Navigation – Plan du site

AccueilNuméros91Varia“The Voice of a Child”: Construct...


The Voice of a Child”: Constructing a Moral Community through Retteng Poetic Argumentation in Toraja

« La voix de l’enfant » : la construction d’une communauté morale à travers le genre retteng, une forme de poésie argumentative toraja
Stanislaus Sandarupa
p. 231-258


Le sujet de cet article concerne le retteng, un genre de poésie rituelle argumentative. Il intéressera les ethnographes et les chercheurs en anthropologie linguistique car il constitue un type d’argumentation verbale hautement dialogique. Par le biais de citations et de métaphores élaborées, il permet aux orateurs de commenter, d’évaluer et de façonner l’exécution du rituel tel qu’il se déroule, et par là-même, de (ré)affirmer les valeurs morales de la société. Après une description du genre et des milieux dans lesquels ce genre poétique est produit, l’analyse porte sur un exemple observé dans la région sud du pays Toraja (Sulawesi, Indonésie) pour finalement révéler les implications plus larges de ces pratiques. L’approche relève de l’anthropologie linguistique et en particulier, de la sémiotique fonctionnelle, ce qui permet de considérer les données comme des « textes dénotationnels » (ce qui est dit), des « textes interactionnels » (ce qui se passe socialement) et des « textes médiationnels » (comment les deux sont reliés de manière indexée). Une telle approche permet de comprendre comment ce que les orateurs disent (par métaphores et allusions) est mieux compris si on observe ce que cela fait sur le plan de la communication et socialement (défier l’autorité d’un orateur, argumenter sur ce qui est vrai, médiatiser un conflit avant qu’il ne s’envenime). Je montre comment les locuteurs utilisent divers dispositifs linguistiques culturels et stratégiques (par exemple, citation de discours, contextualisation et entextualization, jeu phonologique et métaphores) pour faire évoluer leur argumentation poétique, afin d’indiquer le sujet du débat. Une analyse approfondie de l’exécution du retteng montre que le sujet débattu porte sur « la voix de l’enfant », une manière de parler de l’expertise et de l’autorité. En fin de compte, la « voix de l’enfant » fait référence à plusieurs choses : tout d’abord, à « l’honnêteté » et à « la capacité de connecter le sentiment de quelqu’un à ce qu’il dit », « la capacité d’exprimer sa connaissance par l’art de la parole ». Enfin, « la voix de l’enfant » réfère à une société bonne et éthique qui permet à ses membres de parler efficacement de problèmes sociaux par le biais de la parole rituelle.

Haut de page

Notes de la rédaction

At the time we were preparing this article for publication, Archipel’s board was greatly saddened to learn of the unexpected death of Stanislaus Sandarupa, on January 17, 2016, in Makassar. Only 59 years old, he had recently been appointed Professor (Guru Besar) at his university. His long-term collaborations with national and international colleagues were exceptionally generous and fruitful, and he was an equally generous and inspiring mentor for his students. Having defended his thesis in linguistic anthropology at the University of Chicago in 2004 (Sandarupa 2004), he contributed a new perspective on Toraja speech, by analyzing poetry in its social and linguistic contexts. The final version presented here was edited by Elizabeth Coville and Kathleen Adams.

This paper develops ideas first presented in my dissertation (2004) supervised by Michael Silverstein, the late Valeri Valerio, and Danilyn Rutherford, Department of Anthropology, The University of Chicago. Both Kathleen Adams and Elizabeth Coville were also members of the dissertation committee. Versions of this project have been presented at various conferences, including the Third International Conference on Language Education at Menara Phinisi UNM Makassar (6-7 December 2013). That conference paper was subsequently included in the Proceedings of the Conference. Herewith, it is reanalyzed for the purpose of this publication.

Texte intégral

I wish to thank Kathleen Adams, Elizabeth Coville and Dana Rappoport, for their valuable comments, suggestions, revisions, and evaluations, and the editor of Archipel, Daniel Perret, and other reviewers for their most helpful suggestions, herewith, incorporated in this article.


1This article addresses the use of poetry in ritual and sociopolitical life in an Indonesian hinterland society. As has been well documented, Austronesian societies are known for their elaborate poetic forms and draw on poetry in ritual and on formal occasions as well as in daily interaction (Sweeney 1987, Fox ed. 1988, Bowen 1991). These practices are not unique to the Austronesian world. For instance, Lila Abu Lughod (1988) has demonstrated Bedouin poetry’s association with honor as well as how Bedouins use poetry in interactions to covertly and overtly express various sentiments. Likewise, Steven Caton (1990) has shown the centrality of poetry to the sociopolitical and cultural system in Yemen.The use of poetry is similarly central to the Torajan ritual and sociopolitical system, which is the focus of this article. This article uses a functional semiotics approach to shed light on the sociocultural dynamics of Torajan ritual poetics. The present study is in line with the recent ethnographic work on Toraja that considers art as politics (Adams 2006). In short, this article addresses the relationship between poetics and politics.

Research methodology

2The following data are drawn from two years of field work (1994-1996) in the villages of Randan Batu, in northern Toraja, and Balik, in southern Toraja, using the method of participant observation and employing the techniques of field notes, interviews, audiorecording, and videotaping. The interviews conducted were informal. The interviewees consisted of the three men involved in the retteng performance (labeled A, B, and C) and three traditional religious priests or tominaa (Pong Jen, Ne’ Rimma’ and Tato’ Dena’) with whom I discussed the performance after it was completed. These particular tominaa were known to have extensive knowledge and to be well-versed in interpreting ritual speech. I also interviewed other village elders, including Pabisa and Luther Bala. As someone of Torajan heritage and a community member, I sometimes take things for granted. With the help of these experts, I gained a richer understanding of the retteng performance and was ultimately able to access the hidden meanings of retteng poems and the social relationships to which they refer. In this fashion, as a scholar, I was able to obtain a more objective understanding of “what was going on,” enabling me to present the local perspectives.

3In analyzing these data, I focus on the interactive quality of the poetic argumentation. In the performance, the first singer recites a poem directed to an addressee without giving any explicit signal or mentioning the addressee’s name. Moreover, the poem is composed in metaphoric riddle form, thereby concealing the poem’s target. As an interaction that requires a response, one puzzling thing strikes the observer: given the ambiguity of the situation, how does someone in the group decide to become a respondent and take up the role of speaker? What indexical cues are used to infer that the poem is directed to him? And furthermore, how does the ritual speaker turning the role over to a respondent exert an effect on the community?

  • 5 This is not to deny that a given performance is related to a number of other speech events that pre (...)

4These are some of the themes that will be developed here. Although there are different ways to analyze poetry, in this article, I find it useful to understand it as a phenomenon of textual performance. Understanding it in this way requires striking a balance between text and context. Bauman and Briggs (1990) have nicely summarized this problem, citing on the one hand scholars who say that “performance studies seem too much concerned with context and too little concerned with textual detail” and, on the other hand, those who argue that “performance approaches are too caught up in poetics to be able to discern broader social and political contexts” (Bauman and Briggs 1990: 67). However both views separate text and context, treating them as self-contained and bounded objects, whereas I draw on ethnographic data pertaining to retteng performances in order to highlight the interconnectedness between text and context. By adopting the general view that speaking (including composing poetry in performance) involves how one sign points to “the spatial, temporal, or causal co-presence of another [sign]” (Silverstein 2001:73), the analysis begins with linguistic data. In other words, different from the approach that starts from context that affects the selection of poem to be recited in interaction (Abu-Lughod 1988), this approach begins with textual analysis of language.5 As Bauman and Briggs (1990:68) say “communicative contexts are not dictated by the social and the physical environment but emerge in negotiations between participants in social interaction.”

5My analysis focuses on extracting the “texts” from this cohesive poetry. Using the perspective of functional semiotics, this whole poetic discourse constitutes an “ensemble of texts” (see Geertz 1973) that further need specification. In order to show the richness of Torajan culture, I follow Silverstein in making a distinction of three levels of text: the denotational text (saying something), the interactional text (what’s happening socially) and the meditational text (how the structure of cohesive signs indexes the interactional text) (see Silverstein 1976, 1992, 1993).

6In the performance, speakers display the art of speaking both to each other and to the audience. Poetic discourse illustrates what Roman Jakobson (1960) calls poetic function, or message that is focused on its own form, a form characterized in this case by parallelism. Like all systems of signification, poetry can also become an object of metalinguistic focus. Thus such artful acts of speaking can be objectified, scrutinized, and evaluated by others. This allows the performers to create, comment on, and discuss the ongoing discourse. This evaluation relates to rhetorical strategies and the construction of power in performance. In considering the Toraja retteng, it is also important to note the dialectic relation between what linguists term contextualization and entextualization. Contextualization allows a performer to anchor the performance in context and in so doing, the poetic discourse is brought closer to social reality. It becomes an index of its situational surroundings, including participants’ competence and the relation of performance to other events. The reverse of this is entextualization, in which power is built by taking a unit of text from others’ authoritative voice through quotation to support the functional goal of the performance.

7A key aim here is to extract the cohesive and coherent topic or central idea being debated during the interaction. In this particular retteng below, the topic is “the voice of a child” which emerges as a structure of cohesive signs.
It emerges as “text” that indexes the “context” of how social relations among the interlocutors are transformed, how they construct social boundaries, and how they identify relevant voices. Some utterances are associated with certain speakers, an association which is related to Bakhtin’s voice (Bakhtin 1981 [1935]).

8This article is organized as follows: I begin by examining the elements of ritual speech/poetry. I then focus on the genre of retteng and offer one specific illustrative example. Finally, I offer a semiotic interpretation, showing how this theoretical approach offers rich insights into the relationship of text and context, thus deepening our understanding of Torajan culture as an “ensemble of texts.”

Research Setting

  • 6 It is estimated that an additional 2 million Torajans live outside Toraja, spread across the big is (...)
  • 7 This hierarchical division derives from the south Toraja regency, which is slightly different from (...)

9Located in the northern interior of South Sulawesi province, the Toraja homeland has a population of approximately 600,000 people.6 Torajans are traditionally wet-rice farmers with an elaborate ritual system, and within Indonesia they are famous for their funeral rituals. As has been reported by other ethnographers, Torajan society is hierarchically arranged into several ranks, the highest being referred to as “gold stake” (tana’ bulaan), middle as “iron stake” (tana’ bassi), commoners as “skin of palm-tree stake” (tana’ karurung), and the lowest rank as “reed plant stake” (Nooy-Palm 1979, 1986, Volkman 1985, Waterson 2009).7

10Toraja culture is rich with ritual and ritual speech. In general Torajans classify speech style into two types: “straight talk” (kada-kada dipamalolo) is used in daily interaction and “paired words” (kada-kada dipasilopak) are used in ritual. This latter style of speaking uses metaphor, metonymy, indirection, and parallelism. Parallelism refers to poetic verbal constructions that correspond to one another in sound, meter, meaning and grammatical structure. Many ethnographers have analyzed and documented the widespread use of this poetic ritual speech (Coville 1988, Rappoport 2009, Koubi 1982, Lebang 2006, Nooy-Palm 1979, Nooy-Palm 1986, Sandarupa 1989, 2004, Sandarupa, Assagaf, and Husain 2015, Sarira 2000, Veen 1965, Veen 1966, Zerner and Volkman 1988).

11It is in ritual contexts that speaking expertise is displayed by skilled speakers, those known as the “ones who have extensive knowledge” (tominaa). When tominaa perform their poetic speech in various rituals, they adjust their words in accordance with the rank of the person being mourned (for funerals) and the level of ritual being enacted. Various Toraja rituals are enacted in pairs and in sequences. For instance, death-rituals (rambu solo’) and life-rituals (rambu tuka’) are sequentially arranged from lowest to highest level rituals, as various ethnographers have detailed (Nooy-Palm 1986, Sandarupa 2012). One feature that differentiates these hierarchical performances is the use of various genres of ritual speech or sung poetry. Rappoport (2009) has shown that the paired sung poetry used in the paired rituals is systematic. In general, we can say that the higher the order of rituals, the more elaborate and varied the poetry used. For example, Torajan ritual speakers use less elaborate poetic verse in middle-ranking (“seven nights” pitung bongi) than in high-ranking (rapasan) funerals. The most elaborate rituals employ the most distinctive and elaborate poetry.

  • 8 At the completion of a death ritual, it is believed that the “dark spirit” is transformed into a go (...)
  • 9 The expression “to floor the earth” is a metaphor which means to divide and distribute pieces of bu (...)
  • 10 The word tongkonan refers to Torajans’ ranked ancestral houses with arched horn-shaped roofs. Some (...)

12Despite many recent changes in Torajan society, today’s funerals continue to be key events that are pregnant with poetry. As illustration, consider the case of the death of a chief: his demise risks causing competition, political struggles or even physical fights between his potential replacements. The funeral ritual dramatizes this “war” (rari) characterized by tension, struggle, and contestation between the realm of life represented by the living and the realm of darkness represented by the black spirit of the deceased (bombo) which is being transformed by the funeral.8 The elaborate funerals, then, function to transform the potential real fight into a symbolic fight. One can observe this in the structure of the performance of poetic speech which follows the Torajan cultural logic of movement from unity, to chaos, back to unity. The high level funeral begins with “laments” (badong) sung by a group of men and their leader. These laments are followed by “poetic argumentation” (retteng). Finally, the ritual ends with a rite referred to as to “flooring the earth” (massali padang),9 which entails physically and metaphorically reconstructing the values of unity and orderliness (Sandarupa 2004). Through his monologue, the tominaa addresses members of a village tongkonan10 with a eulogy for the deceased. He then distributes pieces of buffalo meat to them according to rank. In so doing, the tominaa encourages listeners to recover from sadness and return to their normal social roles and positions. People receive different parts of meat but from the same buffalo. Parts of the meat, then, metaphorically represent individuals and the pieces in totality represent the whole of society.

The Retteng Performance

13While the final rite is performed as a monologue by a single ritual speaker, the preceding two are performed interactively, although in contrasting ways. In the badong laments, performers hold hands and follow the leader’s cues, developing their many voices into one single voice, thereby marking the value of unity (Rappoport 2009, Sandarupa 2004, Sandarupa 2013, Veen 1966). In the retteng, which follows or emerges from the badong sung laments, in contrast, one speaker composes a poem which elicits a response from a second speaker who creates another poem in opposition (bali). The retteng performance has not been extensively analyzed, despite the fact that various ethnographers have recorded or mentioned these performances (Rappoport 2009, Sandarupa 2004, Tammu & Veen 1972). Veen (1966) includes twelve retteng texts, ten of which are from funerals. In general, his collections of funeral retteng can be classified as retteng malolo lako to mate “poems directed to the deceased” recited by a single speaker. Even though he includes one retteng that constitutes a reply to another, he does not show the interactive quality of the performance. My data, on the other hand, show the dynamics and the interactive quality of the performance and facilitate a better understanding of how Torajans use poetic argumentation in symbolic battles in an effort to resolve sociopolitical conflicts and construct a moral community.

  • 11 In the northern and western regions of Toraja, the word retteng is used in life rituals. Veen inclu (...)
  • 12 Nowadays, we regularly witness the decontextualization and recontextualization of retteng. For exam (...)
  • 13 Veen (1966) explains the context in which they are used and emphasizes the improvised character and (...)

14In general, retteng can be classified as life ritual-retteng or funeral retteng. For the latter, the word retteng is associated with the expression of sad feelings and lamentations.11 Since the ritual is interactive in form, involving performers who argue with each other, the performance is called siretteng where /si-/ is a marker of reciprocal action. As I previously documented (Sandarupa 2004: 229-235), the retteng, “poetic argumentation” may be performed as a part of the badong round dance ritual.12 Thus, all badong round dance performers potentially become participants in the retteng.13

  • 14 Such as “thank you, you’re welcome” or “how are you ? fine and you ?”
  • 15 These norms are captured in the saying tae’nama’din untossok mata bale (“it is forbidden to pierce (...)

15There are several formal characteristics associated with this genre.
The participants are all men of varying ranks (high and low ranking), both young and old. The most significant feature is that, except in some special cases, prior to the performance no one knows who will be participants, and who will be the initial speaker. Even if the performance begins and someone embraces the role of the initial speaker, it is not known publicly to whom the poetic argumentation is addressed. Unlike interaction characterized by an adjacency pair structure
14 in which the speaker makes the addressee explicit, in this ritual the addressee emerges via self-selection. According to the local view, this is due to his “awareness” (nasa’dingan) and to experiencing the feeling of being “hit” (nakanna) by the words of the initial speaker. The first chanter may or may not have a particular addressee in mind. The target may be revealed as the performance unfolds without any explicit reference to him. In this performance, avoiding or at least concealing direct reference to an addressee is a part of the strategy presupposed by norms prohibiting direct reference to the targeted individual.15

  • 16 See also Veen (1966) about the use of the stick, which he calls bandangan.

16The ritual requires at least three participants (pa’retteng “singer of retteng”), and three individuals are considered the ideal number. The first participant is called pa’retteng, “the singer of retteng.” The person who voluntarily initiates the interaction is called to ma’bungka’, “the person who opens,” and the next self-selecting speaker, hereafter, the respondent is to umbali, “person who answers.” In the case where the debate becomes heated, a third person mediates between the first and the second and is referred to as to ussamboi retteng, “the person who covers the poetic argumentation.” When someone starts the retteng, he picks up a stick (lidi), sometimes tipped with goat hair, to signal the beginning of his poetic performance.16 This is usually accompanied by the expression le, le, le, “here I am, here I am” to get the attention of others. When he ends his performance he uses the same expression, but in this case it is an invitation to someone to respond (see below). These features underscore the interactive quality of performance.

  • 17 In introducing the retteng he collected, Veen (1966) lists a variety of different “subjects” : “The (...)

17Informants agree that from the point of view of content, retteng can be divided into “retteng that is addressed to the deceased” (retteng malolo lako to mate), which is sometimes described as a eulogy, and “retteng that generates an opposed view” (retteng sindung). As we will see later, even though the former is said to be an expression of grief for the deceased, it always touches on sociopolitical and cultural problems as well. Hence, they can both be classed as retteng sindung.17 The term sindung refers to “a deep and wide hole” (Tammu and Veen 1972), and the verb massindung means to dig a hole which becomes wider the deeper it goes, thus while it gives the impression of being narrow, in fact it is large inside. This metaphor beautifully compares the performing of retteng debate to constructing a pit and a tomb for the deceased. In retteng performance, the chanters are always advised not to perform retteng sindung because it contains severe criticism. As will be shown below, people who become participants in this ritual bring larger social and political problems that have arisen in village life into the performance. While the ritual leader always suggests that participants direct their poetic speech to the deceased as a eulogy, it is the dialogue between several interlocutors about the event at hand and other aspects of daily life that dominates. As a result the retteng may develop into a heated debate that takes the form of symbolic violence through verbal display, which may lead to real fighting.

  • 18 Here the speaker addresses the local noble chiefs who are metaphorically compared with kayu kalando(...)

18The individual retteng has the following general four-part structure: the opening which greets everyone present (mekatabe’); the disclaimer (called “the fencing of the neck”, mebala kollong), the main content (lise’na) and the closing (dipalele). Openings, disclaimers, and closing all constitute framing devices. Although the opening varies according to contextual features such as who the participants are, their different ranks and so forth, it is a formulaic greeting to everyone present, a polite way of asking permission to talk before the public. For example: 18

  • 19 A related idea is pemali ullutu tombang panda dibolong “it is taboo to create chaos in the funeral (...)
  • 20 The term pia, “child” and its synonym baitti’, “small one” also occurs in one of Veen’s (1966, 85) (...)

19Like the opening, the disclaimer (mebala kollong, “fencing the neck”) that follows is also formulaic. Its purpose is to protect the speaker from any dangerous effects of his words, such as curses (tula).19 A typical example is: 20 The closing of the retteng is also prescribed. In ending his retteng the chanter sometimes repeats the last line as a cue to the audience to say the final line again with him. This also indicates that he has finished. As mentioned above, it is then followed by the repeated syllable le… le… le… as an abbreviation of dipalele meaning “the turn to speak is now moved to another.”

  • 21 One of the late author’s important contributions to the study of ritual speech was to see that para (...)

20Thus we see that unlike the first phase of the funeral (i.e., badong round dance) and unlike the last phase (i.e., “the flooring of the earth”), this optional genre is interactive in the sense that it is crucial that different individuals take on the roles of speaker and respondent. The concept of dipalele reminds us that the genre depends on responses which are not fixed ahead of time. This is reflected in the word translated as respondent (to umbali “person who answers”), which is based on the same word as the “pair” of a verse of parallel lines. Thus we see that the idea of parallelism is not only relevant to an understanding of the denotational text but also is a key feature of the interactive text.21

21At the same time, this ritual speech practice is strongly characterized by concealment and ambiguity at various linguistic levels from the phonological to the pragmatic. One example of this is phonological transformation. It is also poetically structured and dominated by metapragmatic talk, since it focuses and comments on verbal use itself. As a poetic genre, rhythm, meter, and the use of key tropes such as metaphors and metonyms also pervade this ritual
speech practice. Game-playing and contestation, tension and expression of anger are other major characteristics that occur at various linguistic levels, from the phonological to the pragmatic.

  • 22 The sung poetic text comprises poetically compact discourse expressed in parallel lines. The rule i (...)

22In this ritual, the speaker’s ability to create such ambiguous speech is itself an index of his skill as a speaker and his stature as a leader, and it is precisely this that threatens and challenges the respondent to defend his own honor by using his right to speak and answer, while exercising his ability to create a “pairing” to the previous speaker’s poetry. The assumption of the right to occupy the respondent role is itself a dangerous tactic because the first speaker is never explicit about the intended addressee. The respondent’s decision to embrace the role and his creation of a fitting pair (bali) will be perceived as an index of his own ability to interpret ambiguous signs, a dangerous game. One kind of concealment is a marked form of phonological play with words.22 In addition, there is also some morphosyntactic ambiguity in which the speaker’s use of unknown lexical items and special morphosyntactic construction prevent the clear reading of lines, or allows multiple readings. This is so because the lines employ figures of speech such as grammatical parallelism, metaphor, metonymy, and irony.

  • 23 The performance of retteng is the creative work of adults and the elderly. Indeed young people and (...)

23In short, to be a participant in a retteng performance, one must have the ability to chant, must have mastered the complex knowledge entailed in creating paired lines, and must be able to interpret the most ambiguous signs. Because of these characteristics, the exegetical aspect is crucial, and we can say that within Toraja culture it is in this kind of performance that the tradition of interpretation or exegesis is elaborated. A speaker’s ability to read his interlocutor’s signs is crucial, as misunderstanding may lead the interpreter to fall into a dangerous pit dug by the initial speaker (massindung). The whole game here is the ability to create and interpret another speaker’s signs. In this respect, retteng is similar to the genre of riddles (karrume)23. According to local belief, if one becomes recognized as adept at this art via multiple performances in public events, he may enjoy a rise in social rank. He can be promoted from ordinary retteng chanter (pa’retteng) to a higher rank as an authoritative speaker known as “sitting throat” (gora-gora tongkon), who is a decision-maker in conflict resolution, or to a ritual priest, tominaa, who can invoke the flow of fertility and prevent death.

24The following siretteng - poetic argumentation shows clearly both the interactive quality of the performance and the way in which denotational meaning is concealed through the use of metaphors:

Performer, Singer or Speaker A

eee .....eeee.....eeee


(1) tabe’ kupadolo lam-amban siman mata lam-alan le eeee lako

(1) excuse me, to those who are present pardon me to

(2) tounno’ko’ massali am-alang tasikma’kambuno sam-aane’ le leleeee

(2) those sitting on the rice barn floor, those sitting under the protruded ceiling of the rice barn

(3) tang marendengna’ manim-indi tang ganna’ massai lam-aanangko leeee...

(3) just in case I am not protected and do not live for a long time

(4) bendanna’ te umbating ke’de’na’ te ma’rio-rim-rio le eeeeeee

(4) standing up here I lament; rising up here I express this mourning

(5) nakua aku batim iingku pangngo’tonan mariom... ooku leee... leee...

(5) my lament says my mourning says

(6) unno’ko’na’ membungku’ sam-ambu’ leee bintin massalungku’ luam aangan leee

(6) I was sitting, wearing a sarong, sitting, covered by a blanket

(7) denri pia’-pia’ len-eendu’ baitti’ untuleram-aana’ leee

(7) it is only because of a child who passed by, a small one, who told me a story, again

(8) to umpennampa’ duri bam-anga to umpennallon duu-uuri leee

(8) someone using palm tree thorns as his mat, someone using thorns as a pillow

(9) tang malanapa ampa’- aa’na tang solongpa allonam-

(9) his mat is still not (yet) smooth his pillow is still not (yet) flat

(10) tang solongpa allonanna ole

(10) his pillow is still not (yet) flat

Performer, Singer or Speaker B

em lee


(11) langngan kutongan tom-
ooda’ ee
endek kutundu malem-
eeso lee...

(11) I assert it as true

I make it clearly

(12) kendek masapi medum-
uuku’ eee
puarang mantaa pam-
adang leee....

(12) the big eel is rising out of the water
to search for meat
the big lizard is dividing
the meat

(13) pasapu-sapu ikko’
oo’na pasulilang saresem-
eena lelele

(13) its tail is striking us
its sharp dorsal fins are piercing

(14) ulunna pakadakean
ulunna pakadakean

(14) his head is causing damage
their heads are causing damage

Performer, Singer or Speaker A

em lee


(15) buda londong lan te tom-
saungan lan te panglim-

(15) many are the roosters in the village

fighting cocks in this hamlet

(16) sisonda-sonda unnom-
ooni sisolon ma’kua kum-

(16) crowing one after another
interchangeably producing sound

(17) uua lee...eee.. iaku te akum-

tu kale misa-misam-
angku lee

(17) leee ... with regards to me

that myself alone

(18) indara manarang pam-

aandetu tau rangga inam-
aaya leee...

18) who on earth is clever like an expert?
who on earth has such an extensive knowledge?

(19) ungkitta’ simpona mam-
aanuk tangke isinna pa’kum-

(19) can see the missing teeth of the chicken
the branches of the rooster’s teeth

(20) ubanna koro’-koro’
ubanna koro’-koro’

(20) the gray hairs of egrets
the gray hairs of egrets

Performer, Singer or Speaker C



(21) kita mentu’ te to tom-
oongkon mairi’ ma’rio-rim-

(21) all of us who come to mourn

all who are lamenting

(22) tontong meloi bam-
tadende’ maya-mayam-
aai lee lee

(22) make the lament good

make mourning in a good way

(23) dendika lindona sem-
le’ke’ri pa’todingan-
aanta le leeee

(23) are there other faces?

different marks?

(24) den o upa’ tapoum-
paraya tapoparam-
ayana le..eee...eee

(24) hopefully we will have blessings

the favor be with us

(25) tamasakke’solanam-
madadinding sola mentu’

(25) so that we will all be cool

all will be well protected

(26) madadinding sola mentu’

(26) all will be well protected

25Three people are involved in the above poetic interaction, henceforth labeled A, B, and C. Combined together, their poetic totality is a discourse that is divided into four segments: A1, B, A2, and C. Speaker A (a to minaa priest) begins his retteng, which is then self-selectively answered by B (an elder, but not a to minaa). A answers back and C intervenes and mediates the tension between A and B, ending the poetic argumentation (sambo retteng). Here is the synopsis of the text. Speaker A makes an indirect reference to a problem of some kind (i.e., “thorns”). But he doesn’t just use the metaphor, he also quotes the speech of a child, herewith the voice of a child. Speaker B (not a tominaa priest, but a village elder) responds with a metaphor about destructive poisonous eels and lizards. Here B denies that A’s words are “the words of a child” and insists that instead they are A’s own words and that they are destructive. Then A says there is too much conflict among experts (using the metaphor of roosters) and claims that B’s so-called knowledge is false. Then speaker C (a senior tominaa priest or indo’ tondok, “mother of the village”) steps in and brings metaphors of coolness to counteract the hot retteng of A and B. The retteng has come to center on the question of what constitutes true knowledge, which, as we will see, is linked to the “voice of a child.”

26As a denotational text, this four-part retteng does not display a transparent meaning. We have a series of metaphors – e.g., thorny bedding, flailing eels, fighting cocks, toothless chickens – whose very meanings let alone their social significance are obscure. To understand what is being said, however, we need to see the topic that emerges in the course of the performance and the fashion in which it emerges in and through the interaction among the three speakers. In what follows I detail actions the performers are engaged in and show the interactional order of poetic argumentation. Since retteng performances present a problem of interpretation, I will discuss this matter relating it to contextualization, which is defined as an active process of performers using some linguistic cues (Gumperz 1982) to index features of the settings in order to produce interpretive frameworks. We will see that a key element is the use of reported speech, namely, quoting the voice of a child. It is by paying attention to this double-voicing that we come to see that this retteng is talking not (just) about past events but, more importantly, about culturally-shared ideas about authoritative knowledge and community values.

A’s poem and speaker B’s response

27A’s poem (lines 1-10 above) is characterized by two events: the narrating event and the narrated event. The narrating event involves the on-going speech event (lines 1-5 and lines 8-10); the narrated event is the reported speech event (lines 6-7). In the narrating event A composes his poem in the normal pattern discussed above of greeting (lines 1-2), disclaimer (lines 3-4), content (lines 5-9), and closing (line 10). The narrated event is contained within the narrating event.

  • 24 Deixis or deictic words are words used to point or indicate the context of utterances such as demon (...)

28The structure of this poem ties the performance to the speaker’s competence as an expert speaker. The performativity of A’s poem is also anchored in its context through contextualization both in the narrating and the narrated events. Two such poetic cues indexing the feature of the setting used to produce an interpretive framework are the use of pronouns and the deictic24 demonstrative. In line 4, the speaker contextualizes the performance by using the deictic suffix first person singular /-na’/, “I” in bendanna’, “I stand up” and its parallel ke’de’na’, “I rise up”. He also uses the referential index or duplex sign that has two functional modes, te, “this closer to the speaker”. Here the “I” is constructed as an active identity.

29In the narrated event, the contextualization can be observed by both participants and analysts, in A’s description of his interaction with a child. He again uses the deictic suffix /-na’/ or “I,” (in line 6), referring to the same person in a different role. He describes a situation when he was sitting, wearing a sarong or covered by a blanket. Contrary to line 4 above, the “I” in line 6 is a passive identity. The important thing here is that in such a construction, this “I” becomes the theme or subtopic. The two “I”’s are different because they contain different voices and strategies for evaluating the truth of the coming story. Such contextualization is also found in line 7 in the cleft construction where he presents an interaction with the child in the narrated event.

  • 25 These metapragmatic descriptors are sometimes called verbum discendi. It describes instances of lan (...)
  • 26 For more metapragmatic descriptors associated with death, see Sandarupa (2004 : 144-147).

30A related aspect of contextualization is the use of metapragmatic descriptors, i.e., verbs for “saying.” Such verbs are used whenever speakers engage in direct or indirect quotation.25 Like the word “say” in English, the word kua, “say” and untule’, “tell a story again” are neutral. Other words are associated only with death rituals (lines 4-7), such as “to lament” (umbating), “to mourn” (ma’rio-rio), and “to press down”(o’ton), which are nominalized in the next line (line 5) as “my lament” (batingku) and “my mourning” (marioku), where the suffix /-ku/ is the possessive first person.26 Such a transformation signals the boundary of the “I” in general from the “I” who is doing the current lamenting. By uttering “my lament says,” A urges the audience to pay attention to the content and not the speaking “I.” This cue, however, is precisely what the next respondent (B) fails to pay attention to.

31The verb “to tell a story again” (line 7) can be used as a contextualization cue for how to evaluate the modality of truth of the upcoming child’s story. Since A does not mention the name of the addressee, one may wonder what cues are picked up by the next respondent to interpret that he has now become the target. Later when discussing the incident with me, singer B used the terms nakanna and nasa’dingan in explaining the interpretive process (see also above). The tominaa Ne’ Rimma’ of Balik village, an expert on retteng, also uses the same terms to explain the cues. Words are sharp like a spear that hits the target (nakanna), and thus the addressee becomes conscious (nasa’dingan). The term nasa’dingan is derived from the root sa’ding, which has several meanings, of which the most relevant are “to feel” and “to be conscious of the self”; there is also the sense of “recovering consciousness” and being “shocked” (Tammu and Veen 1972). This singer A quotes the
story of a child using sharp metaphors. In Morris’ terms (1971 [1938]), these utterances, expressed in metaphors (lines 8-10), are characterizing signs that contribute some indexical effects on the next respondent.

32Once B surfaces as respondent, he faces a choice between commenting on the interaction between singer A and the child or developing an interaction between himself and singer A. It becomes clear that he opts for the latter. B, immediately agrees to A’s metaphors (in lines 8-10), using the prefix deictic /ku-/ first person singular, “I” (line 11). The second contextualization cue is the use of the metapragmatic descriptor “to say as true” (untongan) and its parallel, “to make it clear” (untundu maleso) (line 11).

33Instead of composing a poem that comments on the child’s story, B’s poem characterizes A himself, using the eel and lizard metaphors in lines 12-14. Thus, A’s narrated event – his interaction with the child – is transformed into an interaction between A and B.

A’s song in response to B

34A then answers B’s poem (lines 14-15) using metaphors which compare the verbal conflict to roosters fighting, thereby evaluating it indexically in the here and now of the narrating event. He uses as contextualization cues the referential index te, “this” to refer to the village in which the performance is taking place. And he uses the metapragmatic descriptor unnoni, “to produce sound” (line 16) as a sign of a symbolic fight in which the two speakers are engaged.

35In line 17, Singer A contextualizes by defining himself using strategic deictics, saying “I, this very I” (ia aku te akunna) and, in the second part of the line, “that myself alone” (tu kale misa-misangku). The use of the referential index “this” (te) indexes the “I-ness” of the speaker. In the second part of the line, Singer A uses deictic tu, which conveys a “closer” relationship to the addressee. In this case, while speaking, he is indexing himself as if he were speaking from the perspective of the addressee. This is a strategic use of deictics in defining the boundary of “I” from two perspectives: one, from the speaker’s perspective, and the other, from the addressee’s perspective. It is a claim of total difference from the addressee. It may index his creativity. He then poses rhetorical questions (in lines 18-20) introducing the agentive subject and using the enclitic /-ra/ to mark such questions.

Singer C’s intervention

  • 27 The power of this semiotic mediation lies in the performative force of C’s ritual. His retteng pres (...)

36Up to this point, the retteng seems to be leading towards a heated debate, and the performance may move from symbolic conflict into a real fight. Singer C then intercedes and mediates the emerging conflict.27 He begins his sambo retteng by using the inclusive “we” (kita), which includes everyone present, defining the community with its members as “those who assist in the funeral ritual” (to tongkon) and “those who are in the period of mourning” (to mario-rio in line 21). He also uses the deictic te, “this” to index members of the community.

  • 28 If lindo, “face” is the metonym of the whole body, the reference to face marks the individual perfo (...)

37Singer C assumes a moral stance, as can be observed in the contextualization cues of the metapragmatic descriptors in line 22, such as “make good or say a good lament” (meloi bating) and “make it in a good way” (dende’ maya-maya). This means there are no others present; they are one group in contrast to the “fighting cocks” of lines 15-16. His argument is that all who are present are none other than family members. This argument is expressed in line 23: “are there other faces?” (dendika lindona senga’?) and its synonym, “are there other different marks?” (le’ke’ri pa’todinganna?). These rhetorical questions imply the unity of the performers as a corporate group and as family members.28


38From the performance above, we have seen that there is an institutionalized order of interaction in which the performers construct their respective identities via poetic speech. The question is how the participant, in this case performer A, introduces the topic that subsequently lends coherence to what is being said? Let us look more closely at how this topic becomes the center of information being debated, which then constitutes the structure of the text.

39Singer A introduces the topic in the narrated event in line 6 by describing himself as someone wearing a sarong (a wrapped, tube-like cloth). Then he introduces a child, the addressee in the narrated event, by employing the topic marker in the special cleft-construction in line 7. As has been shown above, lines 6-7 are important because they provide us with two important events at two levels of text: the narrated event and the narrating events. Line 6 of the narrated event introduces the relevant participants-singer A (depicted as wrapped in a sarong) and the child.

40After introducing the interaction between the two, singer A frames what comes next as a cleft construction, “It is only because of a child who passed by” (denri pia’-pia’ lendu’) and its parallel, “it is only because of a small one” (baitti’ untuleranna’) with two predicates “who passed by” and the metapragmatic descriptor “to tell again” (untule’). The most important cleft-construction is denri pia’-pia’ or baitti’ untuleranna’, “it is only because of a child who told me a story again” (in line 7).

41The child is described using the word untuleran, where the root is tule’ and the prefix is the active transitive /un-/, whose agentive subject is the child. It therefore means, “to tell a story again or to narrate again,” and the benefactive /-na’/, “me,” indexes the present singer A. So, the clause means, “the child narrates again to me.” The telling of the story extends beyond the moment of telling. A keeps this in mind until he brings it into the present performance, a sign of its importance.

  • 29 Abstracting from the text into grammatical form, both of the sentences above have the same grammar. (...)

42The word den (“it is” or “once upon a time”) is the story frame; the enclitic –ri means “only because”. In Toraja culture, this expression is a conventional narrative opening. Singer A introduces a child as an agentive subject, who becomes the focus of information. Compare this with the text sentence, “a child told me a story again” (pia’-pia’ untuleranna’). In this text sentence, a child is the topical referring noun phrase and told me a story again is the maximal domain information focus. The order of the relationship that the topic comes first, and the focal information comes second.29

  • 30 The Prague linguists use the terms theme that which the clause is concerned and the rheme, the rema (...)

43The two have different text segmental relevance. In the cleft text sentence, its proposition is introduced as a focus of information. We have the following message structure: it (theme) was only because of a child (rheme),30 who (theme) told me a story again (rheme). But we can also have the following message structure: it was only because of a child (theme), who told me a story again (rheme). Interpretation of the predicated theme is understood. In the first interpretation, we have implicit meta-talk, where “I am going to tell you about a child” is a focus of information. This story of a child becomes the focus of information.

44In this narrated event we detect another aspect of performativity, the process of entextualization, which is characterized by several features. First, the use of the verb “say,” that is, “to tell again” (untule’) in line 7 indexes the original context in which the story was told, in a narrative style. This piece of text is now recontextualized in poetic form in the present performance. However, because it is in the new poetic context, its forms are indexically shaped by this new context, namely, it must comprise eight syllables, be expressed in parallelism, and use social metaphor. In other words, the child said something in ordinary speech in the past and in conveying this story, performer A has edited and transformed it by making it fill the formal pattern of eight syllables and parallelism. He has condensed, heightened, and transformed it into short poeticized parallel lines that contain no easily interpreted ordinary words but rather metaphors whose meanings are puzzles. Thus, in the process of recontextualization, the transformation that occurs is from a narrative genre into another genre called karrume “riddle.” Indirect discourse is now used by performer A to express his own voice. Such decontextualiation from its originary context and recontextualization of it in the present performance transforms the voice of a child into the voice of an elder inviting dialogue (see Sandarupa 2004: 241-242).

45Furthermore, the neutral metapragmatic descriptor “tell a story again” (untule’) indexes a minimizing intertextual gap between A’s poem and the child’s story. From this perspective, the two voices becomes very close to each other; the voice of a child merges and becomes one with the voice of singer A (see Bakhtin 1981 [1935]). By representing the child’s voice in this way, speaker A takes a positive attitude toward the child’s words. Through entextualization, he speaks powerfully, adopting – rather than simply reporting at a distance – the voice of a child.

  • 31 We could perhaps distinguish chronological, developmental and psychological age, but that is beyond (...)

46Speaker A can do this in part because of the local understandings of childhood and children. The meaning of “child” is open to different possible interpretations. In Toraja culture, seniority31 is considered especially important for the status and quality of talk. Child-talk (ulelean pia) has the qualities of being unserious, untrue, and lacking in attention to etiquette, as it is expressed in the norm dau’ ma’kada pia bang, “avoid talking like a child.” In this interpretation, a story framed as a child’s is not to be taken seriously. This interpretation is supported by the shared cultural knowledge that Torajan culture emphasizes seniority.

  • 32 For more ethnographic material on children and childhood in Toraja culture, see Koubi (2003).

47The second interpretation is precisely the reverse of the first, that child-talk is serious, true, and full of etiquette. This is suggested by the fact that a common disclaimer (see above) protecting the speaker from the danger of misspeaking is sang tintipa’ to mangla tedong, “I am just like the buffalo boy” and its parallel sang tandapa’ to mangla karambau, “I am of the same level as the water buffalo boy” or mane tali-tali bannu’, “I just wear a headband of bamboo strips” and its parallel, songko’ peladaran, “a learning hat.” Expert speakers also use the disclaimer, pia’-pia’pa dadingku, meaning, “I am just a child,” and its parallel, baitti’pa garagangku, “my shape is still small.” Thus when singer A mentions a child, he is using an expression usually used by expert speakers to index themselves. The very fact that pia’-pia’, “child,” and baitti’, “a small one,” become the subject of metapragmatic descriptor untule’, “to tell again” is the reason we associate this child with this disclaimer. In a broader, cultural sense, children are considered to be impartial and honest (malambu’). Thus traditionally, when a meeting organized to solve certain sociopolitical matters comes to a dead end, participants usually agree to ask a buffalo boy to make a decision.32

48But we must look to the next respondent to see how the topic introduced by speaker A is treated. Singer B’s answer gives us a feeling of coherence when he uses the verbs “say,” i.e. “to assert it as true” and “to make clear” in line 11 preceding his metaphors in lines 12-14. A closer study shows that B’s metaphors comment on A’s metaphors in lines 8-10. In other words, singer B takes A’s metaphors as a contextualization cue for an interpretive framework.

  • 33 These metaphors are characterizing signs that index the analogy with social problem a person has fa (...)

49An analysis of singer A’s metaphors in lines 8-10 has shown us that “someone” (to) is in a certain condition. These metaphorical characterizations are “using thorns of palm tree as his mat” (umpennampa’ duri banga) and “using thorns as pillow” (uppennallon duri) (line 8). The ritual speaker asks us to imagine a person “who sleeps on a mat and pillow of thorns puncturing stick his back and head.” This person we are asked to imagine will never sleep and, thus, will die slowly. The metaphors involve sharp, piercing objects, describing a situation that kills slowly and painfully. After further interviews,
I learned that informants understand the metaphors as alluding to a well
known, unresolved murder that haunts the community because the victim’s various mortuary rituals have not yet been done.33 This unresolved murder has become the burden of the local Torajan community. The problems posed by the murder center on how to find the criminal, the proper way of ritually delivering the spirit of the deceased to the “land of spirits” (Puya) and the proper way of reestablishing social relations with the killer and his family members, whether it is in the form of revenge or of carrying out “a ritual that will bury all revenge in order to produce a peaceful life” (ditambuntanai).

50The child’s voice continues (in lines 9-10) through a negative construction tang … pa, “not…yet.” The “mat and the pillow are not yet smooth,” meaning that this criminal problem has ruptured society for a long time but remains unsolved. All agree that the child’s story expresses the idea that conflict resolution is necessary for the (re)construction of a moral society. To recap, A’s metaphors are as follows:

person who sleeps on thorns of a palm tree as a mat
A person who sleeps on thorns as a pillow
His mat is still not yet smooth

51Upon closer examination we find that singer B’s metaphors constitute a comment on singer A’s metaphors. B composes poetic lines that attribute some characterizations to the singer A, which can be seen in lines 12-14 that introduce the subject of his next characterizations (lines 13-14) in the object position. These metaphors liken the previous singer, A, to animals namely a big eel, and its parallel a large lizard. This is an inversion of reality: the enormous eel lives in a river and it is edible. But when this animal comes on land searching for meat, it becomes a bad animal. In contrast, the large lizard lives on land and it is not edible. When this animal is on land and divides meat for others then it becomes a good animal. But when it eats farmers’ plants, it is a bad animal, representing a bad person. This destructive person is further characterized by the metaphors (lines 13-14). The enormous eel has such tail strength that it can dive and swim in the river against the current. The large lizard has a sharp dorsal fin that tickles. Ultimately, both cause bad results, and this is expressed through the repetition (line 14). Several further observations show us that while A uses plants as metaphors, B uses animals. And A’s metaphors imagine the “thorns” of the plants as agents that stick someone’s back and head to death, while B’s metaphors use the tails, backs, and heads of animals as agents that stick someone to death. In other words, B reverses A’s metaphors which provides a feeling of cotextuality, or the way a text comments on and resonates with an earlier one.

52Singer B takes up the metaphors of sharpness, but he ignores the fact that A’s poem is framed as a child’s quoted story. In so doing he rejects A’s claim to have transformed child’s speech into ritual speech. Indeed B asserts that the purported re-voicing of a child’s voice is not a child’s voice but rather a destructive and dangerous voice that is itself causing death. Thus singer A’s metaphorical words about thorny sleeping mats and pillows turn into the eel’s strong tail, the lizard’s sharp dorsal fin, and both their heads – all of which cause death. These characterizations suggest that singer A is an “agitator, provoker” (to pakadakean). According to Tato’ Dena’ these words must have come out of a mouth that produces crooked speech (tomalute puduk).

53To recap we can compare the two sets of metaphors and see the feeling of coherence that links them:



a person who sleeps on thorns
of a palm tree as a mat

a big eel is striking us

a person who sleeps on thorns

a big lizard’s dorsal fins pierce

as a pillow

the heads cause damage

  • 34 This concern about misreading other people and about knowledge that is false is related to the fact (...)

54Singer A responds by posing the subsequent question, another cleft-construction, where the agentive subject is sought in the question (line 18) and the restrictive relative clause in metaphors (lines 19-20). By posing this question, A problematizes B’s characterization about him. Without explicit accusation,
A criticizes B’s poem by implying that such metaphors come from a person who is not knowledgeable. The evidence of his lack of expertise is that he produces a
retteng that misreads A’s.34 With this, B becomes a person who is unwilling to solve the problem and, therefore, he represents the “bad” type of person in the community, a person who produces crooked speech (to malute puduk).

55He becomes a person who embodies the evils of society (to pakadakean).

  • 35 Pande is the general term that means “expert” and manarang means “clever”.
  • 36 Cf. Sandarupa (2004:248-249) observes that grayish-white feathers of egrets do not exist and theref (...)

56Such criticism shows that B’s attempted challenge has been deflected. Although he has tried to speak as an expert, he comes across instead as rejecting the voice of a child which undermines his authority. Speaker A attacks B’s expertise using the expression “person who is as clever as the expert” (to manarang pande)35 and its paired line “a person who has extensive knowledge” (to rangga inawa) (line 18). In contrast, false knowledge is implied by the phrases “can see the missing teeth of the chicken” (ungkitta’ simpona manuk) or “the branches of rooster’s teeth” (tangke isinna pa’kurung) and, “the gray hair of egrets” (ubanna koro-koro’) repeated twice (lines 18-19). Here performer A introduces the category of expert speakers in Torajan society.
All my informants, such as Tato’ Dena’, Luter Bala, and Pabisa, agree that these speakers have the expertise of knowing all things. Culturally performer B’s speaking and knowledge are challenged on several grounds. First, chickens do not even have teeth, so anyone who claims to see them has no true knowledge, only false knowledge. Second, egrets have only white feathers. Someone who maintains they can see the grayish-white “hair” of egrets possess self-evident knowledge, the knowledge of the obvious, known by all.
36 Therefore, such knowledge is not true knowledge (tang tongan).

57With this it can be certain that what was originally an interaction between a child and A has been transformed into an interaction between A and B.
An interaction between a child and a mature person has been transformed into an interaction between knowledgeable and unknowledgeable leaders.
The emergent structure is now reversed.

C’s covering retteng

58C’s metaphors underscore the values of a moral society. He suggests building a solid retteng community, in which everyone is on good terms with one another, thus invoking again the voice of a child – that the child’s urgings, to build up a moral community, be enacted. Only via this path will the retteng community society become a moral society and, more broadly, only via this path will Toraja become a moral society. Only within such a community can life ensue with blessings and health, as opposed to a thorn-based life.
Thus good poetic argumentation becomes the model of a moral society.
retteng is performed in a good way, that is the only way to get the flowing of blessing, coolness, protection and wealth into the community (Zerner and Volkman 1988, Waterson 2012).

59Performer C’s retteng is a metapragmatic evaluation of the retteng of the previous two performers. He made this metapragmatic judgement by offering direct advice “to always make the lament good” (tontong meloi bating) and its synonym “to mourn in good way” (tadendei’ maya-maya) in line 22.
These two metapragmatic descriptors
meloi, “make well” and dendei’ maya-maya, “mourn in a good way,” have effects on social relations among participants, inducing and inciting them to be on good terms with each other.

60There must be a single voice (misa’ kada) which contrasts with those of “many roosters” (buda londong) and “many interchangeable voices” (sisonda-sonda unnoni) in lines 15-16. To achieve a single voice through ritual, aluk, and by observing taboo, pemali, is the realization of one of the ancestors “exemplary models” (sangka’). When things happen along the lines described above, Torajans call this being on the right track or following “the river’s course” (salunna).

61Thus we see that what Silverstein calls “the poetic chunk” or “the emergent structure” (Silverstein 1992, 1993, 1998) in this performance is the voice of a child. As has been shown, it is the child that becomes the current topic or the focus of information, and it is this topic that is debated for the duration of time of the performance expressed through poetic lines. The larger question raised by this poetic dueling is who are the types or categories of persons who are involved. For instance, Pak Lande’ characterizes speaker A as a leader-type person. How has such a characterization emerged in the course of this 4-part retteng? After A replies to B, he is transformed from the voice of a child to a to malambu’, an “honest person,” which is the same as to manarang pande, “expert speaker,” personalized as A and C, as opposed with to malute puduk, “fake speaker,” as personalized by B. In short, the emergent structure is
A:B :: expert speaker : fake speaker.
Thus we see that in the poetic interaction, the participants enact certain roles and relationships which are related to ways of talking about categories of persons. Through such interaction,
we can identify how their roles vary as the relationships unfold. The emergent structure has “social effectiveness” (Agha 2007). The text indexes the social relations among kinds or types of participants: expert speaker: fake speaker which in turn indexes the interactional positioning of singers in the narrating event. We have moved from child : adult, to good person : bad person, to expert speaker : fake speaker.

62Several points can now be considered as evidence of language use and its association with types of speakers. A’s disclaimer, his use of complicated deictics of first person and place, the dialectic between entextualization and contextualization, and vivid metaphors index to manarang pande, “expert speakers.” With this observation we can say that A and C represent the child type associated with expert speakers as opposed to bad speakers who are not willing to problem-solve. My consultants Tato’ Dena’, Pabisa, and Luter Bala elaborated on the meaning of communicative expertise. They explained to me that in Torajan culture, expert speakers are called to manarang pande.

63The word manarang means “clever” and pande means “extensive knowledge” so literally this expression means “expert who has extensive knowledge.” This culture distinguishes two types of expert speakers. The first type has expertise on solving sociopolitical conflicts. The second category is that of ritual expert. The following outlines local classifications of types of expert speakers in Toraja society:

To manarang pande

Sociopolitical Matters

Ritual Matters

anak pare-pare nangka’


gora-gora tongkon



  • 37 Cf. Sandarupa (2004).

64Two expressions, “hard inedible bits of the jackfruit” (anak pare-pare nangka’) and its synonym, the “kind of rush used for braiding mats and handbags” (anak passasaran tuyu) (Nooy-Palm 1979: 48-52) refer to expert speakers skilled at conflict resolution and settling of disputes. The “sitting throat” (gora-gora tongkon) is an expert speaker in meetings regarding “adat-regulations and sacrificial ritual” (Ibid. 52). Nowadays they function at rituals to welcome guests. The “singers” (tomenani) are the ritual speakers in the bua’ ritual that involves the whole community. The “ritual speaker with extensive knowledge” (tominaa) has expertise in the death and life rituals (rambu solo’, rambu tuka’). Finally, the toburake, in the southern part of Toraja, is a ritual priest and ritual speaker who has expertise in life rituals (rambu tuka’), for fertility (lolo).37

65Singers A and C represent both categories of expert speakers with honesty and willingness to solve sociopolitical and ritual conflicts using shared values and local wisdom. On the other hand, B represents a type of speaker known as to malute puduk, where malute suggests verbal facility, and puduk is “lips” so what he says from his mouth is different from what is in his heart. This makes him a “fake speaker”; he is dishonest and unwilling to solve the problem. One might say he merely “pays lip service.” He is only interested in verbal argumentation for its own sake. In short, he is someone who acts as an agitator or provoker.

  • 38 Thus, in this article I avoid using the interpretive approaches of intention or implicature, approa (...)

66The preceding analysis offers a way to interpret cultural phenomena. In his interpretive account of the Balinese cockfight, Clifford Geertz (1973) demonstrated the insights to be gained by approaching people’s culture as an “ensemble of texts.” In his approach, he defines text as “saying something of something” (Ibid. 448), thus emphasizing the denotational aspect of text. But such an analysis is at the expense of the interactive and meditational dimensions of texts. Instead, I have used the approach of functional semiotics, which Silverstein has further refined into the relation between denotational and interactional texts, mediated by indexicality (Silverstein 1993, 2001). In addition, Torajan local wisdom holds that it is difficult to read people’s minds suggesting that this semiotic approach to the interpretation of social life is also more consistent with their own views of language.38

67The local model of interpretive process uses indexical cues to have effects on emotion and cognition that cause someone to take up the role of respondent. Thus, it is the effect of poetry on the hearer’s side that is important. In this analysis, this model is broadened by combining it with the speaker’s side. The speaker constructs the power of poetic speech that involves expertise in composing interactive parallelism, the ability to interpret others’ poetry and the capacity to advance poetic argumentation using a combination of appropriate metapragmatic descriptors, the dialectic between entextualization and contextualization, complicated deictics of first person and place, the special use of cleft-construction, and good metaphors. They are all indexical cues that provide us with interpretive frameworks. From the analysis above, it is clear that the force and power of this performance is the merging of the cohesive structure of signs and the definition of the relationship among interlocutors.

68As in many cultures, Torajan poetry is an exchange or an interactional event. Torajans enjoy using poetry in rituals to construct their social relations. In such interactional settings, meaning becomes the contested terrain, something co-constructed in the performance. The question of what it says about Toraja culture emerges in the poetic structure (A: B :: expert speaker : fake speaker).
It is this poetic structure that constitutes a text mediationally indexing the context of performance populated by expert speakers (interactional). It is the semiotic indexicality that gives us empirical evidence of the validation of our interpretations.

69From this micro-poetic argumentation above we can observe that functional semiotics also helps us to relate this particular performance to other performances and wider events. Through the tools of indexicals, this particular performance can be tied to beyond the present performance of poetic argumentation, the criminal event and the associated rituals that need be performed and the values of the society at large. The strength of this approach is that it gives us empirical evidence for local processes of interpretation and avoids mystical interpretation of a single event.


70We have seen that the siretteng is a form of death ritual poetry containing riddles (karrume) and “hidden things” (kambunni’). This kind of ritual poetry involves the use of interactive parallelisms concerning sociopolitical matters with the ultimate objective of building a moral community. It is in the retteng - poetic argumentation that we see displays of highly valued poetic speaking skills. Such expertise shows the richness of this culture in the arena of speaking, as exemplified by A. It is through performances of verbal arts that one constructs the self as “expert speaker” and becomes authoritative. Displaying authoritative speaking through multiple public performances may improve a speaker’s social standing. Having examined the text and sociopolitical context of this ritual poetry, we can finally arrive at an understanding of the significance of the voice of a child. The child’s voice indicates “the capacity to express one’s extensive knowledge via expertise in speaking” (to manarang pande). Moreover it also alludes to “honesty” (malambu’) and the “ability to connect what is in one’s heart with what one says” (sa’ti). We also see how the interlocutors are busy indexing the characteristics of a good moral society by defining and redefining the voice of a child. The voice of the child of a good moral society is shown in the performance as a “single voice” (misa’ kada) where people are on good terms with each other (kasiuluran), leading to the immediate solution of the sociopolitical and ritual conflicts. Only through the power of such speech in performance can life and wealth, blessing and health flow to the community. Thus, not only is a moral community constructed through performance, but it also becomes a model for the moral society at large. As we have seen, it is an example of poetic argumentation that heavily relies on the cultural values of unity and brotherhood.

Haut de page


Abu-Lughod, L. 1988. Veiled sentiments: Honor and poetry in a Bedouin society. Berkeley: University of California Press.

Adams, K. M. 2006. Art as politics: Recrafting identities, tourism and power in Tana Toraja, Indonesia. Hawai’i: University of Hawai’i Press.

Agha, A. 2007. Language and social relations. Cambridge: Cambridge University Press.

Bakhtin, M. 1981 [1935]. The dialogic imagination. Austin: University of Texas Press.

Bauman, R. and C. Briggs. 1990. “Poetics and performance as critical perspectives on language and social life”. Annual Review of Anthropology 19:59-88.

Bowen, J. R. 1991. Sumatran politics and poetics: Gayo history 1900-1989. New Haven and London: Yale University Press.

Caton, S. C. 1990. “Peaks of Yemen I summon”: Poetry as cultural practice in a North Yemeni tribe. Berkeley, Los Angeles, London: University of California Press.

Conklin, H. C. 1964. “Linguistic play in its cultural context”. In Language in culture and society. Edited by D. Hymes, 295-300. New York: Harper & Row Publishers.

Coville, E. 1988. A single word brings to life: The Maro ritual in Tana Toraja (Indonesia). Dissertation, The University of Chicago.

Fox, J. J. ed. 1988. To speak in pairs: Essays on the ritual languages of the eastern Indonesia. Cambridge: Cambridge University Press.

Geertz, C. 1973. “Deep play: Notes on the Balinese cockfight”. In The interpretation of cultures. 412-453. New York: Basic Books Inc.

Grice, H. P. 1971. “Meaning”. In Semantics: An interdisciplinary reader in philosophy, linguistics, and psychology. Edited by D. Steinberg and L. Jakobovits, 52-59. Cambridge: Cambridge University Press.

Gumperz, J. 1982. Discourse strategies. Cambridge: Cambridge University Press.

Jakobson, R. 1960. “Linguistics and poetics”. In Style in Language, edited by T. A. Sebeok, 350-377. Cambridge: Massachusetts Institute of Technology Press.

Koubi, J. 1982. Rambu solo’, la fumée descend: le culte des morts chez les Toradja du Sud. Paris: Centre National de la Recherche Scientifique.

Koubi, J..2003. Histoires d’enfants exposés, Pays Toraja, Sulawesi, Indonésie. Paris : Presses de l’Université de Paris-Sorbonne.

Lebang, J. B. P. 2006. Samparan pa’kadanna Toraya. Rantepao: Siayoka.

Morris, C. 1971 [1938]. Foundations of the theory of signs. Chicago: The University of Chicago Press.

Nooy-Palm, H. 1979. The Sa’dan Toraja. A study of their social life and religion: Organization, symbols and beliefs. The Hague: Martinus Nijhoff.

Nooy-Palm, H. 1986. The Sa’dan Toraja. A study of their social life and religion. Rituals of the east and west. Dordrecht: Foris Publications.

Rappoport, D. 2009. Songs from the thrice-blooded land: Ritual music of the Toraja (Sulawesi). Paris: Edition de la Maison des sciences de l’homme.

Sandarupa, S. 1989. Tropes, symbolism, rhetorical structure, structure of parallelism and ‘parallelism’ of structure in Toraja. The University of Chicago, USA.

Sandarupa, S.. 2004. The exemplary center: poetics and politics in the kingly death ritual in Toraja south Sulawesi Indonesia. Dissertation, The University of Chicago.

Sandarupa, S.. 2012. Life and death in Toraja. Makassar: Hasanuddin University Press.

Sandarupa, S.. 2013. “The poetry of taking power in Toraja, South Sulawesi Indonesia”. Research on humanities and social sciences 3:113-132.

Sandarupa, S., R. Assagaf and H. Hasyim. 2015. “Heterogeneity in Torajan ritual speech: Metalinguistic awareness and the nation’s character building”. Sosiohumanika, Jurnal Pendidikan Sains Sosial dan Kemanusiaan, 8(1): 29-38.

Sarira, Y. A. 2000. Litani aluk bua’. Rantepao: Pusbang Gereja Toraja.

Silverstein, M. 1976. “Shifters, linguistic categories, and cultural descriptions”. In Meaning in anthropology. Edited by K. H. Basso and H. A. Selby, 11-55. Albuquerque: University of New Mexico Press.

Silverstein, M. 1992. “The indeterminacy of contextualization: When is enough enough?”. In The contextualization of language. Edited by A. DiLuzio and P. Auer, 35-75. Amsterdam: J. Benjamin.

Silverstein, M. 1993. “Metapragmatic discourse and metapragmatic function”. In Reflexive language: Reported speech and metapragmatics. Edited by J. A. Lucy, 33-58. New York: Cambridge University Press.

Silverstein, M. 1998. “The improvisational performance of culture in real time discursive practice”. In Creativity in performance. Edited by R. K. Sawyer, 265-312. Norwood, NJ: Ablex.

Silverstein, M. 2001. “Functions”. In Key terms in language and culture. Edited by A. Duranti, 71-74. Malden, Mass.: Blackwell Publisher.

Sweeney, A. 1987. A full hearing: Orality and literacy in the Malay world. Berkeley: Berkeley University Press.

Tammu J. and H. van der Veen. 1972. Kamus Toradja - Indonesia. Rantepao, Toraja: Jajasan Perguruan Kristen Toradja.

Veen, van der, H. 1965. The Merok feast of the Sa’dan Toraja. s’-Gravenhage: Martinus Nijhoff.

Veen, van der, H. 1966. The Sa’dan Toraja chant for the deceased. s’-Gravenhage: Martinus Nijhoff.

Volkman, T. 1985. Feasts of honor: Ritual and change in the Toraja highlands. Urbana and Chicago: University of Illinois Press.

Waterson, R. 2003. “The immortality of the house in Tana Toraja”. In The House in Southeast Asia: A Changing social, economic and political domain. Edited by Sparks, Stephen and Howell, Signe (eds.), 34-52. London: Routledge Curzon.

Waterson, R. 2009. Path and rivers: Sa’dan Toraja society in transformation. Leiden: KITLV.

Waterson, R. 2012. “Flows of words and flows of blessing: The poetics of invocatory speech among the Sa’dan Toraja”. Bijdragen tot de Taal-, Land- en Volkenkunde 168 (4):391-419.

Zerner, C. and T. A. Volkman. 1988. “The tree of desire: a Toraja ritual poem”. In To speak in pairs. Edited by J. J. Fox, 282-306. New York: Cambridge University Press.

Haut de page


Stanislaus Sandarupa, Makassar, 2008

Stanislaus Sandarupa, Makassar, 2008

Photo: Dana Rappoport

Haut de page


5 This is not to deny that a given performance is related to a number of other speech events that precede and succeed it (cf. Bauman and Briggs 1990 : 60).

6 It is estimated that an additional 2 million Torajans live outside Toraja, spread across the big islands of Indonesia, such as Java, Kalimantan, and Papua. The local religion is called Aluk To Dolo, “The Religion of the Ancestors.” Torajans believe that performance of a death ritual accompanied by animal sacrifices such as buffalo and pig will facilitate the spirit’s travel to the next world, called Puya, “Land of Spirit” (a temporary place of spirits). Only after the death ritual is complete and a subsequent (optional) life ritual is staged can the spirit travel up to heaven to become a god. According to the 2014 statistics, only 5 % of Torajans still follow the Aluk To Dolo religion : 90 % have converted to Christianity and 5 % have become Moslem. The research was conducted in 1994-1996 and at that time conversion statistics were similar. Although most Torajans are Christians, they continue to use local concepts drawn from Aluk To Dolo religion in their conversation and ritual speech.

7 This hierarchical division derives from the south Toraja regency, which is slightly different from the hierarchical division practiced in northern Toraja.

8 At the completion of a death ritual, it is believed that the “dark spirit” is transformed into a god who ascends to heaven accompanied by the pigs and water buffalo sacrificed during the ritual.

9 The expression “to floor the earth” is a metaphor which means to divide and distribute pieces of buffalo meat to people in the community. The piece of meat received reflects the rank of the receiver.

10 The word tongkonan refers to Torajans’ ranked ancestral houses with arched horn-shaped roofs. Some anthropologists have gone so far as to label Toraja a “house society” as these structures are especially important in Toraja social organization, both in the past and today (Waterson 2003, Adams 2006).

11 In the northern and western regions of Toraja, the word retteng is used in life rituals. Veen includes two samples of such retteng in his collection R11 and R12 (Veen 1966 : 8687, see also Rappoport 2009). Rappoport’s collection is available via internet (http ://

12 Nowadays, we regularly witness the decontextualization and recontextualization of retteng. For example, rather than being performed as part of badong, it is performed on the reception day of a funeral ritual (allo karampoan) when groups of guests arrive and are ritually greeted. The representative of the group can perform a retteng while entering a reception hall (lantang karampoan) to affirm social ties among participants and to reveal their social identities.

13 Veen (1966) explains the context in which they are used and emphasizes the improvised character and says it can be recited at “an arbitrary point” in the badong. 

14 Such as “thank you, you’re welcome” or “how are you ? fine and you ?”

15 These norms are captured in the saying tae’nama’din untossok mata bale (“it is forbidden to pierce the eyes of fish”).

16 See also Veen (1966) about the use of the stick, which he calls bandangan.

17 In introducing the retteng he collected, Veen (1966) lists a variety of different “subjects” : “The deceased can be praised ; those against whom he had a grievance can be criticized ; satirical reference can be made to a person with whom he was associated in life ; matters concerning the deceased can be alluded to ; and people may also make verses referring covertly to each other.” See also comments made by Nooy-Palm (1979 : 16 ; 1986 : 234, 318).

18 Here the speaker addresses the local noble chiefs who are metaphorically compared with kayu kalando, “tall trees” whose numerous branches and large leaves give protection called lamba’ paonganan.

19 A related idea is pemali ullutu tombang panda dibolong “it is taboo to create chaos in the funeral ceremony.”

20 The term pia, “child” and its synonym baitti’, “small one” also occurs in one of Veen’s (1966, 85) collection, R10, lines 12.

21 One of the late author’s important contributions to the study of ritual speech was to see that parallelism was not just at the level of the message (“denotational text”) but also at the level of social interaction (“interactional text”), that is, across speakers. Another related contribution was his emphasis on the complementarity of the lines of poetic speech. Others had seen the paired lines as mostly synonyms conveying the same or similar or opposite meanings, but he argued they provided complementary meanings. As he wrote, referring the title of James Fox’s well-known collection, it is not so much “speaking in pairs” as it is “speaking in complements.”

22 The sung poetic text comprises poetically compact discourse expressed in parallel lines. The rule is that each line consists of eight syllables but the speaker sings only seven syllables and drops the last one and finishes with the insertion of various syllables /-am/, /-ma/ and /-um/. The result is that the form of the last word becomes obscure since the last word appears to be a new word that needs guessing, a metapragmatic phonological transformation as figuration, the trope of concealment (Conklin 1964). For example :
kita angga to ma’bam– : we (incl.) all the badong chanters
adong le ... ee anggaki’ mario-rio  : we (incl.) all the mourners
In the first line, the last syllable of the word
to ma’badong is not sung and it ends with vowel /a/ which then gets the additional consonant /m/ so the whole line becomes seven (7) syllables ending with toma’bam-. Then the pause occurs and the last vowel of the syllable of the previous word (toma’bam-) is chanted and repeated again in the next line (-adong) which is then continued with the insertion /le ee../. Since in the next line anggaki’ mario-rio consists of eight syllables we can say that even though the last syllable of the previous line is sung in the next line, it should be counted as a syllabic part of the previous line so that each of the lines is composed of eight syllables. This is a regular pattern of phonological transformation.

23 The performance of retteng is the creative work of adults and the elderly. Indeed young people and sometimes primary school aged children use riddles for fun and to ask questions of one another. However the use of karrume in retteng is the domain of adults and especially old people. See Sandarupa (2004 : 242).

24 Deixis or deictic words are words used to point or indicate the context of utterances such as demonstratives, first and second person pronouns, tense and so on.

25 These metapragmatic descriptors are sometimes called verbum discendi. It describes instances of language use (Silverstein 1976 : 11-55 ; Silverstein 1993). He uses it to characterize his own way of “saying” versus someone else’s way of “saying.” The use of these metapragmatic verbs to characterize another’s speech is an indexical cue and a powerful means of voicing and ventriloquism (Bakhtin 1981 [1935]).

26 For more metapragmatic descriptors associated with death, see Sandarupa (2004 : 144-147).

27 The power of this semiotic mediation lies in the performative force of C’s ritual. His retteng presupposes the cultural norms of pemali ullutu tombang panda dibolong, “it is taboo to create riot in the funeral ritual.”

28 If lindo, “face” is the metonym of the whole body, the reference to face marks the individual performer with respect to the body of the corporate group—be it the performers’ troop (sang pa’badongan), family group, locally labeled sang rapu, “one ramage,” or village group (sang tondok).

29 Abstracting from the text into grammatical form, both of the sentences above have the same grammar. They relate the predicate untuleran in exactly the same way to the agentive subject and benefactive suffix /-na’/.

30 The Prague linguists use the terms theme that which the clause is concerned and the rheme, the remainder of the message in which the theme is developed.

31 We could perhaps distinguish chronological, developmental and psychological age, but that is beyond the scope of this paper.

32 For more ethnographic material on children and childhood in Toraja culture, see Koubi (2003).

33 These metaphors are characterizing signs that index the analogy with social problem a person has faced without defining the content of it. In order to understand the social problem referred to, I needed further ethnographic research.

34 This concern about misreading other people and about knowledge that is false is related to the fact that people say that no one can read another’s mind or their intentions.

35 Pande is the general term that means “expert” and manarang means “clever”.

36 Cf. Sandarupa (2004:248-249) observes that grayish-white feathers of egrets do not exist and therefore the argument is the same as with the chickens’ teeth. Either way, the point is to define how true knowledge can be compromised.

37 Cf. Sandarupa (2004).

38 Thus, in this article I avoid using the interpretive approaches of intention or implicature, approaches which are exemplified by Grice (1971).

Haut de page

Table des illustrations

Titre Stanislaus Sandarupa, Makassar, 2008
Crédits Photo: Dana Rappoport
Fichier image/png, 1,9M
Haut de page

Pour citer cet article

Référence papier

Stanislaus Sandarupa, « The Voice of a Child”: Constructing a Moral Community through Retteng Poetic Argumentation in Toraja »Archipel, 91 | 2016, 231-258.

Référence électronique

Stanislaus Sandarupa, « The Voice of a Child”: Constructing a Moral Community through Retteng Poetic Argumentation in Toraja »Archipel [En ligne], 91 | 2016, mis en ligne le 01 mai 2017, consulté le 29 mars 2023. URL : ; DOI :

Haut de page


Stanislaus Sandarupa

Department of Linguistics, Faculty of Cultural Sciences, Hasanuddin University, Makassar, Sulawesi Indonesia

Haut de page

Droits d’auteur

Tous droits réservés

Haut de page
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search