stomach." Defined by its simplicity, the feast is also generalized and shared.

The entrails of chickens sacrificed at the nale feast are read for auspices of the coming year. If they contain black spots or red flecks, it is a sign of disease or death in the house. If they are twisted or crossed by membranes, there may be stealing or adultery among the descendants of a particular ancestor. But if they are straight and firm, suggesting the straight flow of blessings from previous generations into the present, the signs are positive and everyone can celebrate.

In the most important ritual villages, private household sacrifices are followed by a public meal attended by male representatives of each cult house. The first share of chicken and rice is dedicated to the sea worms, and other portions are set aside at the clan altar, household altars, and on the tombs of important ancestors. Because this food is specially consecrated as ngagha nale or the "rice of the sea worms" it requires a personal purification before it can be eaten. If any participant has committed theft, adultery, trespassing, the stealing of rice souls, or the breaking of food taboos, he must first confess to the Rato Nale before partaking. The price of spiritual cleansing seemed too high to several of the delegated representatives whom I observed, for many chose not to eat their share of the meal. They respected the ritual context of commensality, but denied its power to regulate their lives.

Time, Carnival, and Disruptive Revels

Mikhail Bakhtin, writing about the folk culture of the Middle Ages and the Renaissance, notes that

the feast is essentially related to time . . . to moments of crisis, of breaking points in the cycle of nature or in the life of society and man. . . . As opposed to the official feast, one might say that carnival celebrated temporary liberation from the prevailing truth and from the established order; it marked the suspension of all hierarchical rank, privileges, norms, and prohibitions. Carnival was the true feast of time, the feast of becoming, change and renewal. It was hostile to all that was immortalized and completed.
(Bakhtin 1984a, 9-10)

The nale festivities can be interpreted as a carnival-like feast, bringing an end to the bitter months of prohibitions and marking the ritual calendar, much as carnivals in Europe marked the last chance to indulge in revelry before the privations of Lent. Incorporating elements of obscenity, rivalry, and ribaldry, nale involves a temporary dissolution of hierarchy and the playful defiance of order.

The sharing of the nale feast celebrates a particular temporality—the repetitive, regenerative time of bodily processes—and an ideal of equality and common substance. It contrasts in this respect to the feasts of the dry season held for stone draggings and house buildings, and with celebrations of those who have taken new titles or achieved competitive renown. In Bakhtin's (1984a, 10) terms, the ladder-of-prestige feasts are "consecrations of inequality" focused on the display of wealth and regalia to com-

municate differences in rank and genealogical order. At the nale feast, by contrast, economic differences are invisible, for each family sacrifices only a chicken. The pretense of immutability that is maintained at other occasions is challenged by a festive laughter, and hierarchies are contested by dissident voices. The great leaders of the ritual centers may be knocked off their horses by younger challengers, and arrogant husbands may be prodded and teased by women's clever puns and songs. In playing with apparently fixed categories of social discourse, participants show them to be ephemeral and relative, a revelation that may prove less threatening than reassuring. The ritual framework defines a space within which license is possible, and places parentheses around established verities. The nale festivities are a time of laughter and excitement, when the tensions between men and women, between established ritual centers and peripheral ones, and between old and young are parodied with droll abandon.

There are several possible interpretations of the combination of ribaldry and reverence in the nale celebration. One, dating back to Frazer, is based on the linking of human and plant fertility, with the rite anticipating the harvest and using magical principles to make the young crop more abundant. A second, drawing on the work of Victor Turner, interprets the egalitarianism as expressing "communitas," an intensified sharing and dissolution of differences, as Geirnaert Martin (1987) has argued happens in Lamboya rice ritual. But in these cases the veiled hostility marking the exchanges between riders and between men and women singers seems out of place. The ritual discipline required by the Rato Nale and the respect shown for the sea worms and the spirit of rice contrast sharply with the disrespect and even mockery directed not only at human failings but also at other participants.

I would therefore propose a third option, which is to see the nale festivities as exploring and celebrating the regenerative powers of differences—between men and women, juniors and seniors, past and present. If, as Bakhtin (1984a, 10) says, laughter is the enemy of hierarchical structures, then the ribaldry can be interpreted as asserting the heterogeneity of ritual participants and their resistance to official authority. The Rabelaisian flavor of the exchanges along the beaches reveals a various lines of conflict between source and periphery, between the values of control and those of joyful release. This confrontation is muted in much of the festivities, but surfaces in the songs and combat, and is never finally resolved.

The deference rendered to the authority of the Rato Nale would seem to be an exception to this defiance of ritual order. It is, however, an exception that proves the rule, since his authority is maintained only by

identifying it with immobility and passivity. Because he cannot act, the Sea Worm Priest can serve as the anchor of the traditional calendar. Because he cannot speak, his preeminence cannot be challenged with verbal taunts. The fragile consensus of heterogeneous groups collected together for the occasion can be held together only by an empty center, where power is retained by an image of powerlessness. As we saw in the last chapter, the timekeeper has on many separate occasions been forced to surrender his authority over other concerns in order to preserve his ritual position as one beyond controversy; the nale festival is just such an occasion.

The carnivalesque aspects of the nale celebration do not represent a mere moment of antistructure in a dialectical process of structure-making (V. Turner 1969). Rather, they express divergent political sentiments, bringing opposing forces into contact to affirm their power to generate a new whole. The dualistic opposition of the sexes and of the two sides of the pasola is an effort to order the relationship between "antitheses that cannot be allowed to become antipathies" (Maybury-Lewis 1989). While the principles they represent may seem contradictory, their combination can be creative rather than destructive. A balance of contending forces is assured by allowing them to meet in the bounded arena of institutionalized "play." The authority of the Rato Nale, the ultimate encompassing frame for the calendrical rites, is not challenged; yet his ascetic discipline and withdrawal enable others to enjoy a time of cheerful, chaotic disorder. My argument is not that the carnival-like celebrations of nale reinforce the serious institutions and rhythms of society, but that they make them possible to live with.

A Century of Pasola Performances

The implication of the pasola in larger historical processes is evident if we examine changes in its performance over the last century. At the end of the nineteenth century, when Sumba was still caught up in internal warfare, the pasola was performed on a relatively small scale, with only the participation of those people who lived in the ancestral villages of Kodi Bokol and Bali Hangali.

The fathers and grandfathers of contemporary riders were horsemen whose skills were valued not only as entertainment (for both humans and spirits) but also as part of the practical arsenal of warfare. In the late nineteenth century, Rangga Baki and Ratenggaro were locked in a deadly pattern of murders, feuding, and livestock raiding. The two villages, however, suspended their hostilities at pasola time, when they rode together

against their opponents from Tossi and source villages of the "trunk of Kodi." One of the region's main pasola champions and fiercest warriors, Rato Muda of Rangga Baki, was said to have been approached to serve in the colonial administration. He refused, saying, "I still have my brother's blood to avenge"—meaning that he could not serve as a peacekeeper under Dutch rule.

The season for headhunting, local feuding, and warfare was July through September, after the harvest and toward the end of the dry season, when agricultural activity was at a minimum. Pasola falls at the other polarity of the calendar, in a time of agricultural cooperation, when people live in dispersed settlements near their gardens so they can weed and harvest the crops of the rainy season. Thus, although the ritual contest held near the beach provided a forum for impressive, competitive displays of force, it was stringently distinguished from actual warfare.

The pasola always has two opposing sides, but their composition has shifted over the years. The origin narrative pits the villages of Pola Kodi against Mbali Hangali. In the 1920s, shortly after the Dutch pacification, the people of Bangedo came to join in the fray, riding with their neighbors from Mbali Hangali. During the rule of the first two Kodi rajas, Rato Loghe and Ndera Wulla, the performance of the last battle, on a field close to Tossi, became a celebration of the power and influence of the Tossi rajas. As the largest village in precolonial Kodi, Tossi produced both the finest horsemen and the largest group of riders taking part in the ritualized combat, establishing a reputation that still lives on in the present.

In the 1930s, Dutch colonial officials decided to encourage the growth of the pasola as a "stabilizing" celebration of the unity of the polity and invited participants from the whole region. They proposed new rules that would forbid the use of sharpened spears and substitute blunt bamboo lances. Less blood was spilled, but more riders dared to join in, and a new group of challengers from the regions of Bangedo and Balaghar formed to test the skills of the Tossi riders. Since a separate rajaship also existed on the other side of the river, what was at stake in the combat became a struggle for symbolic domination of the whole area.

In 1945, upon Ndera Wulla's death, power passed into the hands of H. R. Horo of Rangga Baki, Bangedo. At the ceremony inducting him into office, he ordered the first performance of a mock pasola on the field between Rangga Baki and Ratenggaro. In the 1950s, this "new pasola " was repeated in conjunction with the calendrical rites to celebrate his power and demonstrate the horsemanship of the people of Bangedo. The ritual combat held outside Tossi remained, however, the most important one, and the one that has consistently drawn the largest number of participants.

Raja Horo was criticized by many people in Kodi Bokol for imitating the form of the pasola without an appropriate ritual mandate. As the climate of criticism of local rulers grew more intense, a new rhetoric opposing the Dutch-appointed rajas as "feudal" emerged in the newly independent state of Indonesia. When Raja Horo became the district head (camat ) of Kodi, he stopped performing the pasola near his own ancestral village in an attempt to allay attacks by rivals in Tossi. The fires that raged in the early 1960s destroyed the villages of Rangga Baki, Ratenggaro, and Tossi, but did not stop the pasola , which apparently became all the more violent as the region seemed on the brink of significant social change. The bloodshed of 1965 was much more limited on Sumba than in most of Indonesia, but it upset an image of larger stabilities and turned local inhabitants back to their own ritual system to seek an equilibrium.

At the same time that the people of Tossi were rebuilding their cult house and designating a successor to the office of Rato Nale, the pasola was becoming recognized as a form of local "art" that should be "developed" to entertain outside visitors and even attract tourists. In the 1970s, government funds were allocated to establish temporary housing for visitors and arrange for charter flights at the time of the pasola . Official efforts concentrated on the pasola held in Wanokaka in March, since it could be reached in an hour and a half from the regency capital of Waikabubak, while Kodi required a full day of travel and, by the end of the 1980s, the road to Kodi was still not paved. But news of busloads of foreign visitors who came to watch the performances began a subtle transition in the way the Kodinese themselves viewed the procedure, moving it gradually from the category of "ritual" into that of "spectacle." In particular, it shifted the relations between the riders and the onlookers, for in a ritual the onlookers are also participants, whereas onlookers at a spectacle become disengaged and are transformed into a mere audience. At the pasola , the subtle interplay between young men and the women who both mock and praise them with bawdy songs was thus replaced by a greater separation, and the carnivalesque elements of an all-embracing revelry became domesticated into passive observation.

In his analysis of the differences between carnival and spectacle, Bakhtin (1984a, 7) argued that "carnival does not know footlights": its performances are not framed by distinctions between actors and outsiders, performers and spectators. The subversive force of laughter confronts this boundary and dissolves it. In the past, this may also have been the case at the pasola . I was told that skilled women singers would continue to tease the riders with bawdy verses even when they were on the playing field. Men were the main actors on the stage, but women could challenge them.

At present, however, this behavior would be inadmissible, because the pasola has been framed as a spectacle, with official police forces hired to maintain order. Women remain on the sidelines, participating only by ululating (kaghilikongo ) in imitation of the excited cries that used to greet victorious war parties returning from headhunting raids.

The new official staging of the pasola has also changed its relations to calendrical rituals and to the forms of temporality enacted in the performance. Summarizing Bakhtin's notion of the disruptive power of carnival to move through official time, Holquist (1984, xviii) notes that "the sanction for carnival derives ultimately not from a calendar prescribed by church or state, but from a force that pre-exists priests and kings, and to whose superior power they are actually deferring when they appear to be licensing carnival." This force, which we could term a vital physicality, is represented in the narratives and songs associated with pasola performances. The presence of carnivalesque elements was related to their power to evoke regeneration, to assert that because complete repetition is impossible new life must come from the death and rebirth of the body, in all its grotesque forms.

Official development of the spectacle has focused on a more clearly defined "stage,' elaborate costumes, and orderliness but has not yet sought to change the vulgarity of the verses women sing. The songs are described as "traditional merrymaking" in the official publication on the pasola , and visitors have the impression that they, like the priests' prayers, are respectful invocations of the ancestors. The boundary-crossing of the bawdy songs remains a key characteristic of their power and gives them a pivotal role in commenting on male-female relations. Without the ability to disrupt the spectacle and "steal the stage" in a carnivalesque parody, the role of women in the rite is greatly curtailed.

Rites of Regeneration

The stages of the nale celebration reflect the familiar pattern of rites of passage that move from order into a liminal period of disorder and then back to a new, regenerated social order (Van Gennep 1960; V. Turner 1969). They enact the past by creating a context in which people can assume the roles of ancestral personae, but at the same time that they constitute temporal relations they also offer a site for their disruption.

The nale festival presents a conjoining of male and female, where the complex centered on female symbols of rice, sea worms, and fertility is brought to meet male symbols of horsemanship, warfare, and virility. The songs that anticipate the pasola show an insolent lack of respect for male

spheres of achievement, and remind the men that all this prancing about on stallions must be followed by sacrificial offerings to a female deity, the great Mother of the Sea Worms, Inya Nale.

The complementarity between men and women is contested and challenged, but only to a certain extent. As Turner (1969, 176) has argued, most "rituals of status reversal" provide reversals that are only transitory, and may finally reaffirm "those categories which are considered to be axiomatic and unchanging both in essence and in relationship to each other." He maintains that ritualized hostility can actually serve to reinforce a pattern of complementary but fundamentally unequal cooperation: "Cognitively, nothing underlines regularity as well as absurdity or paradox. Emotionally, nothing satisfies as much as extravagant or temporarily permitted illicit behavior. Rituals of status reversal accommodate both aspects." However, since ritual performance heightens an awareness of the past and the heritage of the ancestors, its temporal sequence can lead in either of two directions: it can, in "successful" performances, reconstruct the fragile hierarchy of the original ancestral consensus or, in "confused" ones, contribute to its dissolution.

Writing on Western historiographic practice, R. G. Coilingwood (1946, 282-302) proposes that the historical method is based on an imaginative "reenactment of past experience" in which documentary evidence is reconstructed into a coherent interpretation. In a nonliterate society, where "documentary evidence" takes the form of tombstones, heirloom urns, and sea worm swarmings, ritual provides the locus for an imaginative reflection upon the past. The material traces of the ancestors are assembled for collective examination, their names are repeated in prayers and invocations, and their continuing power is "tested" through propitiatory offerings. This ritual experience is, in the end, a form of self-knowledge. By playing ancestral personae, participants come to rethink their mode of relationship and may find the means to change things.

The conflicts that underlie these festivities surface in the temporal opposition of order and disorder and the hierarchical division of source and executor. The nale season begins with months of prohibitions, restrictions, and ritual silence. All joyous, noisy behavior is suppressed to concentrate the population's attention on the activities of planting. The Sea Worm Priests, as descendants of the elder brother, bear the heaviest taboos and remain immobilized in their houses, while the descendants of Rato Pokilo roam around, training their horses for the pasola but avoiding the sacred field of combat and the sands near the Sea. Breaking any of the first set of taboos on the priests would endanger the coming rice harvest ("shaking the tender grains loose off the stalk"); breaking any of the

second set would be personally damaging to the individual violators (who could suffer illness or injury) but not to the region as a whole. The hierarchically superior elder descent line is identified with the unified territory, while the junior descent line represents the more particular concerns of its separate members.

As the time of the combat approaches, the complementarity of senior and junior houses is contrasted with male/female complementarity. Women offer an irreverent commentary on the male prestige system with their kawoking songs, even as they prepare to travel as spectators to a glorification of the male qualities of bravery, flamboyance, and daring. The ritualized confrontation "frames" an arena within which juniors may challenge seniors and women may tease men. At the end of the game, these differences are reabsorbed into a wider order as the Sea Worm Priest stops the combat and all participants return to their ancestral villages to sacrifice chickens to their ancestors. Male heat is "cooled" by symbolically female ritual officers; the excitement of the play is transformed into an injection of fertility and vitality for the whole region.

Occasionally, however, the ritual transformation does not succeed, and the reconstruction of a fragile hierarchy dissolves into chaos. The pasola can provide an occasion for outbursts of real violence between parties at odds with each other, so it is now heavily patrolled by the local police and army. In 1980, a rock fight broke out between the people from the villages of Rangga Baki and Tossi. In 1981, swords were bared and one person was struck with the edge of a bush knife in a fight that occurred beside the playing field. In 1988, women's teasing songs on the way to the spectacle in Wanokaka provoked the brutal murder of a woman by her former husband, who left her body at the edge of the playing field. Although individual acts of violence are prohibited by the rules of the game, they are (like brawls between players or spectators at football games) expected and hardly a structural anomaly.

The assembly of large numbers of people in communion with the spirits of the dead and the forces of fertility provokes a combustion of vital forces, which can regenerate the ceremonial system for the coming year or degenerate into a chaos of conflicting claims. These dangers are inherent in the performance of "ritualized violence" since it is still, m a very real sense, actual violence. The ceremonial frame that should serve to contain aggressive behavior may also promote and encourage it. Male exhibitionism is intensified by the undercurrent of mockery expressed by female spectators. The daring of representatives of the peripheral, younger-brother villages is exacerbated by the constraints on timing and ritual control invested in the "mother-father" villages of Tossi and Bukubani. The pasola

represents a chafing under certain types of authority and an opposition of principles of ascription and achievement, which can prove explosive.

As a reflection on the past and an enactment of the ceremonial divisions outlined in traditional narratives, nale and pasola celebrations encourage a reimagining of historical relations, in which the continuity of the Kodinese cultural heritage is displayed by the flexible and contested nature of such ritual confrontations. The antagonistic mode of this central ceremony of unity reflects a social conflict between equality and inequality. Although the competitive struggles of warfare and feasting may be provisionally absorbed into a more encompassing ceremonial whole, they always threaten to break out of these boundaries. In fighting fiercely against each other, the Kodinese acknowledge a mutual origin and a set of shared values. Only through this violent and dangerous game can ritual equilibrium be restored and the idea of an unstable but essential cosmic balance maintained.

Disparate Voices in a Unified Calendar

The multitude of voices heard in the bawdy songs sung along the beaches and the many antagonistic confrontations of clashing lances on the pasola field are juxtaposed to a single evocative silence: that of the Rato Nale, sitting immobile on the veranda of his ancestral house. He does not participate in the fray, but has the power to begin and end the combat by circling the field on his horse. He does not sing teasing songs, but he leads the groups who entreat the spirit of the sea worms to return to Kodi.

The exuberant, carnivalistic display of vitality that occurs at the time of the Kodi New Year is made possible by the ascetic withdrawal of the Sea Worm Priest, the "Lord of the Year" whose disciplined passivity contrasts so sharply with the revelry and license of all others. In this respect, despite all the Bakhtinian echoes of subversion, the participants still acknowledge a particular kind of superordinate, all-encompassing authority. This authority derives from the Rato Nale's power to begin and end the year, providing an indigenous control of time and the cycles of the seasons.

The position of the Rato Nale is at present an embattled one. His authority to set the dates of the performance and regulate its occurrence is being challenged by officials of the Indonesian government, who are imposing an external calendar as an arm of a secular, universalistic, bureaucratic state. The complicated ramifications of these changes will be discussed in chapter 12. Before coming to them, however, we must first understand that the ritual reenactment of the past, in all its diverse and

disputatious aspects, is itself a political act. While many people in Kodi are anxious, confused, and divided about what to draw from the past and how to sustain it, when they gather together to make nale sacrifices and ride on the pasola field they are united in a collective effort to keep a part of their past alive. Retaining ideological control over tradition is extremely difficult. It involves an ongoing process of compromise, reinterpretation, and adjustment to new circumstances. The unity of the Kodi people is expressed in the calendar and in the rites of the New Year, which provide the basis of Kodi cultural identity and Kodi polity. With the present threat to such unity, these reenactments become all the more important as feasts of becoming, change, and renewal which continue an ancestral heritage into the future.