Exchange, Friendship and Regional Relations in the Upper Xingu

The aim of this work is to describe and analyze the different types of relationships established between the peoples of the Upper Xingu through their regional rituals. Starting from a description of the uluki exchange ceremony, we will discuss how this ritual, by mobilizing Xinguano ideas about friendship, produces contexts of interaction in which characteristics of intra-village sociality are extended to regional sociality. At the same time that the uluki defines, together with other regional rituals, a certain “Xinguano interiority” (the world of multi-community rituals), it is also one of its main forms of opening, having the potential to attract to the Xinguano world not only singular persons, but entire groups.

As well as common participation in regional rituals, another characteristic of humanity among Upper Xinguanos is exchange, whether through marriage, knowledge transmission, rituals, or, in the case we will analyse here, objects. Among the Wauja, for instance, Piedade (2004: 22) translates putakanau (a term equivalent to "people") as "people who lend". Dole (1958: 125) has described the Kuikuro as being "constantly preoccupied with the exchange of goods and services and with equalizing accounts", a fact also emphasised by Fausto (2016).
Among the Kalapalo, this characteristic is narrated in a version of the creation myth registered by Basso (1987: 29-81). Here, the mother's sister of the twins Sun (Taũgi) and Moon (Aulukumã) was magically impregnated after walking over arrows left on the ground by Taũgi. She gave birth to the ancestors of the "wild Indians" (ngikogo, non-xinguanos), the ancestors of Xinguano peoples and the ancestors of the whites (kagaiha). Taũgi gave each of them distinct objects: clubs for the first, bows, gourds, belts, and necklaces of inhu (Megalobulimus sp.) shells to the second, and clothing, knives, axes, scissors, etc, to the ancestors of the whites. Taũgi also told the ancestors of the Xinguanos that "When you need things in the future / You can visit other villages / You will exchange your objects with each other / When you need to / You will give your things" 6 .
Exchange being a central element in the constitution of Xinguano "people", our aim is to understand how the uluki contributes to the constitution of this "regional system" -that is, to the establishment of a multi-ethnic constellation which identifies as a common humanity in opposition to its neighbours, and which extends out toward the scale of inter-community relations elements that, elsewhere in Amazonia, seem to be the preserve of the sociality of local groups (Hugh-Jones 2013; Andrello, Guerreiro, and Hugh-Jones 2015). If, as Menezes Bastos (1978) proposed, the ritual system is the lingua franca of the Upper Xingu, then it may be possible to think of the uluki as one of the main sources of creativity/innovation (in Leach's [2004] sense) -and of overture -in the regional system. The uluki is, after all, one of the few rituals that allows for the inclusion and active participation of non-Xinguanos, whether these be other peoples who live on the frontiers of the Upper Xingu social world (such as the Kĩsêdjê and Yudjá, for example), or even the whites.
Our discussions here are based on field observations among the Kalapalo, who, alongside the Kuikuro, Matipu, Nahukwa and Angaguhütü, are one of the five Carib-speaking people of the Upper Xingu. The Kalapalo number a little over 900 people, distributed among 11 villages, most of them on the banks of the Culuene River (main tributary of the Xingu). The ethnography that follows focuses on events that occurred at different times 7 among the Kalapalo of Aiha, the largest of their villages, situated on the right bank of the Culuene River. Let us start, then, with the ritual.

The uluki
The uluki can be an intra-village ritual, in which men of different houses exchange with each other and women of different houses exchange with each other, or an inter-village ritual, involving a visiting people in which men and women exchange together. 8 In the latter case, the ritual of exchange occurs alongside rituals which do not occur in the intra-village uluki. As we mentioned above, it is likewise not uncommon 6 Loose translation provided by Jeika Kalapalo, based on the original audio recording, referring to lines 1985 to 1990 of the narrative published by Basso (1987: 78).

7
This article results from long-term research among the Kalapalo of Aiha, begun in 2006, and distributed over various stays in the village. The only inter-ethnic uluki we attended during this period, and which is the basis for the description that follows, occurred in February 2015.
8 Basso (1973: 136) claims that, during her research, only men participated in inter-village events and that female uluki within villages was quite rare. This is very different from the situation we have found at Aiha since our very first visits. At present, women follow men and actively take part of events that engage more than one people. During the ritual, they activate their own circuits of exchange. Furthermore, exclusively female inter-village uluki are frequently organized, throughout the year, whenever a woman is interested in holding one (which, it must be said, is more often than men). The decision to hold uluki may be spurred by a specific event, such as families who were in town returning to the village, or simply because they "like holding uluki" and use it as means to pass the time. This fondness for the uluki is shared by girls, who often hold "play" uluki. Unlike Basso's (1973: 137), we did not witness any uluki held exclusively between boys. On the relationship of women to uluki, see Novo (2018) and Franchetto (1996). for dignitaries of Upper Xinguano people to visit villages that are not integrally part of this system, such as those of the Kĩsêdjê (Suyá) or Kawaiwete (Kayabi) 9 , in order to hold uluki. The main motivation for these events is the chance to acquire specific objects offered by these peoples (particularly feathers and jaguar claws or coats from the Kĩsêdjê, and tucum and annatto necklaces from the Kawaiwete), or else the opportunity to visit their villages, see their rituals, and, perhaps, to dance with their women (in these cases the rituals which are performed depend on knowledge of their respective repertoires). Simplified versions of the uluki are also carried out with whites who visit the village, and they are an important aspect of relations with the non-Indigenous people who work in the Indigenous Territory. Despite these many possibilities, we will here focus on inter-village rituals that occur within the "Xinguano system", so as to engage in comparison with other regional rituals in which different relational forms are stressed.
Unlike other multi-community rituals, uluki has no marked origin, and lacks a designated origin story (akinha) that registers its beginnings. 10 According to the main storytellers (akinha oto) in Aiha "It was Taũgi [the Sun, demiurge who created humanity] who made uluki. It was not itseke [...] Taũgi made uluki so that we might attend uluki". Thus, unlike most of other rituals, uluki was not learned from the itseke, the hyper-beings, nor from other Xinguano people.
Another characteristic that distinguishes the uluki from other Upper Xinguano rituals is the fact that it lacks an owner 11 . The rituals carried out during inter-village uluki (which will be described below) have specific owners, but uluki itself does not. They do have ugihongo (literally "the one who goes on the stool"), who "conduct their people to rituals in other villages" (Guerreiro 2015: 160), a function present in all multi-community rituals.
The implications of lacking an owner will be discussed below.
Finally, like other exchange rituals, such as the dabucuris of the Upper Rio Negro, or the Melanesian kula, the uluki does not only provide a context for exchange, but is, itself, a gift: every uluki demands a return, and the visiting village will later receive the host village (thus inverting their respective positions), even if this takes a few years. The demand for a return generates a virtually endless cycle of relations between villages that alternate as guest and host.
The word uluki has an Arawak origin, huluki, and it refers to the 'swallow' (kugupiso in Kalapalo) 12 . This name registers the fact that inter-village uluki always occur during the rainy season, which is when flights of swallows can be seen above villages. The Kalapalo also believe that the uluki is capable of halting rains, an idea present among other peoples of the region (such as the Yawalapití, cf. Viveiros de Castro [1977: 108], and the Kĩsêdjê, cf. Coelho de Souza, pers. com.).
In every other multi-community ritual, the host group is responsible for sending messengers to formally invite the chiefs of other groups, a key element in Upper Xinguano diplomacy. For the uluki, in contrast, visitors are not invited; instead, they initiate the event. An inter-village uluki is always an initiative of the visiting village, which decides to carry out the event some time before it occurs, allowing for their own wrestlers to get ready. These preparations involve training sessions for the wrestlers, and scarifications followed by the application of remedies, the ingestion of emetics (made from specific plants and roots), as well as rehearsals 9 By this we mean that although these people maintain relations with Upper Xinguanos, to different degrees, they do not fully take part in their ritual system and are not considered kuge, "Xinguano persons". This distinction is important, because while the Xinguano relational network is heterogenous and dynamic, ritual life is responsible for continuously buttressing the difference between Xinguano and non-Xinguano people (kuge/ngikogo in the Kalapalo language). 10 Menezes Bastos (2001: 344) notes something similar for the Jawari: "The ritual (toryp) itself does not have a mythic origin ('from the beginning') plainly given". However, in this case, the 'absence' of a mythic origin is, for the author, linked to a "historical" memory of its acquisition, since the ritual was introduced into the Upper Xingu through the Trumai, and spread throughout the region by the Kamayurá. Uluki, in contrast, seems to have neither a mythical nor a properly "historical" origin.

11
Barcelos Neto (2008: 180) claims that, for the Wauja, the exchange ritual is owned by someone who became ill due to the Huluki spirits. Vanzolini (2015: 114) also describes a similar situation among the Aweti (who call the ritual joro'jyt). However, we have never heard anything similar among the Kalapalo, who, in contrast, explicitly told us that the uluki did not have any owner.
12 Despite the existence of this term in Carib, people rarely use it, always referring to uluki to designate more or less ritualized exchanges. for the ritual that will be performed. It is only a few days before the scheduled date that the hosts are informed, via radio, that the ritual will take place. This catches their own wrestlers unprepared, weak and "smelling of sex" 13 , which makes them into potentially weaker adversaries.
As with other rituals involving more than one people, and, consequently, bringing together residents of different villages, the uluki enlivens those who participate in it. The event we accompanied originated with the Wauja, who wanted to hold a ritual in Aiha. The potential visitors informed their intention about one week before the day they were due to arrive in the village, but a few days later they cancelled the visit because one of their chiefs had to be hospitalized. Kalapalo frustration was widespread, considering that even people from other villages (who identified as being Kalapalo) had already arrived at Aiha only to participate in the ritual. It was precisely because of frustrated expectations that the Kalapalo decided to go to the main Kuikuro village of Ipatse. Having made the decision, and informing the Kuikuro only three days before the date they were due, they began to worry over the logistics of transporting everyone (and all their goods) to Ipatse.
Alongside logistical matters, the preparation also involved "rehearsing" the rituals to be performed. These rehearsals are part of the preparation of all rituals, particularly those involving different peoples, and they can be repeated over a long period of time -up to a year in the case of the Quarup, for example. But, unlike the chiefly rituals that involve longer preparations and frequent rehearsals, the uluki are little-rehearsed. The main reason for this is the short time between when the event is decided upon and when it is set to occur, but also the fact that the uluki is marked by improvisation and little formality, characteristics that mark relations of friendship in the Upper Xingu. Following Houseman and Severi (1998), we understand rituals as specific relational and interactive contexts, marked by a sequence of acts that condense "opposing modes of relationship" (Houseman and Severi 1998: xii).

From hierarchy to symmetry: hosting chiefs and exchanging with friends
We will see how the sequence of uluki sometimes privileges the distinction between groups (by stressing, on the one hand, the hierarchy of chiefs and commoners, and, on the other, the symmetry between chiefs themselves), and sometimes their approximation, thus creating the conditions for the exchange of objects to occur.
In the morning, a small flotilla of five boats and dozens of people left the port of Aiha headed for the main port of the village of Ipatse. In a clearing, close to the entrance to the village, the recently-arrived visitors began to paint and adorn themselves, shouting küu, and thereby informing the hosts of their arrival and indicating joy and casualness. From where we were, we could also hear the same shouts coming from the host village.
When preparations were through, the visitors went in single file along the main path that accesses the village patio, led by three chiefs -the ugihongo. During the walk, these chiefs spoke 14 , referring to the hosting chiefs, saying that they brought with them "a group of friends made up of your brothers" (Guerreiro 2015: 352) 15 .
Upon nearing the centre of the village, the ugihongo were taken by the hands by Kuikuro men to take a seat in stools placed before the men's house (a small house erected in the village patio, situated slightly to the west, and forbidden to women). The remaining Kalapalo positioned themselves behind their chief-dignitaries. The positions of the ugihongo are defined hierarchically by age: the eldest, considered the principal ugihongo, sits in the middle, followed by the one who sits to his left, and, finally, the one who sits to his right.
13 Applying remedies and ingesting emetics involve dietary and sexual restrictions, since the taste or odour of come foods, as well as the odour of sex and menstrual blood, disturb the spirit owners of the plants to be used (which confer "strength" on the wrestlers). Disturbed in this way, the spirits can cause harm on those who do not observe the restrictions, which can culminate in death. On Kalapalo wrestling, see Costa (2013).
14 In the uluki we witnessed, none of the ugihongo knew this speech, which was therefore proffered by Tühoni, the second chief of Aiha.
15 For lack of space, we will not reproduce here the complete speeches, but a version of these speech (which we take as our reference) was published by Guerreiro (2015: 352-353).
Then, the main Kuikuro chief, considered to be the "owner of the village" (ete oto), left his home donning the chiefly insignia -jaguar skin-hat, snail shell-necklace, jaguar claw-necklace, holding a black bow and its arrow -and began to speak, standing before his house, looking towards the village centre (we were unable to register the speech at that time, and we do not have a Kalapalo version). When he finished, the Kuikuro chief walked to the village centre and, turning to the house of another chief, spoke again, beckoning him to the centre. He repeated the procedure again, facing the house of a young chief, who, along with two others, made up the trio of Kuikuro ugihongo-chiefs, which will henceforth be called imünhilo, "opponent", which is how the Kalapalo refer to the ugihongo of the host village. The two imünhilo called up by the Kuikuro chief made their way to the village centre and spoke to the one who called them. After this, each of the imünhilo in turn, beginning with the main one, addressed the Kalapalo ugihongo while bending towards the ground and spoke, while the ugihongo in front of him proferred his own speech. 16 In the Kalapalo version of the speech we registered 17 , which they only perform when they are the hosts of the uluki, the chiefs do not refer to one another as 'friends' or 'brothers', as they do when they are visitors, but rather by the term of address tongisa, a synonym of anetü ('chief ') but conveying a more respectful tone (Guerreiro 2015). This can be explained by the fact that Upper Xinguano people worry considerably about "hosting well" those in their homes (or, in this case, their village), which means that hosts should not only offer porridge and other foodstuffs, but also "speak well" with the guests, displaying "[the] noble demeanor, which in the Xingu centers on the conceptual cluster of shame, respect and humility" (Ball 2011: 96). Thus, for example, every time a guest is welcomed in the home (excepting very close kinspeople), they should be offered a stool or some other object that can function as a bench, which will be placed close to the door of the house, where they should sit, along with the host(s), to talk. Something similar takes place in the uluki: visiting chiefs are placed on stools and the hosts "talk" with them through welcoming speeches, treating them with respect through use of the vocative tongisa.
When the speeches were over, the women of the imünhilo's houses (their wives, daughters, daughters-inlaw, and granddaughters, depending on each case) brought out large pans with porridge, cooked fish, as well as roast fish and manioc bread, which were given to the ugihongo by the imünhilo. Both the performances of the welcoming speeches and the quantity and quality of the food offered are later commented by the visitors among themselves (never directly to the hosts), as an important part of the hospitality of the hosts and, consequently, of the ritual itself.
After they have received the food, the ugihongo handed the gifts they had brought over to the imünhilo, who were obliged to later reciprocate them. Other inter-village rituals contain the expectation of gifts offered from the chiefs of the guests to the owner of the ritual, but always at its conclusion. In the uluki, this offering occurs at the outset, generating an initial disequilibrium which impels partners to exchange. Gifts offered should be reciprocated by something defined by the one who offers the initial gift. This payment (hipügü) is decided on the last day of the ritual, when the farewell speeches take place. The gifts offered by the ugihongo are always items of high value, ideally items that are representative of the regional productive specializations, such as Kalapalo snail shell-necklaces, Wauja ceramics, Kamayurá black bows, Matipu striped snail shell-necklaces, and Aweti salt. If these objects are lacking, however, other valuables can be offered, such as aluminium pots, glass beads, etc. In the ritual we attended, the gifts offered were inhu snail-necklaces and belts, which were gathered and taken home by the imünhilo. Then the wrestling matches (ikindene) began. The main regional Xinguano ritual involving wrestling is the Quarup, but, yet again, there are important differences in wrestling matches in the two rituals. In the Quarup, the owners of the ritual select and present the main wrestlers before the matches involving each of the invited peoples, whereas in the uluki there is no formal presentation. In the latter, the guest and host wrestlers are simply called up by their respective ugihongo, one by one, to face each other. After the main wrestlers have faced each other, all men were allowed to wrestle.
At this point, various wrestling matches occurred simultaneously, as also happens during the Quarup.
There are other differences between the wrestling matches of the uluki and those of the Quarup. In the former they do not last as long as they do in the latter, and, more importantly, they occur within a generalised climate of amity. During the wrestling matches of the Quarup, it is not uncommon for small fights to break out, involving at times the wrestlers' families (particularly their mothers), who accuse opponents of cheating or of being sorcerers. During the uluki, the atmosphere during the matches is quite different. Even though some wrestlers occasionally get hurt, we have never heard of any sort of mutual accusations between guests and hosts. As if to mark the casualness of the event, wrestlers embrace each other at the end of the bout, smiling, before going their own ways to take up new matches with other opponents. Writing of wrestling during the Quarup, Costa (2013) stresses that they almost invariably end with an embrace immediately followed by a push away, pointing to the ambivalence of the relationship between the wrestlers. In the uluki opponents only embrace each other, and, at the end, Kuikuro and Kalapalo men and women greeted each other, one-by-one and protractedly, making a point of demonstrating all the joy and friendliness about them.
After everyone had greeted everyone else, the guests returned to their initial formation, with the ugihongo sitting down and the remaining people positioned behind them. Three further stools were placed before the ugihongo, where the imünhilo sat, each one facing his "opponent". The ugihongo initiate the exchanges, placing on the ground, in front of themselves, the objects that are made available, each one saying what they expect to receive in return (if they do not want anything specific, they may say only uhitseke gele, which means something like "whatever"). These first objects offered by the ugihongo are accepted by the imünhilo, who take them from the ground and thus initiate the exchange cycles. To an external observer, the whole process may look like a great bazaar. Men and women of the visiting group give their objects over to their chiefs, who are sitting before the chiefs of the hosts, indicating whether they expect something specific in exchange. It is the chief who deposits the object on the ground, before his "opponent" (whoever is sitting in front of him), who picks it up and presents it to his group, indicating what the payment should be if this has been specified by the giver. Within this configuration, up to three consecutive exchanges may take place, each one led by one of the chiefs, which compounds the external observers' impression of "confusion".
However, unlike a market or bazaar, the exchanges carried out in the uluki are not limited to the circulation of merchandise. Although they may involve industrialised goods, and even money, uluki exchanges are best understood in light of what anthropology has traditionally called "gift exchange" (see, for example, Gregory [1982]). Exchange focuses not on a (quantitative) relation between exchanged objects, but on the (qualitative) relations established between people. In this case, rather than persons, it is groups which are the units of exchange, personified in the figure of the chiefs who exchange between each other, and who elicit of each other the positions of giver and receiver of objects at different moments. This occurs because, after some time has passed where the visitors offer up their objects, the positions are inverted and the hosts come to offer, in the same format, the objects they want to exchange. In these exchange sessions, groups are constantly evaluating each other and exchanging, not only objects, but also images of themselves (and the uluki itself projects the virtual continuity of exchanges into the future).
Despite the semblance of confusion, with many things happening at the same time, people are always very much attentive to actual exchanges, commenting on the value and quality of the objects offered, cheering whenever an object of great value is exchanged. It is more typical for the more common, and hence "cheaper" (taloki, 'worthless'), objects to be paid immediately. Nonetheless, arrangements can be made between the people involved in exchange, particularly where valuable goods are concerned, payment of which would require some later effort by the receiver. Despite the efforts of the ugihongo, not all exchanges are consummated: some objects which do not generate interest, or which no one is able to pay, are not exchanged, and return to their original owners.
The objects that circulate, offered by men as well as by women, include, among other things: snail shellnecklaces, aluminium pans of various sizes, glass beads, feathers, feather bracelets, dresses, soap, rifle shells, arrows, pequi dough, roast fish, fishhooks, coffee, sugar, wool, motorcycle parts, watches, and even money.
Becker [Basso] (1969: 252) describes that, in the late 1960s, "certain 'valuable' items, namely shell ornaments, hardwood bows, and very large ceramics, rarely appear during an uluki ceremony, and it is usually 'lesser' items which are exchanged here". In the uluki we attended, in contrast, a reasonable quantity of items considered valuable were offered in exchange, particularly snail shell-necklaces and belts, large aluminium pans, glass bead-belts and necklaces with dozens of coils. However, most of the hundreds of transactions that effectively occurred involved items of small value. The most valuable objects exchanged on that day were offered by men, since women store their most valuable property to exchange the following day, in an all-female uluki. As Quain (1966) pointed out, the uluki functions as a sort of game, in which people (and groups) look to present themselves as generous, offering valuable objects, but also retaining part of their possessions as payment for some desired good that may be offered later. It is not exactly what Weiner (1992) has called "keeping while giving", since these objects are not retained because of their inalienable value, as she defines it, but as means for extracting objects of greater value in future exchanges. This applies mostly to women, who hold their own uluki later.
At any rate, the matter we want to stress is that exchanges are not that personalized (individualized) nor is it always possible to even know who offered or paid for the exchanged objects. Exchanges occur between groups personalized in the figure of their ugihongo-chiefs, and not between individuals. And since groups that are on friendly terms exchange between themselves, a relation of equilibrium should be maintained between the exchanging parties. For this reason, it is only when the receiver cannot pay what is asked of an object that an effort is made to find out who its owner actually is, so that a deferred payment can be arranged; only debt (and hence hierarchy, even if temporary) makes singular persons appear in these transactions. Otherwise, exchanges are highly impersonal.

Celebrating friendship through ritual
When exchanges were completed (that is, when people no longer had objects to offer), the visiting group was finally lodged in one of the houses of the village. Unlike most inter-village rituals, where visitors camp in uncultivated areas outside of the circle of houses, in the uluki all of the visiting group is lodged in the same house, within the village circle. The visitors should share the food they received with the owners of the house in which they are staying. The rituals begin that very evening, and will last the whole night and the following two days. They are promoted by the guests, using instruments and ornaments provided by the hosts. The first ritual to be performed is the unduhe (also known as Tawarawanã), which is accompanied by the kagutu flutes, which are forbidden to women. In the following day the takuaga is executed, which is also a ritual involving aerophones. As we will see later, both are "spirit rituals" that have specific owners in each village, and which, despite this common characteristic, are nonetheless quite different from each other.
According to the Kalapalo, there is no special reason why these two rituals in specific should be performed during the uluki. In effect, there does not seem to be any characteristic shared by these two rituals (unduhe and takwaga) that sets them apart other spirit rituals that occur in the region, and which could have linked them to the uluki. However, it may be possible to consider them, alongside the uluki exchanges, as examples of the process of constructing the "Upper Xinguano complex", or of "Xinguanization", as described by Franchetto (2011) andFausto, Franchetto andHeckenberger (2008). According to these authors, this process includes a distinction between "mythical time" and "historical time": "While the 'original' peoples were directly created by the mythical heroes, others made themselves 'Xinguano' by adopting Upper Xinguano ceremonialism and pacific, ethical and aesthetical values, [and] a specific diet" (Franchetto 2011: 16). If, as we have seen, Xinguano peoples define themselves by exchange, and if the Upper Xinguano complex is composed of mythical and historical elements, the contemporary configuration of the uluki emerges as exemplary: a ritual the purpose of which is to exchange and which is accompanied by two rituals, one "mythical", learned from the fish (the unduhe), and other "historical", the takuaga, learned from the Bakairi.
Regardless of the origin of the rituals, the main issue for Xinguano people seems to be the fact that they are carried out so that people can "be happy". Thus, it is common that during sporting events or meetings that involve people from different villages (and/or white people), they perform "ritual sketches" involving dancing and/or singing, and very often the rituals "presented" are, precisely, the takuaga and the unduhe 18 . Let us now turn to how these take place during the uluki.
The first ritual sequence begins with the owner of the trio of kagutu flutes of the host village, handing them over to the ugihongo, who then hand them over to the visiting men who will be responsible for playing them throughout the night and the following day. There are few men in Aiha who know how to play these flutes. In the first night, the flutes are played in the village patio, once the women have gone home, where they keep the doors shut. Throughout the remaining days, the music is played exclusively in the men's house.
In the context of the uluki, the kagutu music is played exclusively as an accompaniment (akongo, "which goes together") to the unduhe.
According to the Kalapalo, there are two versions of the unduhe: the snake unduhe, which is said to belong to the Arawak speaking peoples, and the fish unduhe (which was learnt from the fish), which is the version that belongs to the Kalapalo. 19 This is one of the most frequently performed rituals in Upper Xinguano villages, an obligatory part of the ritual sketches that occur during events in the Xingu Indigenous Territory (particularly in the area of the Leonardo Vilas Boas Local Technical Coordination), as well as those in the city. This more "quotidian" version is a simplification of the complete version of the ritual, which is only sponsored once by its owner, when he decides to (quite literally) burn his unduhugu (which are the ritual objects kept by the owners of the spirit rituals -straw skirts in the case of the unduhe), and thereby ceases to be the owner of that ritual 20 .
The "complete" version of the ritual occurs in the absence of guests, and dozens of songs are sung, referring not only to fish, but also to birds and other animals, just as they were previously sung/performed by fish.
In the simplified version of the ritual carried out during the uluki, the repertoire is shorter (and varies between groups) and requires much less work by the owner of the ritual, when compared to the complete version, since the owner needs a single fishing session to pay all of the dancers. To perform it, a singer stands in the centre, facing the men's house, holding a bow in one hand and playing a maraca with the other, accompanied by another man who, sitting before him, thuds a bamboo on the ground (a utinha, which can also be made of PVC piping), marking the rhythm of the dance. The dancers, adorned with the leaves of a special plant and 18 The kagutu flutes are never played in these events, firstly because they are restricted to women, which would make the event unviable, or at least difficult, in many cases, but also because their execution demands a much more restricted knowledge that that which is necessary for executing the takuaga and unduhe.
19 The Kalapalo claim that the unduhe of the Kuikuro is also different from that which they own, possessing a smaller number of songs that are executed. On Kuikuro rituals, see Mehinaku (2010). Even if the Kalapalo version is distinct from that possessed by the Arawak-speaking people of the region, all of songs of the fish unduhe are in Arawak languages, as indeed are most of the songs of the other rituals. For a musical analysis of the Yawalapiti version of this ritual, see Almeida (2012). 20 The act of burning the unduhugu at the end of the ritual makes the owners of these rituals cease to be owners. The only spirit rituals that cannot be relinquished by burning their respective unduhugu are the kagutu (flute ritual forbidden to women) and the tolo, a female ritual where women sing in words with the kagutu flutes perform in melody. On the Kuikuro tolo, see Franchetto and Montagnani (2011). straw skirts, situate themselves around the singer. The "chief " of the dancers (sanetügü) or "shouter" (ihetogogu) positions himself before the men's house, facing the singer. It is this "shouter" who conducts the direction of the dance, indicating when the dancers are to approach the singer, when they should move back, and when the song should end. At the other extreme of the circle of men there is a second "shouter", who only follows the first, responding to his shouts. Gradually, the women of the host village join the dance, forming a circle outside of the circle of men, and occasionally accompanying them, holding on to the tips of the buriti straw which makes the skirts.
This ritual continues throughout the day, with intervals that mark musical transitions. In the afternoon, the host men always dance, accompanied by the visiting women. But the groups do not mix: the group of visiting men dance with the host women and, later, the host men dance with the visiting women, taking turns. The unduhe ends late afternoon, a little before the final kagutu (both occurring concomitantly throughout the day).
We do not think it is accidental that the first ritual which occurs during the uluki has a "mythical" origin, having been learnt from the fish. According to Guerreiro (2015: 228), the fish present a sociality "closer to the Kalapalo ideal of intra-village sociality", insofar as they do not hunt and are not fighters, behaviour which thus distinguishes them from predatory animals. It is hence a sort of "prototypical ritual"; indeed, unduhe designates not only this ritual, but is also used as a generic way to refer to any "ritual, spirit, mask" (Franchetto and Espírito Santo 2014). If the uluki is intended to mark the close relations between the groups involved, it hence makes sense to perform this ritual.
Through the three days of uluki, the visitors receive fish from the imünhilo and the owners of the rituals.
Unlike other multi-community rituals, however, the visitors receive raw fish, as well as cooked/roasted fish prepared by the women who accompany the ugihongo. Furthermore, the fish is not always accompanied by manioc bread, so that these very same women ask for flour from the village houses. Both the offering of raw fish and the solicitation of flour by the women contribute to bringing the relations between the groups closer to those established in the internal ambience of the village -that is, those established between kinspeople.
At the end of the unduhe, the ugihongo call the imünhilo to the front of the house where they are staying and tell that they will leave the following day. The imünhilo respond that they should stay and dance more the following day. 21 The ugihongo agree and the owners of the sets of takuaga flutes 22 of the host village hand their instruments over to the ugihongo.
The takuaga is performed on the third and final day of the gathering, also by the visiting men, playing the instruments provided by the hosts and accompanied by the host women. As in the previous day, the Kuikuro men also danced, accompanied by the Kalapalo women, but for a shorter amount of time and making use of only one set of flutes (the Kalapalo had been given two sets).
The takuaga is a ritual that the Upper Xinguano people learned from the Bakairi. A set of songs of Yudjá origin have been added to the original Bakairi songs, along with a large set of songs that are invented by players, who have even added new flutes to the original set of five. As with unduhe, this ritual frequently occurs in Upper Xinguano villages "to make people happy", and it also has a simplified version that is carried out within the village, in the absence of guests. The complete version of the ritual, unlike the complete version of the unduhe, requires the presence of guests from other villages, whose wrestlers face off against the host-wrestlers. At the end of the takuaga's complete version, as with the unduhe, the instruments are burned and the ritual ceases to be owned by its owner.
As a "fish ritual", the unduhe underscores a possible identity between the groups that exchange with each other (the simulacrum that foreigners can see themselves as kinspeople to those who live in a same local group). The takagua, however, seems to always remind us of the constant opening up of the Upper Xinguano world to its (relative) exterior, and also of how the uluki has been, and continues to be, a privileged moment in this opening. It is perhaps for this reason, along with the fact that the songs performed in the takuaga are widely known and new ones can even been composed -which means that dancers/musicians do not depend on a specialist 23 -that it is the most regularly performed ritual in Aiha, and perhaps also in the other villages of the region. It is also the ritual chosen to end the uluki, celebrating the "happiness" of all those involved.
In the uluki we attended, which was marred by strong and incessant rains, dances ended far too early, close to midday, but ideally the groups should play the flutes until late into the afternoon. When the ritual ended, the ugihongo called the imünhilo to the centre of the village and made their farewell speeches (which we unfortunately did not register). It is then that the ugihongo tell the imünhilo what they expect in return for the presents offered at the start of the uluki. If they have what is solicited, the imünhilo immediately hand them over to the ugihongo, but this is not required, and the return gift can be given later. The visitors leave at dawn the following day, without ceremony, only with individual goodbyes between kinspeople and friends.

The uluki and other rituals
As we have said, the uluki is part of a complex of multi-community rituals, and comparison with other rituals foregrounds some of its specificities. We will not go into a detailed description of these other rituals, some aspects of which have already been mentioned. However, by paying attention to the relational forms that characterize these contexts, we will see that each produces alternative and complementary socialities on an interethnic scale.
The two main inter-ethnic rituals in the region honour the illustrious dead: the Quarup and the Jawai (called, respectively, egitsü and hagaka in Kalapalo). The first is performed in memory of chiefs or their close kin, and reaches its climax in the production, exhibition and discard of a mortuary effigy that, although decorated, is not anthropomorphic, as well as in wrestling matches between hosts and guests. This is the largest Upper Xinguano inter-ethnic ritual, to which, ideally, all peoples of the region are invited. The first Quarup was held by Sun (one of the twin creators of humanity) in the village of the jaguars and other land animals, in honour of his mother. It marks the beginning of the time of humans, and participation in the ritual defines the limits of Upper Xinguano humanity. It is a ritual in which the difference between chiefs and non-chiefs is made visible in all stages of the ritual cycle, highlighting the asymmetries the characterize, on the one hand, the sociality of the local group, and, on the other, some of the asymmetrical relations between the local groups of a single people. Furthermore, by gathering Upper Xinguano humanity at a time which coincides with the definitive departing of the deceased to the village of the dead, the Quarup is organized along an opposition between the living and the dead, which allows Upper Xinguanos to mutually present themselves as consanguineal kinspeople who, together, mourn the loss of the deceased. This does not, however, mean that other relational forms are absent from the Quarup. Ritual enmity is repeatedly enacted in relations between chiefs, and, above all, in wrestling matches; symmetrical relations between the groups are also marked in the matches and in ritual exchanges; and potential affinity among them is made evident not only in ritual enmity, but also through offering, before the end of the ritual, pequi nuts to the guests by young women who have just left seclusion: in myth, the smell of pequi is produced by the smell of the vagina of women, and the circulation of pequi nuts (which are offered by the chiefs, in their villages, mixed into a broth) is a metaphorical circulation of women between the groups 24 .
The Jawari is also performed in honour of notable dead people, but it is an inter-ethnic ritual which usually opposes only two groups 25 in a javelin-throwing tournament. The ritual has a celestial origin, in which crosscousins from distinct people publicly offend one another, addressing an anthropomorphic, and purposely grotesque, effigy, which they strike with their javelins or wooden staffs. The javelin-throwing tournaments are explicitly compared to war, and the Kalapalo say that they fear them more than the wrestling matches, since it is not uncommon for fishhooks or pieces of lead to be placed on javelin tips so as to seriously injure an opponent. The ritual is performed in honour of a chief, a great archer, or a great singer, on the condition that this chief has had an effigy in a Quarup. Organized around the conflict of cross-cousins, the ritual foregrounds the symmetrical opposition between groups who compete, between people who offend each other and duel, and the possible matrimonial ties that were not realized between groups (brothers-in-law, even if they are also cross-cousins, never face each other in the javelin-throwing tournament). However, as with the Quarup, this does not mean that other relational forms are not also present: the asymmetry between chiefs and non-chiefs is also marked in ritual hospitality; foreign chiefs also see themselves as kinspeople when, at the end of the ritual, the effigy of the dead is burned and they meet with the host-chiefs to mourn the deceased, while the exchange of food marks the relations between cross-cousins (men who were struck should take roast fish to the cross-cousin who hit him with his javelin).
Compared to both of these rituals, the uluki stands out in many respects. As we have seen, the first of these is that it does not have an owner, which makes it differ from other regional rituals and from spirit rituals. In chiefly rituals, the owners are close kinspeople of the honoured chief. In spirit rituals, the owners are people who were ill and who, during the healing process, had a shaman identify an itseke responsible for the theft of the patient's soul. From this moment on, the ill person becomes "owner" of the spirit, and it is his duty to sponsor its ritual, so as to tame it, to domesticate it (Barcelos Neto 2008;Franco Neto 2010). The owner of a ritual should provide food for the participants (dancers or singers, for example) and the guests, when there are guests.
In the Quarup and the Jawari, the presence of the owners of rituals marks a hierarchical distinction between chiefs and commoners, as noted by Guerreiro (2015) and Fausto, Franchetto and Heckenberger (2008). In the uluki, even in the absence of an owner, the figure of the ugihongo makes visible, even if only momentarily, the people who are personified in these chiefs. However, this initial hierarchical differentiation quickly dissipates as the guests are all lodged in one of the village's house and the rituals begin. Any asymmetry that persists between the owners of the rituals that occur during the uluki, such as the unduhe and the takuaga, is necessarily linked to chieftaincy.
Another important difference concerns hospitality. This theme, which is practically absent from ethnographies of the Upper Xingu, is fundamental to supralocal relations and consists in elaborate and diversified practices according to the contexts and the agents in interaction, which can be described according to certain spacetime qualities (Munn 1986). In rituals, hospitality is not merely a "condition" for the rituals to occur; it is through hospitality that the contexts of ritual interaction are constructed and, consequently, ritual is directly affected by the forms that hospitality assumes. According to a spatial perspective, hospitality during the Quarup and the Jawari sets up a "good distance" between hosts and guests. Since their arrival, guests remain in camps outside the village, and their interactions with hosts is almost exclusively through the medium of chiefs. This distance is layered with an evident antagonism. Camps are seen to be dangerous, 24 For a detailed ethnography of the Quarup among the Kalapalo, see Guerreiro (2015).
25 Frequently, the host group is made up of more than one people. Nonetheless, the Jawari put only two teams into competition: hosts and their allies against guests. places that should be avoided by the hosts, and guests frequently intone vocalizations meant to cause fear in the residents of the village. The temporal sequence of these rituals also seem to mark a series of "disjunctive conjunctions": guests arrive and are treated as potential enemies; the chiefs of the guests offer gifts to the host coordinators; the ritual gradually reduces the distance between the groups, which, in turn, increases tensions; the event culminates in a sporting match, from which violence is never excluded; the host chiefs return the gifts of the guest chiefs, and the guests seek out their friends among the hosts, in order to quickly change gifts of little value before they leave.
The hospitality of the uluki takes a different form. In spatial terms, the guests arrive directly in the patio, and are then lodged in a house within the village circle. Once there, they receive visits, and share fully in the day-to-day activities of their hosts, without the need for chiefly mediation. The temporality of the stages of the ritual also differs: the visiting chiefs offer gifts to the host chiefs; men fight, in an informal and casual manner; exchange is intense, and the circulation of objects of great value is anxiously awaited; the rituals that occur during the uluki enact the potential of the exchange of wives between the groups; and when the host chiefs return the gifts, the groups do not go their own ways, but remain a further day, eating together and celebrating.
There is one last difference worth emphasizing, in what concerns the foodstuffs offered in each ritual. In the Quarup and the Jawari, only cooked food ready for consumption is offered: in the former ritual, the main food is smoked fish, a product that results mainly from men's work; in the second, it is said that the aim is to "drink the perereba" (sweet porridge made from manioc broth) made by the sisters of the hosts, emphasizing the consumption of food that is liquid, has a vegetal origin, and is feminine. In uluki, however, hosts often offer raw food (fish and manioc flour), which is to be prepared by the women of the visiting chiefs. The uluki inverts the pattern of hospitality of the other rituals, bringing the visitors into the circle of houses, allowing them to participate in quotidian activities, and supplementing the ritualized exchanged of cooked food (which is also present) with commensality, circulation of raw foodstuffs and sharing of a hearth. It is the largest ritual effort at expanding, to the inter-ethnic scale, the sociality of the local group.
It is also worth noting that each of these rituals is associated with different sociocosmological domains. In myth, the Quarup is a ritual sponsored by the Sun in the village of the jaguars and other terrestrial animals, where the fish are guests and opponents. The Jawari, in turn, is a bird ritual with a celestial origin. The uluki, as we have seen, lacks a mythical origin like the two other rituals, but combines rituals of aquatic origin (kagutu and unduhe).
The following table synthesizes some of the difference between these rituals: In sum, while in chiefly rituals the relations highlighted between groups are either those of consanguinity and hierarchy (Quarup), or potential affinity and antagonism (Jawari), in the uluki they are those of friendship and reciprocity. And the main specificity of this ritual lies precisely in the fact that its main focus is exchange, a type of relation that, in the upper Xingu, does not occur, in the same way, among close kinspeople, where we find relations more like those of sharing, nor among affines, where we find hierarchical relations in which, ideally, the flux of goods is opposed to the flux of spouses.

On Xinguano friendship
The type of exchange that occurs during the uluki -which, as we have been arguing, reproduces and reinforces aspects of friendship -is distinct from how, in day-to-day interactions, close kinspeople and affines make things circulate among themselves. Among close kin (including those who live in the same house and those who live in different houses of the same village), the ideal form of relating is through sharing and the care of older kinspeople for younger kinspeople. The flow of food is intense and it falls to elder kinspeople (whether of ascending generations, such as parents and grandparents, or of the same generation, such as elder brothers) to feed and satisfy the desires 26 of younger kinspeople, so as to ensure that they live a happy and healthy life. It is precisely through sharing that kinship relations are created and perpetuated. Among affines, in contrast, what we see is that the flux of goods continually recreates the hierarchy of the relation, which is manifest as shame and respect. It thus befalls affines who are wife-receivers to work for their father-in-law and brothers-in-law. It is also extremely impolite to solicit anything from those affines whom one should respect.
Transactions carried out between friends (ato), on the other hand, necessarily imply the mutual recognition of those involved, as subjects, capable of extracting from one another the objects that, then, circulate -at least when they are exchanged -as parts of those persons, thus appearing as what Strathern (2006: 270) calls "mediated exchange". Friendship is a type of relation that is very common among Upper Xinguanos and involves same-sex people who exchange with each other, within enduring relations. It is a type of relation that is quite distinct from (if not actually opposed to) that which is established, for example, among Jê formal friends.
Here the relation is marked by distance and avoidance (Carneiro da Cunha 1979), while Xinguano friendship is characterized by the absence of formality and the presence of proximity: friends embrace each other, touch each other and share secrets, as well as exchanging objects with each other.
According to Santos-Granero (2007: 11), in Amazonia relations of friendship "emerge in contexts of great fear of (potentially) dangerous others. It is this fear that endows native Amazonian intertribal friendships with their peculiar character and explains the great importance attributed to trust in such relationships".
This view seems interesting in the Upper Xinguano context, where the establishment of a "peaceful regime" also produces a regime of great distrust, and friendship appears to be particularly marked where security is at a low: in the periphery of the kindred, and between people of different groups. However, in the Upper Xingu relations of friendship can also have a collective character, which distinguishes them from the cases compared by Santos-Granero.
As we have shown, the uluki is an effort at creating the fiction that guests and hosts are part of the same local group: their chiefs address each other by vocatives that preclude the distance and potential antagonism that is present in the speeches of other rituals, friendship emerges as a marked category of the relations of reciprocity, and visitors are brought into the circle of houses, where they cook their own food. It is as if this ritual amplified to the regional scale those qualities of the sociality of the local group -which, according to Stephen Hugh-Jones (2013), is the main characteristic of regional Amazonian systems such as those of the Upper Rio Negro and the Upper Xingu.
By using the category of 'friendship', which, as Coelho de Souza (1995) has pointed out, converges on affinity as well as consanguinity, the uluki engenders a constant game of identification and differentiation which creates the conditions for groups to exchange things between themselves (not only objects, but also women, speeches, food) which present themselves as parts of collective people. Chiefs that address each other as 'brothers', hosts that receive guests in their villages (and not in forest encampments, as in the chiefly rituals), the distribution of raw food, and the rituals that follow the exchanges, all serve to blur, at some moments, the distinctions between the two groups, "making kin of them", masking the difference between potential affines, in a process that never reaches its final stages.
These mediatory role between identification and differentiation, consanguinity and affinity, is manifest in various Amazonian contexts by what Viveiros de Castro (1993, 2002 called 'included thirds', one of the possibilities of actualizing potential affinity. According to Viveiros de Castro (2002: 152), these are "positions that go beyond the dualism of consanguines vs. affines, or kin vs. strangers, and which perform key mediating roles". Santos-Granero (2007: 9), for his part, points to friendship as the point of mediation between "the voluntary and the ascribed, the personal and the social, the altruistic and the selfish, the affective and the formal". Whether these figures be defined as included thirds or as friends, what matters to us here is that they typically appear in an "egocentred" form, that is, as singular persons (such as, for example, in the case of Jê formal friends, or Araweté friends). The uluki case suggests, however, that there is a sociocentric yield to this relational form within a collective scale, pointing to an amplification of the qualities of (para-)kinship relations beyond the remit of local groups. This is perhaps why the uluki played a key role both in the process of "Xinguanization" of the people of the headwaters, and in the event of "pacification" brought about by the Villas Boas brothers during the creation and consolidation of the Xingu Indigenous Park. In Kalapalo narratives about early 20 th Century attempts to pacify/Xinguanize the Jaguma (enemies who spoke a Carib language of the Arara-Ikpeng branch), the uluki was the first ritual that they established with these people, acting as a gateway into the Xinguano ritual system.
A similar process occurs to non-Xinguano people which, while not participating fully in the ritual complex, nonetheless remain within its "orbit of attraction" through the uluki 27 . It is also the uluki that, in its more domestic and informal variants, mediates many of the relations maintained with non-Indigenous people. At the same time that, alongside the Quarup and the Jawari, the uluki defines a certain "Xinguano interiority" (the world of multi-community rituals), it is also one of its main hinges for opening up towards the exterior, with the potential to attract not only singular persons, but also whole groups, into the Upper Xinguano ritual system.
This brief ethnography of the uluki has sought to identify some of its specificities vis-à-vis other rituals, and to suggest other questions that merit further investigation in the future. First, it is necessary to better understand the forms and meanings of Xinguano hospitality, a more variable phenomenon than may first appear. Hospitality is what enables people to circulate through an extensive territory in the knowledge that the chances that they will be well-received are greater than the chances of falling to enemy ambush. Hospitality, as Nancy Munn (1986) has shown, has the potential to expand intersubjective spacetime, creating long-lasting exchange relations between people of different locales (operating passages, therefore, between various scales of sociality). Another question that must still be explored is, evidently, the role of gift-exchange in the regional economy. We know that the exchange of objects is an important part of rituals, but we still need a more in-depth ethnography of Xinguano economy.
27 And also through marriages and shamanic relations. Marriages between the people of the headwaters and people who are not considered "Xinguano" are infrequent (the Kalapalo consider them "wrong", undesirable), but they exist and are important in the inter-ethnic dynamics of the groups who carry them out. The influence of Upper Xinguano shamanism on their neighbours is expressive, and would merit its own ethnography.
We suspect that the conjugation of ethnographies of forms of exchange and hospitality beyond ritual contexts can clarify aspects of the formation of multi-ethnic and multilingual political formations, like that of the Upper Xingu.