Previous Next

Master Poets, Ritual Masters


Introduction: Suti Solo do Bina Bane

This is a study of oral composition. Specifically, it is a study of the way in which oral poets on the island of Rote in eastern Indonesia compose recitations within a tradition of strict canonical parallelism. It is thus a study of particular poets, their specific compositions and the tradition in which they operate. The materials for this study have been gathered from more than a dozen master poets over nearly 45 years. Before it is possible to begin an analysis of their compositions, it is essential to provide some background to this study and the tradition of analysis within which this study is situated.

A Personal Introduction

Soon after I arrived on the island of Rote in April 1965, I took a long walk with Jaap Amalo, the elder brother of the then Lord (Manek) of Termanu and District Head (Camat) of Rote, Ernst Amalo. Jaap Amalo had spent his career in the civil service, initially during the Dutch colonial period and thereafter in the period of Indonesian independence. Having left Rote as a young man, he had returned to the island on his retirement and reimmersed himself in Rotenese culture. He understood perhaps better than most others just what an anthropologist might wish to do on the island and he was anxious to give me thoughtful advice on Rote and its culture. He talked to me specifically about Rotenese ritual language, although at the time I hardly grasped what he was trying to tell me. I had brought with me a tape recorder and it was with this in mind that he advised me to begin my research by recording the chant Suti Solo do Bina Bane. He told me that he regarded Suti Solo do Bina Bane as the most beautiful and moving of Rotenese ritual compositions. Although I had, at the time, no idea what this composition might be like, I had a clear and specific directive to record the chant.

Some days later, Ernst Amalo took me to Feapopi where he had his official residence as Manek of Termanu and introduced me to an assembly of clan lords and elders. Since he had no idea of anthropology, he explained that I had come to be the historian of Termanu and had brought a ‘voice-catcher’ (penangkap suara) to record Termanu’s oral traditions. At this gathering, I met, for the first time, Stefanus Adulanu, known as ‘Old Meno’ (Meno Tua), the senior most elder of clan Meno, who held the title of Head of the Earth (Dae Langak). Old Meno had served as the scribe to the court for at least four decades; he was one of the most respected elders of Termanu and was regarded as one of the most knowledgeable figures in the domain. Fortunately for me, he took on himself the task of teaching me the traditions of Termanu.

After this formal introduction, my wife and I were able to settle at Ufa Len in the house of a clan lord of Ingu-Beuk, Mias Kiuk, who was brother-in-law to the Acting Manek of Termanu (Wakil Manek), Frans Biredoko. From Ufa Len, I made regular visits to Ola Lain where Old Meno resided. Although he admitted to me some time later that he was first troubled and puzzled by my arrival, he explained that he saw in my voice-catcher the means for him to transmit to his grandson some of his most important traditional knowledge. In addition to reciting (or finding others to recite) key oral narratives and genealogies of the domain, Old Meno initiated me into an understanding of ritual language and slowly helped me to translate and comprehend ritual language compositions, many of which were gathered from an array of other local oral poets. At my request, Old Meno recited for me the composition Suti Solo do Bina Bane that Jaap Amalo had first urged me to record. The text of this composition is a superlative example of ritual language and is the first text examined in detail in this study.

This first text initiated a study that has continued for some 50 years—the entire time that I have spent studying ritual language. On subsequent visits over the years, primarily to Termanu, and more recently during recording sessions held on Bali with oral poets from different parts of Rote, I have repeatedly recorded versions of Suti Solo do Bina Bane. I have asked some poets to recite this composition on different occasions at different intervals in their life, but I have also asked poets who are related to one another and are said to have learned their art from the same (or a closely related) source to provide me with their own version of this composition. Initially, I focused my efforts on Termanu, which constitutes a single speech community, but in time, I have recorded this chant from oral poets in other dialect areas.

I have purposely focused on Suti Solo do Bina Bane to understand, in specific detail, how individual poets compose their chants, how their compositions may differ from one recitation to another and how compositions vary from one speech community to another. This study is the result of these investigations. However, these focused, fine-grained investigations of a single named chant relate to wider issues in the study of oral traditions. Rotenese ritual compositions belong to a tradition of oral composition based on strict canonical parallelism. Composition in strict canonical parallelism is—and certainly once was—a widespread means of poetic creation. Thus, understanding a single tradition based on this form of composition—as in the case of the Rotenese—may contribute to a fuller understanding of other traditions of parallel composition.

Parallelism and Canonical Parallelism as Forms of Oral Composition

Rotenese ritual language is a special poetic register that relies on a strict—indeed obligatory—use of semantic parallelism. As such, it belongs to a tradition of composition that is common to much of the world’s oral literatures. The term ‘parallelism’ dates to the eighteenth century and derives from the studies of Robert Lowth. In 1753, in a series of lectures delivered as Professor of Hebrew Poetry in the University of Oxford, Lowth noted that one of the major principles of composition in parts of the Old Testament was a carefully contrived pairing of line, phrase and verse. He described this form of composition as follows:

The poetic conformation of the sentences, which has been so often alluded to as characteristic of the Hebrew poetry, consists chiefly in a certain equality, resemblance, or parallelism between the members of each period; so that in two lines (or members of the same period) things for the most part shall answer to things, and words, to words, as if fitted to each other by a kind of rule or measure. This parallelism has much variety and graduations. (Lowth 1829: 157)

In a later lecture, in 1778, Lowth more explicitly defined his terminology for this form of composition, which he called parallelismus membrorum:

The correspondence of one verse or line with another, I call parallelism. When a proposition is delivered, and a second is subjoined to it, or drawn under it, equivalent, or contrasted with it in sense, or similar to it in the form of grammatical construction, these I call parallel lines; and the words or phrases, answering one to another in the corresponding lines, parallel terms. (Lowth 1834: ix)

Lowth’s observations led to the recognition of similar forms of composition across a wide spectrum of the world’s oral traditions as well as various written traditions based on oral models. Akkadian, Sumerian and early Egyptian texts were shown to be based on parallelism, and when Canaanite texts were discovered, it became evident that particular Biblical texts shared in a wider Semitic tradition of parallel composition. In Europe, the Finns were the first to take up Lowth’s ideas, recognising in their own folk traditions a pervasive use of parallelism. Elias Lönnroth recorded, selected and compiled these compositions to create the Kalevala, perhaps the most frequently cited example of pervasive parallelism. Various other scholars noted and discussed similar forms of parallelism for Ostyak and Vogul folk poetry, for Hungarian, for Mongolian, and for Turkic. Still other researchers have established parallelism as the first principle of Náhuatl and of Mayan poetry, and others have taken up the study of parallelism in the poetry of contemporary Maya groups, in Cuna folk traditions and in Quechua poetry dating back to the time of the Incas.

Early in the nineteenth century, scholars pointed to parallelism in Chinese poetry. Since then, parallelism has been noted in a variety of linguistic forms: in early written documents, in the rhyme-prose of the Han Period, in ‘parallel prose’, in love songs, in proverbs and in popular poetry. Similar usages have been observed in the extensive ritual texts of Zhuang of south-west China, in Tibetan, as well as among the Thulung Rai of eastern Nepal, the Sadar of Jaspur and the Toda of South India. In mainland South-East Asia, the use of parallelism has been documented in compositions among the Kachin, Kmhmu and Khmer speakers, and in written and oral compositions in Vietnamese.

The largest textual literature on parallelism can be found in the Austronesian-speaking world, where extensive traditions of parallelism in poetry and ritual languages have been documented for the Rhade of Vietnam, the peoples of Nias, the Batak, a number of Dayak groups in Kalimantan, among Bugis and Toraja groups in Sulawesi and for numerous populations throughout eastern Indonesia where vibrant traditions of oral composition in parallelism persist, particularly on the islands of Sumba, Flores, Savu, Rote and Timor.1

While there can be no doubt about the prevalence of the use of parallelism in a great diversity of oral traditions, Lowth’s qualification on his initial observations has still to be noted: ‘parallelism has much variety and graduations.’ Parallelism does not simply define a particular linguistic phenomenon, but rather points to a complex of intersecting phenomena. At times, parallelism may be considered to mask as much as it reveals.

The Jakobson Perspective on Canonical Parallelism

Throughout his career, the linguist Roman Jakobson returned repeatedly to parallelism as a special linguistic consideration, each time casting new reflections on the topic (see Fox 1977). It was Jakobson who coined the term ‘canonical parallelism’ to define parallel compositions ‘where certain similarities between successive verbal sequences are compulsory or enjoy a high preference’ (Jakobson 1966: 399). His purpose was to delimit traditions of oral composition in which specific pairings were culturally defined and specifically required for composition and, thus, to distinguish such traditions from others in which such pairings occur as poetic rhetoric that can be highly variable from composition to composition.

Distinguishing traditions of canonical parallelism and attempting to understand them shift the research agenda to a focus on the ‘canonical coupling’ of words—to what Jakobson described, in his own concise formulation, as the ‘paradigmatic axis of selection’ whose function in poetry is to create ‘metaphor’ by means of similarity (1956: 58 ff.). Canonical parallelism thus offers special insights by its explicit expression of this poetic function. Jakobson’s frequently quoted statement that ‘the poetic function projects the principle of equivalence from the axis of selection into the axis of combination’ (1960: 358) has particular relevance in the study of traditions of canonical parallelism. Whereas in other forms of poetry, the ‘poetic function’ may be subtly concealed or entirely implicit, in canonical parallelism this function is given direct and explicit expression. Culturally defined linguistic equivalences, both semantic and syntactic, become manifest.

Many, perhaps most, of the major traditions of composition identified as based on parallelism may well be traditions of canonical parallelism, but the underlying canonical coupling on which they are based may not always be immediately manifest or consistently evident. Some of the oldest traditions of such canonical parallelism have been preserved in written form. This holds for magnificent early examples of parallelism from Akkadian and Sumerian cuneiform sources as well as from Egyptian hieroglyphic texts or from a rich array of early Chinese texts such as the Taoist Yuandao from the period of the Han Dynasty, but equally so for surviving Maya sources such as the Popul Vuh, the Rabinal Achi or the recently discovered Zhang corpus of ritual texts.

All of these materials reflect—some more directly than others—their basis in oral composition. The regular recurrence of specific paired terms can at times be strikingly apparent but often it requires a considerable effort, directed at a large corpus, to decipher the full range of these pairings and to discern their use in a tradition of oral composition. Much the same can be said for many excellent collections of oral compositions, compiled and published without adequate attention to their underlying principles of composition.

The Biblical Scholarly Focus on Canonical Pairs

Biblical scholarship, from the time of Lowth, has focused extraordinary attention on the use of parallelism, but this attention has more often been directed to what Lowth termed ‘parallel lines’ or ‘parallel phrases’ than to ‘parallel terms’. Where the emphasis in this attention is to be placed remains a source of dispute. A considerable scholarship has been devoted to these issues, beginning with George B. Gray’s The Forms of Hebrew Poetry, which was first published in 1915 and then republished in 1972 with a substantial ‘Prolegomenon’ and extensive updated bibliography by David N. Freedman (Freedman 1972: vii–liii). Of equal relevance and comparative importance was Louis I. Newman and W. Popper’s three-part Studies in Biblical Parallelism published between 1918 and 1923 (Newman and Popper 1918–23). Of these volumes, Newman’s Parallelism in Amos (1918) was one of the first studies of its kind to point to comparative examples of parallelism in other Middle Eastern literary traditions.

A major impetus to the study of parallel terms came with the discovery in 1928 at Ras Shamra in the ruins of ancient Ugarit of a considerable corpus of Canaanite poetic texts dating from the fourteenth century BC. The eventual decipherment of these texts led to the recognition of a shared Hebrew–Canaanite tradition of composition based on recognisable paired terms, and this, in turn, opened a new wave of comparative scholarship that has continued to the present.2

Despite these new developments, differences in approach to the analysis of Hebrew parallelism persist. Thus, for example, James L. Kugel, in his book The Idea of Biblical Poetry: Parallelism and its History (1981: 1), asserts emphatically that the ‘parallelistic line’—‘a relatively short sentence-form that consists of two brief clauses’—is ‘the basic feature of Biblical songs’. By contrast, Stanley Gevirtz is equally emphatic in his ‘Prologue’ to Patterns in the Early Poetry of Israel:

More importantly, in the present context, it was found that the parallelistic structures evident in the Ugaritic poets were in all significant respects virtually identical with those known from Old Testament poetry. Still more central to the concerns of this present work was the recognition of a poetic diction common to the two literatures. Specific ‘pairs’ of words in fixed parallel relationship were found to occur in both Ugaritic and Hebrew literature with such frequency and regularity as to preclude the possibility of coincidence, while the differences in age and locale excluded the possibility of direct borrowing. (Gevirtz 1973: 3)

Gevirtz’s analysis of specific pairs is particularly illuminating and has led to further deepening of scholarly research. Thus, following on earlier work by Cassuto, Held and Gevirtz, Mitchell Dahood embarked on an effort to compile a comprehensive ‘thesaurus’ of what he identified as ‘Ugaritic–Hebrew parallel pairs’. His first compilation, prepared with assistance from Tadeusz Penar and included as an Appendix to his translation of Psalms (Dahood and Penar 1970: 445–56), consisted of 157 pairs of parallel words.3 His next compilation (Dahood and Penar 1972: 71–382), again with Tadeusz Penar, added a further 609 parallel pairs. This was followed by the addition of another 66 pairs (Dahood 1975: 1–39) with supplementary information on 18 entries, and then a further 344 pairs and supplementary information on 105 entries (Dahood 1981: 1–219), bringing his total lexicon to some 1,176 parallel terms. In compiling this lexicon, Dahood’s intention was to identify the semantic resources that poets of a common tradition relied on for their parallel compositions. He stated this clearly:

The present work aims to recover from all of the published Ugaritic tablets and from the Hebrew Bible, including Ecclesiasticus or Ben Sira, the Canaanite thesaurus from whose resources Ugaritic and Hebrew poets alike drew. (Dahood and Penar 1972: 74)

The Focus on Word Pairs in Mayan Languages

The Náhuatl and Mayan language-speaking area of Mesoamerica is another region of the world that is notable for its traditions of canonical parallelism. In his Pre-Columbian Literatures of Mexico, Miguel León-Portilla (1969), the doyen of pre-Columbian language studies, noted the similarities to Biblical parallelism. In his words:

[S]ome of the most frequent stylistic procedures … were more or less alike in all the various early Mexican literatures and show a certain similarity to the forms of expression used in other ancient compositions also preserved by an unbroken tradition, as in the case of the Bible and other texts from the Eastern cultures. Anyone who reads indigenous poetry cannot fail to notice the repetition of ideas and the expression of sentiments in parallel form. (León-Portilla 1969: 76–77)

What makes this region different is a remarkable continuity of oral traditions. Local social and religious activities combine with a vibrant living tradition of composition in parallelism. This is particularly so among the Mayan-speaking populations of Mexico and Guatemala.4 And nowhere is this continuity more significant than in the various studies of the Popul Vuh, or the Book of Counsel, of the Quiche Maya. With his collaborator, Earl Shorris, León-Portilla has described this extraordinary volume as one of the great treasures of literature:

The Popul Vuh is the best connection we have to a Mesoamerican civilization that has lasted for more than a thousand years, and continues still; it is a compilation of a way of being in the world, a book of gods and humans, a work for all ages. (León-Portilla and Shorris 2001: 401)

Munro S. Edmonson has recounted the history of the survival of the manuscript of Quiche in the introduction to his translation of the Popul Vuh: its composition about 1550, the copying and translation of this manuscript by the Dominican parish priest Francisco Ximénez, between 1701 and 1703, and the succession of nearly a dozen translations of this work into a variety of different languages. In offering his translation, Edmonson was explicit in his conception of the underlying basis for the composition of the work:

It is my conviction that the Popul Vuh is primarily a work of literature, and that it cannot be properly read apart from the literary form in which it is expressed. That this form is general to Middle America (and even beyond) and that it is common to Quiche discourse, ancient and modern, does not diminish its importance. The Popul Vuh is in poetry, and cannot be accurately understood in prose. It is entirely composed in parallelistic (i.e., semantic couplets). (Edmonson 1971: xi)5

The opening lines of the Edmonson translation of the Popul Vuh illustrate this strict canonical/semantic parallelism:

This is the root of the former word

Here is Quiche by name

Here we shall write then,

We shall start out then, the former words,

The beginnings

And the taproots

Of everything done in Quiche town,

The tribe of the Quiche people.

So this is what we shall collect then,

The decipherment,

The clarification,

And the explanation

Of the mysteries

And the illumination

By the Former

And Shaper;


And Engenderer are their names,

Hunter Possum

And Hunter Coyote

Great White Pig

And Coati,


And Quetzal Serpent,

The Heart of the Lake

And Heart of the Sea,

Green Plate Spirit

And Blue Bowl Spirit, as it is said,

Who are likewise called,

Who are likewise spoken of

As the Woman with Grandchildren

And Man with Grandchildren


And Xmucane by name,


And Protector,


And Great-Grandfather

As it is said

In Quiche words …

In a critically important paper, ‘Semantic Universals and Particulars in Quiche’ (Edmonson 1973), following the publication of his translation, Edmonson examined in detail the semantic parallelism in the first 94 lines of the Popul Vuh. Forty-nine pairs occur in these lines, of which there are five repeated pairs, thus reducing the number to 42 semantic pairs. These pairs—a small number in a work of more than 8,500 lines—are illustrative of the canonical pairs that form the basis of Quiche composition.

In his analysis, Edmonson grouped these pairs that he had identified into three classes: those that he considered ‘universal’ such as word//name or heaven//earth, root//tree or mother//father; those he considered ‘widespread’ such as bowl//plate, plant//root, lake//sea or tribe//town; and those that he considered ‘particular’ to Quiche culture such as possum//coyote, pig//coati, majesty//quetzal or heart//breath. He also noted that certain terms, as, for example, the term in Quiche translated as ‘word’, formed several different pairs: with the noun ‘name’, but also with the verbs ‘to say’, ‘to do’, ‘to describe’ and ‘to be’ as well as with the word for ‘clear’.

Edmonson’s short exploratory illustration of possible ways of analysing Quiche semantic couplets led to a much more thorough examination of these pairs in a Leiden PhD thesis, ‘The Poetic Popul Vuh: An Anthropological Study’, by Robert de Ridder (1989). With close attention to Edmonson’s translation, de Ridder provides an interpretation of the Popul Vuh based on a thoroughgoing analysis of specific word pairs, grouping them together and contrasting them with one another. Like Edmonson’s paper, de Ridder’s thesis is an exploration of the possibilities open to analysis by careful focus on culturally defined semantic equivalences.

Dennis Tedlock, a student of Edmonson, has produced his own translation of the Popul Vuh (Tedlock 1996) but he has done this in consultation with contemporary Quiche ritual officials whose interpretations allowed him to consider the Popul Vuh in terms of the possibilities of performance.6 In a long paper, ‘Hearing a Voice in an Ancient Text: Quiché Maya Poetics in Performance’ (1987), Tedlock has described this cooperative ‘ethnopaleographic project’ in which he even attempts ‘a reconstruction of the oral delivery’ of particular passages from the Popul Vuh. His paper is a model of nuanced ethno-poetic analysis in which he is able to show similarities between passages in the Popul Vuh and the prayers recited by Quiche ritual specialists. For him, however, Edmonson ‘let the search for parallel couplets dominate his entire reading of the text’ (Tedlock 1987: 147). Tedlock is interested in Quiche triplets as well as couplets and in all sorts of syntactic shifts that occur between couplets. As he makes clear, recognition of canonical pairs is essential for an understanding of Quiche composition, but the use of the pairs and triplets can occur in varying syntactic contexts. As in the case of Biblical scholars in their examination of both Hebrew and Ugaritic texts, it is essential to provide a closely focused analysis of the various patterning within individual texts and variations in this patterning across texts.

Research on Mayan parallelism, which goes well beyond the research on the Quiche language, and similar research within Biblical scholarship represent two distinct fields of investigation—largely independent of one another—that are directed towards comparable ends and are suggestive of processes of oral composition. Both broad fields of research relate to and offer insights on similar processes that continue to be used by contemporary oral poets in eastern Indonesia.7 This study of the oral composition among the oral poets of Rote is intended to carry forward this analysis.

A Focus on Semantic Parallelism on Rote

Since 1971, I have produced a succession of papers that examine the use of semantic parallelism in Rotenese oral compositions, and I have continued to rely on a number of conventions that I adopted in the first paper I published, ‘Semantic Parallelism in Rotinese Ritual Language’ (Fox 1971). Because all oral compositions based on canonical parallelism occur or are intended to occur in a ritual context and generally allude to matters of ritual significance, I have consistently referred to these oral compositions as a form of ‘ritual language’. In Rotenese, the term bini is used to describe all poetic compositions in ritual language. These bini are then differentiated according to how they are performed—whether chanted, sung or recited with the accompaniment of a drum or according to their ritual context: an origin recitation, a funeral recitation, a lament and so forth.

From the outset, I have adopted, from Biblical scholarship, the use of the conventional ‘//’ to identify canonical pairs and refer to all such pairs as ‘dyadic sets’. A dyadic set consists of two semantic elements, x and y, but any element can occur in more than one dyadic set.8 In semantic analysis of the lexicon of Rotenese ritual language, it is important to distinguish elements that form only a single dyadic set from those that form a ‘range’ of dyadic sets.

The following short poem from the dialect of the domain of Termanu provides an illustration of the use of pairs in this canonical parallelism. Analysis of this poem provides an opportunity to introduce the basic terminology and notation system that I use in this volume:


Lole faik ia dalen

On this good day


Ma lada ledok ia tein na

And at this fine time [sun]


Lae: tefu ma-nggona lilok

They say: The sugar cane has sheaths of gold


Ma huni ma-lapa losik.

And the banana has blossoms of copper.


Tefu olu heni nggonan

The sugar cane sheds its sheath


Ma huni kono heni lapan,

And the banana drops its blossom,


Tehu bei ela tefu okan

Leaving but the sugar cane’s root


Ma huni hun bai.

And just the banana’s trunk.


De dei tefu na nggona seluk

But the sugar cane sheaths again


Fo na nggona lilo seluk

The sheaths are gold again


Ma dei huni na lapa seluk

And the banana blossoms again


Fo na lapa losi seluk.

The blossoms are copper again.

This poem is composed of just eight semantic pairs or ‘dyadic sets’. These dyadic sets are listed here together with a simple notation to distinguish them:










sugar cane//banana:















The poem also has various connectives, emphatics, time markers and verbal elements that are not subject to pairing: ia (‘this’), ma (‘and’), tehu (‘but’), de (‘but that’), fo (‘that’), heni (‘away’, ‘off’), bei (‘still’), bai (‘also’), seluk (‘again’), lae (‘they say’).

Using the notation (a1//a2) to designate the various dyadic sets in this poem, it is possible to discern the poem’s formulaic ‘orderedness’:

Lole faik ia dalen

  a1  b1        c1

a1, b1, c1

Ma lada ledok ia tein na

        a2    b2       c2

a2, b2, c2

Lae: tefu ma-nggona lilok

       d1           e1       f1

d1, e1, f1

Ma huni ma-lapa losik.

       d2           e2    f2

d2, e2, f2

Tefu olu heni nggonan

  d1  g1             h1

d1, g1, h1

Ma huni kono heni lapan

       d2     g2            h2             

d2, g2, h2

Tehu bei ela tefu okan

                    d1   h1

d1, h1

Ma huni hun bai.

       d2    h2

d2, h2

De dei tefu na-nggona seluk

           d1           e1

d1, e1

Fo na-nggona lilo seluk

            e1       f1

e1, f1

Ma dei huni na-lapa seluk

            d2          e2

d2, e2

Fo na-lapa losi seluk.

           e2    f2

e2, f2

A key to understanding these pairs and the way they function in composition is to recognise the varying range of the elements of which they are composed. For example, lole meaning ‘good, beautiful’ can also pair with (ma-)na’a, meaning ‘striking, handsome, pretty’. Similarly, fai(k) (‘day’) also pairs with leoda’e (‘night’); the verb olu (‘to shed’) forms a pair with tui (‘to drop leaves’); while both oka and hu occur in multiple pairs: oka (‘large root’) with samu (‘small root, tendril’) and with polo (‘shoot, growing tip of a plant’), and hu (‘trunk, base’) with boa (‘fruit’) and do (‘leaf’). Only tefu//huni (‘sugar cane’//‘banana’) forms a single, exclusive dyadic set. That sugar cane and banana are icons of male sexuality adds a further cultural dimension and sheds light on this short poetic passage, which is a metaphorical statement about male succession within a lineage.

The semantic pairs used in Rotenese compositions are not a collection of terms that are exclusive to a single pair. Underlying a large percentage of these pairs is a semantic network of connections. These connections are limited and culturally circumscribed. The knowledge of such pairs is a requirement of composition and learned within a particular speech community. Unacceptable pairing is quickly noted in a performance and usually leads to hissing a poet into silence. Previously, false pairings were considered to cause a poet sickness and even death for a ritual mistake. This poem is short and thus provides a simple illustration of the composition process. Most poetic compositions range from 100 lines to many hundreds of lines and correspondingly require the proper pairing of several hundred specific terms in all grammatical categories. The composition entitled Dela Kolik ma Seko Bunak that I analysed in detail in ‘Semantic Parallelism in Rotinese Ritual Language’ (Fox 1971) consisted of 224 lines that utilised 123 distinct dyadic sets plus seven dyadic chant character names and two dyadic place names.9

The Speech Communities of Rote

The island of Rote—referred to in ritual language as Lote do Kale—is located off the western tip of the island of Timor. It extends roughly in an east–west direction so that what the Rotenese call the island’s ‘head’ (langa) is ‘east’, and its ‘tail’ (iko) is west. These directional coordinates provide the basis for all ritual action in Rotenese poetic discourse. The island’s population at present (2015) has grown to approximately 100,000.

The ‘language’ of the Rotenese constitutes a complex dialect chain stretching from one end of the island to the other. While neighbouring communities can understand each other, differences increase in terms of the distance of these communities from each other. At the two ends of this dialect chain intelligibility is a challenge.10

Rote’s political history has contributed to this dialect diversity. In 1662, the Dutch East India Company began establishing contracts of trade with local rulers, thus recognising their authority in particular areas on the island. Initially, 12 domains were recognised. This process of Dutch recognition, however, continued through the eighteenth and into the nineteenth centuries, eventually dividing Rote into 17 separate domains plus a domain that embraced the offshore island of Ndao.11 As a result, most Rotenese domains, referred to in Rotenese as nusak, have possessed a distinct social and political continuity since the seventeenth century.12

Each domain maintains its own traditions, claims its own special narratives and asserts the superiority of its rituals and its practices. Until the late 1960s, each domain had its own court to adjudicate on disputes based on its domain-specific customary law. It is a dogma among Rotenese on the island that all but one of these domains possess their own language. As a consequence, individuals often go to great lengths to elevate minor variations as evidence of major differences among domains. On an island with subtle dialect variation, the domains constitute distinct speech communities.

Each domain has its own ritual name—in fact, multiple ritual names—most of which are specific to a particular location within the domain. Among the poets, there is general, but by no means universal, consensus about the principal name of a domain. Although these names may vary slightly, as articulated by poets in their domain’s dialect, they are broadly recognised throughout the island and are indispensable to ritual recitations.

Figure 2: Domain Map of Rote

Source: © The Australian National University CartoGIS

This is the list of the principal names of the domains according to Termanu dialect.

The Principal Ritual Names of the Domains of Rote


Soti Mori ma Bola Tena


Fai Fua ma Ledo Sou


Londa Lusi ma Batu Bela


Pengo Dua ma Hilu Telu


Pele Pou ma Nggafu Lafa


Lenu Petu ma Safe Solo


Meda do Ndule


Tunga Oli ma Namo Ina


Koli do Buna


Tufa Laba ma Ne’e Fe’o


Pila Sue ma Nggeo Deta


Pena Pua ma Maka Lama


Dae Mea ma Tete Lifu


Nggede Ke ma Danda Mamen


Ninga Ladi ma Hengu Hena


Tada Muli ma Lene Kona


Tasi Puka ma Li Sonu


Dela Muli ma Ana Iko

All of this is highly significant for understanding traditions of composition and performance. All poets identify with their particular domains and regard themselves as responsible for conveying the ‘true’ ancestral traditions of their domain. Thus, each poet’s composition is firmly situated in a particular speech community and is judged by members of that community. When poets from one domain encounter poets from other domains, they expect differences in the content of a composition (and would probably be upset were the composition from another domain to mirror too closely that of their own), but they are particularly attuned to the dyadic sets used in composition. Rotenese ritual language transcends the dialect differences among domains by utilising these differences as an essential component of its lexicon.

A considerable number of dyadic sets, particularly for synonyms, are composed of words taken from different dialects. Which word is the ‘dialect’ term and which word is ‘local’ depends on the particular speech community. A few examples may suffice. Adopting the perspective of the speech community of the domain of Termanu, in the dyadic set for ‘human being’ or ‘person’, hataholi//daehena, hataholi is the local term, whereas daehena comes from various dialects in the east of the island. In the dyadic set for ‘cave’ or ‘grotto’, lea(k)//lua(k), lea(k) is the local term in Termanu, whereas lua(k) comes from the dialect of Korbaffo. Similarly, in the dyadic set for ‘enough’ or ‘sufficient’, henu//sofe, henu is the Termanu term, while sofe comes from the dialect of the south-western domain of Thie. It is possible that as much as 20 per cent of the lexicon of the ritual language used in any one speech community is made up of words from other dialects.13

In this study, an initial focus is on composition from the domain of Termanu. From Termanu, analysis shifts to similar compositions in other domains: Ba’a, Ringgou, Dengka and Thie.14 This allows further insight into the variation in composition across dialects and speech communities.

The Poet as Custodian

This is a comparative study of the compositions of specific poets. The individuals to whom I refer as poets are, on Rote, known as manahelo (‘chanters’ or ‘those who chant’). Whereas this general designation privileges one modality of performance, chanters can also sing (soda), speak (kokola) or tell (tui) their compositions. Sung recitations may be accompanied by the sasandu, a stringed bamboo instrument encased in a surrounding lontar leaf, or by the drum (labu). There is a variety of modes of performance, some specific and appropriate to the particular composition, and a general recognition that certain poets are more talented than others in singing or chanting.

Although the style of performance is important, the emphasis among all poets is on their substantive knowledge. Poets are primarily the custodians of a coded ancestral wisdom. They are all ‘men of knowledge’ (hataholi malelak) whose principal task is to communicate this ritual knowledge without alteration. As manahelo, they are judged by their fellow Rotenese on the depth of their ritual knowledge and this, in turn, is most evidently communicated by the quality and coherence of their compositions and their extensive citation of ritual names. As a consequence, with few exceptions, master poets are elders and the knowledge they convey in their compositions is a cultural accumulation acquired over decades. Relatively few individuals have the talent and inclination to accumulate the knowledge and fluency necessary to recite but there are individuals, usually prompted and facilitated by older relatives with whom they associate, who embark on this path and begin to emerge as promising poets when they are in their 40s or 50s. Gradually, by the time they are 60 or 70, they come to be regarded locally as capable manahelo. Recognition is locally conferred and invariably limited to a select number of individuals. At present, on Rote, there are only a few master poets in any of the island’s speech communities.15 This study will examine the compositions of some of these master poets and compare them with each other.

The Canon of Ritual Knowledge

Rotenese parallel poetry relates to the rituals of life on the island. All poetic compositions are referred to, in Rotenese, as bini. In turn, these bini are described according to a combination of subject matter, ritual context or performative mode, which usually identifies a specific ritual context. Bini can be used on all ceremonial occasions of the life cycle as well as on occasions of formal interaction, such as the greeting or farewelling of guests, making requests to superiors, installing officials or the negotiation of bride-wealth. Formerly, bini were also recited at the initiation of new activities: the planting and harvesting of fields; the beginning of tying, dyeing and weaving of a new cloth; the start and ceremonial conclusion to the construction of a house; and, most importantly, at the annual ‘origin’ ceremonies (known as hus or limbe) in each domain. Most of these rituals have now ceased to be performed, but the knowledge of the composition associated with them continues to be maintained in poetry.

There are two major categories of bini: origin bini and mortuary bini. These categories are related to each other in terms of the rituals of life and death. Distinguishing them, however, provides the basis for an initial understanding.

Knowledge among the Rotenese is identified with the recognition of origins. To know and to be able to trace the origin of things constitute true cultural knowledge. All Rotenese activities as well as all of the key items of daily use, from rice and millet or water buffalo to the tools for building a house or weaving a cloth, possess distinctive ‘origins’ that are referred to and recounted in the poetry of the bini. The knowledge and recitation of this poetry upholds and ritually enlivens these Rotenese activities and links them properly to an ancestral past.

There is coherence to the recitation of these origins. Although never recited as a single successive recitation, most origin bini relate to what might be described as an epic account of relations between the Sun and Moon and their descendants and the Lords of the Sea with their adherents. These heavenly creatures and their counterparts in the ocean engage with each other, hunting together, exchanging goods, threatening war and intermarrying. This engagement occurs on the ‘dry land’ that divides them from each other and generates the cultural goods and practices that define Rotenese life. All of these goods, including fire for cooking, come from the sea. Hence most Rotenese rituals were concerned with the celebration of these oceanic origins.16 After acknowledging these origins, a particular feature of many Rotenese poetic compositions involves the systematic recitation of the transmission and propagation of these goods throughout the island.17

In contrast with the various origin bini, there is less coherence and more diversity among the mortuary bini. All of these bini, however, follow similar ritual formats. In one such format, the spirit of the deceased is addressed directly, his or her life alluded to and the sorrow and distress of grieving relatives are emphasised; then the deceased is told to depart—to board the ship of the dead and sail westward—but at the same time reminded that it will return in spirit form to come among relatives. In another ritual format, the deceased is compared with a particular named chant character and the stereotyped life of this character is recounted. In some of the more elaborate of these compositions, the deceased is given voice to admonish his relatives before sailing to the west.18 The use of these different formats varies from domain to domain and is now heavily influenced by Christian rituals and the use of Christian poetic parallelism. The domain of Termanu, in particular, has preserved a significant canon of mortuary bini identified with specific chant characters whose life course is the focus of the chant. The repertoire of these chants fits different social categories. There are set mortuary chants for nobles and for commoners; for young nobles or rich commoners; for girls who have died as ‘unripe’ virgins; and, above all, for ‘widows and orphans’.19 ‘Widow and orphan’ is a general category that invokes a particular view of life and can be used to fit virtually all human circumstances. There is thus a variety of ‘widow and orphan’ chants and, to make matters more complex, some origin bini can be recomposed and transformed into ‘widow and orphan’ mortuary chants. The chant Suti Solo do Bina Bane that is to be considered exhaustively in this study is a prime example of this potential for poetic transformation.

Suti Solo do Bina Bane as an Oral Composition

In an oral tradition, there is no one definitive composition. On Rote, each composition is judged by internal qualities and coherence supported by the authority of the particular poet who stands as custodian of an ancestral canon. There are many versions of Suti Solo do Bina Bane to be considered in this study. They are clearly related to one another but there is also variation not just in their narrative structure but also in the ritual focus of their composition.

Suti Solo do Bina Bane belongs to the class of Rotenese origin narratives and recounts the arrival of further beneficence from the sea. Suti in the name Suti Solo refers to a nautilus shell and Bina in the name Bina Bane refers to a bailer shell. In the poem, these two shells are personified as creatures from the sea who are washed up onto the tidal flats, gathered by two women while fishing and then carried from one location to another. In the origin versions of this composition, these creatures are eventually made into specific cultural objects. The nature of these cultural objects varies according to the speech community (nusak) in which this composition is told. In origin versions of the poem, there exists an esoteric dimension surrounding the reasons why the nautilus and bailer shells are ‘expelled’ from the depths of the sea. Poets who recite Suti Solo do Bina Bane as an origin narrative variously allude to this hidden aspect of the canon—some more fully than others—but in the end, this dimension remains shrouded in mystery.

Suti Solo do Bina Bane can also be recited as a ‘widow and orphan’ mortuary chant. In these versions, the displacement of the shells—generally as the result of a violent storm—and then their quest to find an appropriate social setting are taken to represent the human condition. Versions of this kind emphasise the search for companionship and generally end with the return of the shells to the sea. This kind of composition, when well composed, makes an ideal mortuary chant.

Figure 3: Suti Solo and Bina Bane

‘Then I, Bina, with whom will I be

And I, Suti, with whom will I be

With whom will I talk

And with whom will I speak?’

It is in the nature of Rotenese oral composition for a poet’s recitation to be ambiguous. The poet, as the custodian of a traditional knowledge, must allude to this ancestral knowledge but he does not expound at length nor make explicit all aspects of that knowledge. No composition is therefore entirely explicit. Indeed, the best poets are those who cleverly hint at what is to be known without openly revealing it.

In my experience, there is hardly a recitation that does not prompt discussion afterwards, with various interpretations and attempts at exegesis of its meaning. I was fortunate, when I first arrived on Rote and began trying to understand ritual language, to have the elderly poet Old Meno to guide me. I suspect that several of his recitations for me were possibly more elaborate than would have been the case if I had not been so naive and unknowing. For some of his compositions, even Old Meno would leave things unsaid, but he would hint at what he had omitted or left ambiguous and would then gently lead me towards some understanding of what this might be. Rotenese poetry is thus based on a different kind of pedagogy: one in which meaning is progressively encircled until, for some but not all, key ideas are grasped, although they may never be articulated.

By this same token, no recitation is a complete recitation.20 Often among a group of poets, when one has finished his recitation, another will speak up and provide, from his perspective, other elements of the composition. Oral composition is thus a continuing process but can also be seen as part of a joint effort by different custodians of knowledge to contribute to the weaving of a larger whole.

1 See Fox (1977) for detailed documentation on the variety of the world’s traditions of parallelism and Fox (1988, 2005) for detailed summaries of the Austronesian traditions of parallelism.

2 Freedman, in the bibliography of his ‘Prolegomenon’ (1972), provides an excellent listing of useful publications on Ugaritic–Hebrew comparative scholarship to that date. In a valuable PhD thesis submitted to the University of Otago, Margaret R. Eaton (1994) examines the development of Biblical research on parallelism and word pairs and provides a further bibliography on this topic.

3 Dahood distinguishes between ‘strict parallels’, ‘collocations’ and ‘juxtapositions’, and includes all three of these categories in his Ugaritic–Hebrew parallel pair compilations. This is an important methodological distinction. By strict parallels, he refers to pairs found in both Ugaritic and Hebrew. Collocations and juxtapositions refer to terms that are found as pairs in one of the two languages but nevertheless occur in close association or juxtaposition in the other language. Most of his citations are strict parallels.

4 For a classic study of these traditions, see Gossen (1974).

5 It is important to note that Edmonson (and various other authors writing on Mesoamerican parallelism) uses the term ‘couplet’ for what Dahood refers to as strict pairs or parallel pairs or what other Biblical scholars have called ‘fixed pairs’. In his work on parallelism, Edmonson—as indeed León-Portilla—was inspired by K. Garibay (1971), who focused on the importance of parallelism in his Historia de la Literatura Nahuatl. Garibay coined the term ‘difrasismo’ (diaphrasis) to describe the pairing of two metaphors to express a single thought.

6 There are more than 750,000 ‘Quiche’ living in Guatemala today, many of whom continue to maintain their traditional religious practices. Current classification divides ‘Quiche’ among at least six different but related languages, of which Central or South-Western Quiche has the largest number of speakers, with more than half a million.

7 For a compilation of papers on parallelism in different languages in eastern Indonesia, see Fox (1988).

8 By this convention, any so-called triplet that occurs in a composition can be represented simply as (a, b) + (a, c); such sets can be extended if particular elements form a variety of different pairs.

9 Each of these sets is identified and translated in the paper Fox (1971: 252–55).

10 One of the first Rotenese to write about his own language, D. P. Manafe, a schoolteacher attached to a middle-level school in the town of Ba’a on Rote, prepared a document, ‘Akan Bahasa Rotti’, in 1884. This was eventually published by the linguist H. Kern in the Dutch journal Bijdragen tot de Taal-, Land- en Volkenkunde in 1889 (see Kern 1889). In his paper, Manafe was at great pains to point out the considerable dialect variation on the island. He insisted, however, that Rotenese in the east of the island could, despite difficulties, still understand Rotenese in the west of the island. Some 120 years later, this proposition is questionable for the dialects at either end of the island and may have already been so at the time that Manafe wrote. Certainly some of the dialects within Manafe’s east–west division of dialects are mutually intelligible but not necessarily all of them.

11 The population of the island of Ndao possesses a distinct language that has been heavily influenced by Rotenese but is more closely related to Savunese. Ndao figures prominently in the traditions of Rote; its ritual name is Ndao Nusa do Folo Manu. Most Ndaonese are bilingual and participate in Rotenese culture. A number of Ndaonese are master poets who recite in Rotenese ritual language.

12 For a discussion of the history of these political developments on Rote, see Fox (1971, 1977, 1979a, 1979b).

13 It is essential at the outset of this study to emphasise the pervasive importance of dialect variation for the compositions of any particular speech community. However, from an island-wide perspective, the use of dialect variation is a more complex subject for analysis, involving a concatenation of dialect forms. Thus, to take a simple example, the word for ‘person’ in Termanu is hataholi//daehena; in the eastern dialect of Ringgou, it is hataholi//laehenda; whereas in the western dialect of Dengka, it is hataholi//andiana.

14 In this study, I rely on the historical (Dutch) spelling of the names for Rotenese domains—a convention I have followed in other publications. There is a variety of ritual names for each domain as well as other names. Use of the historical spelling for specific domains allows reference to a literature on the island that dates to the middle of the seventeenth century.

15 Nineteenth-century Dutch literature on the island suggests that there was much more mobility among chanters than at present. Large funeral ceremonies would attract poets from other domains who would compete with each other in mortuary chanting and would often be handsomely rewarded with pieces of gold placed on their tongues.

16 I have discussed some of these origin bini in a number of publications: 1) the origin of fire (Fox 1975); 2) the origin of textile patterns (Fox 1980); 3) the origin of the house (Fox 1993); and I have tried to outline the structure of the Rotenese ‘epic’ narrative in Fox (1997a).

17 I use the term ‘topogeny’ to refer to the ordered recitation of a succession of place names. Such topogenies are common in chants that recount the origins of particular goods that come from the sea, such as rice and millet. See Fox (2007a).

18 I have discussed one such chant in Fox (2003).

19 The Rotenese concept of ‘widow and orphan’ carries with it a weight of understanding that will be discussed, at various points, throughout this study.

20 It is possible that previously in the performance of critical rituals, there may have been stricter rules of exposition in communing with the ancestors. Now that the compositions are recited outside their ritual context, this is no longer the case. Disembodied from their ritual context, recitations take on new significance.

Previous Next