Long-Term and Short-Term Cycles

Both kinship and alliance link individuals, with their own emotional attachments and needs, and social groups. Individuals may develop matrimonial strategies that stress short-term tactical "moves" related to practical goals—assembling the bridewealth, getting assent for a desired match—but these individuals are also part of houses and villages with long-term developmental projects. In Sumba (as in most of Eastern Indonesia) these projects are expressed in terms of the desirability of closing a cycle: the "reunion" of the descendants of the brother and of the sister, the "return" of heirloom valuables that have circulated outside of the house, the "bring-

ing back" of a matriline so that a grandfather and his grandchildren will be of the same walla . These projects are articulated within the context of long-term affinal debt, with a temporal span almost always longer than the life span of any given individual.

It is, I believe, the extended temporality of ties based on blood and marriage that makes their regulation more one of long-term projects than short-term strategies. Bourdieu's (1977, 34) opposition of "official kinship" and "practical kinship" criticized anthropologists for basing their theories on the retrospective analysis of genealogies rather than the ongoing strategies of local agents. But the strategic advantage of a delayed return means that the "result" of an alliance in exchange politics often emerges only after the fact . Hence, if we accept his position that temporal units achieve power only through their play within a given structure of relations, most of the societies of Eastern Indonesia would have to be said to "play the present with an eye on the past," because they stress the durability of exchange relations over time and through multiple individual relations. An intergenerational model of exchange is needed, in which not individuals but "houses" emerge as moral actors, the transactors in the often rivalrous arena of marriage.

If the "house" is the locus of strategic calculation, it is perhaps not surprising that the Kodi represent this fact by seeming to attribute volition and even calculation to the house and the objects within it. A gold pendant that leaves the house, for instance, is said to "yearn to return home" (kareiyo balingo ) when it travels to a distant region, and thus to influence human actors to seek new brides from the daughters of its former masters. A buffalo cow received in bridewealth bears a calf who "wants to retrace her footsteps" and go back to the corral left behind, through a son who marries his father's sister's daughter.

The curiously animated objects of Kodi exchange reflect a displaced agency, in which collective strategies are attributed to the valuables and their location instead of to the actors who manipulate them in particular instances. Agency is also sometimes displaced onto the ghostly influences of ancestors, who may tell their descendants they want a cycle to be completed or (most often) want their own place filled by a close relative. The pressure from a deceased wife is cited as a prime reason to renew the alliance by marriage to her younger sister.

Different projects have, of course, different time scales. Just as people must assemble in order for events to "take place," these people must also "take time" from their own lives and devote it to a particular activity. Some institutions and activities, however, carry more "weight" than others (in a metaphor that is as common in Kodi as in English). "The greatest

burden" any Kodinese will tell you, "is that imposed by the house." Its duties are the "heaviest" (rehi mboto ), surpassing even the onerous demands of affinal relationships, and its sanctions the strictest.

Because the house is the temporal center of gravity in Kodi social life, its continuity provides an anchor for more free-floating notions of individual life projects and obligations. Alliance provides the primary means of testing that center and of "weighing" the import of a particular house against its rivals. Thus, it is a mark of prestige and achievement for a house to be filled with "heavy" valuables, which cannot be easily moved along alliance paths. The truly wealthy are those who are able to withdraw their goods from circulation, to sanctify them as "inalienable wealth" (Weiner 1985) and keep them as an exclusive treasure. Few houses ever attain that status in the fierce exchange rivalries of Kodi, and in fact only a very small number of cult houses have supposedly "immovable" objects within them. The accumulation of "lighter" goods, which circulate at marriages and funerals, gives a temporary, shifting sense to status that is always contestable.

Houses strive to appear immobile in time, to hold on to their wealth, and to elaborate and embellish the house site with new objects that are also unmovable. Because of the shifting requirements of swidden cultivation, the locations of gardens and hamlets are constantly being changed; correspondingly, a complex ritual apparatus has been erected to legitimate each change. Movement is required in order to produce new crops, as the movement of women from one village to another is required to produce new descendants. But the striving for immortality is a striving against movement and toward stasis—the always desirable but elusive stasis of the unchanging patrimony.

The objects and the location of the house must define it as enduring through time, since its physical shell—the structure of thatch and bamboo—has to be rebuilt every ten years or so. Because Kodi genealogies are relatively shallow, they do not reflect ideas of the continuity of the house, which always extend further back than four or five generations. In a sense, the objects that the ancestors left behind take their place, "standing in" for the passage of time.

Alliance in Kodi is particularly unstable and shifting, since it serves not as an overarching political structure for the whole society but as a status mover and converter. The directionality of exchanges is not fixed, nor are the social classes in which these negotiations are carried out. Instead, everything is open to contestation and reformulation, and the possibility of status reversals and shifts always looms close at hand. Marriage nego-

tiations are themselves "tournaments of value" (Appadurai 1986), where new power relations are decided and social groups are reordered.

For these reasons, the sense of time that is expressed when people locate themselves in the various categories of descent and alliance is quite different from that experienced by many other Eastern Indonesian societies. In contrast to the extended time frame of Rotinese genealogies (Fox 1980c, 99), Kodi li marapu are elliptical litanies of ancestors, objects, and a few place names. And in contrast to the enduring directionality of alliance obligations in Rindi, East Sumba (G. Forth 1981), Kodi marriage ties turn this way and that, ruled more by the shifting balance of rivalrous exchange than by an idea of enduring cycles. The theme of a "return to the source," so central to the ceremonial systems of the Mambai of Timor and the Tana Ai' of Flores (Traube 1986; Lewis 1988), surfaces in Kodi only as the "conservative strategy" of certain wealthy families; it is not given the encompassing religious and metaphysical significance that it has on neighboring islands.

In Roti, Fox (1980c, 100) has argued, the immense depth of genealogies is related to the "social use of history as a means of differentiation," in which the numerous petty rulers who were recognized by Dutch colonial authorities are legitimated and their titles used to maintain claims to power. In Kodi, a much shorter colonial history and a less hierarchical oral tradition has resulted in a time line that lacks a "master narrative" in the form of a royal genealogy that orders the past for all the inhabitants. The result is not a timeless past, but a past that simply is not chronicled in the terms of genealogical charters. It is chronicled, instead, in more loosely connected narratives, dispersed through a number of houses and storytellers, and the objects that they use to remember and commemorate historical events. The main ordering device used is the calendar, which divides the year into ritually marked intervals and synchronizes social activities. The priest of the calendar sits at the center of a complex island-wide ceremonial system, in which the ancestors of the Kodi "lords of the year" (mori ndoyo ) established the temporal divisions recognized throughout Sumba (see chapter 12).

Ricoeur (1988, 105) has argued that "The time of the calendar is the first bridge constructed by historical practice between lived time and universal time. It is a creation that does not stem exclusively from either of these perspectives on time. Even though it may participate in one or the other of them, its institution constitutes the invention of a third form of time." In the following chapters, we explore Kodi perspectives on the construction of the past, beginning with imported elements brought by

the ancestors when they migrated from the west. Time, as part of lived experience and an attribute of the natural world, is said to have a local origin, while the past is made up of many parts, some expressed in narrative, others in objects, and still others in ritual action. While the past is diverse and often contested, it can stand apart from daily life and become an object of reflection. As an abstraction, separated from everyday entanglements, it permits the totalization of hindsight, the retrospective vision of the whole. This vision, then, is called back into question by its involvement in continuing chains of exchange transactions and the encounter with the new, progressive, linear notion of "history."