Navigation – Plan du site

AccueilNuméros en texte intégral97-1ArticlesDossier : Discours rituels en Ama...As Spirits Speak: Interaction in ...

Dossier : Discours rituels en Amazonie

As Spirits Speak: Interaction in Wauja Exoteric Ritual

Christopher Ball
p. 87-117


Paroles d’esprits: interaction dans un rituel exotérique wauja. Cet article aborde les relations entre rituel et discours dans une société amazonienne à partir de la micro-analyse d’un acte de parole: la cérémonie wauja (Arawak, Haut Xingu) permettant « d’attirer les esprits ». Le patient y est soulagé par une conversation avec les monstres-esprits qui l’ont rendu malade: il espère par ce moyen les transformer d’esprits pathogènes en esprits gardiens. Cette conversation est menée avec un groupe de représentants du village, des compagnons qui acceptent de parler en tant qu’esprits et, ce faisant, d’entrer dans une relation contractuelle d’échange avec la personne malade et les esprits, relation qui implique diverses obligations et compensations en maladie ou en santé susceptibles de s’étendre sur une vie. Compte tenu de la gravité des circonstances et du poids de l’obligation créée, l’interaction rituelle est étonnamment ordinaire et peu élaborée en regard des pratiques chamaniques ésotériques et des complexes rituels collectifs, comportant danses et musique, de cette société. En examinant comment l’idéologie wauja rend possible une communication verbale avec les esprits de la part de personnes ordinaires, je suggère que les idéologies culturelles wauja relatives à la maladie, au langage et à la cosmologie émergent justement dans ces textes interactifs. La pratique wauja « d’attirer les esprits » éclaire ainsi la relation entre agentivité et performativité rituelle dans la mesure où elle permet de montrer comment des agentivités multiples, de personnes et d’esprits, sont liées à travers des textes construisant la réalité sociale.

Haut de page

Entrées d’index

Palavras chaves :

interação, ritual, semiótica, xamanismo

Géographique/ethnique :

Wauja, Amazonie, Brésil, Arawak, Alto Xingu

Thématique/disciplinaire :

Anthropologie, Ethnologie
Haut de page

Notes de la rédaction

Manuscrit reçu en octobre 2009, accepté pour publication en janvier 2011

Texte intégral

« The world, in truth, is a wedding ».
(Goffman 1959, p. 36)

« Felicity’s Condition: to wit, any arrangement which leads us to judge an individual’s verbal acts to not be a manifestation of strangeness ».
(Goffman 1983a, p. 27)

  • 1  I use the hyphenated term « spirit-monsters » to highlight the supernaturally dangerous aspect of (...)

1This paper approaches the connections between ritual and discourse in an Amazonian society through a microanalysis of one instance of talk. I examine a Wauja (Xingu Arawak) ritual session of « bringing spirits », wherein a sick person is aided by a conversation he has with the spirit-monsters1 that are afflicting him, whom he hopes by virtue of this very conversation to turn from pathogenic to guardian spirits. The conversation is facilitated by a group of representatives from the village, consociates who agree to speak as spirits, and in so doing enter into a contractual relationship of exchange with the sick person and the spirits that involves various obligations and rewards that may span a lifetime, in sickness or in health. In view of the gravity of the circumstance and the weight of the obligation it creates, the ritual interaction is a surprisingly inelaborate and quotidian counterpart to more regimented esoteric Wauja shamanic practice (Barcelos Neto 2004), and elaborate collective rituals involving music and dance (Mello 2005; Piedade 2004). I question how it is that Wauja ideology supports the practice of regular folk verbally channeling spirits, and suggest that in fact Wauja cultural ideologies of illness, language, and cosmology emerge in precisely such interactive texts as this. The Wauja practice of bringing spirits highlights issues of agency in relation to ritual performativity because it demonstrates how multiple agencies, of persons and spirits, are linked through texts that construct social reality. The ritual aspect of this linguistic act makes its performative potential reproducible in recognizable and unmarked iterations (Keane 1997).

2My analysis can be located among discourse-centered ethnographies of Amazonian societies (Sherzer and Urban 1986; Urban 1991). I take seriously Urban’s (1996) example that analysis of apparently minute elements of discourse such as pronouns as they function in ritual and quotidian life contributes to our understanding of how collectivities are constructed in Lowland South America. I follow Graham’s (2003) linguistic anthropological analysis of the ways that expressive performance by Xavante people in Central Brazil builds indexical continuities with the past and facilitates control over present circumstances. My orientation to discourse is also grounded in the work of Erving Goffman, whose ideas have seen previous application in the anthropology of Amazonia, specifically in Gregor’s (1973) work among the Wauja’s close neighbors the Mehinaku.

3Consider the two quotations from Goffman (1959, p. 36) presented above. In the first statement, « The world, in truth, is a wedding », I take Goffman’s use of « wedding » as opposed to « marriage » to invite the inference that he refers to the ritual aspects of performance that permeate everyday interaction, his way of saying that all the world’s a stage, that reality is ceremony and vice versa. The quote raises the connections between ritual performance and performativity properly speaking, as it reminds that social relationships are interactional achievements founded in pattern and authority. I hope to pivot on Goffman’s observation, now widely attested, that interaction is thoroughly ritualized, to show that Wauja ritual, and ritual writ large, is also thoroughly interactional. I guard as fundamental that language use is dialogic, and I show the same to be true of ritual.

4The second quote comes from Goffman’s reformulation of Speech Act Theory’s insight, in spite of its other limitations, that meaning something by saying something is a performative act. Goffman suggests that at the base of the very possibility of any felicitous language-mediated interaction lays the presupposition that participants must constantly and mutually display to one another that they are not insane. This he calls « Felicity’s Condition: to wit, any arrangement which leads us to judge an individual’s verbal acts to not be a manifestation of strangeness » (Goffman 1983a, p. 27). I add to this that what conclusions actors can derive in terms of an interlocutor’s sanity from talk and comportment, measured by standards of what it means to be strange, are matters of culture that display documented variability. I aim to elucidate something of Felicity’s Condition among the Wauja as it applies to this case, such that people speaking as spirits may not for a moment be presumed a manifestation of strangeness.

  • 2  Goffman (1983b) theorized this as « the interaction order ». I use the term « dialogicity » and th (...)

5These two quotes frame what I take to be Goffman’s relentless toggling back and forth between the conditions of possibility of social order, evidenced at once in the institutional and role and status categories that are the usual objects of sociological analysis, and also in the conditions of possibility of real-time lived experience, of the intersubjective unfolding of dialogicity as the basic fact of social being2. I hope to elaborate upon the intellectual contributions gestured at in each of these quotations, and I do so through consideration of the same sorts of data that occupied Goffman; interpersonal interaction mediated by talk.

6My attention focuses on two aspects of this Wauja practice. First I focus on the actual micro text and context of the spoken interaction that unites the patient and his family with the spirit representatives and thus the spirits. In this part of the analysis I am concerned with demonstrating the ways in which person deixis and voicing structure the transformation of people into spirits, as men take the perspective of others and voice the intentions of those others. I also focus on the discourse features that frame this ritual act, those qualities such as laughter, brevity, etc. that make this ritual moment exoteric as opposed to esoteric. Second I focus on the web of Wauja social relations that supports spirits speaking through men as a possible condition of reality. I develop a semiotic analysis of the Wauja process of illness and healing to show how spirits come into contact with humans, and how humans manipulate the conditions of such contact through talk, consumption of drink and smoke, through dance and song. The techniques that make it possible for humans to be(come) spirits, techniques that I analyze as involving indexical iconicity, are essentially the same techniques that make discursive interaction possible. That is to say, indexical linkage between iconic frames is the basic structure of interdiscursivity and it is the basic structure of ritual as interactional text (Silverstein 2005). By showing the semiotic parity of these forms, and in addition their actually existing, and therefore indexical, connections to one another in Wauja practice, I hope to show how Wauja speak as spirits.

  • 3  A challenge by Pierre Déléage to critique ritual efficacy in concert with emphasis on the interact (...)

7I argue that this analysis of Wauja interactional ritual may teach us two comparative theoretical lessons. First, that ritual structure is to a large extent emergent in interaction, and thus that it is formally and causally in dialectic with interactional textuality (Bauman 1996). I pursue formal connections in terms of iconicity and causal connections in terms of indexicality. My second point is that because ritual and interaction are thus structurally related, analysis of their connections helps to clarify, helps to explain even, the particular cultural ideological conditions that make any sociohistorically specific speech acts, conversations, pronouncements, prayers, plays, etc. possible. By this I mean to question the utility of ritual « efficacy » as an analytic by turning attention to ritual « possibility », the normative cultural presuppositions about subjectivity, cosmology, and the principles (also principals for that matter) of intersubjective business that underlie dyadic communication and the accomplishment of creative, transformational performance3. The ideological presuppositions that possibilitize interaction of certain cultural kinds are reproduced and themselves transformed in discourse and ritual (Silverstein 1976). So in investigating how Wauja ritual interaction that purports to bring spirits is made possible ideologically, I am also concerned with how it is effective as a performative act, but not so much in whether or how it effects a cure by its channeling, but rather how it effectively (re)produces the possibility of its own discursive normalcy.

8As Keane (1997, p. 24) points out « performatives seem to express the intentions of speakers (it is not just anyone but I who hereby wed thee). But their institutional force depends on certain conventions (“I hereby do thee wed” is by social and legal convention how one weds somebody) ». In order to be felicitous ritual language must express the agency of the speaker and of the agents invoked therein such as ancestors, spirits, or divine authority (Austin 1962). Ritual discourse exemplifies the difference between performance and performativity. Kulick (2003, p. 140) poses the distinction as follows, « performance is something that a subject does. Performativity, on the other hand, is the process through which the subject emerges » (see also Butler 1993). I consider how speaking as spirits is not just a performance, but a ritually authorized performative act, an act through which subjects emerge and are transformed. Because Wauja, like other Amazonian peoples, attribute agency to beings like spirit-monsters, the speech of these subjects is just as performative as ours. Humans speaking as spirits is not mere performance, it meets Felicity’s Condition and goes further to the constitution and potential transformation of agentive subjects.

Bringing Spirits

9Wauja people participate in the Upper Xingu multilingual culture area located in the Xingu Indigenous Park in Central Brazil. The Upper Xingu contrasts with other Amazonian multilingual culture areas such as the Vaupés where a system of language group exogamy helps to produce thoroughly plurilingual communities and individual speakers (Franchetto 2001). The Upper Xingu, instead, operates on a preference for language group endogamy that reinforces an ideology of individual and language group monolingualism. This produces a unique situation where genetically unrelated languages from the South American Arawak, Tupí(-Guarani), and Carib stocks are spoken in communities of people who actively maintain regular social contact with one another, who share a core mythical, ritual, and material culture, but who keep linguistic purity as a distinctive badge of identity within the system (Franchetto and Heckenberger 2001). Intergroup factional alliances and rivalries crosscut this unity. The Xingu Indigenous Park should not be seen as merely a container for a system that pre-dates contact however, since the constitution of the park has contributed historically to some of these salient regional sociological and cultural features through a process Menezes de Bastos (1989) refers to as pax Xinguensis.

  • 4  In this treatment of Wauja ritual practices of healing I do not have space to attend to two import (...)

10The Wauja speak an Arawak language related to but different from the Arawak languages spoken by other members of the Upper Xingu such as the Mehinaku or the Yawalapiti. They regularly host and travel to intergroup ritual events that work across the language and marriage boundaries in the social system to reproduce exchange relationships and to reinforce a mix of social solidarity and competition. Ritual exchange events also serve as public reminders that Upper Xinguans share the basic outlines of an ideology of ritual agency and performativity. Upper Xinguan intergroup meetings operate on the principle that this ideology is put into practice in the same way by all Upper Xinguans when they are back in their own communities. It is this practice and the ideology that emerges with it in a Wauja setting that concerns the rest of this paper4.

11In the Wauja context, as in the rest of the Upper Xingu, illness brings a patient into conjuncture with spirits, kin, and shamans. Spirit-monsters may not necessarily intend to harm humans, but they are attracted to us, and when they come into contact with people they cause a part of the soul to become separated from the person. This is the root of Wauja disease, and curing requires convincing the spirit to leave the human, which makes shamanic retrieval of the separated soul possible. At the base of illness and curing then is dialogic interaction between humans and spirits that has the goal of reestablishing the proper separation of their agencies. These interactions, mediated by specific forms of talk and exchange, are part of a ritual process that projects relations across ontological levels as it reproduces political structures and accomplishes psycho-social repair. The ritual process from illness to health, or death, also calls for the constitution and employment of a group of people who are neither kin nor shamans, but who represent afflicting spirits through the curative and celebratory phases of ritual performance. I focus on the interactional moment when this group first comes together, assumes spirit identities, and introduces itself to the victim by speaking as pathogenic-cum-protecting spirits. This act is called atuwata apapatai in Wauja, literally meaning to bring spirit-monsters.

12The first researcher to describe the Wauja practice of bringing spirits was Barcelos Neto (2004, 2009). Barcelos Neto (2004, p. 145) recognizes that the social actors nominated to the role of what I call representative of the spirit-monsters have the capacity to « personify » the spirit-monster that have taken fractions of the patient’s soul, and in such a role, to return the fractions to the patient. The spirit-monsters are brought through the representatives to the exact site and time of the meeting with the patient, wherein the representatives embody the spirit-monster and have the agency of these. They speak to the patient as spirit-monsters, not interpreting their messages, but animating their voices. As such the representatives stand in for the afflicting agents and are seen to (ideally) turn the direction of the « attack » for the benefit of the patient. Barcelos Neto (ibid., italics in original) summarizes that with respect to the patient, representatives are « indices of a therapeutic power, the maximum concentration of which is found in the apapaatai/yerupuhu ». Barcelos Neto’s (2009) discussion of bringing spirits among the Wauja argues that material objects involved in this and related rituals are the foci of exchange relations between humans and humans and spirits. I complement this insight with attention to the iconic and indexical function of transacted material items, and I argue in addition for the discursive construction of the actors and axes of these relations of exchange.

  • 5  Women can in principle be shamans, but I know of no historical cases of female shamans in the Wauj (...)

13Bringing spirits is only one of the curing techniques Wauja people may employ during illness. As a remedy, bringing spirits is part of a cluster of curative actions involving both shamanic treatment and public ceremony such as dances accompanied by masks, singing and music that celebrate spirits. While it is connected to these and complements them, it also differs from visionary shamanism and other varieties of individually practiced shamanism involving a charismatic healer operating with esoteric knowledge, as well as ritual musicianship and singing that require long and expensive apprenticeship in a store of proprietary song. In contrast, bringing spirits is a kind of healing that may be practiced by anyone, even young people and women5, and while it does have a ritual script, the words are simple, quotidian, and no secret is made of their content. Bringing spirits is an opportunity for regular folk in the village, mostly but not exclusively men, who may not be ritual experts by virtue of age or experience, to aid an ill person by channeling the afflicting spirits, encouraging them to leave the victim alone.

14Three parallel transformations are the desired effect of this act. First the kauki-tsupa « (spirit-monster’s) victim » should be transformed into o-wekeho « (spirit-monster’s) owner or master ». This is in shorthand a change from victim to master. The spirit-monster should change from apapatai « (pathogenic) spirit-monster », to o-kawoka-la « his (the master’s) guardian spirit ». We can summarize this as a change from pathogen to guardian. The representatives undergo their own transformation from human to spirits, and from representative to performer as they act as the victim/master’s o-kawoka-la-mona, which in the first instance refers to « the victim’s pathogenic spirit-monster’s representative », and in the second instance to « the master’s guardian spirit representative ». This can be characterized as a change from representative to performer. The illness stage of the ritual healing process brings together the victim, the representative, and the pathogen. Through this act and possibly individual sessions of shamanism, the spirit-monsters afflicting the victim may be convinced to leave him alone. The victim is cured, and in his new role as master he will owe the formerly pathogenic spirit-monsters, now his guardian spirits, a debt he repays through later celebration. The performers of a contracted celebratory dance will ideally be the same individuals who were contracted to bring the spirits when the victim was still ill. The interaction between victim, representative and pathogenic spirit-monsters, correctly articulated in bringing spirits, moves eventually to the celebration phase turning these into the spirit master, his spirit performer, and his guardian spirit (Figure 1).


from the illness phase through

transformation performed in bringing spirits

to the celebratory phase

Remains human as…


« victim »



« its (the spirit-monster’s) master »

Alternates between

human and spirit as…

o-kawoka-la-mona « his (the victim’s) pathogenic

spirit-monster representative »



« his (the master’s) performer of dance, song, flute, etc. »

Remains spirit as…

apapatai, pathogenic « spirit-monster »



« his (the master’s) guardian spirit »

Fig. 1 – Ideal process of subject transformation through bringing spirits.

  • 6  I treat connections between Wauja grammatical marking of (in)alienable possession and Upper Xingua (...)
  • 7  See Viveiros de Castro (2002) for a description of the related Yawalapiti modifier.
  • 8  In addition, kaukitsupa, the term for victim is based on the stem kau « pain, illness, or fear » p (...)

15Terms referencing the representative/performer as well as the guardian spirit are typically but not exclusively grammatically possessed, as indicated in Figure 1 with o-, the 3rd person possessive proclitic. In addition, possessed forms based on the noun stem kawoka « guardian spirit » also take the suffix -la, which marks possession on the Wauja grammatical class of alienable nouns. This references the victim/master’s relationship to both of these roles, he contracts them and persuades them to help him, and in exchange he is bound to sustain them with gifts including food, drink, and tobacco6. Note also that the only difference between the Wauja linguistic forms for guardian spirit and guardian spirit representative is the addition of the clitic -mona, which denotes a mere copy of an original, a token of a type7. This supports the way in which Wauja describe the representatives as individual and admittedly necessarily degenerate manifestations of generalized and purer spirit subjectivities8.

16As can be seen in Figure 1 the victim/master role remains filled by a human subject throughout, and the apapatai/kawoka (pathogen/guardian) role remains filled by a spirit-monster subject throughout, while the representative/performer role, as mediator, moves back and forth between these subjective positions. It straddles the boundary between them. This happens in relatively tight and relatively loose cycles, as the kawoka-mona moves between human and spirit identities over the course of one ritual session of bringing spirits or dancing with masks, but he also moves between those same identity positions over the long term because bringing spirits and dancing with those spirit’s masks are activities that may be separated by months or years as part of the same ongoing exchange relation.

17I now describe the scene of one ill man’s call for spirits. I begin by outlining the basic contours of the players involved, combined with description of the ethnographic context. I then focus on presentation of texts recorded that day that make up the session of bringing spirits. Peheke is a shaman. As a shaman he has gone through the process of guarding within his body a ball of illness. This ball, called yalawo in Wauja, is the source of any shaman’s power. Not unlike a vaccine, it is a defeated pathogen now under its master’s control. It represents a bridge between the spirit and the human domains. Originally the material effect of Peheke’s contact with a malevolent spirit-monster, it is now put to positive use, giving him the power to intervene in other people’s bodily encounters with spirits, the cause of illness. Wauja shamans sacrifice wellness for this power, they are described as perpetually sick. The ball gnaws at them, and while it is normally under control it is always a discomfort at best, agonizing at worst. In extraordinary circumstances, the black matter may fester and the shaman who once was a patient may relapse and become ill again. This has happened to Peheke. He was seen by another more senior shaman the night before who confirmed that Peheke’s yalawo was rising in his throat, threatening to exit through his mouth, which he told me, would cause Peheke to die. The shaman identified a group of spirit-monsters that were the cause of this condition. These included Yamurikuma, a female spirit, Yuwejokui-xuma, Yakui, and Kojoma. Even though Peheke is a shaman, in his current state he is a victim like anyone else and his family members seek help in the community to make him well.

18At this point in a diagnosis, the patient and his family face a choice. They may choose to continue contracting the services of an individual shaman beyond diagnosis to treatment, and/or they may call for regular village folk to assist in curing by representing the spirits that the shaman has identified. This is bringing spirits, and it is relatively inexpensive in the short run, as the family only has to provide drink and tobacco to representatives immediately, whereas treatment by an individual shaman or shamans requires payment of valuables such as shell ornaments, feather diadems, ceramics, cash, etc. The representatives channel the spirits and work to get them to leave the victim. As Peheke’s son Walaku put it to me while he was in the center calling his father’s representatives, people call spirits so that they will go away, the representatives smoke the tobacco given to them, the representatives drink the porridge given to them, and the spirits simultaneously smoke and drink. They are happy so they go away. The fact that the representatives speak as spirits is bound up with the fact that they receive drink and tobacco from the ill person’s family. They smoke, drink, sing, and speak as spirits. Later they will dance as spirits in celebration of recovery and in prevention of relapse. The tobacco and drink offerings are first pair parts directed to the spirits through the representatives. The spirits are pleased to smoke and drink, and the representatives, speaking as spirits when they vocalize their identity as guardians, complete a second pair part reply. This is an exchange relation between the victim/master and his pathogenic/guardian spirits mediated by material and verbal indexical signs.

19Among the spirits the shaman saw that are afflicting Peheke and that Walaku needs to call are spirits called Yuwejokui-xuma. I followed Walaku to the center of the circular Wauja village and recorded in digital audio as he called to and interacted with the representatives. Walaku also offered explanations of the process to me on tape as it was going on. He describes the spirit Yuwejokui-xuma to me as a family; there is Yuwejokui-xuma proper who is an old man, his wife who is an old lady, and their two children. Standing in the center of the village, prepared with an offering of porridge, he calls for a young man named Puixa to represent one of the children now.



Puixaaaaa... maniu toyaaaaa...

Puixa come here friend


P-atuwa-ta apapaataiiiii...

(You) bring a spirit




What is it?



O-tain pitsu-wiu

You are the child



Katsa o-tain?

Whose child?



Yuwejokui-xuma o-tain

Yuwejokui-xuma’s child



Yuwejekui-xuma o-tain?

Yuwejokui-xuma’s child?



All right



Siyai kaliu-han

Siyai there


Opawa pitsu-wiu

Is your sibling







Eh heh


  • 9  It is a stylized version of a general cultural permissiveness with regard to volume and distance o (...)

20Walaku stands in the center of the circular village and calls out to the people he seeks as representatives. The call in line 1 is broadcast in a very loud voice with extremely elongated final vowels. This is a summons, a maximally public announcement in the manner of a town crier9. Although it is available to anyone in earshot, this call does not have the same generalized audience as its object that a public announcement might. Rather the call is always directed at a specific addressee. In this way the social relation initiated by this summons is from the beginning available for public scrutiny. Village members know who is calling for spirit representatives, they may deduce who is ill, and they may hear directly who the representatives are.

21What is not divulged at this point to the audience of public overhearers is the identity of the spirits. That is the subject of the intimate conversation that follows the summons. In this case, Walaku summons Puixa, who was conveniently hanging around near the men’s house in the village center. Note that even if the distance to the addressee does not require yelling for the whole village to hear, it is still formally required that the summons be broadcast to all houses. Puixa arrives after about 20 seconds, and greets Walaku in line 3 with the utterance natsi? « what is it? », the standard Wauja reply to being hailed. Walaku replies in line 4 by telling Puixa that he is assigned the identity of a child. « Whose child? » asks Puixa in line 5. Walaku completes the assignment by stating in line 6 that Puixa is to be Yuwejokui-xuma’s child. Puixa quickly agrees in line 8, and Walaku informs him in lines 9 and 10 that Siyai will be his sibling.

22In this initial phase, Walaku uses the second person singular pronominal clitic p(i)- in line 1 and the second person singular pronoun pitsu in lines 4 and 10 to address the representative. Here he addresses him initially as a human, but in the act of assigning a spirit-monster identity to him, the second person pronoun also comes to reference that spirit. Before being called and having his spirit identity assigned, saying « you » in reference to Puixa references the man. However, after the first equation of « you » with the identity of a spirit-monster in line 4 with the statement « you are the child (of Yuwejokui-xuma) », when Puixa is next addressed as « you » in line 10 the referent is the spirit. We know this because Puixa and Siyai are not brothers, so when Walaku says to Puixa that Siyai is « your sibling » he refers to the relation between the children of Yuwejokui-xuma, not the relation between the two young men. This is the first evidence that person deixis, particularly its inherent indexical component, is key to creating the interactional fact of people speaking as spirits. Deixis is a fundamental component of the participation frames that bridge interaction and ritual (Hanks 1990). As a paragon of ritual performativity, the act that Walaku and Puixa have loudly, and then quietly, completed in the center of the village is a baptism.

23Walaku continues calling out to various people and assigns spirit roles to them as they arrive in the center. At this point Walaku offers porridge to the gathered representatives. On this occasion it is akain-ya, a drink made from fruit of the pequi tree. Drinking this offering is the first act the representatives perform as spirits. They must drink in this role so that the spirits may also drink. Since the female spirit Yamurikuma was included in the diagnosis, a group of women has already been called and gone to the victim’s house. The women visit separately from the men, so the representatives assembled now are all male and most are young. That does not mean that they are all novices. The youngest Wauja visionary shaman is in this group, but here he is contracted not for his more expensive treatment based in esoteric knowledge, rather he is asked as a spirit representative to help in the same ways as the laymen. Bringing spirits has a leveling effect, because no prior special relationship with spirits is assumed, needed, nor even useful in channeling. When the representatives have all arrived and drank porridge, they casually organize themselves and proceed in more or less single file from the center to Peheke’s house (Figure 2).

Fig. 2 – The representatives proceed from the village center to the ill person’s house. The leader sings while some of the younger men imitate monkey vocalizations and laugh.

24Some carry walking sticks and some tie reeds into makeshift crowns that they wear on their heads. These are rudimentary decorations, appropriate to the exoteric feeling of this ritual, but they aid in the transformation of men into spirits. Leading the men is a respectable young adult who is apprenticing with his father in shamanism, while the visionary shaman brings up the rear of the procession. The leader sings one of the afflicting spirit’s songs, and some of the young men laugh and affect monkey noises in their role as spirits. The rowdy young boys walk with the young shaman and they joke together, perhaps already invoking unruly spirits. In the course of their movement from the center of the village to the patient’s house on the periphery of the village, the representatives cross from a public space to a relatively private one.

25Upon their arrival at Peheke’s house, a senior family member named Ajapu greets the spirits in line 14 by encouraging them to come into the house.



Yuhiyu hiyu hiyu hiyu…




Maniu, maniu, maniu

Come, come, come







P-iye-ne p-uku-la-uya Yuwejokui-xuma

(You) take your arrow Yuwejokui-xuma






P-iye-ne p-uku-l-iu Yuwejokui-xuma

(You) take your arrow Yuwejokui-xuma

26As the representatives enter the house in single file, they stop singing, and mark the end of the song with a collective cough and sigh chorus in line 15. This is a typical Wauja utterance that signals a transition between ritual phases of performance, used especially when a task has just been completed. The spirits have arrived, and they are prepared to smoke tobacco and speak with the victim. Ajapu distributes tobacco cigarettes that he has pre-rolled. He says in lines 16 and 18 for the spirits to accept their « arrow », here a metaphor for the cylindrical cigarette. He addresses the spirits in the second person using the proclitic p(i)- to reference both the subject of the verb « take » and the possessor of the noun « arrow ». He reinforces the identity of the receiver and smoker through direct addresses as Yuwejokui-xuma. The spirits accept their tobacco with a ritual sighing vocalization of hawauyaaaa, and they begin to smoke. The manner in which they exhale the smoke is also stylized. When they approach the victim in turn and smoke over him, they simultaneously blow the smoke out and vibrate their vocal chords to produce a « voiced » exhale. It is at this point that the representatives speak to the patient while they smoke.

27There is still a sense of levity, shuffling around, engaging in small talk. A few of the younger men who are not regular tobacco users cough and sputter as the older men tease them, not for performing the ritual incorrectly, but for being inexperienced smokers. As the representatives move towards the victim, Ajapu offers me a cigarette and instructs me how to manually and verbally receive it. He moves quickly between explicit technical instruction in the mechanic aspects of the ritual moment to a casual conversational footing. Ajapu explains that he has less tobacco than he would like, having smoked and shared much of what he had on a recent trip to the park’s post. He jokes a little bit with me about some Brazilian nurses who have recently come to the post, laughingly teasing that I would like to meet them. Other representatives awaiting their turn laugh along. This sort of banter fulfills the common Wauja and wider Upper Xinguan conversational expectation of sharing news obtained from intervillage travel. It also is a common type of male gendered talk that presumes salacious motives as it reinforces masculine camaraderie. It is typical in the staging and especially rehearsal of ritual, marking the backstage space-time of preparation, waiting, and rest that punctuates entrance to and performance on the stage. Such offstage joking is a necessary counterpart to the concurrent curing happening one meter away onstage at the victim’s hammock. The men enter that arena one by one, each hovering over the victim as they blow smoke and identify themselves as spirits.



Pi-xawoka-la natu-wiu

I am your guardian spirit


Yakui natu-wiu

I am Yakui




All right



Pi-xawoka-la natu-wiu

I am your guardian spirit


Yowejokui-xuma natu-wiu

I am Yowejokui-xuma




All right



Pi-xawoka-la natu-wiu uwa

I am your guardian spirit uncle


Kojoma natu-wiu

I am Kojoma







mmm au















Yuhiyu hiyu hiyu hiyu…


28The statement that the representatives make in lines 19-26 explicitly identifies both speaker and addressee through person deixis. The representative refers to himself using the first person singular pronoun natu « I » in lines 20, 23, and 26. The use of « I » here figures the representative not only as a spirit, but specifically as guardian spirit. The representative states that he is a kawoka. This positions the representative as non-human, and it specifies the spirit-to-victim relation as that of a guardian spirit to its master. The victim is addressed using the second person singular proclitic p(i)- « your » in lines 19, 22, and 25 marking that the guardian spirit who is speaking belongs to the addressee. This positions the patient as a master, for the spirit identifies himself as no longer malevolent, he is now a protector. This utterance signals entry into a new social relation whereby the spirit breaks contact with the victim’s body, in exchange for contact mediated by the master’s giving sustenance in the form of drink and tobacco to the spirit. The victim-turned-guardian-spirit-master verbally consents to each proclamation in lines 21, 24, and 27. The spirit’s statement that it is now a guardian would seem to mark the transition accomplished in the whole ritual context of drinking, singing, and then smoking as spirits, but of course this statement is crucially not merely a report on a given state of affairs, however immediately accomplished. It is a performative statement that works to create the social relation it describes (Austin 1962).

29I appeal to Urban’s (1989) scalar analysis of the first person pronoun « I » in discourse in the Brazilian Gê language Shokleng. Urban builds on Benveniste’s (1971) recognition that the true pronouns you and I blend indexical and semantic properties to refer to whoever is uttering « I » at the present moment. Urban adds a number of uses attested in Shokleng discourse and in other languages that form a cline from this prototypical indexical reference to the everyday self to the assumption of a non-ordinary self. In Urban’s scalar scheme, simplified and quantified here for ease of explication, the types of « I » are 1) the indexical-referential used to index the everyday self; 2) the anaphoric and 3) the de-quotative « I » used to index a reported self; 4) theatrical « I » that hides the everyday self; and finally 5) the projective « I » where a non-ordinary self is assumed. The « I » of type 1 represents the normal case where « I » refers to the present speaker. The « I » of type 2 is found in direct quotation as in « He said “I am leaving” » where « I » refers to another speaker. Urban states that the « I » of type 3 can be seen in Shokleng when narrators lapse into voicing characters in myth without explicit quotatives. The theatrical « I » of type 4 is found in performances such as when a Shokleng myth narrator says « It was I who emerged first », voicing the original ancestor. The final projective « I » of type 5 can be exemplified by the same statement « It was I who emerged first » in the case that the narrator is not merely focused on the audience and the quality of his performance. Rather, if the speaker is sufficiently inwardly directed the use of « I » reflects and supports what Urban calls « subjective changes » whereby the narrator becomes the ancestral hero. Urban describes this projective « I » as common to ritual moments of possession and trance. Course (2009, pp. 305-306) has used this model to suggest that use of « I » in Chilean Mapuche singing that accomplishes « habitation within the subjectivity of another » falls somewhere in between mere quotation and total possession.

30I argue that the difference between the theatrical « I » and the projective « I » parallels the difference between performance and performativity in that projective « I » is not just something the subject does, but that it is a process through which the subject emerges. It is important to note that Urban’s (1989, p. 38) scalar model is also a circle or loop, since « the actor can become so immersed in another “I” that that other “I” becomes once again virtually indexical referential ». The use of the Wauja first person pronoun « natu » in bringing spirits is just that type of projective « I » that is at once the invocation of another subjectivity and a return to the basic indexical referential function of person deixis. It appears here as both highly ritually regimented, like more familiar instantiations of projective « I » in trance and shamanic possession, but also a moment in quotidian interaction wherein indexical referential « you » and « I » talk to one another. This is one important way that the ritual of bringing spirits is interactionalized.

31Immediately after the last representative speaks to Peheke, the host Ajapu asks if he is finished in line 29. The representative of Kojoma replies that he is finished, at which point the men once more emit the simultaneous cough and sigh that signals completion in line 30. The group begins singing the same song as before in line 31 and they exit the house to return to the center where the session of bringing spirits concludes, marked once more by a cough and a sigh.

32At every stage of the ritual there is a mix of quotidian and ritualistic linguistic elements. Ritualistic elements I consider to be those that have a marked intonation, volume, pitch, morphology, etc. that either cannot be accounted for by idiosyncratic circumstance, by regularly observable discursive function such as emphasis, clandestine whispering, etc., or that, where possible, can be accounted for by testimony or metapragmatic reflection of speakers. For example, speakers identify some of the utterance partials of these texts as only possibly usable in ritual. For example, the stylized cough that marks entry into and exit from the house and the termination of the total ritual sequence is clearly described by Wauja whom I asked, although like a cough in form, as clearly not a functional cough. Speakers recognize it as an index of metricalization, in this case marking the end and beginning of a song sequence. The huwauyaaaa sighing performed upon receiving tobacco, and the monkey call pantomime that the young men engage in while walking are specific to ritual. People do not sigh like that when leisurely sharing a smoke in a moment of rest, it is specific to receipt of tobacco offered in an exchange relation with a spirit. The initial calling of representatives in a long, drawn out, and loud voice has similarities to chief’s speech, which involves a more elaborated version of ritualized Wauja summons from the village center.

33Contrast these elements with the relatively quotidian and conversational moments as in the exchange between the newly contracted representative and his sponsor. Conducted immediately after the hailing, the assignment of a spirit identity uses normal voice, pitch, and no special morphology. It is even brief to a point, a fully colloquial and matter of fact affair. Perhaps we might say that in this very brevity it too is ritualistically scripted or regimented. Ajapu welcomes representatives on behalf of the victim’s family in a stylized way inside the house, intoning a welcome as he distributes tobacco. Elements of tobacco reception, both inhaling and exhaling, involve highly ritualized vocalizations such as huwaoyaaaa. When the representatives speak to the patient they do so in hushed tones but with colloquial syntax. The phrase that they utter is relatively set; « I am your guardian spirit, I am X », but it is hardly ritualized to the point of denotational or obvious interactional or sociolinguistic unintelligibility. This ritual is verbally approachable and intelligible for average Wauja speakers. Even though there are ritualized elements of discourse throughout, it uses what we might deem an entry-level ritual Wauja language. Unlike the genres of chief’s speech, song, or curing by cigar and prayer, all of which require individual verbal expertise achieved through intensive formal training and apprenticeship, this discourse exemplifies formally and interactionally common Wauja language, with an accompanying well known set of intonational, phonic, gestural, and onomatopoeic overlays. It is so simple that children can do it. So they do.

  • 10  Men are much more likely to be frequent kawoka-mona than women, in part because the spirit pantheo (...)

34They may be a little nervous, they may mess it up, and they laugh throughout, and this attitude of slight risk and levity is totally acceptable. Like young men who apprentice on what are called tankuwara flutes for weeks at a time before performances, they stumble and joke throughout their rehearsal with a sense of whimsy. In contrast, more serious and regimented practice is appropriate in connection to other Wauja sacred flutes that may only be played by experts and out of the sight of women. Bringing spirits is a thoroughly verbal ritual, it is a ritualized discourse. It is clearly exoteric because it is anything but inaccessible and unintelligible. One must only be called to stand in as a spirit to participate fully, and chances are for any Wauja person past puberty that this request will come soon and often throughout adulthood10.

Wauja Semiotic Ideology of Ritual

35I now want to fit this ritual interaction into the wider nexus of relations between victims of illness who become masters of guardian spirits, spirit-monster representatives who bring spirits and who dance in ritual performances afterwards, and spirit-monsters who attack humans and then become their guardians. Let me clarify that after bringing spirits occurs, the victim’s health may or may not improve. If it does not then shamanic curing and/or celebratory dance for the spirits may also be performed. Shamanism is outside the scope of my analysis here. Furthermore, whether or not the victim’s family contracts the representatives to perform celebratory dance before he gets better, if he does not die, the victim who becomes a guardian spirit master will sponsor celebratory dance for his guardian spirit after recovery. Thus, although celebratory dance can occur before a spirit has become a guardian, when I discuss celebratory ritual performance here I have in mind the prototypical case where the dance is contracted by a guardian spirit master who has been made well, at least in part by bringing spirits. In this way I locate bringing spirits in the full cycle of illness through to ritual performance in celebration of guardian spirits that can span decades. My goal is to show how the specific act of bringing spirits is framed by a Wauja semiotic ideology that underlies long-term connections between humans and spirits, and to suggest that this ideology is reproduced through ritual discourse of just the sort as bringing spirits. Wauja cosmological structure is invoked in instances of quotidian talk and exoteric ritual discourse in addition to full-blown esoteric ceremony.

  • 11  The third trichotomy of signs in Peirce’s theory includes the possibility that the interpretant of (...)

36In order to consider the specific semiotic sign modalities that establish and reproduce this process in terms of social relations, it is first necessary to distinguish between icons and indices as sign types. Both are considered here under the Peircean definition of signs, where a sign is any thing that stands in some relation between an object and its interpretant, itself another sign in the mind of a conscious observer. I emphasize here the relation between sign and object. An icon is defined as a sign that stands for its object (to its interpretant) through a relation of formal similarity or « likeness ». An index is defined as any sign that refers to its object by being actually modified by it. An index occurs in contiguity with its object and in fact is always in « dynamical (including spatial) connection both with the individual object, on the one hand, and with the senses or memory of the person for whom it serves as a sign on the other » (Peirce 1932, p. 170). These categories may combine, such that a photograph would be analyzed as an index primarily, as it is causally produced through the spatiotemporal contiguity of light and film. It is, however, an index of its object that may be interpreted by the viewer as an icon of its object. Thus the interpretation of the image as signaling through formal similarity or likeness is secondary, and we may designate such a sign as a photograph as an iconized index11.

37Figure 3 below presents three pairs of semiotic relations that mark the social relations observed in the previous descriptions of the illness-to-ritual process. It is crucial to distinguish between semiotic relations and social relations here. In this analysis « semiotic relations » are modes of connection between sign and object (and interpretant), whereas « social relations » are the objects of signs. When I use the term « relations » without the modifiers « semiotic » or « social », this refers equally to both. Social relations are defined here as relations between social roles, roles that may be potentially inhabited by different subjects or biographical individuals in different places and times. This separation of relations between roles and the subjects occupying those roles is evidenced by the switching of roles, say from victim to master, by the same inhabiting biographical subject, or by the passing of role responsibilities and prerogatives from one biographical subject to another, as when a father who was a spirit representative years ago may pass the responsibility of being a spirit performer in celebratory ritual to his son (Figure 3).


Semiotic Relation




symptoms (fever, dreams, trance, nasal bleeding, etc.)


social relation of victim


pathogenic spirit-monster


alimentation exchanged for



social relation of master


guardian spirit-monster



alimentation exchanged for



social relation of victim


pathogenic spirit-monster representative


alimentation exchanged for performance


social relation of master


guardian spirit-monster performer



speech, crown, comportment


social relation of pathogenic spirit-monster representative


pathogenic spirit-monster


dance, music, song, paint, decoration, dress


social relation of guardian spirit-monster performer


guardian spirit-monster

Fig. 3 – Signaling of social relations in illness-to-ritual process.

  • 12  The example of symptoms as indexes of disease is a classic example in the semiotic literature on i (...)

38Figure 3 shows how specific signs signal social relations in Wauja ritual. In this way we may read 1a, for example, as describing an indexical semiotic relation whereby symptoms index a social relation between the victim and a pathogenic spirit-monster12. Symptoms index a pathogen in spatiotemporal contiguity with the sign causing it directly. In this case, the Wauja theory of illness recognizes pathogens as spirit-monsters, and more precisely, the approximation of spirit-monsters to humans. In other words, what is indexed by symptoms here is the positioning of spirit-monsters in a specific social relation with human persons. This demonstrates why I analyze the objects of the indexes here as social relations as opposed to mere terms in social relations. It also shows the double causative property of indexicality, because when the cause of a symptom is seen to be the social relation of a spirit-monster to a patient, this relationship’s existence has been semiotically caused by the interpretation of the symptom’s significance. The indexing of social relations can effectively bring them into being then, and their iterative indexing or changes in the patterns of their indexing can serve to maintain or transform these social relations.

39In the pair of social relations in Figure 3 boxes 1a and 1b the victim or patient and then the master stands in an evolving indexical relation to the spirit-monsters. First he is afflicted by them and later they become his guardians. In the illness stage in 1a the sign is in connection with the senses of the victim and other observers as manifested in fever-induced dreams and other symptoms of the social relation of pathogen to patient. The social relation between master and guardian spirit in 1b is indexed most saliently by exchange whereby the victim, through the assistance of his family members, offers tobacco and drink to the spirits through the representatives. After the victim has become a guardian spirit master in 1b the social connection is maintained in memory and this spirit-monster may be recalled after years when the spirit’s master decides to honor the spirit-monster in dance. There is risk involved in not honoring the guardian spirit-monsters through ceremony, a potentially harmful misrecognition of the second indexed relation that threatens reversion to the first. Also, masters may feel the intervention of their guardian spirits in « saving » them from harmful situations (see Barcelos Neto 2004, p. 143), thus there is a sense that these spirit-monsters remain in potential contiguous proximity.

40Consider the connection between the victim and the representative in Figure 3 boxes 2a and 2b constructed during and after the cure. The first relation in 2a consists in the exchange of food for the work of curing, whereby, when the victim is sick, a family member of the victim offers food in the center of the village and requests and assigns spirit-monster roles to contracted representatives in bringing spirits. This index stands for, as it conjures, an exchange relation between victim and representative. At this point I reiterate that indexes are reflective as well as potentially creative signs, so that where some social relation is the indexical object of a perceivable juncture of social action between participants, the signaling serves to reinforce and to create the social relation itself.

  • 13  The passing of role inhabitance prerogatives can occur, especially when an owner keeps the right t (...)

41The social relation in Figure 3.2b is indexed through the asymmetric flow of resources from the master to the performer in payment for the return of the spirit-monsters who are invoked by ritual dance in their honor. This basic exchange relation of payment is reproducible whenever the ritual sponsor should decide to pay his former representative(s) with food, drink, and tobacco for their ritual performance of music and masked dance. Payment mediates the relation of master to performer as it did in the relation between victim and representative. This economic relation is an indexical relation because it is established in and successively mediated by spatiotemporal co-presence in the act of giving. For example, during a section of the performance of the matapu celebratory ritual individual performers (former representatives) carry bullroarers depicting the matapu fish spirit from the center of the village to the patio in front of their master’s house. This is done for every matapu guardian spirit master (former matapu victim) who chooses to pay. When the bullroarers have been spun in front of the masters’ houses, the representatives/performers return to the center where they stand with their sponsor who then personally offers a lit tobacco cigarette from his mouth to theirs. This is an indexical relation of spatiotemporal contiguity precisely because the sign-vehicle, the tobacco payment, marks or makes visible the victim/master to representative/performer social relation through a co-present exchange. Looking within the pairs comprised by Figure 3 boxes 1a and 1b and also in boxes 2a and 2b note that each contains similar semiotic relations in that they are almost all signs embodied in the exchange of alimentation, including tobacco. Further, each contains similar social relations in that the identities of the inhabiting subjects tend to remain constant while the roles switch13.

42I have described the relations between the master and the spirit-monster and the master and the representative as at base indexical. I next describe the relation between representative and spirit-monsters in Figure 3 box 3 and the particular way that the representatives stand for and speak as the spirit-monsters. I argue that this is at base an iconic relation of similarity in form or likeness. Evidence for this is varied. First consider the very Wauja term for representatives, kawoka-mona. In this nominal construction, kawoka refers to the guardian spirit, and the modifier -mona denotes an instance or likeness of the guardian spirit.

  • 14  See Barcelos Neto (2004, p. 142 et passim; 2009) for an alternative application of Viveiros de Cas (...)

43Viveiros de Castro (2002) analyses four modifiers in Yawalapiti, an Arawak language spoken by Upper Xinguan neighbors of the Wauja. In my interpretation of Viveiros de Castro’s analysis, the four modifiers, including the Yawalapiti equivalent of Wauja -mona, indicate points on two axes. The Y-axis is a scale of super- to sub-, to use the Standard Average European prefixal terminology. On this axis -kumã (Wauja -kuma) « excessive », or « superlative » is opposed to -malú (Wauja -malun) « inferior », or « deficient ». The X-axis distinguishes between type and token. Here -rúru (Wauja -yajo) « authentic », or « archetypical » is opposed to -mína (Wauja -mona) which designates the mere instantiation of a type, a « token », an iconic relationship of similarity in form, a « replica ». Note the exact parallel in terminological modification between the Wauja terms apapatai(-yajo) « (true) spirits » and apapatai-mona « animals » that take their image. Also, Wauja people often joked with me that while they were Wauja-yajo « true Wauja » I was Wauja-mona « like a Wauja », or a « token Wauja » person. The application of -mona to kawoka « guardian spirit » in the term kawoka-mona for guardian spirit representatives similarly signals that these are like the spirits or that they are token spirits14.

44Evidence from the forms of practice associated with this role supports the evidence for iconicity provided by linguistic coding. The guardian spirit-monster representatives are called and paid to iconically reproduce the spirit-monsters. They voice the spirit-monster, reproduce a simulacrum of spirit-monster talk, promising to leave. Barcelos Neto (2004, p. 138) emphasizes the iconic, onomatopoeic character of the verbalizations of the representatives in the practice of curing.

The content of the phrases converges on the enunciation that the visiting apapaatai will no longer harm the victim. These are the few words that the apapaatai pronounce during the ritual visit, all the rest of the oral communication is onomatopoeic. In the majority of cases, the victim does not perceive what was said to him, as the intention of the visit is not to establish communication at the linguistic level, the speech here is practically rhetorical without any explicit therapeutic effect.

45As I have demonstrated, the denotational message consistently communicated during such visits is the identity of the representative. This enunciation usually takes the form of two successive expressions juxtaposing the first person pronoun with the second person possessed form of kawoka, « guardian spirit », followed with the name of a particular spirit-monster and the first person pronoun, e.g.; pi-xawoka-la natu-wiu, Kojoma natu-wiu « I am your guardian spirit, I am Kojoma ». Recall that the victim does perceive what was said to him if he is conscious, and may indicate such perception with a sigh of agreement. That the representative voices the identity of the spirit-monster as though he were the spirit-monster is clear from the following points. First of all, the representative identifies himself as named spirit-monster X, as the case may be. Second, the representative does not say that he is the victim’s kawoka-mona, which would indicate that he was merely a representative, but he states that he is the victim’s kawoka, indicating that he is a guardian spirit. The representative voices a role occupied by spirits rather than a social role occupied by humans. Third, the representative should not address the victim with any terms indicating his kin relation to the victim. On one occasion I witnessed a representative address a victim who was a young boy with the appropriate statement of identity followed by the kin term (used as a vocative) nowa, « my nephew », the term that he would normally use with the boy in human interaction. This produced much laughter as it was seen as a mixing of the expression of spirit-monster with human, kin based relations. The representative had to do it over in this case, without the explicit reference to his humanity, in order to produce the appropriate voice. In this light the language practice in such curing sessions is not merely « rhetorical » without therapeutic effect. Rather the positive and linguistically coherent voicing of a spirit-monster identity works to the therapeutic end of producing a transition from pathogenic to guardian roles in relation to the victim.

  • 15  Simulacra and diagrams are icons in that they both represent their objects through similarity in f (...)
  • 16  Consider some distinguishing characteristics of the social role of representative in comparison to (...)

46After the verbal exchange in bringing spirits, the representatives subsequently diagram the spirit-monsters’ exit with their own15. So rather than « bring » and « send » the spirit-monsters away, we may say that the representatives at this stage come and leave as images of spirit-monsters. Likewise, when this iconic relation is repeated in the production and performance of relatively esoteric celebratory ritual, great attention is paid to presentation of the performers in the form of spirit-monsters through elaborate body decoration, dress, and dance. That this is an independent and iconic relation is attested by the sheer amount and importance of the iconography and aesthetics of spirit-monster performance in Wauja ritual16.

  • 17  In Peircean terminology this should technically be referred to as rhematization (Gal 2005).

47Up to this point I have referred in some instances to the victim and the master, or the representative and the performer, or the pathogen and the guardian as though they were separate roles. Sometimes, however, I have referred to combined victim/master, or representative/performer, or pathogen/guardian roles. This reflects an important tendency in the Wauja semiotic ideology I describe here to see these roles as equivalent and to subsequently laminate them on top of one another. Since the victim and owner role are inhabited by the same individual at successive stages the roles are read as alike, and since the pathogen and guardian roles are inhabited by the same subject, these are read as alike. Similarly, the role of representative and performer are read as alike. This can be thought of as a process of iconization of the roles, a process that leads to iconization of the relations between them17. I represent this tendency in Figure 4 with pair-wise collapsing of participant roles into one inhabiting subject and the concomitant collapsing of four indexical relations from Figure 3 boxes 1 and 2 into two indexical relations (Figure 4).

Fig. 4 – Dynamic system of relations in illness-to-ritual process.

48In Figure 4, the boxes represent social roles, the connecting lines represent social relations. The social relations remain the objects of signs here. Note that Figure 3 boxes 1a and 1b have been collapsed and are represented together in the connection between the left and the upper right of Figure 4, and that Figure 3 boxes 2a and 2b have been collapsed and are represented in the connection between the left and the lower right of Figure 4. The set of relations in Figure 3 box 3 have likewise been collapsed and now are represented vertically on the right side of Figure 4. The reformulation of the relations in Figure 3 into the schematic representation in Figure 4 is meant to show the results of the first process of iconic lamination just described, and further to allow for a subsequent analysis of the way this semiotic system as a whole dynamically produces the apparent reality of humans personifying spirits and vice versa.

49The performance of representatives/performers using talk, song, flute, dance, clothing, painting etc. is not mere performance. It is performative because it is productive of subjects and is potentially transformative action. I argue that what makes it performative is the mutual integration of iconic and indexical elements in this system of relations. On the one hand, the two indexical relations in Figure 4 are read as like one another and on the other hand the likeness in the iconic relation on the right of Figure 4 is read as spatiotemporally grounded and causal.

50Recognition of iconic similarity in the spirit-monster and representative/performer roles leads to interpretation of the relations between them and the victim/master as alike. Because they are iconically related, as depicted in the vertical relation on the right of Figure 4, they may be seen to enter together into the indexical relations depicted from left to right in Figure 4. The culmination of the Wauja ideological tendency to treat these roles and relations iconically results in the two indexical relations in Figure 4 collapsing into the same relation between the same roles. Indexical acts of exchange and speaking performed between people come to be interpreted as identical to and simultaneous with acts performed between people and spirit-monsters. This simply means that Wauja ideology treats exchange between a victim/master and representative/performer as the same as exchange between victim/master and pathogen/guardian. What is important to see is that it is a complex semiotic accomplishment that is produced through ritual.

  • 18  In line with the technical term rhematization to describe taking a sign as an icon, I term taking (...)

51Deictic pronominal person reference in particular grounds the iconic similarity between spirit-monster and representative/performer roles in the addresser and addressee frame and makes the similarity an actuality18. When the representative says « I » and « your » as a spirit this is interpretable as having the potential to transform the victim into a master. The iconic setup of people as -mona versions of spirits is fully leveraged by the indexical properties of person deixis to make this performance of spirits a performative statement by agentive spirits. The full performativity of bringing spirits and performance of celebratory dance later in the illness to ritual cycle depends upon the interplay between the iconic similarity of roles and the indexical creativity of relations between them.

  • 19  Much of the analysis in this section is reproduced from my dissertation (Ball 2007) where I formul (...)

52The victim/master’s exchange activities in indexing and thus maintaining his relations with representatives/performers simultaneously index and thus maintain his relations with spirit-monsters. In the case of celebratory ritual sponsorship, the master pleases the representatives through payment and through encouraging communal artistic and spiritual activity. Likewise he pleases the spirit-monsters through feeding them and making them visible and grand. The performers smoke tobacco, eat, and drink at the same time the spirit-monsters do. The converse follows from this, when a master fails to maintain an appropriate relation with either spirit-monsters or representatives he fails to maintain the appropriate relation with the other. Thus, when an owner does not feed his performers, the spirit-monsters suffer and may become displeased, which risks an undesirable mode of co-presence. Recall that this is the type of relation that produces sickness in the first place, and in such a case of poor sponsorship by a master the danger is that the indexical relation of master to guardian spirit performer might in fact revert to the initial indexical relation between victim and pathogen. This is a semiotic ideology within which a particularly Wauja notion of « Felicity’s Condition » applies to the act of bringing spirits. This ideology underwrites a frame in which persons speak as spirits in the same way that they smoke and dance as spirits. Speaking performatively as a spirit under the right circumstances is a reaffirmation of Wauja sanity and enacts a hoped for restoration of human and spirit subjectivities19.


53Boas (1940) reminded in 1902 that anthropological research on esoteric ritual in the indigenous societies of the Americas, valuable as it is, must be complemented by analytic attention to the exoteric doctrines from which esoterica emerges. This prefigures the later view in linguistic anthropology that ritual of the stereotypical sort is a by degrees maximization along a gradient of ritualization present in all social interaction. Even fleeting, apparently private « self-talk » such as uttered upon righting oneself after a stumble, or the basic dyadic achievement of shared correct reference between two partners in talk should be seen, not merely as thoroughly social, but specifically ritual(ized), as institutionally buttressed instances of practice that sketch as they invoke in the here and now shared conventional regimes of knowledge about the world. If this line of research has increasingly shown that daily interaction is more or less ritualized, then the complementary approach to explicit ritual forms might ask to what degree these are « interactionalized ». And what better place to investigate the by degrees creep of principles of interaction into ritual performance than in exoteric mini-rituals that fall out of prescriptive inclusion in the sphere of ordained sacred rites? These may seem for actors to be somehow more structured and thus available for reflexive engagement than everyday talk, while they share with the colloquially mundane a sense of consequential contrast to the high stakes of risky ritual cosmic manipulation. In the Wauja case, bringing spirits prefigures but is distinct from larger and more elaborate celebratory rituals. It is situated at a midway point on a gradient from quotidian interaction to song and dance accompanied by masks and musical instruments.

54The Wauja practice I have examined wherein lay people interact with victims of illness as spirits is exoteric in the sense that it is done by non-specialists and its ritual script is not proscribed knowledge. In distinction to the contingency of the curative success of the overall ritual illness process, Wauja describe this particular mini-rite as one that is not at risk of failure (Barcelos Neto 2004). Furthermore, the success of people standing-in for spirits is assumed throughout Wauja ritual, and it has been my goal here to explore this as a semiotic accomplishment. I have analyzed the verbal interchange between subject positions in terms of transpositions in voicing (Bakhtin 1981 ; Hanks 1990) and person deixis (Urban 1989). I utilize the concept of interactional frame (Bateson 1972 ; Goffman 1981), along with the semiotic principle of indexical iconicity (Peirce 1932) to describe ritual efficacy in terms of how Wauja people come to speak as spirits and how this is perceived as possible. At stake is the realization of social relations in a culturally specific universe and the reproduction of causal relations that underlie all such embedded action. Anthropological theorization of ritual requires an understanding of the micro-communicative economies that frame what can so often appear to be the seamless invocation of macro orders. Ritual is then seen as a key site of the practice of diagramming-while-creating interaction, a space where cultural ideologies of communication, and semiosis more broadly (Keane 2003), emerge and become institutionally validated.

55My analysis has taken a brief conversation between a victim and spirit representatives as an example not of ritualized interaction, but rather as an interactionalized ritual. The difference may seem trivial, but I hope to pull out some important consequences from this alternative characterization. These consequences exist at two levels of abstraction. First, and most obviously, interaction as a modifier of ritual allows us to see the plastic, goal oriented, and purposive aspects of rituals as social action. Second, by casting ritual moments as interactional partials, we may see how spatiotemporally disjunct ritual acts, such as people bringing spirits and the same people dancing the same spirits years later, may fit together into larger structures. Here I avoid accounts that might cast ritual sequence across seasonal permutations for example as mimetic representations of a natural or social order. This avoidance is based in the same rationale that rejects theoretical organization of discursive interaction at units above the individual utterance in structural models derived from syntax or other formal or grammatical linguistic bases. Instead, I appeal to the sui generis character of interaction as a collaborative co-construction of dialogic text, grounded in context and connected to other interactive texts through indexicality. Seeing ritual as interaction then, links ritual to the linguistic anthropological category of text, and treats the performance of ritual as entextualization, the connections between rituals and ritual moments as intertextuality.

56The Wauja practice of bringing spirits is successful as a curing ritual – not that it always cures, but that it always has the perceived ability to cure – because it is successful as interaction between humans and as interaction between humans and spirits. It counts as a possible cure because it counts as a conversational channel between humans and spirits. This is all to say that the ritual of curing embodied in bringing spirits is an example of how interaction works to reproduce the conditions of its very possibility, and that these conditions are in intimate connection with ritual structure. Recall that Goffman called this Felicity’s Condition, as he implied that interaction is ceremony and vice versa. In this Wauja case, the sanity, the utter presupposability of persons speaking as spirits is supported by an indexical iconic semiotic ideology in which likenesses take on causal power, and the actualization of those presuppositions in interaction becomes the causal reproduction of likenesses: ritual.

57Acknowledgements: The research upon which this article is based was supported by a Fulbright-Hays Dissertation fellowship and by Doolittle Harrison and Tinker travel grants from the University of Chicago. I thank my Wauja hosts for allowing me to share their lives. I owe thanks to Pierre Déléage for soliciting this piece and helping to guide its direction. I also thank two anonymous reviewers for the Journal de la Société des Américanistes for their insightful feedback. Any errors are of course my own.

Haut de page


Austin John L.
1962 How to do things with words, Harvard University Press, Cambridge.

Bakhtin Mikhail
1981 The dialogic imagination, University of Texas Press, Austin.

Ball Christopher
2007 Out of the park: trajectories of Wauja (Xingu Arawak), language and culture, PhD. dissertation, Linguistics and Anthropology, University of Chicago, Chicago.

2010 Take it as an index: dicentization in ritual, paper presented at annual meeting of the American Anthropological Association, New Orleans, 17 November [manuscript].

2011 « Inalienability in social relations: language, possession, and exchange in Amazonia », Language in Society, 40 (3), pp. 307-341.

Barcelos Neto Aristóteles
2004 Apapaatai: rituais de máscaras no Alto Xingu, tese de doutorado em Antropologia Social, Universidade de São Paulo, São Paulo.

2009 « The (de) animalization of objects: food offerings and subjectivization of masks and flutes among the Wauja of Southern Amazonia », in Fernando Santos Granero (ed.), The occult life of things: native Amazonian theories of materiality and personhood, University of Arizona Press, Tucson, pp. 128-151.

Basso Ellen
1973 The Kalapalo Indians of central Brazil, Holt, Reinhardt and Winston, New York.

Bateson Gregory
1972 Steps to an ecology of mind, University of Chicago Press, Chicago.

Bauman Richard
1996 « Transformations of the word in the production of a Mexican festival drama », in Michael Silverstein and Greg Urban (eds), Natural histories of discourse, University of Chicago Press, Chicago, pp. 301-327.

Benveniste Émile
1971 « The Nature of Pronouns », in Émile Benveniste, Problems in General Linguistics, University of Miami Press, Miami, pp. 217-222 [1956].

Boas Franz
1940 « The ethnological significance of esoteric doctrines », in Franz Boas, Race, language, and culture, University of Chicago Press, Chicago, pp. 312-315 [1902].

Butler Judith
1993 Bodies that matter: on the discursive limits of « sex », Routledge, New York.

Course Magnus
2009 « Why Mapuche sing », Journal of the Royal Anthropological Institute, 15 (2), pp. 295-313.

Fiorini Marcelo and Christopher Ball
2006 « Le commerce de la culture, la médecine rituelle et le Coca-cola », Gradhiva, 4, pp. 96-113.

Franchetto Bruna
2001 « Línguas e história no Alto Xingu »,
in Bruna Franchetto and Michael Heckenberger (eds), Os povos do Alto Xingu: história e cultura, Editora da Universidade Federal do Rio do Janeiro, Rio de Janeiro.

Franchetto Bruna and Michael Heckenberger (eds)
Os povos do Alto Xingu: história e cultura, Editora da Universidade Federal do Rio do Janeiro, Rio de Janeiro.

Gal Susan
2005 « Language ideologies compared: metaphors of public/private », Journal of Linguistic Anthropology, 15 (1), pp. 23-37.

Ginzburg Carlo
1989 « Clues: roots of an evidential paradigm », in Carlo Ginzburg, Clues, myths, and the historical method, Johns Hopkins University Press, Baltimore, pp. 96-125.

Goffman Erving
1959 The presentation of self in everyday life, Anchor Books, New York.

1981 « Footing », in Erving Goffman, Forms of talk, University of Pennsylvania Press, Philadelphia, pp. 124-159.

1983a « Felicity’s condition », American Journal of Sociology, 89 (1), pp. 1-53.

1983b « The interaction order », American Sociological Review, 48 (1), pp. 1-17.

Graham Laura
2003 Performing dreams: discourses of immortality among the Xavante of Central Brazil, University of Texas Press, Austin.

Gregor Thomas
1973 Mehinaku: the drama of daily life in a Brazilian Indian village, University of Chicago Press, Chicago.

Hanks William F.
1990 Referential practice: language and lived space among the Maya, University of Chicago Press, Chicago.

Keane Webb
1997 Signs of recognition: powers and hazards of representation in an Indonesian society, University of California Press, Berkeley.

2003 « Semiotics and the social analysis of material things », Language and Communication, 23 (3-4), pp. 409-425.

Kulick Don
2003 « No », Language and Communication, 23 (2), pp. 139-151.

Mello Maria Ignez Cruz
Iamurikuma: musica, mito e ritual entre os Wauja do Alto Xingu, tese de doutorado em Antropologia Social, Universidade Federal de Santa Catarina, Florianópolis.

Menezes Bastos Rafael José de
1989 « Exegeses Yawalapiti e Kamayurá da criação do parque indígena do Xingu e a invenção da saga dos irmãos Villas Bôas », Revista de Antropologia, 30-31-32, pp. 391-426.

Piedade Acacio Tadeu de Camargo
O canto do Kawoká: música, cosmologia e filosofia entre os Wauja do Alto Xingu, tese de doutorado, Universidade Federal de Santa Catarina, Florianópolis.

Peirce Charles S.
1885 « On the algebra of logic: a contribution to the philosophy of notation », The American Journal of Mathematics, 7, pp. 180-202.

1932 Collected works, Harvard Belknap Press, Cambridge.

Sherzer Joel and Greg Urban (eds)
1986 Native South American discourse, Mouton de Gruyter, Berlin.

Silverstein Michael
1976 « Shifters, verbal categories and cultural description », in Keith Basso and Henry Selby (eds), Meaning in anthropology, School of American Research Press, Albuquerque, pp. 11-56.

2005 « Axes of evals: token vs. type interdiscursivity », Journal of Linguistic Anthropology, 15 (1), pp. 6-21.

Urban Greg
1989 « The “I” of discourse », in Benjamin Lee and Greg Urban (eds), Semiotics, self, and society, Mouton de Gruyter, coll. « Approaches to semiotics » 84, Berlin, pp. 26-51.

1991 A discourse-centered approach to culture: native South American myths and rituals, University of Texas Press, Austin.

1996 Metaphysical community: the interplay of the senses and the intellect, University of Texas Press, Austin.

Viveiros de Castro Eduardo
2002 A inconstância da alma selvagem, Cosac & Naify, São Paulo.

Haut de page


1  I use the hyphenated term « spirit-monsters » to highlight the supernaturally dangerous aspect of these beings in Wauja and Upper Xinguan cosmology (see also Barcelos Neto 2004; Basso 1973). I also use the term « spirit » as a shorthand for this.

2  Goffman (1983b) theorized this as « the interaction order ». I use the term « dialogicity » and the adjectival form « dialogic » in this paper to include not only physical co-presence in dialogue, but also the more general process of social negotiation of multiple perspectives or voices (Bakhtin 1981).

3  A challenge by Pierre Déléage to critique ritual efficacy in concert with emphasis on the interactional aspects of ritual aided this formulation.

4  In this treatment of Wauja ritual practices of healing I do not have space to attend to two important questions related to the fact that the Wauja are members of Upper Xinguan society. First I do not discuss further how these practices and ideologies compare to and differ from other Upper Xinguan groups with similar shamanic and ritual systems (Basso 1973). Second, I do not address the important part of Upper Xinguan ritual practice that regulates, not only human and spirit relations, but relations between Upper Xinguan groups. Elsewhere I address Wauja ritual practice as oriented towards alters in general, be they spirits, Upper Xinguan friends and rivals, or Whites (Ball 2007).

5  Women can in principle be shamans, but I know of no historical cases of female shamans in the Wauja context. Also, as is the case with Wauja public ceremonial dancing and musical performance, women tend to represent many fewer kinds of spirit-monsters than men, centering principally on the spirit named Yamurikuma. For a detailed analysis of the Yamurikuma complex, see Mello (2005).

6  I treat connections between Wauja grammatical marking of (in)alienable possession and Upper Xinguan ideologies of personhood, collectivity, and exchange in a forthcoming article (Ball 2011).

7  See Viveiros de Castro (2002) for a description of the related Yawalapiti modifier.

8  In addition, kaukitsupa, the term for victim is based on the stem kau « pain, illness, or fear » plus the nominalizer -ki and a classifier -tsupa. Some informants suggested to me that kau as a term for pain, illness, or fear is a contraction of kawo, and in turn of fearsome and potentially harmful kawoka « guardian spirits », making an explicit lexical connection between the state of victimhood and contact with the spirit-monster. Though these terms may be diachronically or even derivationally related, I have no more formal evidence than homophony and speculation.

9  It is a stylized version of a general cultural permissiveness with regard to volume and distance of acceptable hailing of potential interlocutors. That is to say that Upper Xinguans have less reservation about hailing consociates or even strangers in a manner that members of Euro-American society might deem more appropriate for vehicles, as in calling out « Taxi! » or perhaps calling animals such as pigs, for example with the cry « sooie! ». I was often surprised to be yelled at from large distances, summoned and expected to drop everything and proceed to engage the caller, not in any urgent matter, but often in simple chit-chat.

10  Men are much more likely to be frequent kawoka-mona than women, in part because the spirit pantheon is mostly male, but anytime Yamurikuma or a few other exclusively female spirits is involved, women will be called.

11  The third trichotomy of signs in Peirce’s theory includes the possibility that the interpretant of a sign’s relation to its object may represent it to be other than it actually is and he gives names to such interpretations such that the technical term for an index taken as an icon would be a « rhematic index ».

12  The example of symptoms as indexes of disease is a classic example in the semiotic literature on indexicality. See Peirce’s (1885) statement that in indexicality « the sign signifies the object solely by virtue of being really connected with it. Of this nature are all natural signs and physical symptoms ». See also historian Ginzburg’s (1989) essay on the developmental relationship of hunting, divination, Hippocratic medicine, history and the interpretive (social) sciences to the principle of generalizing to the unobservable through indexical sign (including symptom) interpretation in the individual case.

13  The passing of role inhabitance prerogatives can occur, especially when an owner keeps the right to celebrate his guardian spirit-monsters for many years and the individuals who served as representatives during his illness have since grown old or died, requiring able replacements.

14  See Barcelos Neto (2004, p. 142 et passim; 2009) for an alternative application of Viveiros de Castro’s analysis of these modifiers, where the emphasis is placed on the distinction between -kuma and -mona as representing « spirit » versus « body » respectively.

15  Simulacra and diagrams are icons in that they both represent their objects through similarity in form. In the case of a diagram the likeness is in the formal, possibly dynamic, arrangement of constituents.

16  Consider some distinguishing characteristics of the social role of representative in comparison to that of the shaman, in particular that the role of representative is defined by access to largely aesthetic knowledge related to the practices involved in the manufacture, design and use in music and dance of ritual paraphernalia, and that anyone, even children, can perform in this temporary capacity, as opposed to access to the relatively permanent and restricted class of esoteric knowledge that defines the role of shaman (Barcelos Neto 2004, pp. 144-146). While shamans are « translators/mediators/negotiators », the kawoka-mona are what Barcelos Neto calls in Portuguese « presentificadores », what we might render as « presentators » in English.

17  In Peircean terminology this should technically be referred to as rhematization (Gal 2005).

18  In line with the technical term rhematization to describe taking a sign as an icon, I term taking a sign as an index to be dicentization. The terms rheme and dicent come from Peirce’s third trichotomy. For simplicity I refer to these processes as iconization and indexicalization here. I develop the comparative semiotics of dicentization in ritual elsewhere (Ball 2010).

19  Much of the analysis in this section is reproduced from my dissertation (Ball 2007) where I formulated it to analyze transformations of semiotic ritual function in a Wauja performance of the Atujuwa mask dance in France. The basic frame for understanding how Wauja representatives and performers channel spirits remains the same in this analysis, and so I apply it here. In the case of ritual show abroad I claim that the indexical connections grounding the embodiment of spirits in the here and now are potentially de-linked, allowing for some performers to claim that cosmological risk of ritual failure is irrelevant to mere show. Other aspects of the same Wauja performance in France are described in Fiorini and Ball (2006).

Haut de page

Table des illustrations

Légende Fig. 2 – The representatives proceed from the village center to the ill person’s house. The leader sings while some of the younger men imitate monkey vocalizations and laugh.
Fichier image/png, 356k
Légende Fig. 4 – Dynamic system of relations in illness-to-ritual process.
Fichier image/png, 11k
Haut de page

Pour citer cet article

Référence papier

Journal de la Société des Américanistes, 2011, 97-1, pp. 87-117

Référence électronique

Christopher Ball, « As Spirits Speak: Interaction in Wauja Exoteric Ritual », Journal de la Société des américanistes [En ligne], 97-1 | 2011, mis en ligne le 10 décembre 2014, consulté le 07 décembre 2021. URL : ; DOI :

Haut de page


Christopher Ball

McKennan Postdoctoral Fellow, Dartmouth College, Department of Anthropology, HB 6047, Hanover, NH 03755 []

Haut de page

Droits d’auteur

© Société des Américanistes

Haut de page
  • Logo Latindex
  • Logo Centre National du Livre
  • Logo CNRS - Institut des sciences humaines et sociales
  • OpenEdition Journals
Search OpenEdition Search

You will be redirected to OpenEdition Search