The Long Conversation: Fieldwork Conditions and the Study of Time

I did not come to Sumba to study the perception and organization of time. My dissertation proposal, formulated after reading what few materials on Kodi had been published by missionaries and visitors, focused on the relation of mythology and social organization.^[1] Rather, I was only slowly initiated into the topic over the course of seven different periods of ethnographic research that spanned more than a decade (1979-89). I did, though, come with an interest in the cultural construction of the past, and particularly in the narrative definition of precedent, which ethnographic studies had indicated would be expressed in elaborate genealogies in parallel verse (Adams 1969, 1970; Fox 1971).

This past itself was more alive and influential than I had expected, because in the early 1980s the traditional religion of marapu worship was still practiced by three-quarters of the population. Spirits of the dead were invited to all important ritual events and fed sacrifices, as well as regularly blamed for inflicting misfortune on their descendants when promises or obligations were not fulfilled. Precedence was invoked as an ordering principle but continually contested and renegotiated. Genealogies were short and alliances unpredictable; hence, the unity of the Kodi people took shape mainly through adherence to the yearly calendar. In sharp contrast to the stratified societies of East Sumba, the Kodi people had no royal

[1] When I arrived on Sumba in 1979, the only ethnographic materials concerning the island available in the United States were the collected essays of the missionary-linguist Louis Onvlee (1973), a series of articles by art historian Marie Jeanne Adams (1969, 1970, 1971a, 1971b, 1974, 1979), and some older missionary writings (Kruyt 1921, 1922; Wielenga 1911-12, 1916-18). Kodi was described in a fascinating but brief article by Van Wouden ([1956] 1977), and some Kodi fables had been published by Needham (1957b). Since that time, however, there has been an explosion of interest and research, including the essays in Fox 1980d and new ethnographic studies by G. Forth (1981), Kuipers (1990), Renard-Clamagirand (1988, 1989), Geirnaert Martin (1987, 1989, 1992), Keller (1988), I. Mitchell (1981), and Keane (1990).

genealogy that ordered collective memory or produced a single master narrative of regional history. Over the next ten years, subsequent field trips led me to question whether the period they called la mandei lama ulu ("the past") was one thing or many different things.

Despite their initial standoffishness, Kodi people proved remarkably open and hospitable once they heard that I had come to study their language and culture. In Java, I had met a Kodi student who arranged for me to spend the first three months of my stay with his aunt, Gheru Wallu, in the market center at Kory in Greater Kodi. A rigorously traditional woman who spoke almost none of the national language, she was also a skilled herbalist, masseuse, and midwife. As my skills in the language gradually improved, I shifted to a new location closer to the ancestral villages along the coast—ritual centers and site of the calendrical festivities. I moved into a small house in Bondo Kodi that had been built for a nurse and set up my own home, living with two local girls, Maria Rihi and Fenina Manu, who helped with cooking and washing. Finally, toward the end of my stay, I spent two months in the distant river valley of Balaghar as the guest of my teacher Maru Daku.

During most of my fieldwork, men served as my "teachers" while women were my "companions." The important men who consented to become my instructors and guides in the arcane world of ancestral custom called me their "student" and "daughter." The women who lived with me in three separate households and helped me with the practical matters of life called me "sister" and "friend." Because of the way my gender and relative youth were culturally construed, my relations with men were cordial but hierarchical. They knew I wanted to collect and compile socially valued forms of knowledge, and they respected my work because I also respected them. Women, however, usually excluded from competitive claims to possess such knowledge, teased me about my earnestness; it was they who gave me the Kodi name by which I am still known there: Tari Mbuku, the name of a female ancestor but also, interpreted in Indonesian, having the double meaning of "looking for a book."

When I first began "looking for a book" I expected it would be an analysis of genealogies and traditional narratives in relation to the system of kinship and alliance. The material's I finally brought together for my dissertation, however, focused on the feasting system and its basis in spirit worship (Hoskins 1984). In turn, feasting was an arena for achieving renown and playing out the politics of exchange, which, I came increasingly to realize, was connected to a very different form of temporality.

The puzzles that eventually became the subject of the present study emerged in conversations with four Kodi "teachers" each of them a "man

of knowledge" who controlled a different aspect of local temporality. The first "knew stories," the second "knew history" the third "knew how to sing the words of ritual" and the fourth "held the rites of the year."

Maru Daku, a famous bard and respected elder, was unrivaled in his mastery of Kodi verbal lore ("The book of our customs lies underneath his skin," a friend once said of him), and he was a figure of great authority, if a controversial one. Maru Daku's command of traditional narratives held his listeners in thrall, even when they disapproved of many aspects of his own life. An early convert to Christianity, he eventually repudiated the church leadership and returned to marapu worship. His great inventiveness with words was both praised and suspected and (as happens so often in Kodi life) won him an audience but not always followers.

Hermanus Rangga Horo, my second teacher, had been not only the last Dutch-appointed raja of Kodi but also the head of island government for a period after independence. He believed strongly in scholarship and recordkeeping; indeed, the personal notes and journals he kept over the eighty-seven years of his life were the only surviving local archive. I draw on them, and on his own vivid memory of past events, frequently in these pages, as well as on many discussions of custom and local litigation, the problems of governing a remote island only now "breaking into history," and the transformations that had occurred since independence.

My third teacher, Markos Rangga Ede, was a priest and singer who carried out marapu rites in his homeland of Bukambero and throughout the district of Kodi. Boasting a baritone so forceful it could "tear apart houses" with its tones, he was a traditionalist leader who served at times during my stay as the ward clerk and ward headman of Bukambero; he was always a strong local personality. For eight years I have followed him to rituals, recorded his songs and the dialogue of orators and diviners, then spent days going over the material with him to understand it in all its complexities.

The highest-ranking priest in Kodi was a "Father Time" figure, the Rato Nale, or "Lord of the Year." In Tossi, the ritual center of the domain, this office was held by Ra Holo, a patient and considerate host to me on my many visits. His own modesty and self-effacing demeanor contrasted strongly with the heavy obligations of his task; as the embattled nature of his position grew ever clearer to me, I came better to understand why he seemed reticent and taciturn on particular occasions. His counterpart in the outlying village of Bukubani, Ra Ndengi, agreed to narrate certain myths and allow me to observe the ceremonies of the new year; he was not, however, as reflective and questioning of his own task as Ra Holo.

Death interrupted the long conversations I began, taking away many

of the strongest voices in my fieldwork. Maria Rihi, my "younger sister" and companion in my first home, was tragically killed in an accident at the end of my first year of research. A year later, Maru Daku fell seriously ill and recited a touching last testament to family and friends in my living room (Hoskins 1985). He later recovered and was able to bid me farewell, but I received news of his death as I was preparing my dissertation. His nephew, Ndara Katupu, who helped to compile and transcribe his words, died some months later. My host for return visits in 1984 and 1985, H. R. Horo, died before we returned on a film project in 1986 and 1988. And a graduate student from the Australian National University, Taro Goh, who attended the funeral I describe in chapter 9, died in the hospital of East Sumba six months after beginning his own field research on the island. The loss of these people, sudden and unpredictable because many were so young, haunted me during my writing and analysis, and no doubt contributed to the focus of this work on temporal notions as a response to human mortality.

The "sitting work" of transcribing texts, analyzing them, and interpreting them always took much more time than the "walking work" of traveling to distant villages and attending rituals.^[2] It was while sitting with women, chewing betel, stringing cotton thread onto a loom, and untangling recently dyed yarns that I first stumbled on the idea that different temporalities could be the key to understanding much of the ceremonial system and its transformations. The shrewd and often sarcastic commentaries of women who watched from the darkened hearths at the center of Kodi houses encouraged me to look more closely at the ordering of events and lives in time. Their narratives also alerted me to the special significance of objects, because whenever I asked a woman for the story of "her life" I was given instead the story of an object, an exchange valuable, or a domestic animal. Whereas men narrated accounts of how they became the owners and exchangers of wealth, women confided indirectly, telling stories of possessions that had been tied to their own identities and then been taken away.

Because this study examines the public, shared world of the calendar and the annual cycle, it only touches on the more private world of domestic objects, persons, and gendered selves. The wider temporal perspective of the calendar, the exchange cycle, and the encounter with history is not an exclusively male one, but it does give important men a greater voice than

[2] Among those who worked with me on the transcription and evaluation of a number of texts, I would like to thank A. W. Bulu, Gregorius Gheda Kaka, Lota Mahemba, Yusup Ndara Katupu, Gideon Katupu, Ndengi Yingo, Radu Yingo, Rehi Pyati, Andreas Ra Mone, and Deta Raya.

their wives or daughters. The more hidden world of household politics and personal memory deserves separate treatment (Hoskins 1987b, 1988c, 1989c, in press). The conversations I had with men centered on time as an encompassing system of order; in those with women it was presented more as a dimension of biographical experience. Eventually, I hope to write about both perspectives.