Textualization of Oral Epics 9783110825848, 9783110169287

Table of contents :
Preface
Introduction
Text as process and practice: the textualization of oral epics
European epics
The writing of the Iliad and the Odyssey
The textualization of South Slavic oral epic and its implications for oral-derived epic
Performance, textualization, and textuality of “elegy” in Old Norse
Turkic and Siberian epics
Silencing the voice of the singer: problems and strategies in the editing of Turkic oral epics
Textology and epic texts from Siberia and beyond
“I lift you up, the dry throats” – on Nanaj shamanic epic
Indian epics
The textualization of the Sanskrit epics
Text and context in the textualization of Tulu oral epics
African epics
Authenticity and oral performance: textualizing the epics of Africa for Western audiences
The textualization of Swahili epics
Creating an epic: from apprenticeship to publication
The narrator as an editor
North-American and Oceanian epics
Sung epic and Native American ethnopoetics
Generic models, entextualization and creativity: epic tradition on the Southern Cook Islands
Index
List of contributors

Citation preview

Textualization of Oral Epics

W G DE

Trends in Linguistics Studies and Monographs 128

Editor

Werner Winter

Mouton de Gruyter Berlin · New York

Textualization of Oral Epics

edited by

Lauri Honko

Mouton de Gruyter Berlin · New York

2000

Mouton de Gruyter (formerly Mouton, The Hague) is a Division of Walter de Gruyter GmbH & Co. KG, Berlin.

© Printed on acid-free paper which falls within the guidelines of the ANSI to ensure permanence and durability.

Die Deutsche Bibliothek

— Cataloging-in-Publication

Data

Textualization of oral epics / ed. by Lauri Honko. - Berlin ; New York : Mouton de Gruyter, 2000 (Trends in linguistics : Studies and monographs ; 128) ISBN 3-11-016928-2

© Copyright 2000 by Walter de Gruyter GmbH & Co. KG, 10785 Berlin All rights reserved, including those of translation into foreign languages. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording or any information storage and retrieval system, without permission in writing from the publisher. Cover design: Christopher Schneider, Berlin. Printing: Werner Hildebrand, Berlin. Binding: Lüderitz & Bauer, Berlin. Printed in Germany.

Contents Preface

vii

Introduction Lauri Honko Text as process and practice: the textualization of oral epics

3

European epics Minna Skafte Jensen The writing of the Iliad and the Odyssey John Miles Foley The textualization of South Slavic oral epic and its implications for oral-derived epic Performance, textualization, and textuality of "elegy" in Old Norse Joseph Harris

57 71 89

Turkic and Siberian epics Karl Reichl Silencing the voice of the singer: problems and strategies in the editing of Turkic oral epics Arthur T. Hatto Textology and epic texts from Siberia and beyond Juha Pentikäinen "I lift you up, the dry throats" - on Nanaj shamanic epic

103 129 161

Indian epics John Brockington The textualization of the Sanskrit epics Lauri Honko Text and context in the textualization of Tulu oral epics

193 217

African epics John William Johnson Authenticity and oral performance: textualizing the epics of Africa for Western audiences

237

vi

Contents

Jan Knappert The textualization of Swahili epics Dwight F. Reynolds Creating an epic: from apprenticeship to publication Dan Ben-Amos The narrator as an editor

247 263 279

North-American and Oceanian epics Dell Hymes Sung epic and Native American ethnopoetics Anna-Leena Siikala Generic models, entextualization and creativity: epic tradition on the Southern Cook Islands

343

Index

371

List of contributors

389

291

Preface The task of textualizing oral epics in writing is a mission impossible. Oral performance cannot be captured in letters and words. Too many essential features are simply left aside in the written codification of a speech event which normally employs a wide array of paralinguistic means of expression from gesture to music. Even the most meticulous notation of all dimensions of the original oral performance does not reflect the intertextual construction of meaning, the core process of reception, manifest before the eyes of the outside observer but not seen by him because of a lack of traditional knowledge. The potential of the oral epic performance to open up several channels of communication simultaneously and mould the shared tradition into a novel experience packed with relevance for the participants has been largely underestimated or neglected. The written codification of oral expression creates a kind of epic different from that experienced by the original audience. The route from an "incomplete and unfinished" oral epic manifest in performance to a complete written codification of its story manifest in a book does not, however, represent a process of decay but an intersemiotic translation. The focus on oral verbalization may be said to liberate the oral form in an important dimension, namely, the language. What we get is a new coherence of the story, a full exploitation of the local epic register through the vision of preferably one singer utilizing his/her epic idiolect and interpretive skill. The miracle of the process is that what we experience as literary value or beauty is there in the original oral textualization and is merely magnified, not created, in the written codification. The linguistic power of the oral genre becomes accentuated in the new non-oral form capable of living on as a piece of literature proper. These facts have dawned on epic scholars only recently. The present volume reflects the awakening among top epic scholars, a process of new understanding taking place slowly along converging routes and with slightly different emphases. The days are past when a scholar sought for a "master form" by combining elements from different singers of epics, sometimes from different regions and eras, too. Such composite texts were in danger of gliding outside the local poetic system. Their connection to sung performance was lost or skewed. The reaction of modern scholarship has been to stay as close to the oral rendition of an "epic text" as possible and to listen carefully to the poet's voice. In practice, this has led to an emphasis on the singer's vision as a unifying force in the sequencing of traditional elements and his/her construction of meaning for the epic. The new demands for accuracy and open reporting on methods in field documentation are a corollary of the source-critical concern, especially when

viii

Preface

it has emerged that long oral epics are obtainable in induced contexts only, i.e. outside or in the margin of the social processes which normally encircle and set the limits to the epic performance. The whole epic is available only as non-transferr able mental texts in the minds of individual singers. Mental, not composite text may function as the frame of reference for modern research, because it is the element uniting the different performances of a particular epic by a particular singer. Yet it is neither fixed nor stable and develops throughout the singer's performing career. Since the mental text is manifest only in varying forms at different performances, the problem of editing cannot be eliminated totally, although the accurate documentation and publication of one performance is a basic method for fieldwork. The scholar will have to justify in detail the form in which he believes the epic may be presented as a reflection of the singer's vision. The present volume is based on a selection of papers given at a conference on "Textualization of oral epics" in Turku, Finland, in June 1996. The conference belonged to a series of annual seminars arranged since 1991 at the University of Turku for the Folklore Fellows in Oral Epics, a scholarly network consisting of about 70 active epic scholars in all parts of the world. I wish to thank the sponsors, the Alfred Kordelin Foundation, Helsinki, and the Academy of Finland, Helsinki, for financing a major research programme on oral epics in southern India as well as the international seminars in Turku. The Kalevala Institute, newly established at the University of Turku, kindly hosted the final editing of the book and will sponsor comparative epics research in the future, too. I am most grateful to Anneli Honko, Editorial Secretary of the Kalevala Institute, for harmonizing the manuscript, and to Susan Sinisalo for her linguistic assistance with the editing. Turku, Christmas 1999 Lauri Honko

Introduction

Text as process and practice: the textualization of oral epics Lauri Honko The concept of oral text has experienced a revolutionary development in recent years. What used to be an innocent object of research, a verbal transcript of an orally performed traditional song (with or, more commonly, without musical notation), has been problematized from a variety of angles by questioning its boundaries, apparent fixity, performative representativity, situational and cultural contextuality, co-textual and intertextual environment, discursive function and ideological bias. The modest transcript has undergone acute source-criticism: its textual origin and linguistic accuracy, its methods of documentation, transcription, translation, editing and publication have been subjected to scrutiny, not forgetting the singer's "voice" (always in danger of suppression), the collector's purposive role in the making of the text and the editor's impact on the final form. The demand is that the contours of the singer's and the scholar's interpretations be made clearly visible and reviewable apart from each other. Direct quotations from the singer's speech, preferably in vernacular transcription, are more dependable than scholarly resumes based on failing memory and defective textualization. Paradoxically, the quest for authenticity has not necessarily led to "purer" texts but to a relativization of the concept of authenticity and, hopefully, to a better understanding of the multifaceted processes of oral and written textualization. Scholars, perhaps more inquisitive than ever, tend to demand a full history of textualization for texts claiming orality and traditionality. The new wave of interest in textualization must be seen in the light of changes of attitude toward the concept of text in particular research traditions. Comparative research on oral epics is a good field of experimentation in textualization for several reasons. It is multidisciplinary, highly dependent on textual documentation and faces, e.g. in the analysis of long oral epics, very concrete problems of documentation and text-making. There are scholars who doubt that long oral epic is even possible. As a young scholar, I myself was educated to believe that long epic, based on folk poetry like the Kalevala, is a product of written compilation. In 1949 an expert of Vladimir Propp's stature could still claim that "the people never create an epic... the true epic always consists of isolated songs which the people do not join together..." (Propp 1984 [1976]: 73-73). Soon after that the breakthrough of long oral epics documented in the Soviet Union, Central Asia, Africa, India, Oceania and other areas where the long format was found alive in oral tradition led

4

Lauri Honko

to a paradigm shift which in my case resulted in intensive fieldwork among the Tulu speaking people in southern Karnataka, India, and eventually to the publication of the Siri epic (Honko 1998:18-19). Since oral epics research is multidisciplinary, the changes in attitudes toward textualization need not be fully identical or simultaneous in the participating disciplines, such as sociolinguistics, different branches of philology, folkloristics, cultural anthropology, comparative literature and so on. One of the reasons for inviting representatives of several disciplines interested in oral epics to meet in Turku in June 1996 was to let them compare their text concepts and report on changes which they had experienced in their own field of research. Somewhat fittingly, the conference became a platform for rarely heard reports on textualization, oral and written, with all its methodologically problematic implications. The purpose of this introduction is not to focus on the papers in this volume, which are fully capable of speaking for themselves. (Abrief characterization of their contribution will be given at the end.) My main task will be twofold: first, to survey the shifts of text paradigms in my own field, folkloristics and comparative religion, and second, to relate their impact on research on long oral epics by offering a processual model of textualization. The situation of folkloristics as regards textualization is as problematic as that of printed oral poetry. Both the discipline and its material are based on oral culture but bound to written texts, i.e. archived or published documents on oral performance. Written texts constitute a necessary precondition for scholarly analysis on orality. Before profiling the conceptualization of text within folklore research it may be healthy to explore the line of demarcation between folklore and literature through the eyes of earlier scholarship. This line is of considerable interest for all disciplines dealing with oral tradition, mostly classified as the hazy marginal zone of literature proper, a kind of pre-literature or "unwritten" literature.

Problematic stereotypes: oral tradition and literature The sharp divide between orality and literacy, once drawn by Romanticists frowning on the impact of a literary, individual hand on folk poetry, the product of collective creation, is a thing of the past. In recent decades it has been heatedly defended by scholars such as Albert B. Lord (1960) and Walter J. Ong (1982) but more generally the quest for purely oral cultures, uncontaminated by the art of writing, has given way to views which avoid the opposition and establish forms of cohabitation for orality and literacy (Goody 1987) or disclose literary influence and/or individual authorship behind many an "ancient" oral tradition (Finnegan 1988). What has emerged is an abundance of "interactive" forms and, at least in the case of long and complex

Text as process and practice

5

oral performances, an individual impact to the effect that reciter becomes author. Oral textualization may accommodate literary pieces and literary works may contain traditional elements in oral style. Illiterate singers are capable of handling information stemming from literary sources. Literary and oral performance traditions interact in most cultures and have done so for thousands of years in ancient civilizations such as India's. Yet certain differences remain. If we look back to Roman Jakobson and Petr Bogatyrev, who defined the boundary between folklore and literature in 1929, a few of the criteria to which they adhered still persist. The literary work is unique, complete and permanent, whereas A folklore work is extra-individual and exists only potentially; it is only a complex of established norms and stimuli; it is a skeleton of actual traditions which the implementers embellish with the tracery of individual creation, in much the same way as the producers of a verbal message (la parole, in the Saussurian sense) act with respect to the verbal code (la langue). A literary work is objectivized, it exists concretely apart from its reciter. Each subsequent reader or reciter returns directly to the work; ... whereas for a folklore work the only path leads from implementer to implementer. If all bearers of a given folklore tradition die, then a resurrection of that tradition is no longer possible; whereas, on the contrary, the reactualization of the literary works of a distant past is not uncommon... (Jakobson and Bogatyrev 1978 [1929]: 91-92.)

The "potentiality" of an oral work leads to a fluidity of its manifest form, a far cry from the fixity of the literary work. Another decisive criterion for Jakobson and Bogatyrev was the fact that works of folklore must be accepted by society, whereas the literary author may go against society and formulate his thoughts free of any "censorship". For the folk singer solidarity to tradition and social convention is obligatory, otherwise his works would not qualify as folklore. Needless to say, this does not exclude social protest in folklore. The consonance of a song with the values of at least one group promoting recognition of the singer in question seems to constitute the kernel of "collective creativity" for Jakobson and Bogatyrev, who deem it typical of folklore but not of literature. This mild restoration of Romanticism may raise fewer objections than the talk about expressions of folklore as "works", i.e. thing-like entities. I will soon return to this problem, but it is interesting to note that instead of a master text Jakobson and Bogatyrev postulate "a skeleton of actual traditions" and a language-like production of oral discourse. Both statements apply fairly well to the model of textualization of oral epics presented below. What is meritorious, too, is their warning against the use of literary stereotypes in the characterization of oral texts:

6

Lauri Honko ...one must beware of the mechanical applications of methods and concepts obtained in the elaboration of literary history... the difference between a literary text and a recording of a folklore work must be taken into account. The typology of folklore forms must be built independently of the typology of literary forms. For example, compare the limited set of fairy tale plots typical of folklore with the diversity of plot characteristic of literature. Like structural linguistic laws, the general laws of poetic composition which result in a spontaneous likeness of plots are much more uniform and strict in their application to collective creativity than in regard to individual creativity. The immediate problem facing synchronic studies of folklore is the characterization of the system of poetic forms which make up the actual repertoire of a given community (geographic, ethnic, professional, coeval, or other similar unions). The relationship of the forms within the system, their hierarchy, and the degree of productivity of each are to be investigated. (Jakobson and Bogatyrev 1978 [1929]: 93.)

The need for synchronic, empirical studies on hitherto unknown poetic systems of oral tradition is spelled out with clarity, as is the risk of applying high-Western literary concepts to low-Western or non-Western predominantly oral cultures. Avoiding Western stereotypes is probably the most important methodological problem of comparative research on oral epics. The warning was issued seventy years ago but it has not been internalized by all researchers. Even such a basic distinction as the primary oral textualization (in performance) and its secondary written codification (in documentation and publication) is alien to certain scholars, who prefer to use "text" only for the act of writing, not for speech. Instead, we get euphemisms like "oral verbalizations" (Ong) or "standardized oral forms" (Goody). This attitude is unfortunate in four respects: first, it leaves the field of textmaking in the hands of scribes and bypasses the wealth of oral textualization; second, it corroborates the stereotypes of written culture in an area where they do not primarily belong; third, it prevents the scholar from discovering how orality deals with text; and fourth, it may invite him to a detour of mystification of Romantic or some other kind, i.e. making oral tradition a collective endeavour void of individuated texts. Softening the boundary between orality and literacy does not solve or eliminate the problems of relating the two. Looking at the prime object of our present interest, oral (printed!) epics, we may ask: what is their relationship to primarily literary epics. Oral, tradition-oriented and literary epics The taxonomy of epics cannot manage with the dichotomy of oral and literary epic alone. A transitional category between the two is needed. If we char-

Text as process and practice

7

acterize the purely literary epic as a text without "anterior speech" inscribing "directly in written letters what the discourse means" (Ricoeur 1991:106; cf. Siikala in this volume) and the oral epic as speech only secondarily codified into written letters, we have denoted the main difference between Milton's Paradise Lost and the Tulu oral epic of Siri. In the former case the primary codification of text takes place by writing, whereas in the latter it is effectuated by oral means (song, recitation, dictation), the exceptional written form resulting from the non-written one. Another obvious difference between oral and literary epic is that the literary author, even if he utilizes traditions (as Milton did the Christian tradition), does not let these elements determine his choice of plot or form, whereas the oral singer of epic is bound by tradition and the horizon of expectations of his audience. Putting the primarily literary epics aside for a moment, we find that the remaining epics are not purely oral in the sense that they only reflect oral epic tradition but do not accurately follow the "anterior speech", i.e. the oral performance. To this category belong the great epics from Sumerian Gilgamesh and Indie Mahäbhärata and Rämäyana to the Homeric epics, from Iranian Shähnäme and Central-Asian Manas, Gesar and Jangar and African Bani Hiläl, Mwindo, Lianja and Sunjata to Beowulf, Roland, El Cid, Nibelungenlied, the Eddas and the Kalevala, i.e. mostly epics which have consolidated the genre as a crowning chapter in the history of world literature plus hundreds of lesser-known scripts of epics caught in oral performance, written down and edited. Admittedly, we do not know the history of textualization of many of these epics in detail, and when we do, the histories differ quite dramatically. Yet we may conjecture that as a rule the final product is as much dependent on the scribe as on the singer. Most of these epics, or rather, the forms in which they have come to us have a long history of writing, copying and editing behind them and it is the task of scholarship to try to reconstruct that history on the basis of mostly scanty textual evidence and minimal knowledge about their cultural context, use, performance and documentation. Yet the epics in question undeniably possess, though in different ways, an intimate relationship to oral epic tradition. I propose to use "tradition-oriented" (or simply "traditional") to denote this vast category of tradition-bound epics which have been moulded, if not created, in the hands of performers, scribes and editors. They possess "anterior speech" in the form of oral epic registers internalized not only by their performers but also by their scribes and editors, but they are not direct documents from oral performances. Decisive here is the distance between the originally performed material and the final text of the epic. Poetically, the final text may be an "improvement", but even that relegates it to the category of "tradition-oriented", i.e. not directly oral. The typology of ways in which the final epic text may "deviate" from its oral materials and models is still to be created but it will be broad indeed.

8

Lauri Honko

First of all, not many scribes and editors had the urge to stay close to what they heard. They internalized the poetic language, because they had to, dictation being not the best way to document a song. To take just one example, the compiler of the Kalevala, Elias Lönnrot, developed a method which allowed him to let the singer continue without those interruptions which could often shorten or damage the song. He wrote down only the first letters of words. Since much of oral poetry is formulaic and repetitive, he soon commanded a wider repertoire of authentic lines than the singer whose performance he was writing down. He "recognized" the line, gave it a form while it was still lingering in the air. Phonetic accuracy could not be achieved, of course. Instead, the competent scribe participated in the process of singing. In fact, he "sang" the line at least three times, first, when it was still in the air and had to be noted down quickly, second, when he wrote a fair (supplemented) copy of his notation soon afterwards, and third, when he placed the line in the final epic text at his desk much later. This textual process gave the line a form consonant with the epic idiolect of the scribe, the best form according to the compiler's linguistic competence, learned and developed while listening to the singers but also having other goals of unification and commensurability which the singers could not dream of. The goal was not to collect accurate texts for archiving but to publish a long epic, a format not automatically at hand. It was Elias Lönnrot's own poetic system and his epic idiolect which constituted the basis of the final epic text. The system was admittedly traditional but not a copy of the poetic system of any oral singer or singers. When the compiler continued his work, it was not a mechanical procedure. We must recognize in Elias Lönnrot the creative scribe whom we find behind so many great epics. He obviously differed dramatically from most other collectors (he collected 25,000 lines when all the previous collectors had got only 10,000) and he had a goal, a long epic, and a narrative competence comparable to the oral singers'. Yet the epic he created is "tradition-oriented" not "oral" in the sense that the distance from the original materials as regards plot and format grew through the five versions of the Kalevala which he "performed" in writing during 1833-62. Here is another criterion of orality: the five versions of the Kalevala are reminiscent of an oral process where the work is always performed upon request, always different yet retaining the basic story. A literary work cannot be transformed at every recital; its text, once completed, does not change. There are many other ways of handling the "anterior speech" in the process of written codification. The most conscientious scholarly attempts to preserve, not change the oral epic are only one category apparent in the articles published in the present volume. Jan Knappert, the erudite connoisseur of Swahili and other epic traditions, describes the process of "collation" in terms of comparing several available texts based on oral performances and patching up the final epic text with elements from different sources. Who, then, is the

Text as process and practice

9

singer? The oral performer who did not even know all the "versions", i.e. the renditions by other singers? Or the learned scribe who knew the whole material and wanted to be loyal to it, eliminating only glaring inconsistencies and patching up the most obvious gaps? Is the final text consonant with any text an oral singer might produce? Should we call the epic "tradition-oriented" or, recognizing the attempt to follow the tradition, "oral"? These are hard questions to answer and they dot the entire history of textualization of oral and traditional epics. On the other hand, the absence of these questions in the case of primarily literary epics shows that we are dealing with truly different categories of epics. Literary epics may seem easiest to set apart, even if the "literariness" of not only scribes and editors (i.e. occasional singers in disguise) but also of, alas, the illiterate oral singers who compile long epics not shared by other singers in the same form, easily confuses the picture. Yet, it is worth trying to draw clear lines of demarcation even though they may have to be compromised from time to time. The term "oral epic" would thus be limited to cases where "anterior speech" has been directly inscribed and published as such. Understandably, knowing the difficulties in documenting long oral epics and the compromises that even the best epic scholars have been forced to make (cf. Honko 1998: 169-217), the number of codifications of long oral epic which qualify as "true to the original" is limited and depends on where we want to draw the line. Thus far the comparative research on oral epics has accepted the attempt to preserve the oral form as sufficient for including a textualization in the "oral epic" category, even if the result is not always the best possible. It should be noted that certain epics listed above as "tradition-oriented" may occasionally have been textualized with an accuracy which qualifies the result as "oral". The development of audiovisual documentation in fieldwork during the latter half of the 20th century has made a new level of accuracy available to growing numbers of epic scholars. The next millennium may witness a breakthrough of oral epics research based on improved quality of materials, not only textual but audial and visual as well, enabling new kinds of questions to be posed and analyses to be made which the uneven quality of earlier materials did not allow. The paradigms of oral text Before presenting a processual model of primary and secondary textualization of oral epics developed on the basis of empirical research, we must confront certain ambiguities concerning the "oral text". By way of a backdrop, a glance at the shifts of paradigm relevant to our subject may not be out of place. As pointed out elsewhere (Honko 1998: 154), the interesting thing is that even if a new paradigm challenges the previous one in a most dramatic

10

Lanri Honko

way, it does not replace it totally. Thus the conceptualization of "text" contains several layers of intellectual heritage, mostly conditioned by the nature of developing scholarly curiosity. For a scholar, the "text" is a corollary of what he wants to know. Looking back, we may discern at least three different text paradigms in folkloristics plus a contemporary debate on the role of oral text in modern research. The first phase of interest in oral tradition may be characterized as "pretextual" (Honko 1998: 44). Originally folklore was seen as a source of information, a kind of archive of the wisdom of ancestors. Its form was of no importance. The discipline recognizing oral tradition as its prime research object, folkloristics, was not yet born. Writers and scholars coming from different fields of learning culled folklore materials for positive knowledge about historical, mythological, linguistic, topographical, demographical, sociological and other facts. The artistic and literary value of myths, narratives and folk songs was occasionally noted. Poeticians writing treatises on various genres in literature sought their beginnings in oral traditions; this development peaked during Romanticism in the late 18th and early 19th century. Since the early performances of literature were oral, the written form was for a long time less dominant than it is today. The idea of folklore works as presentable pieces of folk wisdom and art developed earlier than the idea of folklore as texts to be studied. The first folkloristic concept of text in the proper sense of the word is coeval with the birth of the discipline during the latter half of the 19th century. Forerunners can be found. In Finland, Henrik Gabriel Porthan, Professor of Rhetorics at Turku University, in his book De poesi Fennica (1766-78) outlined "text-critical rules" for the editing of normalized forms of oral poems on the basis of variants. This task was taken up by the historic-geographic method of folkloristics a hundred years later, in Finland by Julius Krohn and his son Kaarle, who began to cooperate with Scandinavian and German scholars and created the "Finnish School" in folkloristics (Krohn 1926, Engl, trans. 1971). For the new discipline concentrating on oral traditions, the text concept turned out to be a pervasive methodological criterion of identity. It called for a new kind of exactitude in relation to folklore materials and constituted the basis of scholarly analysis. This was partly a reaction against the sweeping generalizations of the contemporary quasi-historical, mythological and evolutionary theorizing which occasionally referred to folklore materials as evidence. The new discipline demanded that, before any scholarly conclusions could be made, as many text documents of the expression of folklore under study had to be amassed as possible and organized in geographical and chronological order. The main task was to trace the variation of form in the texture of expressions of folklore, i.e. at the language level of text. It was essential to know the boundaries of text, where it began and where it ended,

Text as process and practice

11

because without a definition of the textual unit its variants could not be identified. The final goal was to describe the "archetype" or textual core of an expression of folklore and explain its development by looking at the variation found in its renderings thought to stem textually from the same root. The principles of text-criticism were borrowed from methods developed in the comparison of manuscript variants. It seemed possible to assume genetic dependence between certain folklore texts and place them in a stemma, a hierarchy of variants showing their mutual relations and derivation. The method made folkloristics a text-oriented discipline, and little attention was paid to the fact that folklore variation was basically different from manuscript variation, i.e. variation of a fixed text in the process of making copies of it. Neither "fixity" nor "copying" were good metaphors for understanding the real variation of living folklore. Yet the methodological stringency brought about plentiful systematic work and during a hundred years the paradigm produced monographs on the textual development and history of individual folktales, ballads, proverbs, etc., on a scale massive enough to fill the shelves of folkloristic libraries around the world wherever the discipline had established itself. The development was different in Europe and North America. The collecting of variants from the entire distribution area of the folklore items to be studied and linguistic textual accuracy were the hallmarks of the method in Europe and led to the creation of folklore archives and series of text publications, whereas in North America the lack of proper folklore archives and the publication of popularizing anthologies of folklore led to a situation where the level of accuracy in the texts available to researchers was low (Honko 1998: 44-45; cf. Halpert 1947: 355-60). Thus when Elizabeth C. Fine labels the concept of text advanced by certain American proponents of the historicgeographic method as the "literary model of text" (Fine 1984: 28-30), it is mainly based on the opinion of leading U.S. folklorists in the 1940s stating that textual accuracy is less important than the fluency of discourse, an attempt to make the best of the situation where folklore texts were proven inaccurate. European folklorists, however, following the model of comparative Indo-European linguistics, maintained linguistic stringency, and soon the textual base grew to millions of items in the best archives. In Europe, then, we cannot speak of a "literary model of text". A more adequate term for the European concept of folklore text is "text-critical". Its closest transAtlantic counterpart is the "ethnolinguistic model of text" (Fine) developed by anthropologists such as Franz Boas and Edward Sapir, who saw language as the main tool in understanding cultures and based their analysis on accurate transcripts of oral discourse. Both European folklorists and American anthropologists felt that cultures could be objectivized as texts. "The text is king" period lasted about a hundred years. Around 1970 a new paradigm was ready to question the basic tenets of the text-centred folk-

12

Lauri Honko

lore research. It was dominated by the concept of "performance": the focus shifted from the words of the song to the singer and his/her performance, and in fact to the entire situation of performance, interaction with the audience and the processes of the construction of meaning in a particular cultural context. Folklore ceased tobe impersonal text documents and became a social matter, a tool of political will and power, an indicator of creative potential inherent in tradition and a testimony of the cultural and performative competence of the performer. The interest in verbal form did not die out; instead, it was expanded through "ethnography of speaking" and "ethnopoetics" to cover new dimensions of vernacular art. In the field of oral epics, the empirical work by Milman Parry and Albert Lord made a strong impact. Their oralformulaic theory inspired many, and before long old theories of memorization of oral text were replaced by fresh analyses on composition-in-performance. (Honko 1998: 46; Foley 1995a: 605-14.) "The performance is king" characterizes this third stage in the development of the folkloristic concept of text. The paradigm relativized text and declared that texts are misleading, because there is no stable text in folklore. It placed new requirements for texts made on oral performances. The verbal part is only one part of the text, not necessarily its core. An extension of "text" is needed to cover the verbal and non-verbal interaction between the performer and the audience, paralinguistic expressions such as gesture and body movement, the utilization of space and artifacts and, above all, collateral action (dance, pantomime, ritual, song, orchestra). The reading of "extended text" presupposes two documents: a "performance record" and a "performance report". As Elizabeth Fine (1984: 95) explains: "...a record, unlike a report, attempts to record systematically at least one level of signal, such as morphemes or phonemes, from the beginning to end of a performance". A report includes the elements which it is not possible to present by way of notation but which are necessary in order to understand the text, be they performer history, contextual information, cultural background, genre characterization or whatever. An "extended text" with multiple and complex notations surrounding the verbal text is not a text but a libretto. Understanding the notation of changes in intonation, stress, pitch, pausing, gesture and spatial movement, etc., alongside the description of phonetic, morphemic and prosodic features of the text supplemented by audience reaction requires deconstruction with subsequent reconstruction not of the text but of the performance event itself. When the method succeeds, it opens up the individuality of performance and performers and shows in detail how transient, connotative meanings are processed in a particular poetic system. When it fails, it offers a jumble of details and particles of speech without any novel insights into the dynamics of performance. The cumbersome notation apparatus may be suitable for small genres but hardly possible to apply to, say, oral epics lasting six days.

Text as process and practice

13

Yet the performance paradigm has made it perfectly clear that the oral performance is as a medium totally different from the printed text. Its spectrum of expressive means is much wider than that of print, and it effectively utilizes contextual, allusive understanding of the verbal message, often supported by the invisible presence of traditions not expressed verbally but influencing the processing of meaning. What we have here, in fact, requires intersemiotic translation, i.e. "the transference of a message from one kind of symbolic system to another" (Nida 1964: 4). I doubt that the complex and typographically cumbersome notation of performance will point the way to future documentation and textualization of oral epics. My own experience from southern Karnataka, India, speaks for multiple audiovisual documentation (we simultaneously used 1-2 video cameras, 2 audio recorders and took photos) of complex performance situations (in our case, ecstatic possession rituals) which provides (1) the verbal text of the epic, invocations, prayers, dialogues, etc., (2) the integration of the epic singing in a wider ritual process, and (3) a continuous profile of the performance event lasting 10-14 hours per night (during 2-4 nights). The same technique was applied to simpler interview situations to allow continuous documentation and to eliminate pauses due to cassette change during singing. For the publication of our text and contextual information there are better ways than the notation discussed above. The shift of paradigms may be in the making as we turn to the new millennium. "The performance is king" paradigm relativized text, the next paradigm will probably relativize performance. Any performance is a compromise, an intelligent adaptation of tradition within unique situations structured by a confluence of several factors. It can be understood only against a broader spectrum of performances of the same integer in similar and different contexts. A single performance cannot witness for other performances, just as one singer can not represent other singers. What are the joining links between commensurable but variegated performances is a question to be discussed shortly. During recent years a more general debate on the nature of "text" and "oral text" has been taking place. I have briefly reviewed the discussion in another context (Honko 1998:142-52) and will confine myself here to a few remarks. As suggested above, oral text represents speech only secondarily codified into written letters. Should we accept Paul Ricoeur's formulation (1991:106) that "a text is really a text only when it is not restricted to transcribing an anterior speech, when instead it inscribes directly in written letters what the discourse means", we would have to relegate oral textuality somewhere outside "real texts". Leaving aside the complexities of "purely" literary creation, one possibility is to speak, as we have done, about primary and secondary written codification. The primary codification would then include Ricoeur's "real text", i.e. writing without anterior speech, and oral tex-

14

Lauri Honko

tualization would primarily take place in speech events and various performance situations and assume a written form only secondarily and exceptionally, because most oral textualization is unattainable after its performance and there is no guarantee that a later appearance will manifest itself in the same form. There is no need to deny speech text-like qualities, and the status of a "text void of speech" may turn out to be problematic. Yet it is not a problem to be discussed here. What does deserve some attention is the status of traditional or traditionoriented epics. They are based on "anterior speech" but do not reproduce it slavishly. They may contain ideas and expressions not found in any independent documentation of anterior speech. In this respect they constitute a category between "oral" and "literary", sometimes called "semiliterary". The decisive criterion, however, is the throughgoing dependence of these epics on oral models and traditional rules. This dependence takes the form of a solidarity to oral tradition which overrides all claims of individual authorship. Just as oral singers disclaim any individual creativity in passing on tradition, the compilers of tradition-oriented epics tend to disclaim their personal contribution. That our analysis cannot accept these disclaims uncritically (normally they are not true), does not deprive them of their significance. Describing the textualization history of traditional epics becomes a complex task which in many respects resembles the analysis of oral textualization (cf. the scribe as singer and the enigmatic elements taken from oral tradition). Another observation concerns the model of oral-based textualization applied by scholars who have little or no access to fieldwork on oral tradition, such as Homerists, medievalists, philologists (of dead languages) and literary historians. Understandably, they tend to emphasize the written word in the process and mystify the pre-written textualization. Faceless rules of tradition and expressive powers stemming from oral sources, sometimes simply called "the voice", a prototypical agent, are apt to replace different types of singers known from the ethnographic experience gained from living oral-poetry cultures. For Paul Zumthor, for example, the human voice is not only a carrier of articulation and lexical meaning but also of transformal, non-linguistic meaning: "the voice goes far beyond the spoken word". Voice thus assumes a function separate from language, "the voice is assigned the task of serving a protective function: it safeguards a subject matter that its language threatens; it checks the loss of substance which a perfect act of communication would entail." (Zumthor 1990: 74.) In Old English studies, A. N. Doane sees in the meagreness of the textual base a reduction of "the reverberation of many voices from the past in a single present one, an audible metaphor for all performances" (Doane 1991:103-04). The mystification of orality may thus lead to a pessimistic view on the visible text and fixed linguistic form in general. In the spirit of Derridean

Text as process and practice

15

grammatology the oral text is seen as stretched out or "tortured" into a linguistic form in danger of sacrificing its potential meanings to the "overdetermination" of language. There is a nostalgia for the multifaceted and "unfinished" oral poem and the "oral meaning" prior to textualization, both oral and written, the latter adding a "wrong" medium to the drama. The original multivocity of expression is sacrificed to artificial textual fixity in textualization. Somewhat similar pessimism about oral text prevails in the terminological triad "entextualization / decontextualization / recontextualization" proposed by Richard Bauman and Charles Briggs. They ask: how is it possible that verbal art, so deeply bound to its context, can be detached from it at all? The answer is, roughly, that detachment is made possible through entextualization, which leads to decontextualization, which leads to recontextualization. In this perspective, textualization in general becomes a suspect phenomenon, a process able to force oral discourse or "a stretch of linguistic production into a unit - a text - that can be lifted out of its interactional setting". The performer's task then becomes "to render stretches of discourse discontinuous with their discursive surround, thus making them into coherent, effective, and memorable texts". (Bauman and Briggs 1990: 7374.) It seems questionable, however, whether the singer's work in oral textmaking can be characterized as "discourse rendered decontextualizable" or "discontinuous with its discursive surround", when, at the conscious level at least, exactly the opposite takes place (cf. Honko 1998:150-52). A nostalgia for pre-textual meanings does not give us much to work with. For a fieldworker, oral text-making is not shrouded in mystery. The "voice" becomes audible through the singer and the "text" takes shape by oral/aural means. It is his/her singing praxis embedded in its social ambience which seems to open up vistas on textualization. Oral text as process and practice There are two aspects from which the dynamics of oral text-making may be elucidated, one diachronic and the other synchronic. Folklore texts are constantly on the move, they become born and disappear to be reborn again in the seemingly endless line of consecutive performances. If we wish to visualize the entire process through which a particular integer of folklore passes on its way from a pre-performance existence into oral textualization and further, presuming that documentation takes place, to a written form to be published and circulated to wider audiences, we must draw a profile of what we characterized above as "an intelligent adaptation of tradition within unique situations structured by a confluence of several factors". Since there is no master text for the oral integer in question and since its fixed form is

16

Lanri Honko

made permanent only by exception and does not represent other renditions, we cannot really say in any detail what the oral text should contain. Yet, since there is textuality and cohesion in the different renditions, we do not normally have any difficulty in recognizing the oral text. It is observable as an individual text, even if we are unable to agree on its "correct" form or fixity (major "deviations" from the form may be valuable intertextual clues and should not simply be eliminated). Thus what we witness is not so much a particular text as a textual process through which a story moves via "high" moments of manifest performance and "low" periods of latent existence in the mind of the performer. If we are able to gather information on all the factors which influence the performance, we may order our knowledge in a processual profile of the textualization of a particular story. In so doing we must critically assess - and fight against such stereotypes as "one story", "variant" and "fixed form". The story may be modulated in ways for which we possess no textual evidence. "Variant" raises the questions of inertia, continuity and invariant in oral poetry (what is the "thing" that varies?); to avoid the problem we may try to use such terms as "telling", "rendition" or "performance" instead of "variant". The moment we see a verbal form we tend to impute connotations of fixity into it, unfortunately. The diachronic aspect of textualization is visible mainly to the analyst, who gathers the necessary information and sketches a history of textualization. Even if most of the facts have been elicited from the singer telling the story, the processual profile may not be of much interest to him. His attitude to the story is dominated by his singing practice, which represents the synchronic aspect of textualization. The oral text is not played out in a vacuum, as a separate item or expression. The story has individuality but its verbal form does not exhaust the richness of life available in and around the performance. First of all, the verbal form as expression refers to so many unspoken traditions and implies so many latent channels of communication that the communicative thrust tends to expand from the verbal text to the entire event of performance and social interaction. The oral text is inseparable from the flow of information and the construction of meaning taking place inside and outside it. In a way, the oral text is occasionally lost from sight as a "thing" and replaced by something larger, a communicative event. In other words, the utterance is no longer readable in its verbal form only. At this point we must either abandon the text as a unit or start thinking in terms of an "extended text" as the true object of our synchronic analysis. It belongs to the narrative practice that a full verbalization of an episode or a narrative sequence may be replaced by allusions consisting of few words but retaining the most essential symbols guiding reception.

Text as process and practice

17

Table 1 (Honko 1998:167): THE MAKING OF ORAL EPICS Pool of tradition: Coexistence of traditional multiforme, genres and registers in the human mind Epic register: common storylines, descriptions, multiforms, phrases, formulas Reception 1. The singer's repeated exposure to and gradual internalization of traditions V The singer's thematic competence: orientation toward particular genres and narratives Epic idiolect: the singer's individual selection of shared expressions as language \7 Reception 2. Internalizing a particular epic: intertextual interpretation Adaptation of tradition: mental editing and permanent change (diachronic variation) Mental text: storyline (flexible), rules of sequencing, textural cues The singer's performative competence: traditional rules and conditions of performance Performance strategy, mode of performance, performative style Adaptation of tradition to situations of performance (synchronic variation) The processing of traditional meaning by the singer in action: manifest text Observation of audience reaction and its immediate impact on performance Defensibility and truthfulness of performance: the last performance is the best Reception 3. Intertextual interpretation Interest in documentation: its origins by the audience (processing of inside /outside tradition community multiple meanings) Documentation strategy Feedback to the performer (optional) V ~7 Looking forward: modification of Context of documentation (natural, tradition in view of future performances induced, artificial) Collected text 1 Editing strategy: textual choice Transcription, translation, commentary 1 Publication strategy: collateral forms The oral epic as a book (AV options)

18

Laari Honko

The previous model (see Table 1) attempts to draw attention to both aspects of textualization, the diachronic and the synchronic. It endeavours to m a p the confluence of factors present in performance. The number of these variables should enable us to see why there cannot be two identical renditions of one and the same folklore integer. The generalizing form of the table should not mislead the reader: the evidence for the argument stems from fieldwork on a particular genre, oral epic, viewed in its performative and cultural context over eight years. The necessarily brief commentary will occasionally specify and explain the statements through examples from the field in question. The model in Table 1 attempts a realistic inventory of factors influencing all textualization processes from oral pre-textual beginnings to printed book. It subsumes different concepts of text, i.e. it does not depend on particular text paradigms such as the historic-geographic or the performance-record model. The empirical background is mainly two poetic traditions, namely, the Baltic-Finnic lament in half a dozen languages and the long oral epic in the Tulu language in Karnataka, South India. Let us briefly examine the key terms. Pool of tradition. In what form is oral tradition preserved in h u m a n minds? What are the shared elements of folklore, and what is the contribution of the individual to the formation of folkloric expressions? It seems that most elements of epic discourse are older than their users, i.e. convention dominates over invention. Yet the elements are free to vary and combine, and it is in the variation and combination of multiforms, themes and formulas that the individual novelty can be found. That is w h y we cannot postulate a wellarranged library of earlier performed oral texts in the mind of the individual b u t rather a "pool" of generic rules, storylines, mental images of epic events, linguistically preprocessed descriptions of repeatable scenes, sets of established terms and attributes, phrases and formulas, which every performer may utilize in an imaginative way, vary and reorganize according to the needs and potentials present at a new performance. As stated elsewhere, it may be "realistic to seek a model applied in chaos theory, of narrative elements bubbling freely in a pool of tradition, ready to float in m a n y directions and to fuse in novel ways. I deliberately use the word pool, meaning both a body of water and a f u n d to which many contribute and from which m a n y can draw." (Honko 1998:70.) The rules and models for the fusion of traditional elements vary by genre, and the performers normally specialize in a limited n u m b e r of genres. In long and complex genres such as oral epic the art of sequencing is a special skill cultivated by the singers. Whatever is shared by more than one singer belongs to the pool of tradition. The pool holds a multiplicity of traditions, a coexistence of expressive forms and genres, mostly in a latent state, only parts of it becoming activated

Text as process and practice

19

by the individual user. That it is not merely a technical abstraction denoting "shared elements" of folklore but a real platform of human exchange affecting a number of individuals is made clear through the interaction and communication within traditional groups by way of the social display of traditions in special performance arenas. Epic register. Performances utilize special languages in a generic context, "a performance arena" as John Foley (1995b: 47-49, 79-82) calls it. "Epic register" denotes a special language, a "way of speaking" valid for performers and audiences of epics in a particular setting. Dell Hymes's definition of epic registers as "major speech styles associated with recurrent types of situation" (Hymes 1989: 440) admirably fits the special institution of Siri possession cults as a platform where discourses based on the Siri epic dominate all behaviour. If "pool of tradition" implies the availability of shared traditions, "epic register" represents the organization of the available elements by genre and speech event. Learning the language of epics usually begins by learning a few oral epics one after another, whereby memorization may play a part. Later, the internalization of the rules of composition and the repertoire of repeatable expressions frees the singer from problems of wording and enables him to concentrate on the path of composition, i.e. the movement of plot and the construction of meaning. Multiforme. The role of repeatable expressions in oral epic art is paramount. Epic discourse consists of descriptions of standard events (receiving guests, having a grand meal, sending a letter, etc.). These again contain a large number of relatively short formulas, elaborated phrases, standard images and minor episodic elements all of which vary in length, degree of embellishment and emphasis. We denote as multiforme those "repeatable and artistic expressions of variable length which are constitutive for narration and function as generic markers" (Honko 1998:100). Multiforme always contain linguistic elements, keywords and syntactic structures which help us to recognize them, in other words, they are defined by their texture, not by their structural position or actual meaning. Multiforme should, at least in theory, be recognizable to the singer, too, who is aware of his poetic use of the "same" expressions. Multiforme are thus not to be equated with the "type scenes" or "themes" of the oral formulaic theory. For syntagmatic purposes we use such units as subepic, description, episode and step (Honko 1998:117-25), from larger to smaller, leaving "multiform" outside the syntagmatic coverage of content so as not to force the concept to be present everywhere, also in contexts where there ie no empirical evidence of the einger's recognition of a poetic unit. Multiform is designed to help ue to explore the einger's claseification of repeatable expressions (which may contain "gaps" and need not cover all text produced).

20

Lanri Honko

Reception 1: Internalization of traditions. The essential thing in learning traditions seems to be early exposure to them. Complex genres such as oral epic require specialization, which normally proceeds in three phases: 1) exposure to narrative traditions in their performative contexts, 2) a call to become an active performer and the subsequent period of learning as a novice and 3) the emergence of a self-reliant singer recognized by the tradition community and able to assume leadership at performance events. The first phase implies broad learning, a construction of cultural competence, the performance career being not yet in sight. In Baltic-Finnic lament traditions early exposure to dirges, for example, at a young age at the funerals and memorial ceremonies arranged for near relatives, will leave a strong impact on the mind of a young woman, but this potential will not be converted into a full-fledged performance until much later, perhaps at the age of 40 or more, when members of her own family begin to die (Honko 1974: 23-26; Honko et al. 1993: 57277). In the case of the Tulu oral epic Siri, the early exposure of a boy to epic singing by women as part of work in the paddyfield was coupled with a mental disturbance indicating an intervention by spirits which eventually led to an early call for the teenage youngster to become first an assisting Kumara at Siri possession rituals (the singing of the epic as a central requirement) and then, at the age of 20, to assume the leadership of a group of Siri women (Honko 1998: 519-34). Tradition-orientation. In the process of learning oral traditions there is always choice. The availability of what could be learned is overwhelming compared to what is eventually internalized by an individual. Several factors direct the choice: the talent one may show at an early age, the family around which cultivates particular traditions, a genre specially valued by the community, a practical need to acquire professional knowledge and skill, the impact of teachers and guides respected by the individual in question. The list could be continued. In the end, however, there is logic in the construction of a store of traditions in the human mind. To use Anna-Leena Siikala's term "tradition orientation" (Siikala 1990: 202) we may characterize the process in question as a life-long series of adoption and rejection of available traditions in accordance with the individual's interests and worldview. In tradition orientation it is not only the positive selection which becomes visible; specific interpretations are also developed simultaneously for the integers selected. This means two things: first, the internalized elements are not "just stories" but become amalgamated in the person's view of life, and second, even if the same story is internalized by several storytellers, an interpretive adaptation is likely to make the final result (as may be witnessed in the performance of a story) different from one storyteller to another. All elements acquired in tradition orientation need not be developed to the level of performative competence, yet even if they remain at the level of cultural cognition

Text as process and practice

21

and knowledge, they constitute a sounding board for the active performance of traditions. The focus of tradition orientation becomes manifest in the genres and narratives actively performed by the individual. In the case of complex genres, such as oral epics, an apprenticeship may ensue which may last for years and lead to a full specialization and performance career within basically one genre. In most cases, however, an individual's tradition orientation represents competence in many genres and domains of culture. It reflects his social roles, personality and worldview. Epic idiolect. Concentration on learning a genre, say, oral epics, opens u p a pathway to a special language, constitutive for a large number of narratives. The expressions shared by many singers within that language we just called "epic register". The command of that register is the main object of active learning for a novice, because therein he finds the poetic rules and repeatable expressions present in all epics, a kind of core or identity of the genre. He also learns to see how special characteristics of particular epics relate to that common core. On the basis of the epic register and the accumulation of knowledge resulting from the internalization of particular epics, the receptive capacities are transformed into performative competence. A novice can be said to transform into a singer when the memorizing and repeating of heard expressions give way to more imaginative and fluent use of the epic language in performance. The singer's competence lies in the producing of his own version according to a heard performance, not its mechanical reproduction. The metaphor of language is essential for the understanding of epic composition and the singer's ability to produce cohesive narrative in freely flowing lines without any observable difficulty in wording. An epic may take several days to complete, yet the fluency of oral textualization does not seem to be a problem. What we witness here, in fact, is not the memorization of earlier textualizations but the production of a story in a particular language, viz. the singer's epic idiolect. The utilization of the epic register implies individual choice. The supply of narratives normally exceeds the demand: there is more in the tradition than one singer can consume. Learning the epic register does not imply its full utilization. Some of its elements become more important than others, whereas some do not seem to affect the singer's tradition system at all. The idiolect creates its own linguistic universe, the pillars of which are the epics known to the singer, their performance contexts and the poetic means he has absorbed and is able to take into creative use. All textualization takes place in the idiolectal universe of shared expressions. Reception 2: Internalizing a particular epic. Stories are as individual as their singers, if not more so. In our Tulu material oral epics are easy to distinguish from each other. They clearly belong to the same genre and similar expressions

22

Lanri Honko

are utilized. Yet each epic possesses an individual melody which discloses its identity before w e have understood its wording. It also has a particular refrain that beats regularity a n d cohesion to oral textualization f r o m beginning to end. The names of personages, often poetically elaborated with attributes, constitute networks of implied social and cultural references not to be f o u n d in other epics. Repeated images and key symbols m a y represent u n i q u e formulation, etc. What is more, the worldview and system of values p r o m o t e d by the story m a y vary f r o m one epic to another. In other words, there m a y be ideological "conflicts" within the genre. Thus the internalization of epics presumes choice of a second degree, viz. one particular epic m a y be closer to a singer's heart than another. The Siri epic t u r n e d out to be the key to the worldview of the singer, Mr. Gopala Naika, with w h o m w e worked for so long. It represented the worldview of the female cult group which adhered to the epic a n d annually renewed its b o n d a g e to it. For our singer at least, each epic and its ritual context seemed to constitute a world of its own. "I am living in the Siri world", h e confessed after having sung another epic (Kooti Cennaya) which propagates the virtues of warrior-heroes. His system of values was anchored to the Siri epic with its female heroines a n d virtues of non-violent resistance. In other words, he h a d m a d e a choice between the Siri world and the Kooti Cennaya world. He could not have both, because the value systems of these epics are so different. H e w a s able to understand and even sing several epics, six epics altogether, b u t it w a s just one of them, the Siri epic, which w a s closest to his heart and p e r v a d e d his philosophy of life. The learning of a n e w epic normally creates an intertextual situation. The key to learning is not "learning by heart" b u t understanding w h a t is being sung. Epics learned earlier and even short passages from different genres or formulaic expressions found in different contexts will function as intertexts a n d facilitate the w o r k of reception. It is here that the pool of tradition will show its force. The essential "work" of reception with a performative interest ranges f r o m listening carefully to the individual features (storyline, melody, refrain, structuring, etc.) of the epic to the making of an interpretation of w h a t is sung. The interpretation forms the basis of mental editing, adaptation of the story a n d its poetic means to the narrative competence and tradition system the singer already has. Intertexts have a role to play in this creative process, which does not result in a mechanical replica b u t a partly " n e w " version of the story. In a way, the first performance of the n e w epic is the culmination of w h a t w e here call Reception 2. Mental text. Long oral epics are rarely, if ever, performed in full. There m a y not be any cultural locus for a performance lasting several days. The time a n d need for performance m a y be defined by collateral action. In our case, the Siri epic, or fragments of it, w a s used as a w o r k song in the paddyfield or

Text as process and practice

23

as a myth recitation at possession rituals. What unites, then, the actual performances of the epic in its cultural context, is something we will never see, namely, the mental text in the minds of the singers and, probably in simpler forms, their audiences (Honko 1996: 4-5). We need a term to denote the "whole story", which must exist and can indeed be elicited in induced contexts created for the purpose of documentation, and to be able to understand the production of text in actual performance. It seems realistic to postulate a kind of "prenarrative", a pre-textual frame, i.e., an organized structure of relevant conscious and unconscious material present in the singer's mind, not fixed as a written text, yet linguistically pre-processed by way of expressions and sequences which are easy to activate in performance. Mental text seems to contain such elements as a storyline scheme (basic plot, not fixed but open to elaboration), a number of textual elements, i.e., episodic patterns, images of epic situations, multiforme, etc., and their generic rules of reproduction (including rules of sequencing) as well as contextual frames such as remembrances of earlier performances, yet not as a haphazard collection of traditional knowledge but a prearranged set of elements internalized by the individual singer. This variable template is an emergent entity, able to be cut to different sizes and adapted to various modes of performance yet preserving its textual identity. It is not as fixed as all its documented manifestations tend to be. Yet it is only through its fixed manifestations that we can try to construct components of a particular mental text. Therefore it must be stressed that mental texts do not refer to fixed wordings of expressions kept in the memory and reproduced in performance. We may speak of an oral text's "fixity" only after its phase of emergence in performance is over and the text has attained its form, regardless of how temporary or stable that form may prove to be. Mental editing and diachronic variation. Stories are not copied as such in the transfer of tradition. Any story which the singer begins to consider for adoption requires a kind of deconstruction in view of his repertoire and tradition system, i.e. it must be made to fit the earlier conventions of standardized expression internalized by him. The same applies to short passages, images and phrases often taken over by the singer from another performer and tentatively stored in his pool of potential expressions for use in later performances. The process of adaptation involves milieu-morphological and tradition-morphological changes (Honko 1981: 19-33) in order to localize and familiarize the story. This leads to diachronic variation, i.e. the changes made are preserved in later performances of the story, as opposed to synchronic, situational variation, which is likewise adaptive but temporary and which will not be visible in the next performance. Researchers have been largely unable to differentiate between diachronic and synchronic variation because of thin materials not reflecting the real variation of folklore as

24

Lanri Honko

manifested in the verbal behaviour of a singer or singers within a community or region where true interaction and exchange between performers takes place (cf. Honko 2000). I propose to call "mental editing" the adaptive processes of oral textualization taking place between actual performances, i.e. all changes made outside the actual composition-in-performance. The latter also causes changes but they are dominated by situational factors, most of them valid only in one performance. Mental editing works mainly by way of adoption and rejection of motifs and expressions offered in the performances of other singers. More importantly, it is responsible for the creation of a long oral epic. Mr. Gopala Naika never acquired the Siri epic as a whole from anyone. The sources were many and the process of editing long. A mental text of the whole epic was not needed to sing it anywhere. Instead, it served as a mythical charter for the Siri possession ritual led by the singer. The "translation" of all human behaviour manifest at the ritual into a cohesive entity reflecting key aspects of the Siri story, a narrative left almost untold yet constantly alluded to, required from the singer a linear concept of the entire plot. The plot helped him to relate to each other the "live" events based on divine intervention by the epic characters at the possession ritual. His command of the whole epic enhanced his authority over his assisting Kumaras and the Siri women whom he educated to command key parts of the epic also during visits to their homes, i.e. outside the ritual meetings. In interviews Mr. Gopala Naika admitted that he had combined certain episodes heard from different Siri women on diverse occasions to make them continue from one incident to another. Occasionally he had added some joining lines of his own. The creation of a long oral epic was a result of practice, an experiment in sequentiality, developed piecemeal over the years of his performance career (Honko 1998: 527). Performance strategy. The cohesion of the actual epic performance cannot simply rely upon the main plot of the entire story. There are many parallel subplots in the narrative, thus the decision when to follow which must be repeatedly taken. The number and order of possible units is not stable: some units seem obligatory whereas certain others are optional. (Honko 1998:139.) To be able to perform, the singer must design a performance strategy based on the potential and limits of the performance situation, audience, time-frame, collateral action (work, ritual, etc.), which often include quite unique elements and may determine what parts of the epic must, m a y or need not be performed. Since long epics can never be performed in their entirety in a multi-factor cultural context, each performance poses the question of how to cut the epic into a size and form relevant from the point of view of the situation.

Text as process and practice

25

Performers and audiences have their preferences and part of the performance strategy is defined along the way, as a response to what happens in the actual situation. Yet there is no need to overstate the impact of situation or audience on the "path of composition". Its problems seem to be more innate difficulties in joining subepics, letting the life-stories of epic personages cross or deciding the relative emphasis of certain episodes in a plot sequence. There are junctures in the epic where the best linkage between certain subplots has not yet been found. Despite the experimental dynamics of pathmaking, our empirical evidence shows that the singer always tends to conceptualize his actual performance as a "truthful", "correct" and "defensible" presentation of the Siri story, and even more, as probably the best presentation so far. From this point of view, an "abridgement" of the story is no vice. (Honko 1998:134.) Mode of performance. Oral epics may be performed in a variety of ways. The singer may perform in solo, with or without instrumental accompaniment, in linear monovoiced narration or with shifts to different tones (as different personages "speak"), switching melody and poetic category (prose, poem, song, recital, dance, pantomime or a combination of these) or by one or more ensembles (lead singer, accompanying singers, dancers, orchestra, etc.) utilizing all available means of expression from dialogue to full drama and incorporating members of the audience in the performance. The phenomenon may be called "mode of performance". The term denotes the technical and artistic setup governing the entire epic performance, its "external form". An essential part of performance strategy is the decision on the mode of performance. Only after it has been decided may the often lengthy preparations begin. The singer will consider the available time frame, necessary assistants, the nature of audience and collateral activities. Different contexts may require different modes even if the story remains the same. It is clear that the choice of mode affects almost every aspect of performance and makes a strong impact on what will later be viewed as the "epic text". If we are unable to visualize and imaginatively, if retroactively, coexperience the mode of the particular performance from which our epic text originates, our ethnopoetic analyses may be led astray. (Honko 1998: 75-76.) Performative style. Another term is needed for the intra-textual shifts observed in the singer's epic discourse. I propose the term "performative style" for different forms of discourse used within a particular mode of performance. In the Tulu tradition, for example, each epic has a melody, refrain and recitation pattern of its own. This shows that individual epics may have specific performative styles which govern the performance. Within one and the same epic it is possible to find linear singing in the 3rd-person singular as well as a variety of other performative styles, such as emphatic singing in

26

Lanri Honko

the lst-person singular at the peak of the ritual process, or the praying style with slightly different rhythmic patterns and occasional rapid recitation. The correct understanding of oral text presupposes that we recognize the performative styles used and the points at which the shifts from one style to another are made. Rhetoric means of emphatic adhortation, questioning as well as raising or lowering the voice, pausing and using mime, gesture and body movement as well as operative space, i.e., kinesic and proxemic means, may all contribute to various designs of performative styles. Looking at the external criteria of style, we must not forget that the content of discourse is part and parcel of style. Thus such moods as respect, humour, irony, anger, sorrow, etc., clearly affect the performative style in ways which may or may not be readable in the produced oral text. Style is obviously the field where keen documentation and close analysis of even the minor details pay off. (Honko 1998: 77-78.) Construction of meaning and synchronic variation. The production of a particular

oral text in a more or less unique performance situation thus represents a confluence of multiple factors. The basic premises of performance have been fulfilled in the much earlier processes of internalizing epics, their language, performative modes and styles, collateral activities, and so on. What is at stake in the actual telling of a story told so many times before is the construction of its actual meaning, never quite the same from one telling to another. Every performance may be seen as an attempt toward the "truth" of the story, its core meaning, once again made to shine in the sacred behaviour of the worshippers in a possession ritual acting out their epic identities, or in the likewise exalted recitation of women plucking paddy seedlings to the rhythm of the sung epic. According to the singers, practically every performance achieves the goal by displaying the truthful story in a form more perfect than ever. For the observing scholar, the achievement lies in the ingenious adaptation of the story to the particular situation of performance. On the textual side, this brings about variation which we have called "synchronic" because it is uniquely adaptive and produces forms determined by the situation. In a sense, it shows the shifts of emphasis in the standardized expressions serving the unique construction of actual meaning. There is no need to expect exactly the same expressions and emphases to repeat themselves in the next or later performances. Flexibility is the hallmark of the work of adaptation and construction of meaning by the singer. Audience interaction. The singer's work in performance does not take place in a vacuum. It is directional and anticipates reception by its audiences. The idea of more audiences than the physical one present at the performance springs from the observation that beyond the real listeners the singer also

Text as process and practice

T7

has ideal audiences. The very act of performance presupposes the creation of a correct place for it, a "performance arena", which standardizes the speech event regardless of the actual audience. This arena is more spiritual than physical, yet there may of signs of a symbolic construction of an ideal platform, such as the construction of a temporary altar for gods who constitute the main audience and interlocutor with the singer. In the singing of the Siri epic this is accentuated by the singer's facing the temporary altar with icons, his back turned to the Siri women and the rest of the village audience. It is quite clear that the singer is more occupied with making contact with the divinities than with the devotees behind him or the common village people watching the ritual drama. This constellation consolidates the performance as a "service to god", the corollary being that divine epic characters are invited to use the bodies of the present worshippers as their mundane vehicles and enter the scene. Eventually, the divinities take over the cultic responsibility and a perfect, immaculate service ensues. In many analyses of audience interaction the singer's need to please his/ her audience takes a prominent role. In the previous example the interaction is focused on gods, not on the physical human audience, but in many cases people around the singer constitute the sounding board crucial for his inspiration and motivation. The audience can be activated to the role of coperformer; dialogues between the singer and the audience are common, a feature that emphasizes the centrality of the construction of the actual meaning of a performance by the individual and the social ambience at hand. Persons in the audience may be identified with personages in the epic, also in profane contexts and temporarily, not only, as in our Siri case, in the systematic identification of possessed women with their spiritual alter egos. The latter form effectively eliminates the line between performers, the leading and assistant Kumaras, and the front-line audience, the Siris. A wider circle of relatives and spectators embraces them with a similar, although less ecstatic involvement. Reception 3: Intertextual interpretation. So far we have dealt with the reception of traditions and particular epics in the learning process of the singer. A third and important form of reception is the network of multiple meanings attached to the epic by the spectators and listeners. Somewhere here we find the group which makes the story a real epic, a tradition community which accepts the narrative as a song of truth and recognizes in it features of its cultural identity. Our definition of epic speaks of "exemplars" and "identity representations" (Honko 1998: 28). There is no guarantee that the collectively experienced identity feeling can be matched with identical interpretations at the semantic level. On the contrary, most groups are heterogeneous as to the traditional knowledge of their members. Correspondingly, interpretations held by group members vary

28

Lanri Honko

from person to person and, qualitatively, from intellectually meagre and superficial to rich and detailed. It is not uncommon for conflicting interpretations to generate debate within the community of reception and lead to an identity negotiation, a rather normal phenomenon in the case of strong symbols with multiple meanings. All reception is dependent on intertexts, i.e. a store of similar or antagonistic storylines, narrative structures, hero types, episodes, phrases and formulas known from other contexts but possessing interpretive potential for the individual receiving the epic. The intertextual universe varies from person to person depending on their education, social status and role, ideological stance and, above all, previously internalized traditional expressions. That is why we get multiple meanings in Reception 3. Feedback. The reception always guides the singer, however independent he may seem to be. During the performance a continuous reading of receptive cues informs the singer about the effectivity and appeal of his performance. It may bring about changes in the strategy of performance, some modification of the path of composition and adding or diminishing emphasis and embellishment of certain episodes. What the audience obviously wanted to hear or became impressed by may have a long-term effect on future performances. Mental editing always contains an evaluative aspect and memories of earlier singing may affect the planning of later renderings of the story. Typically, the singer has no need to look back to his earlier tellings for their own sake but, as an inquisitive outsider-interviewer will soon find out, where there are some new ideas to be drawn from them, mental editing is the place to sort out their impact. The same applies, for example, to new expressions and episodes of the epic heard from other singers. Some of them will be tried out on the singer's own textualization whereas others will be rejected as incommensurable, wrong or ugly. Documentation strategy. As a separate line of textualization the above model concludes with a profile of documentation which leads to a publishable text. Interest in such documentation is rare inside the oral cultures which produce epic performances as part of their annual cycle of feasts, rituals or working techniques. However, it does occasionally exist, as in our case. Mr. Gopala Naika expressed in the early 1980s his wish to have his Siri epic taken down by dictation. Dictation took place later, in 1985-86, in many sessions weeks and months apart, mainly by a student in longhand and Kannada script. In December 1990 our Finnish-Tulu team of scholars audio- and videotaped his (first!) sung performance of the whole epic during seven days. In February 1999 a three-volume work was published concentrating mainly on this performance and contextual information (Honko 1998; Honko et al. 1998a, b).

Text as process and practice

29

In a way, we became involved in a project instigated by the singer. More normally, the original interest in the documentation comes from outside the traditional community. That is why we have so few oral textualizations published by the communities who produced them. An inevitable intercultural imbalance pervades most oral epic work published so far. Our documentation strategy consisted of several points: (1) we would prefer the singing mode, not yet explored and an interesting point of comparison with the dictated rendering; (2) we would postpone lengthy interviews after the whole epic had been completed in order to minimize our impact and let the singer concentrate on his singing; (3) we would eliminate pauses or gaps due to cassette change by using multiple taping (1 video, 2 audio recorders as a minimum) at all times from morning till evening; (4) we would allow the singer as much time as he wanted and let him decide when it was suitable to make pauses in singing, i.e. we would follow his segmentation of the oral text; (5) we would let the singer decide on the best environment for our work and the presence of people other than our team members; (6) we should comply with his concept of the "whole" epic, which, as it turned out, would mean the inclusion of introductory invocations (the element left out of the dictated Siri text); finally, (7) we would not consider any "collation" with the singer's other performances of the same epic but publish just one text, the oral text as it was performed in the imminent situation of performance. (Honko et al. 1998: xxxiii-xxxix; Honko 1 9 9 8 : 1 6 3 64,261-71.) Documentation strategies vary from case to case, as I have shown in a survey of eleven productions of an oral or tradition-oriented epic (Honko 1998: 169-217). More often than not the strategy is simply a compromise made in a field work situation on the basis of what seems possible, a pragmatic optimal choice among available limited alternatives. Recipes for "correct" documentation techniques are in this sense meaningless. Yet the comprehensiveness of the oral text, its unbroken continuity, linguistic accuracy and, in general, the heeding of the singer's wishes as regards the details of documentation and the context of performance are high on the list of requirements for a good oral text. Since long oral epics are practically never performed in full in their normal culturally determined contexts, an induced "natural" context or a laboratory environment are the main alternatives available for the documentation of the whole epic. By creating an environment according to the wishes of the singer it is possible to approach an induced performance context which is novel but natural, at least to a degree. Some of the best texts may come from situations to be classified somewhere between "induced natural" and "laboratory context". These situations may involve motivating and guiding the singer. Sometimes performance and documentation strategies intertwine: the singer moulds his act to suit the recording and expressed goals of col-

30

Lauri Honko

laboration. All this makes it imperative to get a full report on the documentation process, including the discussions with the singer. The reader must be made able to judge how the method of documentation may have influenced the textualization. (Honko 1998:161-63.) In our project, an extensive report on every step, minute by minute, in our documentation process was published, and everything included in that report (Honko 1998: 261-388) can be corroborated by audiovisual evidence from our audio and video tapes and photographs. Every syllable of the epic and every small remark expressed outside singing can be made accessible to the reader who is willing to plunge into the Siri archive of RRC in Udupi, Karnataka, India or the TKU archive in Turku, Finland (about 250 hours of videotape, 350 hours of audiotape and some 6,700 photographs). Collected text. What constitutes the "text" to be offered to the reader is one of the hardest questions in the textualization of oral epics. Publication requires a strategy: how to go about the transcription, translation and interpretation of an oral epic suddenly epitomized and petrified through scholarly documentation. The fate of the original oral discourse is often to become compromised in the process. The absence of music, gesture and transformal meaning changes the traditional concept of an "epic" into something else, not likely to be recognized by the performers. (Honko 1998:163.) The situation poses a number of problems, partly addressed above in the context of the debate on performance records and reports. What is basically needed is an intersemiotic translation able to convey the experience of a speech event in writing or in audiovisual media. What the academic community normally expects is a book, a readable text. If we wish to heed this expectation we must decide what goes into that text. Our decision to publish just one performance of the Siri epic by Gopala Naika was taken so rigorously that a part which he sang for us much later a second time, and which contained some new information, was not integrated into the earlier text. In other words, we refrained from "collation" or "patching u p " with the help of other "variants", a method so widely used by editors of epics. On the other hand, when Gopala Naika interrupted his singing, saying that he had forgotten an episode prior to the passage he had just completed and asked for permission to sing that too, we consented and the episode was inserted in its place while editing, mainly because all this happened within the same singing session. Generally, however, we rejected the possibility of comparing the singer's several performances of parts of the Siri epic in view of selecting the "best" passages for the patchwork final text. Patchwork represented a bad word for us, a plague in the history of the editing of oral epics, and there was no pressing need to consider it, because the published version turned out clearly the most comprehensive and "good enough" for

Text as process and practice

31

orientation. The comparative work on different performances of the Siri epic was postponed to a later volume focusing on the interpretation of the epic. Phonetic transcript. Published oral epics may be divided into two groups: those which include the text in a phonetic transcript of the original language and those which do not. There is no doubt about which group is scientifically more adequate. Without a phonetic transcription, the linguistic interest is lost, the prosody cannot be studied, the poetic means of the original language are largely hidden and the translation becomes suspect. Yet, even when the phonetic script is available, its form may be problematic: should it follow the idiolect of the singer, i.e. should it reflect the sung forms or the normalized transcription found in lexicons? In our case, the whole process of editing was structured by our decision that the Siri epic would be published in the original Tulu, following a modified Sanskrit phonetic transcription, and in English translation, both texts preferably side by side allowing for immediate checking by those who know Tulu and for viewing the phonetic structure by those who do not. The latter is by no means irrelevant, because much of the poetic texture can be "read" even without knowing the language, i.e. alliteration, assonance, repetition, word order in relation to word length, line structure, figura etymologica, etc. The accuracy of translation automatically increases when the translator knows that the original text will be available to the reader. Because modern documentation technology provided impeccable accuracy of the original, this accuracy was not to be compromised in the process of editing the oral text for publication. For once, orality was to be carried as far as possible in the written media. (Honko 1998: 583.) It was decided to write down what was heard on tape without any editing or normalizing. As we know, the hearing differs between scholars even if they are used to transcribing. We based the transcript on multiple hearing by all four academic members of the team. As more ears were included in the work, the final form partly became a compromise of different hearings. Our transcription follows the phonetic forms actually used in the song more closely than a transcript using normative lexicon forms would do. For the reader, this has the advantage of a more accurate prosodic appearance of the text, but the deviation of the sung epic idiolect from the lexicon form requires some attention when single words are quoted and should then be normalized. (Honko 1998: 584.) Poetic line. The most conspicuous feature of the transcript is the division of singing into poetic lines. This affects both transcription and translation. If clear-cut metric patterns dominate the poetic discourse, they can be followed. Quite often, however, the metric patterns are too complex and their use so flexible that they cannot be solely relied upon in creating the lines. In the Siri

32

Lanri Honko

epic, for example, there are passages which show a uniform prosodic pattern, but the patterns are many and their constancy is low. A metrical analysis would obviously require a simultaneous melodic analysis. Yet ethnomusicologists tend to regard the "music" of the Siri epic as mainly speech-like recitation, although the rhythmic patterns are so accentuated as to exclude mere speech analysis. We are clearly working in the grey zones of research where there are no experts able to tell when intonation turns into rhythmic recitation and recitation into melodious song. A method which can be applied as long as the final metrical analysis is not available is to follow the breath breaks of the singer as systematically as possible. They mostly create line-like structures which include the impact of rhythm and melody. Occasionally, however, rapid "runs" may produce overlong lines, and breathing may occur due to factors irrelevant from the prosodic point of view. Punctuation is another tool in creating the poetic appearance of an oral work. In oral discourse there is no punctuation; the presentation may be "additive", "repetitive", "direct (speech)", "reported (speech)", etc., but it usually avoids subordinate clauses. At least in Tulu, the poetic discourse enjoys nominal or participial constructions and other ways of substituting for sentence hierarchy. That is why we did not use any punctuation in our transcription of the Siri epic in spite of the fact that it reads almost like a script of a drama at times, full of direct speech and sentences framing it. The suggestion by Gene Roghair (1982:52) that in the epic discourse, for example, the "filler" words would have the same function as punctuation has in writing sounds interesting but cannot be systematically applied. The traditional discourses seem to possess strategies for sequencing that differ from one genre to another. Translation. Even if punctuation can be avoided in the rendering of the original language in phonetic script, for translation, however, punctuation is allimportant regardless of language, or, to put it more precisely, its importance depends more on the language into which the text will be translated. This is probably the critical shift from orality to literacy in the process of textualization. As absent as commas, full stops, semicolons and exclamation or question marks are in the original Tulu transcript of the Siri epic, as crucial they become in its English translation. Without them, the readability of our translation would sink dramatically. With their help, quite paradoxically, much of the orality can be saved and certain characteristics of Tulu, such as the relatively free word order or the multitude of participial constructions, as such untranslatable, can be preserved for viewing by the English reader, too. The lack of copulative conjunctions does not disturb in oral Tulu but in the English translation we could not manage without a comma denoting the

Text as process and practice

33

coordinative arrangement of adjacent words. In the process of translation, punctuation became a means of expressing oral sequentiality. Our philosophy of translation may be coined as a simple maxim: the translation should be "as verbatim as possible and as readable as necessary". We exclude, at least for the most part, the strategy of "functional equivalence", which would send us into the depths of the English language to seek historical, even archaic forms of vernacular, popular and poetic at the same time, which would "correspond" to the language of Tulu oral epic lays and provide us with not verbally or structurally similar but functionally equivalent expressions. (Honko 1998: 586.) For obvious reasons, the principle of "getting as close as possible to the original text" must be prioritized in the comparative research on oral and traditional epics. The target group of translation consists of informed scholars who will, it is hoped, be able to read through occasionally heavy and clumsy English the structures of expression in the original text. The presence of the original wording on the same double page invites comparison and in a way supports the verbatim translation, an effect which would be lost in the case of however impressive but far-fetched cultural equivalents. Vet, the readability of translation must be defended. This is best achieved by relying upon very basic vocabulary in English, even at the risk of simplicity, leaving some of the poetic impact of the text on the other side of the language barrier, i.e. untranslated. When the translation has not been "decorated" with elements for which there is no base in the original, the poetic merits remain where they belong, and the reader may feel relatively safe using the translation even without knowing the original language. The translation of a long oral epic requires even more thought on linguistic and ethnographic equivalence than the translation of short narrative poems. The text abounds in descriptions, multiforme, phrases and formulas, which are repeated dozens of times in slightly varying contexts (cotextual environments) in slightly varying forms, yet always easily recognizable. In our case, the translations made during the long textual work process tended to differ even in cases where the original text was practically the same. During the repeated discussions on the alternative translations of one and the same phrase or formula it was decided by our team that the principle of equivalence must apply here and that the translation should remain the same for a particular phrase or formula unless the cotext forces us to seek a modification of the standard translation. (Honko 1998:586-89.) Editing strategy. The decision to use one single performance of an oral epic saves the editor from many problems. First of all, it guarantees that as much as possible will be preserved from the orality and cohesion of the story. The oral text will directly reflect the "tradition system in use" and the compositional work of the singer. The performance of a whole oral epic in an

34

Lanri Honko

induced or artificial context eliminates certain factors which in the culturally ascribed contexts would exert quite considerable textual influence upon the oral work. Most important, it offers a possibility to visualize the mental text in its entirety by way of one application. Yet every oral text represents just one performance. The documented telling cannot claim primariness in view of other renderings. This multiplicity of form has made folklorists rush to the next singer and the next village to get more "variants" for comparative research. That we did not choose to do so when documenting the Siri epic was partly due to the complicated performance contexts of the epic, which we did not satisfactorily command, and partly to the exceptional resourcefulness of the singer we worked with. We had to deepen our "textual ethnography" in order to illuminate the variety of traditional performance contexts in ritual and daily labour, of collateral genres such as invocations and prayers, and above all, the renditions of the epic in varying lengths and different situational adaptations. Simultaneously, it was clear that Mr. Gopala Naika was more cooperative and able to discuss difficult questions concerning the composition and performance of oral epics than most singers. It was natural to concentrate on his repertoire of epics and also on his interpretation of them. We wanted the material from within one and the same tradition system where the elements of tradition interact with and influence each other. At the same time, we wanted to avoid the editorial dilemma which arises when diverging oral performances by the same singer are placed side by side and the construction of a textual mosaic becomes the task of an outsider. (Honko 1998: 596-98.) There are other strategies of editing oral epics which cannot be discussed here. They often arise from the necessities of the available materials and informants and usually involve stronger interventions by the editors than in our example. But even in our case we cannot say that orality has been transformed into printed text without any editorial intervention. What happened in transcription and translation made an impact on the original oral poetics. Our policy of linguistic verity also contained decisions on form and meaning. And what had to be left outside the text was no less than the full multivoiced, even noisy drama of live oral performance. Publication strategy. The book is the last link in a chain of compromises constituting the process of textualizing an oral epic. Our singer, Mr. Gopala Naika, although very warm to the idea of transmitting the epic to Tulunaadu people through books to be read in schools, had considerable difficulty in visualizing how the Siri epic could ever be put into book form. The performative and referential elements abolished in the process seemed so essential to him. Yet, the book represented for all of us a heroic attempt to save one oral epic from oblivion and end the textual poverty surrounding it. The epic deserved, we felt, the attribute "great" because of its size and narrative value. What could

Text as process and practice

35

be saved and what would be lost in the process of documentation and subsequent publication was a problem that kept haunting us. Yet our team produced the book, in three volumes, because it seemed to be the only way to approach the academic community interested in epics. If the future academe is more audiovisually oriented, we may consider a CDROM or multimedia presentation of our materials for Tulu speakers and, with translation and commentary, for classroom use at universities. A hard decision had to be made over notes. They were eliminated from the epic text simply to protect the readability of the Siri epic, a beautiful piece of narrative art even in translation. Culturally based expressions and metaphors are sometimes difficult to understand, but by keeping the translation close to the wording in Tulu, we deemed it possible for the reader to understand something on the basis of context, especially as the expressions keep repeating themselves. An occasional short reference to a ritual detail would precipitate a lengthy explanatory note but would simultaneously lead the reader beside the point. In other words, we favoured a flowing traditional epic texture instead of an encyclopaedia where the keywords fall rather haphazardly. We realized that the huge text mass will assist the reader in interpretation and bring forth an initiation rite whereby the epic register will become familiar to him. (Honko 1998:165-66,593-94.) The oral/literary paradox It should be clear by now that the publication of an oral epic is paradoxical. Mr. Gopala Naika's well-founded doubt that a book could convey the performance of the Siri epic underlines the fact that we are here facing acute difficulties of intersemiotic translation. The oral performance is stripped of a vast array of external sign systems (human voice, melody, instrumental music, dance, expressive movement, mime, gesture, dress, make-up, ritual objects, etc.) and internal resources of meaning (traditional referentiality, intertextual universe, religious belief, situational frame of mind, eventoriented behavioural patterns, etc.) not to speak of the psychological forces conjured up by the suggestive word power of the epic and collateral genres in their dramatic acting-out of not only the content of the story but also the shifts of identity among the participants. It is under such premises that the primary text is orally conceived and focused on a particular interactive audience. If losses are so profound, what is there to be gained through secondary written codification of an oral text? There are some gains, too. Documentation, preferably by audiovisual means, will save the oral text, which would otherwise disappear after performance, for multiple reviewing by members of the community around it, by scholars and interested audiences elsewhere

36

Lanri Honko

including future generations. Documentation paves the way for the oral work to the worldwide literature of epics where ethnic communities and nations are placed side by side in a comparison of their poetic contributions to the human oral literary heritage. Yet documentation provides only the base for publication. The written codification of the text represents a further step to be handled by intersemiotic translation. The stripping of the collateral sign systems present at the production of an oral text in favour of written letters always means decontextualization of the oral text regardless of how well contextual information has been collected and the speech event itself documented. Reception of a text changes as dramatically as the text itself in the process. The speech derived from a performance is redressed in letters which will perform in the reader's reception under totally different circumstances from those valid at the speech event. The form of written codification depends on the goals of the scribe and editor. If the goal is to revive the original speech event and enable the reader to take the responsibility for reconstruction of the situation by experiencing the text rather as a tool for that reconstruction than as a readable text, then we are heading toward a performance record and supporting performance report as described above. The goal is then "extended textualization". It is by no means easy under the tremendous pressure of the systemic force of notation techniques. A "total" description may prove unreadable, which is not necessarily against the philosophy of this operation aiming at the preservation of the oral at the cost of the written. Yet a meaningful description including only the most marked elements of the performance may help the analysis more than a perfectionist description with too many irrelevant details. Most efforts to create performance records focus on small genres. So far no attempt has been made to notate the performance of a long oral epic in full paralinguistic detail. The goal of most publications of oral epics hitherto has been another, namely, a readable text reflecting the original oral text with optimal linguistic accuracy but without sacrificing relatively easy comprehension. This goal implies a will to render the oral text as a reading experience, not as a reconstruction of singing, i.e. under the terms of literary reception, not under the terms of recreated orality. From the point of view of intersemiotic translation this alternative is defensible. If the documentation has been good and the editing careful, as much of the orality will be preserved as possible under the prerequisites of the receiving medium, i.e. writing. The oral singer in fact becomes an author for a text able to join literature proper. A source critical check is necessary if the reader of an oral epic wishes to estimate the nature and degree of orality of the work. Shortcomings in documentation, inadequate transcription or lack of the original speech, drastic editorial interventions and inaccurate translation can easily annihilate the

Text as process and practice

37

authorship of the oral singer. The work then glides to the category of traditional epic, preserving a solidarity to oral tradition, an adherence to traditional rules of expression, but becoming a product of "secondary singing" at best. In most traditional epics, the history of textualization is problematic. An assessment of the process ideally presupposes good reporting and publication of primary source materials, so splendidly achieved in the case of the Finnish national epic the Kalevala but sadly lacking in the majority of tradition-oriented epics. There is another paradox concerning the copyright. An oral work is much better protected against improper use than its written codification. The mental text of an oral epic goes to the grave with its author and cannot be effectively stolen even during his lifetime. Even if there is no master form for an oral text, its authorship is well protected, since its oral manifestations quickly disappear. Even the singer is unable to "quote himself" extensively; he must start afresh at each performance and create a new telling of the story. The copyright of traditional epics is as complex as their history of textualization and could be dealt with at considerable length (cf. Honko 1989:3-12). Suffice it to say that the roles of singers, scribes, editors and publishers must be assessed prior to any attempt at applying the controversial rulings of copyright laws in the case of expressions of folklore. For most older epics the question is academic and mainly of source-critical interest. One more paradox is that the textuality of an oral epic may rise to new heights thanks to a competent written codification. The medium of writing offers the possibility of reviewing and, consequently acquiring a much better understanding of the basic oral text, often left unsung and replaced by allusions, sometimes almost lost in the heat of a multifaceted oral performance. The published text does not allow such shortcuts, it requires the reader to follow the story in a neat sequence of episodes deemed important by the singer. It is impossible to escape the textuality so systematically nailed to its own premises rather than on those of the performance situation. If the oral text has any literary value, it will, after all, bloom, not wither, in writing. The medium itself is helpful in creating such value, even if the basic merit goes to the representative of oral verbal art, the singer. A literary work is normally separated from its author and begins to lead a life of its own. Most traditional epics, decontextualized and distanced from their shady origins, have fully enjoyed this literary advantage. Even accidental textualizations of oral epics may have achieved a textual monopoly as representatives of their poetic culture in world literature. In some cases such as the Finnish Kalevala, a process of continuous written performances has opened up new vistas across the oral/literary interface to the production of traditional epics. As pointed out, a solidarity to tradition by the compilers, occasionally backed up by Romantic theories of collective creation, has helped

38

Lauri Honko

to depersonalize the question of authorship of traditional (and oral) epics to the detriment of a realistic understanding of the life of oral poetry. The mission of contemporary research on oral epics is to localize the "voice" heard in the epic and its social ambience. Proper textualization, especially its second, written phase is an important tool in preventing the voice from getting suppressed. I mentioned that the possibility of adding footnotes to the translation of the Siri epic was rejected. The weightiest argument for this policy is what may be called the "economy of voice" (Honko 1998: 141-51, 594). As stated elsewhere, the integrity of the Siri epic as an oralliterary achievement and a great narrative would suffer each time the sophisticated voice of a scholar intervened and interrupted the narration, for which the continuous "flow" is essential both in its composition/performance and reception. The scholar would draw the reader's attention to himself and his knowledge, perhaps fairly irrelevant for the reader's actual needs. The voice of the singer would be subdued, to the detriment of understanding. The art of explanation may belong to the scholar's domain, although even there the singer's voice should be heard, if available; yet when it comes to performance, the word power belongs to the epic itself and the singer, not the scholar. We should remember that "explanation" is a discourse alien to the epic genre. A great epic does not explain anything, but as soon as you have heard the whole narrative, you will understand everything, including the details. (Honko 1998: 594.)

The contributions of this volume The problems of textualization are not regional but global in the sense that the same list of questions is valid regardless of the period or culture from which a particular epic under study derives. That is why Homerists, or some Homerists at least, are interested in African, Indian or Central Asian epics: empirical findings about the textualization of living oral epics are relevant for scholars trying to envision the textual process through which the Iliad and Odyssey came to us. On the other hand, the scholars assessing the nature of oral epic traditions in Africa, India or Central Asia are very much, some say too much, dominated by the impact that the Homeric epics have made upon the comparative research on epics. The powerful Western stereotype about "great epic" tends to mould our observations made in other parts of the world and the characteristics of non-European epics are in danger of becoming skewed or misjudged. One central task of modern research is to alleviate and relativize the inbuilt conceptual and methodological conflicts between research traditions based on dead and contextless text documents on the one hand, and on the vividly varying textualization of incomplete or unfinished oral epics on the other.

Text as process and practice

39

The studies on European epics largely belong to the former category, i.e. the textual base of European epics is complete, not to be augmented by empirical research. Contextual and text-historical information is scarce as a rule and the study of oral variation almost impossible. Yet the plausibility of the oral origins of the Iliad, Odyssey, Beowulf, Roland, El Cid, Nibelungenlied and the Eddas makes scholars search for analogues from fields and environments in which the epic textual universe is still alive and open to observation. The following statement in this volume by the classicist Minna Skafte Jensen reflects the interest in comparative knowledge and admirably summarizes the global "list of questions" to be answered by scholars textualizing oral epics: However, when classicists turn to the fieldwork of folklore and anthropology in order to find the relevant comparative information, they are often disappointed, even though the times are past when singers and storytellers were considered just "informants", bearers of traditions they perhaps hardly understood. Modern editors are becoming increasingly careful to contribute a wealth of detail about the singers, the performances, and their social contexts. However, since oral epics are flexible, since the audience has a central influence on the text performed, and since the fieldworker is a main part of the audience which is present at a recording - technically or in writing - it is of paramount importance to be informed of the scholar and his/her role. What was behind the scholar's choice of field, poetic form and individual singer(s)? What was the overall purpose of the fieldwork? Which were the questions the scholar wanted to find answers to? What training linguistic as well as methodical - had he/she gone through in advance? What were the singers asked for? How was the recording made? Who were present? How did they react? Where did the singer make his pauses? How many sessions were held? (Homerists as well as those who study other epic traditions from the past are of course especially eager to hear of dictation processes.) After the recording, what happened to the text between the first draft and the printed book? What kind of interpreting, commenting, copying and standardizing was undertaken? Such information is regrettably seldom given. (Below, p. 60.)

What Jensen calls for is not just more history of textualization but, in fact, a deconstruction of the entire process which led to the published epic in our hand. Not only the circumstances of documenting and editing an epic text but the entire strategy of the project, its motives, tools and resources, technical as well as ideological, are at stake. The particular problem, however, on which Jensen focuses her article, is the prime cause of written textualization in ancient Greece. Our stereotype concept of the oral/literary interface needs considerable fine tuning when applied to the Homeric epics. No external force put writing on the cultural map of ancient Greece; the art of writing was introduced for minor functions only and for centuries it neither dominated nor replaced oral performance. The reason for writing down the orally performed epic was excep-

40

Lauri Honko

tional, almost an anomaly, and the documentation did not have much consequence for performance traditions for a long time. The prime cause of textualization was an act of taking possession of the authority of epic by way of creating an epic text and its later use was more religious and political than performative. 1 One of the few places where empirical work on living European oral epic was possible until recent times is Bosnia/Hercegovina. John Miles Foley examines the methods of textualization in two interesting cases, namely, Vuk Karadzic's editing of Srpske narodne pjesme (Serbian Folk Songs, 1841-62) and the fieldworkby Milman Parry and Albert Lord which resulted in SerboCroatian Heroic Songs (1953-), an edition of oral epics still to be continued. Their comparison provides for a spectrum of different epics (Christian vs. Muslim), formats (short vs. long), narrative modes (cyclic vs. unified), target audiences (national, popular vs. international, scholarly), assisting persons (a network of collectors vs. a field assistant), choices of collected text ("representative" vs. strictly oral), methods of editing (light harmonization vs. strict accuracy) and topographical representativity (national vs. regional/ individual). Moving along the axis "what gets recorded" - "what gets published" "what gets received" Foley rectifies the misconception of Karadzic's alleged "freedom" in editing (being much less than in the Grimmian case) and notes the implied intertextualism in the reception of Srpske narodne pjesme due to an informed audience and points out the lack of similar intertexts in the Parry-Lord case. Interestingly, this observation casts a new perspective on the "difficulty" of reception in the case of Beowulf, for which the modern reader needs an extensive commentary where its contemporary audience managed with relevant intertexts. In Beowulf, according to Foley, "the apparently 'partial' nature of the text is not a paleographical or artistic blemish but betrays another way of accessing background information" (p. 84 below). As Foley exposes, Parry and Lord are the true pioneers of comparative epic research in general and oral-poetry fieldwork in particular. Their oral formulaic theory has indeed bred more interdisciplinary research on epic composition than any other theory in the 20th century although, it may be added, the fieldwork by epigones has been scanty or absent. Yet it is precisely in fieldwork that Parry and Lord excel in imagination and accuracy. Their three modes of documentation, "song", "recital" and "dictation", their multiple recordings of the same and different epics by the same and different singers and their experimentation with interrupted performance and a new start from the beginning are way beyond the customary level of collecting oral poetry. As so often with imaginative collecting and research, the problem is the richness of the material: how to create the cadres able to utilize the available resources in full.

Text as process and practice

41

In the opposite corner of Europe, Old Norse-Icelandic poetry provides no long epic, according to Joseph Harris, only "an arrested development in that direction", i.e. from lay to epos, in a few late eddic poems. Yet Iceland is blessed with rich prose epic. The exact nature of authorship and textualization is controversial, because the oral stages and oral performances are described parsimoniously at best. What seems fairly well attested is, however, an interplay between prose and verse. The prose sagas display copious verses, probably a reflection of their oral mode of performance. In Norna-Gests dattr there is a description of an itinerant non-professional entertainer who mixes eddic verse with sagas. Harris examines the elegy, a genre which we find just as eulogy - incorporated in long epics in many cultures. The eddic elegies deal with death, loss, and consolation in a manner which discloses a woman's voice at their core, possibly stemming from funeral laments and "distantly reminiscent of the Geatisc meowle whose lament for Beowulf includes her own fears" (below, p. 95). The function of the elegy as well as of Snorri's Prose Edda is scrutinized from the point of view of textualization. The cultural area phenomenologically closest to Europe as far as epics are concerned is probably Central Asia. Both formats, long and short, are found in many linguistic groups, and the modes of epic performance well correspond to what we know and conjecture about the performance of European epics. Karl Reichl, a specialist on Turkic oral epics who has worked with the last singers of long epics among the Karakalpaks, devotes his analysis of textualization to the most elusive components, the living sound of epic and the paralinguistic features of composition contributing to the construction of meaning in performance. He emphasizes the fact that "the edition of an orally performed poem, song, narrative, sermon or speech is by no means the simple conversion of the spoken word into the written word" (below, p. 105). It is the entire speech-event which requires transcription and "the edition of an oral text must be supplemented by at least a partial documentation of the elements making up the speech-event" (cf. performance record cum report above p. 12). The key problem, says Reichl, is what to include in the written representation of the speech-event: should the "text" be a transcription (documentation) or rather an edition of the event? Even the most complete replica of the event "is radically different from the event itself" and "the audience of a speech event takes part in the synchronous event... while the audience of a replay watches a past event without being a part of it" (p. 106). That is why a replica may not be the ultimate goal of scholarship in the field of epic studies but rather a singer's "mental text" as reflected in its various manifestations in performance. Reichl invites us to look beyond performance and exercise critical activity also implying selection and not only complete documentation. His exemplification of how to treat various modes of performance also allowing for their musical aspects, paralinguistic features (bow strokes, drinking tea, etc.), phonetic or phonemic

42

Lanri Honko

structure, essential and non-essential features of prosody and the unexplored grey zones between intonation, recitation and singing is rich indeed. Any edition of text will transform the living event and silence the singer's voice, Reichl says, but we must develop methods of rescuing as much as we can of that primary epic voice. Arthur T. Hatto approaches another field of oral epics, probably less known than Central Asia but part of the northern periphery of Turkic-related epics, namely, Siberia. He addresses neglected types of heroism, i.e. solitary scholars as heroes of textualization fighting their philological battles under onerous conditions and oral bards as custodians of traditional expressions even in cases where their meaning has waned away. The endeavours of a single explorer and collector of linguistic specimens, say, the Hungarian Antal Reguly (1819-58), represent a prototype: taxing travel through unexplored territory, writing down cumbersome dialectal language in a phonetic system simultaneously developed for the purpose, little or no contextual information on the texts thus created and demanding posterior editorial work by generations of scholars mostly outside any contact with the culture from which the text documents derive (cf. pp. 149-51). The nature of "musical text", the problem of stichic or just rhythmic recitation and the peculiarities of genre-specific discourse abound to make the seemingly easy transformation from oral to written word a series of compromises. On the side of native custodians of oral epic we have an even broader register of textual heroes ranging from the "reluctance of Burjat singers to let their words be fixed in writing" and "zealots of clan traditions" demanding "faithful transmission of essential content from the bards" (p. 134) with "very stable" epics as a result (p. 135) but small chance of their documentation to the great Evenk singer-turned-collector N. G. Trofimov (1915-71), who wrote down the Sodani epic "under canvas and by candlelight when each hard day's work herding the reindeer was done" incited by the "responsibility for a tradition to which very few had access" (p. 139). The ailing bard had to be flown to a hospital, where he died on the operating-table. An exercise-book addressed to A. N. Myreeva, an epic researcher, was found in his hold-all. It contained his record of the Sodani epic. Hatto's screening of Siberian oral epics from one language area to another is an impressive testimony of the riches of tradition and the immense difficulty in securing at least small slices of that heritage for posterity and scholarly viewing. The story remains essentially the same when Hatto crosses the Pacific and examines Alfred Kroeber's textualization of Inyo-kutavere's Mohave epic. Kroeber faced all the difficulties of interview and interpretation through a native field assistant and their impact on textualization, a subject seldom discussed as openly as by him. Hatto's own editing of Kirghiz heroic epics disclosed that the Kökötöy MS was not a field-record but a fair copy, a frequent but often by-passed phenomenon in the history of oral textualization.

Text as process and practice

43

The complexities of diacritics in texts codified in different (Arabic, Cyrillic, etc.) scripts were alleviated, somewhat unexpectedly perhaps, by the existence of a supra-dialectal epic diction to be compared with "the special operational jargons in warfare that ease the cryptographer's task" (p. 153). It is this diction which seems to survive and penetrate even the poorest transcriptions and translations in fact, as Hatto wittily points out, making oral epics hard to destroy. Juha Pentikäinen's example of oral textualization concerns just one ritual for the dead documented by him among the Nanaj in Siberia, in the lower Amur region. The text given in full does not represent an epic yet reflects a narrative idea about the journey to the Beyond, Buni. It is a combination of addresses, invocations, prayers and incantations sung or recited by a saman of a special kind, a master of the death ritual. It is surrounded by a peculiar commentary due to the fact that the ritual in question had been banned in the Stalin purges of the late 1930s and had survived dormant as part of memory culture for over 50 years, until it was revived in 1991 upon the request of scholars. A continuous negotiation on what to do next in the ritual frames the performance and the liturgy proper expresses uncertainty and fear of ritual error culminating in parodical remarks on the "fake" character of worship. In spite of this, the archaic ritual was able to impose the role of a psychopomp on most participants who felt that they were leading their relatives to Buni, a long suppressed wish which now found its fulfilment. A small but interesting piece of tradition revival reminds us of the role that short-format ritual genres may play in the textualization of a narrative idea. India is the motherland of epics. Nowhere has the literary tradition of epics interacted with abundant oral epic traditions for as long as in India. The oral origins of the great Sanskrit epics, the Mahäbhärata and the Rämäyana, have been recognized despite the fact that both epics possess a personal "author", Vyäsa and Välmiki. Another paradox is the statement, shared by the specialist John Brockington in this volume, that "oral tradition can be a far more accurate medium of transmission than writing" and that "the agency for fixing a text" can be found in "the Indian concept of oral transmission rather than writing" (cf. p. 209). The classical epics have "resisted the levelling effects usually associated with oral transmission" and retained original linguistic and stylistic features. Even if certain passages have been reworked by reciters, "it seems also that admiration or respect for Vyäsa and Välmiki has been a powerful influence in retaining much of the Sanskrit epics in their earliest form" (p. 210). Instead of replacement, accretion has been the method utilized by bards, reciters, scribes and copyists, and the result has been "inflation of the text". On the other hand, development and change leading to multiple variation have characterized the textual universe of the Sanskrit epics spreading over to numerous other cultures and languages during the past two mil-

44

Lanri Honko

lennia. The variation has grown so wide that over seventy years ago a scholar, frustrated in his text-critical attempt to reconstruct a plausible original for the varying recensions, denied the existence of an original Rämäyana (Hopkins 1926), and as late as 1991 A. K. Ramanujan gave the title "Three Hundred Rämäyanas" to his survey of the Rämäyana traditions in Annamese, Balinese, Bengali, Cambodian, Chinese, Gujarati, Javanese, Kannada, Kashmiri, Khotanese, Laotian, Malaysian, Marathi, Oriya, Prakrit, Sanskrit, Santali, Sinhalese, Tamil, Telugu, Thai and Tibetan, each language hosting numerous tellings of the epic in different formats and modes of performance and sometimes retaining in oral form sets of themes unknown to Välmiki (Ramanujan 1991: 24, 37). Is there a common core to the Räma stories, Ramanujan asks, "except the most skeletal set of relations" between Räma, his brother, his wife and the antagonist Ravana? Such a relational structure ascribes the name Rämäyana to tellings which on closer look are not similar at all. "Like a collection of people with the same proper name, they make a class in name alone," says Ramanujan (1991: 46). Conservatism and innovation, stability and flexible adaptability, seem to be two sides of the same coin, the perseveration of oral epic traditions. Development and change constitute the backbone of Brockington's analysis, too. Examining the occurrence of formulaic expressions and other poetic means in detail, he has been able to delineate the textual history of the Sanskrit epics in five stages. Although initially distinct, "by the third stage in the growth of the Rämäyana, the originally separate traditions of the Mahäbhärata and the Rämäyana were coalescing" (p. 211). It was in that period, "beginning probably in the first century AD", that the two works were committed to writing. After that written "borrowing of formulae" also took place. The increasing frequency of formulaic expressions in the latter parts of both epics "seems to be not an index of orality but rather a sign of the decay of the genuine oral tradition". Other features such as ring-composition in speeches, certain patterns of repetition and the employment of standard themes "are likely to be better indicators of oral composition and transmission" (p. 211). For a long period of time, the singers of the Sanskrit epics continued to perform, but at one point or another the scribes who wrote the manuscripts superseded them. One might add that the voiced performances were always partial and situational. They could very well survive, as even contemporary theatrical tradition shows, without competing in any way with the scriptural performances aiming at greater length and better cohesion. A special feature of the epic scene in India is the abundance of living oral epics. Not only are the classic Sanskrit epics still very much alive in both oral and written performance, they also interact with a multitude of local epic formations with or without affinity of plot, poetic means, mode of performance or performative style. India is a place where scholarly discoveries can be made and practically unknown epics of a longer or shorter format

Text as process and practice

45

documented in oral composition and performance. An example is the documentation of the Siri epic of the Tulu-speaking people in southern Karnataka by the present writer and a team of Finnish and Tulu scholars in December 1990 based on a single sung performance by Mr. Gopala Naika, an illiterate agriculturalist and possession priest. The 15,683-line oral epic is the longest and textually most accurate version of the Siri epic published so far. It represents the fullest manifestation of the singer's mental text. The topic of the article is the impact on oral text exerted by the context of its performance. We are dealing with an epic in a "strong context" capable of breaking our stereotype of singing oral poetry in special sessions arranged for the singer and his audience. Hie performance of the Siri epic and other Tulu epics is embedded in a dominant action, be it a sacred possession ritual at a religious festival or a profane work process of plucking seedlings in the paddyfield. The singer is not only a singer of epics but, for example, a possession priest conducting a complex ritual. Possession brings the divinized personages of the epic to the scene and enables them to occupy the bodies of participants and take responsibility for the worship directed at higher gods. The epic is no longer narrated, it is danced and played out in the flesh. The three generations of epic personages mix and interact in a confluence of divine powers and human concerns. If the epic is myth, its performance is ritual. The basic story remains untold as a linear narrative but functions as a mental backdrop filling the minds of the participants. What is said and sung, however, is more representative of the situation than the long epic as a story. The strong context almost annihilated textualization in the ordinary sense of the word. The evasive text of the "whole" and coherent epic had to be sought in an induced context of performance where the disturbing collateral action was eliminated and the singer was given the reins of the situation. The broken stereotype had to be revived in order to allow the singer to fulfil his dream of the total epic, impractical or unwanted in his culture but representing religious and oral-literary values for its composer. Academic skill was invited to save a precious heritage of Tulunaadu, perhaps to be read in schools one day. Empirical research on oral epics has been very successful in Africa in recent decades. John William Johnson scrutinizes the shift in the premises of textualization which occurred parallel to the breakthrough of modern audiovisual documentation technology and the new emphasis on cultural context and ethnographic authenticity. In earlier times literary polishing of oral text was not frowned upon, and composite texts synthesized from elements stemming from different performances could be taken as a textual improvement. The fact that these texts were never recited by anyone did not disturb the comparatists using them in their epic analyses. Mainly from his experience on Somali and Mande epics Johnson summarizes the modern requirements as consisting of fieldwork conducted with the people who perform the epic,

46

Lauri Honko

collecting of text with a tape recorder in a natural or induced context, transcribing the text verbatim, and translating the text with the aid of local informants, "who help write extensive notes and annotations that help translate the culture as well as the language of the epic" (p. 243). The idea of an "invisible collector" is deemed unrealistic by Johnson. Collectors affect the process of textualization and easily become a part of not only the narrating event but the narrative itself as many epic singers tend to incorporate remarks and observations on their audience in their song. According to Johnson, texts produced in "laboratory" conditions or in an "induced natural context" are "rich in creativity, narrative development, and cultural information, though unrelated to any specific social event" (p. 240). The book is still the main means of publishing the results of documentation, says Johnson, but envisions that the lack of the all-important musical accompaniment, which affects language, prosody and performative style, will force us to consider other forms of publication (CD-ROM with Hypercard, etc.) more adequate for conveying the form and meaning of epic expression in performance. He also refers to "mental text" as the point of departure for documentation and defends the authenticity of "full" or "whole" oral epic texts even when they are performed for the first time in a fieldwork situation. Jan Knappert, a connoisseur of African, and especially Swahili epics, of which he presents an impressive list with 75 titles, used a tape recorder in his collecting but still clearly represents an older stratum of textualization methodology. The key statement (p. 254) is that "of all the different versions that I could collect of one epic, hand-written or on tape, not one was ever complete". The quest for a master text among the versions, to be supplemented with elements from other versions in order to "reconstruct" an "ideal text", inevitably leads to a composite epic which never existed in the minds of oral singers. This leaves us with two alternatives: either the compiler has produced something non-traditional and thus artificial, or he has risen to the level of singers by way of commanding the traditional rules of composition and accomplishing a written performance of the epic which compares well with the oral ones. In the modern view, the existence of a "complete" epic outside the individual mental texts of the singers is an illusion, and thus a "reconstruction" is not possible, even in cases where the "versions" stem from one and the same singer (whose mental text is not stable but develops over his/her performance career). For Knappert, however, such a reconstruction seems possible with the help of the best native experts: "Later, the complete text will have to be constructed, with the help of a scholar who remembers the outline but is too old to recite or sing it" (pp. 250-51). This brings to mind Elias Lönnrot's encounter with Vaassila Kieleväinen, an old man unable to sing anymore but willing to narrate his outline of parts of the Kalevala story. The

Text as process and practice

47

quest for the "best" and/or "complete" form is a recurring phenomenon in the history of epic textualization. Yet it leaves the compiler at the mercy of his helpers, who may not be easy to reconcile. As Knappert points out, "the researcher is completely dependent on a small number of soi-disant experts who will tell him: this singer is charlatan, that one is talented and knows the text and the tunes, and: he can sing" (p. 251). The risk is, simply, that a composite text may not satisfy anyone and will give misleading information on the local poetic system. Another not very modern feature is that Knappert warns against duplication in collecting oral texts. "Singers or reciters will offer the same titles over and over again... [yet] priority has to be given to unknown material" (p. 251). The statement goes against the need for "thick" materials for the study of organic variation and rules of oral composition for which multiple performances of the same song by the same (and other) singers are needed. Most collectors have, however, followed Knappert's line, which is shown by the fact that thick corpora of material are so difficult to find. Paradigmatic differences aside, Knappert has several good recommendations for younger researchers. One is the emphasis on the importance of understanding the epic language (which is not identical with the command of language in general) and the suggestion that grammars be written for great oral epics in order to analyze their language. Dwight F. Reynolds's article is one of the best field reports on oral textualization and its written codification in the present volume. He describes in detail his documentation of the Sirat Bani Hiläl epic in northern Egypt in the 1980s focusing on "the interaction between theoretical models on the one hand and the pragmatic choices and tough compromises we must all make if our work is to get published on the other" (p. 264). The epic was conceived as a single story by the singers even if only episodes or strings of episodes could be performed in the contexts available for the singing of epics. Reynolds does not use the term "mental text", yet he goes even a bit further in his formulation of oral fixity: "One way of conceiving of this is to say that the text has a definite form in the minds of the poets, but is realized in sections which are structured as 'performance units' "(pp. 265-66). In performances lasting 1-3 nights episodes were sequenced by "bridges", short units carrying the plot from one episode to the next. The performance tradition, however, was developing toward one-episodic mode and the "bridges" were becoming obsolete and difficult to document. This shift of performative format seemed to contribute to the disintegration of the whole narrative into separate and more independent stories. Being an ethnomusicologist as well, Reynolds could take the extraordinary step toward apprenticeship and learn the epic tradition from the inside, too, both textually and musically. His teacher found him "astonishingly adept at imitating him on the two-stringed spike fiddle, the rabäb, and

48

Latiri Honko

astonishingly slow when it came to assimilating text" (p. 268). While teaching his apprentice, the singer master did not change the oral text over several days, but in public performance he could sing the same passages with substantially different wording. In other words, "he was quite capable of both engaging in and refraining from textual variation" (p. 268). Interestingly, Reynolds was able to test his teacher's tolerance on variation by singing passages on different melodies to see which he would find acceptable and which he would reject, or by altering the rhyme-scheme and by summarizing or adding details "to discover his sense of a well-told tale" (p. 269). Active apprenticeship proved "one of the most effective tools for comprehending the parameters of variation within a tradition" (p. 268). As his documentation and editing strategy Reynolds chose concentration on one singer's repertory and became "very firmly convinced of the centrality of the individual poet's vision of his art" (p. 269). Textually, his singer had greater literary cohesion in his epic than other performers using heavily transformal, extra-linguistic means of poetic expression. In this sense he was easier to textualize for publication. The documentation took place at evening gatherings in private homes according to the wish of Shaykh Tähä, the singer, who took active responsibility for the process of editing, too. For example, he supplied missing episodes in later performances, as the narrative required. Reynolds is lucid on his strategies of transcription, translation and publication of "the projected Arabic edition and English translation" emphasizing the dialogic methodology of textualization he utilized in his work not only with Shaykh Tähä (who soon passed away) but other singers, too. For Reynolds, the deciphering of traditional expression was a central element of his cooperation with the singers at the editing phase of epic text. Dan Ben-Amos views "the narrator as an editor" from a slightly similar angle. His fieldwork experience among the Edo in Nigeria (in 1966 and later) confronted him with the "aesthetics of linguistic mystery: Edo divi, 'deep Edo'": "Spellbound by a familiar plot, the listeners resigned to the fact that master narrators use idioms that they neither recognize nor fully comprehend." Digging deeper, Ben-Amos found that his field assistant was ignorant of the meaning of certain expressions and tried to circumvent them by offering translations "intended to smooth out difficulties rather than to expose the complexity, or obscurity of meaning" (p. 284). It was only the singer, Iditua, who could clarify that certain phrases denoting father and mother, for example, actually referred to his own father and mother in a way not understood by the field assistant or the audience in general. That a "national epic" of the sacrosanct past could contain personal elements of a private nature is well-nigh impossible according to our (faulty) stereotype, but BenAmos produces evidence that "when the narrator is introduced into the process of interpretation and explication, the performance of the epic becomes very much part of the present" (p. 286).

Text as process and practice

49

Ben-Amos also surveys the famous battle around whether "there is an epic in Africa" or not. The four contributions on African epics in this volume testify to the affirmative. As a possible explanation for the conflict of opinion the definition of epic has been used. Another factor seems to be a certain insensitivity among scholars to the existence of a long epic narrative in traditions they have studied. An example is R. E. Bradbury, an anthropologist who conducted field research among the Edo for nine years from 1951 to 1960 and, it appeared, covered every aspect of their life but who did not give a clue to the existence of long epics. In 1966 Dan Ben-Amos documented the Agboghidi and Odologie epics from a master singer who had been one of Bradbury's key informants. Another area which seemed to be lacking a long epic is North America. In his article (actually a small monograph) Dell Hymes explores the problem in Native American tradition and identifies three oral forms which seem to approach the sung epic but concludes that in Native North America "narrative is essentially unsung" (p. 292). He then expands his analysis to the poetics of oral narrative in general, delineating his theory of oral composition, often referred to as ethnopoetics, a theory which he has been developing for several decades. Hymes makes the all-important distinction between the internal and external organization of poetic lines, the former being metrical and dominating song and the latter being "measured" and typical of spoken narrative. The key units in the analysis of the latter are line and line order, "defined by number and kind of relations among lines, rather than within them" (p. 296). For larger units other terms will be employed (verse, stanza, act, etc.). The focus is on basic kinds of recurrence of units such as relations involving the numbers three and five, or two and four. These "cultural pattern numbers" dominate the architecture of narratives, "mostly out of awareness by those who use them, like so much of grammar" (p. 297). The exemplification of his thesis is Native American but on the basis of a few Old World samples Hymes is ready to presume that "such relations are not particular to one part of the world, but common to human languages" (pp. 297-98). The exciting thing about the approach is that it reveals (hidden) structure on the surface level of language, i.e. texture, and makes form and meaning inseparable. Hymes is critical of the way in which the phonemic representation of language conquered the scene and made the presentation of texts in the phonetic detail of their original transcription outmoded. "Here is a sphere," he says, "in which the requirements of recovering and understanding performance conflict with mainstream linguistics, proceeding as it mostly does as if language has but a single function" (p. 300). He believes that his kind of work "must become the foundation of an understanding of Native American mythology and spoken narrative" and concludes (p. 295):

50

Lauri Honko As for linguists and folklorists, not many are comfortable working with both linguistic detail and "poetry". Yet such work is fundamental to the philology, or ethnopoetics, or whatever one chooses to call it, of recorded traditions. Whatever the name, such work heightens the meaningfulness of the texts and the respect in which they can be held. It reveals much of beauty that now lies hidden. I like to say that it "liberates" the original form.

How to save the beauty of oral textualization in its written codification without compromising the accuracy of expression is a question which does, or should, haunt every conscientious epic scholar. The task is well-nigh impossible, but Dell Hymes puts the emphasis on a key point: what is at stake is our ability to read critically the text offered to us and our ability to listen to the voice behind it. In the concluding contribution Anna-Leena Siikala explores the neglected margins and backgrounds of the making of epics. Her fieldwork experience is from the Southern Cook Islands, especially Mauke, and the material she analyzes lies at the oral/literary interface, at the crossings of myth and history, of tradition and individual experience cum interpretation and at the boundaries of formal epic genres and more general but epic-related discourses. First of all, she questions "the European model of legend collection" as separate stories which "does not provide room for representations of inner relations of narrative corpus or alternative ways of creating coherence for it" and mostly leads to loss of original meaning (pp. 314 15). Secondly she warns against forcing our Western view on textualization upon non-Western cultures which possess other tradition-based strategies of written codification, notably hand-written manuscripts used as aids for oral performance but without attempting to replace oral discourse with the written one. The qualified narrators of "mythic history" in Mauke master the entire (local) history of the world as a continuum. They are able to produce a shorter or longer version of that history as a narrative which takes the cosmogony and genealogy of chiefs as the basis for linking the ancestors of the storyteller as well as his community to the sacred and mostly secret origin in a politically favourable way. " A proper knowledge of the genealogies and connected narratives can be used as a valuable weapon in disputes over land rights and titles", Siikala says and adds that the script of the performance always presumes wider (often untold) structures of knowledge "concerning the internal connections and mutual hierarchies of narrated and not-narrated" (p. 351). This intertextuality and the implied meanings are poorly if at all reflected in the published collections of narratives. As an example Siikala scrutinizes a kind of "personal epic", a 36-page oral narrative by Samuela-TeRei-More-Taunga-o-te-tini who performed it to Stephen Savage in 1916. Samuela was 62 years old and fluent in both Samoan and Rarotongan Maori (the most important names were given in both). The narrative was created to replace a lost hand-written manuscript of the world

Text as process and practice

51

and family history owned and performed by Samuela's father. According to Siikala, such "manuscripts crystallize the internal relationships of individual texts and give alternative models for the composition of larger entities" (p. 357). Memorizing his father's narrative, Samuela actually composed an epic about the world and his own life, ultimately a kind of "invention of self". It may not qualify as a national epic, yet the discourse employed is clearly a "superdiscourse", a medium for cultural, not merely personal, identity and an analysis of social hierarchy and power relations. Individuality of interpretation does not exclude traditional expression and social adequacy.

Note 1. Two peripheral remarks provoked by Jensen's analysis may be allowed. While she states that fieldwork evidence shows "without exception" (p. 59) that the idea of recording an oral epic in writing does not originate with the singer or his audience, the present volume offers examples to the contrary (e.g. the Siri epic, p. 221). This is just one more case illustrating how difficult it is to generalize about epics, even if we agree that mostly the reasons for documenting the epics are external to the culture maintaining the tradition. - Another detail worth noting is the fact that Jensen deals with the "first" written codification of Homeric epos, not with the textual history of Homeric manuscripts (131 of which were devoured by flames with the Library of Alexandria) and its famous gap of more than a thousand years until the first complete text surviving to us saw the light of day (Venetus Marcianus 454, see e.g. Foley 1990: 26-27). The long history of editing and copying makes us relegate the Homeric epics to "traditional", not oral epics.

References Bauman, Richard and Charles L. Briggs 1990 Poetics and Performance as Critical Perspectives on Language and Social Life. Annual Review of Anthropology 19: 59-88. Bauman, Richard and Joel Sherzer (eds.) 1989 [1974] Explorations in the Ethnography of Speaking. 2nd ed. Cambridge: Cambridge University Press. Clayton, Jay and Rothstein, Eric (eds.) 1991 Influence and Intertextnality in Literary History. Madison: University of Wisconsin Press. Doane, A. N. 1991 Oral Texts, Intertexts, and Intratexts: Editing Old English. In Clayton and Rothstein 1991: 75-113. Fine, Elizabeth C. 1984 The Folklore Text. From Performance to Print. Bloomington: Indiana University Press.

52

Laitri Honko

Finnegan, Ruth 1988 Literacy and Orality. Studies in the Technology of Communication. Oxford: Basil Blackwell. Foley, John Miles 1990 Traditional Oral Epic. The Odyssey, Beowulf, and the Serbo-Croatian Return Song. Berkeley, Los Angeles and London: University of California Press. 1995a Folk Literature. In: Greetham 1995: 600-26. 1995b The Singer of Tales in Performance. Bloomington and Indianapolis: Indiana University Press. Greetham, D. C. (ed.) 1995 Scholarly Editing. A Guide to Research. New York: The Modern Language Association of America. Goody, Jack 1987 The Interface Between the Written and the Oral. Cambridge: Cambridge University Press. Halpert, Herbert 1947 American Regional Folklore. Journal of American Folklore 60: 35566. Honko, Lauri 1974 Balto-Finnic Lament Poetry. Stadia Fennica 17. Helsinki: Finnish Literature Society. 1981 Four Forms of Adaptation of Tradition. In Honko and Voigt 1981: 19-33. 1989 The Final Text of the Recommendation for the Safeguarding of Folklore. Nordic Institute of Folklore Newsletter 2-3/1989: 3-12. 1996 Epics along the Silk Roads: Mental Text, Performance, and Written Codification. Oral Tradition 11/1:1-17. 1998 Textualising the Siri Epic. (FF Communications 264.) Helsinki: Academia Scientiarum Fennica. 2000 Thick Corpus and Organic Variation: An Introduction. In: Honko 2000: 3-28. Honko, Lauri, Senni Timonen and Michael Branch 1993 The Great Bear. A Thematic Anthology of Oral Poetry in the Finno-Ugrian Languages. Poems translated by Keith Bosley. Helsinki: Finnish Literature Society. Honko, Lauri in collab. with Chinnappa Gowda, Anneli Honko and Viveka Rai 1998a, b The Siri Epic as Performed by Gopala Naika. I, II. (FF Communications 265,266.) Helsinki: Academia Scientiarum Fennica. Honko, Lauri (ed.) 2000 Thick Corpus, Organic Variation and Textuality in Oral Tradition. (Studia Fennica Folkloristica 7.) Helsinki: Finnish Literature Society. Honko, Lauri and Vilmos Voigt (eds.) 1980 Genre, Structure and Reproduction in Oral Literature. Budapest: Akademiai Kiado.

Text as process and practice

53

Hopkins, E. Washburn 1926 The original Rämäyana. Journal of the American Oriental Society 46: 202-19. Hymes, Dell 1989 [1974] Ways of Speaking. In Bauman and Sherzer 1989: 433-51,473-74. Jakobson, Roman and Petr Bogatyrev 1978 On the Boundary between Studies of Folklore and Literature. In: Ladislav Matejka and Krystyna Pomorska (eds.), Readings in Russian Poetics: Formalist and Structuralist Views. Michigan Slavic Contributions 8. Ann Arbor, Michigan: Michigan Slavic Publications. Krohn, Kaarle 1926 Die folkloristische Arbeitsmethode begründet von Julius Krohn und weitergefiihrt von nordischen Forschern. (Institut for sammenlignende kulturforskning B5.) Oslo: H. Aschehoug & Co. 1971 Folklore Methodology. Austin and London: University of Texas Press. Porthan, Henrik Gabriel 1983 [1766-1778] Suomalaisesta runoudesta [On Finnish poetry]. Trans, and introduction by Iiro Kajanto. Helsinki: Suomalaisen Kirjallisuuden Seura. Lord, Albert B. 1960 The Singer of Tales. Cambridge Mass.: Harvard University Press. Ong, Walter J. 1982 Orality and Literacy. The Technologizing of the Word. London and New York: Methuen. Nida, Eugene A. 1964 Toward a Science of Translating. Leiden: E. J. Brill. Propp, Vladimir 1984 [1976] Kalevala kansanrunouden valossa [The Kalevala in the light of folk poetry]. Kalevalaseuran vuosikirja 64:67-79. Pieksämäki: Suomalaisen Kirjallisuuden Seura. [Presented in 1949 and posthumously published in V. J. Propp, Fol'klor i deistvitel'nost' (Folklore and reality) in 1976.] Ramanujan, A. K. 1991 Three Hundred Rämäyanas: Five Examples and Three Thoughts on Translation. In Richman 1991: 22^49. Richman, Paula (ed.) 1991 Many Rämäyanas. The Diversity of a Narrative Tradition in South Asia. Berkeley & Los Angeles & Oxford: University of California Press. Ricoeur, Paul 1991 From Text to Action. Essays in Hermeneutics, II. Trans, by Kathleen Blarney and John B. Thompson. Evanston, Illinois: Northwestern University Press. Roghair, Gene H. 1982 The Epic ofPalnädu. A Study and Translation ofPalnäti Vtrula Katha, a Telugu Oral Tradition from Andhra Pradesh, India. Oxford: Oxford University Press.

54

Lanri Honko

Zumthor, Paul 1990

Oral Poetry. An Introduction. (Theory and History of Literature 70.) Minneapolis: University of Minnesota Press.

European epics

The writing of the Iliad and the Odyssey Minna Skafte Jensen Epic comparisons There is a long tradition for comparative studies in the field of epic poetry. Over the centuries, Homeric scholars have fetched inspiration from fieldworkers who collect and analyse living oral traditions, and these have, on the other hand, been inspired by Homeric scholarship. F. A. Wolf, who with his Prolegomena ad Homerum (1795) initiated modern theories of Homer, was deeply influenced by the great folklorists of his day, and when he concluded that it would be impossible ever to establish the original form of the two Greek epics, it was mainly because he was impressed by the fact that oral transmitters of a tradition tend to change their texts all the time. He was engaged in making a new edition of the poems and was concerned with the basic question which text he was supposed to restore, Homer's original or the version established by Hellenistic scholars at the Museum in Alexandria c. 250-150 B.C. Since the art of writing was considered unknown in Greece in Homer's time he chose the latter alternative as the only one possible. During the following generations not least Karl Lachmann, no less authority on the Nibelungenlied than on Homer, was building his Liedertheorie on analogies of how oral epics were known to function. His views, on their side, were having their impact on fieldworkers, notably on Elias Lönnrot, who did to his collected Finnish Lieder what he thought Homer had done once upon a time: built one great epic out of briefer lays (Honko 1990). The "oral-formulaic theory" of Milman Parry and A. B. Lord stimulated new interest in comparisons.1 If, as they maintained, all literature composed orally has certain characteristics in common as distinct from literature composed in writing, it is important to investigate such features across boundaries of time and space. A detailed knowledge of common working conditions, techniques, aesthetic ideals etc. is important for any study of a given oral tradition, poetic form or individual text, if we want to understand them in an appropriate frame of reference. Since Western scholars are sharing a background that is in all kinds of ways dominated by writing, the risk of cultural misunderstandings in the study of oral poetry - both fieldwork in living oral traditions in our own time and study of texts from earlier periods transmitted in writing but considered to have been composed orally - is close to hand and a factor that researchers must be carefully aware of. Homeric scholarship since Milman Parry has been divided over the issue. There have been periods in which the oral-formulaic theory seemed to offer

58

Minna Skafte Jensen

the most central modern re-formulation of the classical Homeric question, that of single or multiple authorship of the Iliad and the Odyssey (Lesky 1963 [1957/8]: 8) and others in which a majority of scholars rejected the theory as irrelevant. These days some of the most dynamic Homeric scholars are trying to obtain renewed inspiration from comparisons with living oral traditions, whereas many feel uneasy about it and state that the theory has in actual fact brought very little useful insight. In a solid volume from 1991 edited by Joachim Latacz with the purpose of mapping modern Homeric studies, the opinions about the oral-formulaic theory vary from some scholars feeling that there is nothing more to learn from it to another one criticizing Germanspeaking Homerists for never really having taken it into proper consideration. And the scholar entrusted with surveying this part of Homeric scholarship speaks in a highly condescending way of comparisons with "Yugoslav or Old English or Bantu or what have you". 2 This is a revealing phrase; many classical scholars feel that to compare the great Homer with illiterate singers of our own time comes close to blasphemy. I have discussed the ideological aspects of the matter in some detail in a former study (Jensen 1998). However, since to my mind the oral-formulaic theory not only changed the form of the Homeric question, but actually offered a solution to it (Jensen 1980), I consider it imperative that Homeric scholarship should stick to the theory and take it as the necessary base on which any further study must be built. I admit that it makes some familiar kinds of research impossible, and that some problems which used to seem simple have become difficult. On the other hand, it has been the great stimulus in Homeric scholarship of the 20th century; especially, it has opened the way for new readings which have revolutionized the aesthetic and historical understanding of the two great epics. 3 And not least, the advantage of a theory that actually makes it understandable why the Iliad and the Odyssey are as they are, expressed in the highly formalized epic dialect, with previously unexplained shifts in linguistic, religious, material and other levels, and exhibiting the strange mixture of firm overall composition and inconsistencies in detail, is so important that we must accept as our job to solve as best we can the new problems that the theory has brought in its wake. One of these is that it is becoming increasingly difficult to understand why the two poems were ever written. In the give and take of epic comparisons, Homeric scholarship has both offered and received inspiration. For the topic I am concerned with here, however, I am afraid that there is little to be learned for fieldworkers and everything for Homerists: the epic tradition of ancient Greece is known only through the texts which medieval scribes transmitted to us, whereas the social context is largely hypothetical. Considered under the headings Lauri Honko has suggested, we must realize that the sources we have for studying mental text, performance and written codification (Honko 1996:4) are few and dis-

The writing of the Ilias and the Odyssey

59

torted. There are some; we have both written and material sources (including a large and fascinating amount of vase-paintings) which offer information of the sociology of early Greek poetry; but the poetic texts themselves remain our only reliable entrance to the pool of epic tradition (Honko 1996: 8) in ancient Greece, with all the complications involved in reading extratextual information out of poetry. And not only are we unable to confront our object directly, being separated from it by some two and a half millennia; we are all the time risking to contemplate it through anachronistic filters. The questions here discussed were the topic of a seminar of classicists at the Danish Academy in Rome in May 1996, 4 whereas the present paper is addressed especially to fieldworkers and mainly concerned with putting questions.

The question of the recording in writing of the Iliad and the Odyssey Lord's experience that the idea of having an oral epic recorded in writing does not originate with the singer or his audience (Lord 1960:124) has been confirmed in fieldwork from many parts of the globe over the decades which have passed, to my knowledge without exception. And in the meantime, historians have been studying the role of writing in ancient Greek society and reached conclusions which underscore the importance of orality in all fields of life in the city-state, not only during the early centuries, but right through antiquity. William Harris has argued that nothing like general literacy in the modern sense was ever achieved, not even in classical Athens, the centre of ancient Greek culture; and Rosalind Thomas has demonstrated how even after literacy had been firmly established among the educated, orality remained more prestigious than writing as the two coexisted in everyday life. On this background it is difficult to understand why anybody should have wished to undertake the writing of these huge texts, a job which must have demanded considerable technical efficiency and a fluent supply of expensive papyrus. What would be the purpose of it? The singers will have felt no need of it, since it was part of their professional pride to be able to transmit without errors the poems they had learned. Nor will audiences, since the usual way of enjoying epic poetry remained through oral performance, even long after the poems were written. Books and readers as normal factors of the poetic institution came into being only gradually towards the end of the 5th century B.C., whereas the writing of the Homeric poems is estimated to have taken place at least a century before, somewhere between 725 and 525 B.C., with most scholars tending to prefer the earlier date. In our days, oral epic traditions which have coexisted with much more efficient forms of writing for centuries are still recorded in writing only at the wish of some

60

Minna Skafte Jensen

intruder from outside the traditional circle. But a similar cultural distance between an oral and a literate culture does not seem to have existed in ancient Greece. Where would such an intruder come from, and what would be his aim? Nevertheless, the fact remains that the poems were written. It is difficult anyway to compare ancient Greece before c. 525 B.C. with the 20th century where the relationship oral-written is concerned. Almost all aspects of literacy were different from our times: the alphabet was given its Greek form as an integral part of the culture, not as something imposed from outside; there were no connotations of underdevelopment attached to illiteracy - it was rather the other way round, since writing was mostly a slave's job; for those who learned to read, the written texts accessible were simple and well-adjusted to their newly acquired knowledge, with nothing like the problems a new reader faces these days, such as having to cope with the Bible or the Koran; for the same reason, the various types of semi-literacy which have attracted the interest of scholars in recent times cannot have existed; there was no book-trade, libraries, schools, or other bookish institutions. We might justly ask: if that is so, why then try to do it at all? And the answer is that the difficulties we are faced with in comparing early Greece with oral cultures in our times, are even more overwhelming if we compare with modern Western society. However, when classicists turn to the fieldwork of folklore and anthropology in order to find the relevant comparative information, they are often disappointed, even though the times are past when singers and storytellers were considered just "informants", bearers of traditions they perhaps hardly understood. Modern editors are becoming increasingly careful to contribute a wealth of detail about the singers, the performances, and their social contexts. However, since oral epics are flexible, since the audience has a central influence on the text performed, and since the fieldworker is a main part of the audience which is present at a recording - technically or in writing - it is of paramount importance to be informed of the scholar and his/her role. What was behind the scholar's choice of field, poetic form and individual singer(s)? What was the overall purpose of the fieldwork? Which were the questions the scholar wanted to find answers to? What training - linguistic as well as methodical - had h e / s h e gone through in advance? What were the singers asked for? How was the recording made? Who were present? H o w did they react? Where did the singer make his pauses? How many sessions were held? (Homerists as well as those who study other epic traditions from the past are of course especially eager to hear of dictation processes.) After the recording, what happened to the text between the first draft and the printed book? What kind of interpreting, commenting, copying and standardizing was undertaken? Such information is regrettably seldom given (Slyomovics 1987 is a notable exception).

The writing of the Ilias and the Odyssey

61

Therefore the conference in Turku in June 1996 with its emphasis on the "scribe"/ the scholar w h o has brought about the recording of an oral epic, was a real breakthrough. But the lectures and discussions by n o means m a d e the problem of the recording in writing of the Iliad and the Odyssey easier to solve: the overall agreement which gradually established itself among those present seemed to be that not only the idea of recording, but that of making a long, coherent poem at all originated with the scholar rather than with the singer. Dan Ben-Amos even offered the thought-provoking observation that it seems a strange coincidence that whereas rich epic traditions are f o u n d in the former French colonies of Africa, there are none in the British ones, and suggested that the explanation might be not that the traditions are unequally distributed, but that French and British researchers have h a d different interests. In a way, it might even be maintained that if the ancient Greeks h a d not recorded the Iliad and the Odyssey for us, we should not have h a d the idea that it is an integral aspect of an epic poem proper to be very long, nor would Lönnrot have felt inspired to compose his Kalevala out of the m u c h briefer folksongs he collected, nor would fieldworkers of the 20th century have asked singers to perform for days on end to accomplish something different from the songs of their normal repertory in order that it might truly be called an epic. In an intriguing way, not only scholarship but the very epic form as we find it documented in editions came to appear as a gigantic logic circularity.

A catalogue of problems The points at issue may be described under seven headings: 1. Cooperation of writing and oral tradition. Many Homeric scholars d o not see a problem at all. They feel at ease with the model that H o m e r w a s composing on the basis of an old and rich oral tradition, but was himself literate; therefore he understood the advantages writing offered, both for grand-scale composing and for preserving what had been composed. It might even have been part of his genius that he had this insight. A n d anyway, since the texts which Parry and Lord published from their field work in Yugoslavia do not achieve anything like the qualities of the Homeric poems, they may be compared with the (lost) forerunners of Homer, b u t not with the master himself. The qualities of the Iliad and the Odyssey w o u l d be unthinkable within a purely oral tradition. Besides, the Greeks might have got the idea of writing these poems from their Eastern neighbours; Babylonians and Hittites were writing their epics long before Homer. But these considerations do not solve our problem. If singers in the modern world, where writing is almost omnipresent and writing facilities much easier available than in ancient Greece, do not feel a need of having oral

62

Minna Skafte Jensen

epics written, w h y s h o u l d the Greeks, even in a period w h e n there w e r e n o readers to appreciate the written text? Besides, these opinions are generally built o n a m u c h too n a r r o w comparative material a n d an u n d e r e s t i m a t i o n of t h e potentialities of oral traditions. In this connexion, Lauri a n d A n n e l i H o n k o ' s I n d i a n project is highly i m p o r t a n t in d o c u m e n t i n g the existence of large-scale epic. 5 2. The written poems as the result of a gradual development. It is a w i d e s p r e a d h y p o t h e s i s that the Iliad a n d the Odyssey s o m e h o w crystallized out of the tradition. Their very excellence w a s evident to the singing community, a n d therefore they were given preference as against all other epic p o e m s . The d e v e l o p m e n t is envisaged as a m o v e m e n t f r o m multiplicity to singularity:

Iliad &

Odyssey

H o w e v e r , the very m e t a p h o r of crystallization reveals that the practical side of the evolution is n o t considered. H o w could such a process h a v e taken place? Are w e to u n d e r s t a n d that the p o e m s were written a n d then revised a n d enlarged in a continuous chain of copying? If that were so, I find it difficult to u n d e r s t a n d w h y the various stages of the d e v e l o p m e n t should h a v e left n o single trace in the f o r m of other written versions of the t w o epics. 6 If, on the other h a n d , w e are to imagine that the process took place within the oral tradition this contradicts the general experience of h o w oral epic traditions w o r k . Competition is b e t w e e n singers, n o t b e t w e e n songs. Each singer tries to achieve the highest possible perfection, a n d b o t h s p o n s o r s a n d audiences g o for the best m a n accessible. Traditions d o n o t h a v e as their aim to p r o d u c e the ideal s o n g to outshine a n d possibly s u b s u m e all other songs. O n the contrary, the m o d e l above seems to m e to be an example of w h a t I call optical illusions: because w e k n o w that this w a s the outcome, w e think that it w a s also the p u r p o s e . A n o r m a l m o d e l w o u l d be instead:

Iliad &

Odyssey

The writing of the Ilias and the Odyssey

63

3. Performance studies. The oral-formulaic theory has inspired a great interest in the question of performance in ancient Greece. Not only the epics, but also choral lyric, drama, and perhaps prose history are known to have been performed in public, and scholars try to achieve the most detailed and precise knowledge of the modalities of such performances. Already in 1964 James Notopoulos suggested that the 24 songs into which each of the two epics are divided might mirror actual performance, and he published interesting studies of the length of epic performances in modern Crete to confirm his hypothesis. More recently, Oliver Taplin has argued that the Iliad and the Odyssey fall into three parts each, meant for continued performance over three days (Taplin 1992: 39-41). Such studies are, of course, very important for our understanding of the two poems as oral poetry. However, I do not find them directly relevant for the question of how the texts we have came into being, for they can by no means be the result of a normal performance. No scribe would have been able to record in writing while the singer was actually performing; an oral Homer can have reached us only as the result of a dictation. Neither do I believe that a dictating singer and his scribe would have been able to estimate a suitable length of a written poem to fit an oral performance; that presupposes, I think, a quite detailed training in writing for oral delivery. Add to this that the Iliad and the Odyssey, even divided into three parts, are far too long for any normal performance. What we must look for, according to me, is a special session of writing, a "performance" made with the aim of producing written poems, not texts for rhapsodes to learn by heart. 4. Writing as a means of preservation. It is generally taken for granted that the purpose was to have the text fixed and thus preserved for posterity. But in this I think we are the victims of another optical illusion: because we know that that was actually the result, we think that it was also the purpose. On the contrary, it is unlikely that either singers or the community at large doubted the possibility of reliable oral transmission. That is, the same that Lord emphasized for the singer holds good for everybody: the truthfulness of a given rhapsode and his performance might be doubted, but not that of the transmission as such. Not even the best educated and most bookish person in archaic Greece can have noticed the fact that oral transmission is flexible; in a way this was only really noticed when Parry and Lord went to Yugoslavia with their technical equipment and could make pedantic comparisons between different performances of the same text. In this respect the distance between singer and scribe in early Greece must have been much less than what it is today. 5. Writing as a means of political control. It has also been considered if the

purpose could have been that political authorities - and here the tyrants of Athens, Pisistratus (in power 560-28 B.C.) and his sons, are the main candidates - wished to control the content of the text.

64

Minna Skafte Jensen

The idea of political control is certainly important, but as a reason for the writing it is still unsatisfactory, considering that the item to be controlled would necessarily have been the performance, not the written text. Even if rhapsodes were compelled to learn a specific, written version by heart, the authorities would have to control the performances, since that was where the text was published. No written, authoritative text could guarantee that singers did not deviate from it in their performances. 6. The so-called Panathenaic Rule. We have information that Homeric epics were performed at the Great Panathenaic festivals in Athens, and that either Solon (in power 594 B.C.) or Pisistratus' son Hipparchus (killed 514 B.C.) ordered that rhapsodes should perform one after another so that where one stopped the next one should take over. This looks like the clue: here we have a reason for having the poems written; if the rhapsodes should take over one after another they had to have a fixed text, we think. But we also know that the Panathenaic recitals were competitions in which a winner was selected, and how could that work if the text was decided upon in advance and remained the same from one festival to the next one? Here it seems to me much more likely that rhapsodes were competing with songs proper, the content rather than the way of performing. If we compare with the dramatic competitions in 5th century Athens, we have texts which are almost all of them versions of known myths, and what the poets compete with is their individual handling of the known story. I suggest that this might be a direct further development of the epic competitions, and that what a rhapsode was estimated for was his more or less convincing version of the known story, not his way of repeating a fixed text. This calls for another interpretation of the "Panathenaic Rule" one in which the continuation was not of chapters of a written text, but of a storyline in the common pool of tradition: it is reasonable to assume that any singer and listener would know the main parts of a storyline of the events in the Trojan war, so that when one rhapsode finished, the next one was able to take over in a way that convinced the audience. This resembles what Gene Roghair has reported of the Palnaadu epic in Andhra Pradesh, and that is how the poet of the Odyssey imagines the entertainment at the court of the Phaeacians to be: a person in the audience asks for a specific episode and makes it clear from where he wants the singer to begin - the two of them seem to agree on the storyline (Roghair 1982: 26-31, 62-64; Odyssey [8.492-98]). 7. Chronographie vs. episodic epic. Many years ago, Johannes Kakridis pointed out that it is a strange characteristic of both the two epics that they do not tell of the events in a chronographic way, following the order of what Genette has in the meantime termed the story (histoire) as opposed to the narrative {recit) (Genette 1980 [1972]: 25-27). The Iliad is concerned with a single brief episode which takes place in the tenth and last year of the siege, and the Odyssey begins when Odysseus has almost come home. The Iliad then ex-

The writing of the Ilias and the Odyssey

65

pands its story with all kinds of other episodes that strictly speaking do not belong to the one concerned, and the Odyssey begins in the middle and then moves backwards, having the hero himself tell of his adventures. This observation has been very much discussed. We know quite a few titles of epic poems and names of poets, and there even exists a summary from late antiquity of the so-called Epic Cycle, poems which recount the events of the war and its aftermath in an orderly fashion. It is thought that the Iliad, and the Odyssey outconquered all other epics because they were so much more elegantly composed. However, it is difficult to form an opinion of these lost poems just from the brief surveys of their contents. As a matter of fact, in the summary which the mythographer Apollodorus made of the Iliad and the Odyssey this special feature of their composition does not come out (I owe this observation to Martin West). Had we had only his version of them, we should have thought that these two poems had been chronographic too. A model built on comparison There is an ancient tradition, already briefly referred to above, which says that the tyrant Pisistratus or his son Hipparchus collected the Homeric poems which had previously been scattered. This looks like one of the myths that various communities have to explain how poems were originally written (Senner 1989), and it has been considered with shifting interest in the history of Homeric scholarship. For my part, I agree with a famous dictum of Rhys Carpenter's: Had ancient sources not preserved this story for us, we should have had to invent it for ourselves (Carpenter 1946:12; Jensen 1980:128-71). I have discussed the question in my book from 1980, and the present paper may be considered a continuation of that work. I think that the story actually reflects a knowledge of the recording in writing, not fully understood, and that it mirrors an awareness that the two famous poems resulted of this writing. My own hypothesis of what took place is the following: for reasons of chronology that I cannot go into here I consider that the hero of the story was Hipparchus. He wished for political purposes to have an authoritative Homer at his command. "Homer" to him meant the kind of songs the itinerant rhapsodes - some of them called the Sons of Homer - were performing; such songs might be heard on various occasions and regularly every fourth year in Athens, when the city-state arranged rhapsodic competitions at the Great Panathenaic festivals. But when the competition was over and the singers had sung their songs, they moved on to other city-states to perform at other festivals. In order to keep the songs, materially, he decided to have them written. After one of these festivals (presumably that of 526 B.C.) he

66

Minna Skafte Jensen

engaged the winners of the first and second prize to dictate to his scribes (again, the reasons why I think there were two singers to dictate are too complicated to be explained here). He wanted to impress the world with having not only a written Homer, but also a truly large-scale version, so he asked the rhapsodes to dictate their songs as fully as possible and perhaps even offered them a payment per line. The singers, who were used to sing one episode in any given performance, dictated the episodes with which they had been successful at the festival, but expanded their mental texts in all possible ways. In their minds they still had the audience which had been present at the festival, so that the addressees of their dictation were both the tyrants, the scribes and the people gathered for celebrating the Panathenaic festival. Hipparchus had first the Iliad, then the Odyssey written. The scribes - 1 actually think there was a whole team of them - were writing the original copy on waxen tablets, and each of the 24 songs represents the part of the text dictated during a day. Afterwards, the tablets were carefully copied into papyrus and the texts we have today have their orderly form, consistent orthography, and clear distribution in verses and songs from this edition.7 Afterwards the singers went away, but the poems remained. They were carefully stored in an archive as a present for the patron goddess Athena, and nobody would think of reading them. But in critical moments they were there for the Athenians to refer to, and the historian Herodotus actually records instances of such political uses of the poems. "Homer" continued to be a living oral tradition, unaffected by the event. The two poets as well as all the other rhapsodes were available for any occasion which called for epic performance, and after their deaths others took over. The influence of the written Iliad and Odyssey only began when a book culture had become widespread enough for people to want to be able to read their Homer, and not least with Homer's entrance into the school as a training text for pupils. Instead of having new recordings made from dictation, readers preferred the much simpler procedure of copying the already written official Athenian text, and this was the one which reached the library in Alexandria and from there became the basis of the medieval transmission. I have tried to build a hypothesis that exploits the various sources we have from antiquity while taking into account all the above mentioned problems. It considers the Iliad and the Odyssey to be oral poems in the full sense of the term - orally composed, performed and transmitted - , thus maintaining for us the invaluable key to answering the classical Homeric questions. I understand the two poems as examples of how individual rhapsodes handled the common pool of tradition; the relationship between the Iliad and the Odyssey on one hand and the tradition on the other is analogous as that between parole and langue in natural languages. It offers a concrete vision of the "performance" which is at the root of our written transmission, and considers the role of writing in archaic Greece while trying to peel off modern

The writing of the Ilias and the Odyssey

67

ideological connotations of development. It features a person who had both the motive and the money for having the job undertaken, and it considers the overall episodic composition of the two poems as a quality intrinsically resulting from the process of dictation. With this paper I wish to provoke field workers' reactions. Does the hypothesis seem convincing at all? Are the impressions I have of what is possible, what not, in oral tradition correct? Or do I emphasize the problems more than necessary? Also, I appeal to editors who have an experience of the process of dictation to give us all available information about what goes on during such sessions. Not least, when scholars stimulate singers whose normal performances are episodic to perform in other terms than what they are used to, I wonder how this new text compares to those of their normal performances? We think that they have relatively fixed mental texts of the songs in their repertory. Do they choose to keep the storyline of one mental text and expand it, or do they add texts, one to another, to achieve a chronographic composition? Notes 1.1 continue to consider Parry and Lord's achievement a theory, although Gregory Nagy has argued against in order to emphasize that their views were built on field experience (1996:19-20). In my vocabulary the term means a set of well-argued assumptions concerning an overall theme; that it is built on careful and detailed experience is no problem, on the contrary. For a survey of comparative epic studies inspired by the oral-formulaic theory see Foley 1988. 2. Latacz 1991: compare Wolfgang Kullmann (427-28) and James Holoka (475) with Fritz Graf (349); the quotation is from Holoka (476). 3.1 refer mainly to Kirk 1962, Austin 1978 [1966], Gaisser 1969, Nagler 1974, Rossi 1978, Aloni 1986 and 1989, Martin 1989, Nagy 1990 a and b, Foley 1990 and 1995 and Segal 1994, and to the great new commentaries edited by G. S. Kirk and Alfred Heubeck, for both of which the oral-formulaic theory is a general frame of reference. 4. Contributors were Antonio Aloni, Mario Cantilena, Carmine Catenacci and myself; the proceedings have been published in Analecta Romana Instituti Danici 1997. 5. For a brilliant criticism of some Homerists' attitudes to the oral-formulaic theory, see Cantilena 1983. In an interesting recent paper, Therese de Vet discusses a possible interplay of written and oral transmission of Homer around the Alexandrian library in the Hellenistic period. 6. The ancient alternative readings which we know from scholia and the so-called "eccentric" early papyrus fragments are too small and unimportant to be understood as resulting from various oral versions of the poems. Cf. Parry (1930) 1971: 268 and Jensen 1980:106-11. 7. For a recent description of the complicated job of the "scribe", see Revel 1996:11821.

68

Minna Skafte Jensen

References Aloni, Antonio 1986 1989

Tradizioni arcaiche della Troade e composizione dell' Iliade. Milano: Edizioni Unicopli. L'aedo e i tiranni. Ricerche sull'Inno omerico α Apollo. (Filologia e Critica 59.) Roma: Dell'Ateneo.

Austin, Norman 1978 [1966] The Function of Digressions in the Iliad. In: J. Wright (ed.), Essays on the Iliad. Selected Modern Criticism,70-84. Bloomington and London: Indiana University Press. Cantilena, Mario 1983 Oralisti di ieri e di oggi. Quaderni Urbinati di Ciiltura Classica 13:165-86. Carpenter, Rhys 1946 Folk-tale, Fiction and Saga in the Homeric Epics. (Sather Classical Lectures 20.) Berkeley and Los Angeles: University of California Press. Foley, John Miles 1988 The Theory of Oral Composition. History and Methodology. Bloomington and Indianapolis: Indiana University Press. 1990 Traditional Oral Epic. The Odyssey, Beowulf, and the Serbo-Croatian Return Song. Berkeley, Los Angeles and Oxford: University of California Press. 1995 The Singer of Tales in Performance. Bloomington and Indianapolis: Indiana University Press. Gaisser, Julia H. 1969 A Structural Analysis of the Digressions in the Iliad and the Odyssey. Harvard Studies in Classical Philology 73:1-43. Genette, Gerard 1980 [1972] Narrative Discourse. An Essay in Method. Translated by Jane E. Lewin. Foreword by Jonathan Culler. Ithaca and New York: Cornell University Press. Harris, William V. 1989 Ancient Literacy. Cambridge, Mass. and London: Harvard University Press. Heubeck, Alfred (ed.) 1988-92 A Commentary on Homer's Odyssey. I—III. Oxford University Press. Honko, Lauri 1990 The Kalevala: the Processual View. In: Lauri Honko (ed.), Religion, Myth, and Folklore in the World's Epics. The Kalevala and its Predecessors, 181-229. (Religion and Society 30.) Berlin and New York: Mouton de Gruyter. 1996 Epics along the Silk Roads: Mental Text, Performance, and Written Codification. Oral Tradition 11/1:1-17. Jensen, M. S. 1980 The Homeric Question and the Oral-Formulaic Theory. (Opuscula Graecolatina 20.) Kabenhavn: Museum Tusculanum.

The writing of the Ilias and the Odyssey 1998

Kakridis, J. Th. 1949 [1944] Kirk, G. S. 1962 Kirk, G. S. (ed.) 1985-93

69

A. B. Lord's Concept of Transitional Texts in Relation to the Homeric Epics. In: Lauri Honko, Jawaharlal Handoo and John M. Foley (eds.), The Epic: Oral and Written, 94-114. Mysore: Central Institute of Indian Languages. Homeric Researches. (Skrifter utgivna av Kungliga Humanistiska Vetenskapssamfundet i Lund 45.) Lund: Gleerup. The Songs of Homer. Cambridge: Cambridge University Press. The Iliad: A Commentary. 1-6. Cambridge: Cambridge University Press.

Lachmann, Karl 1837-41 Betrachtungen über Homers Ilias. Berlin. Latacz, Joachim (ed.) 1991 Zweihundert Jahre Homer-Forschung. Rückblick und Ausblick. Stuttgart: Teubner. (Colloquium Rauricum 2.) Lesky, Albin 1963 [1957/8] Geschichte der griechischen Literatur. Bern, München: Francke. Lord, Albert B. 1960 The Singer of Tales. (Harvard Studies in Comparative Literature 24.) Cambridge Mass.: Harvard University Press. Martin, Richard P. 1989 The Language of Heroes. Speech and Performance in the Iliad. Ithaca and London: Cornell University Press. Nagler, Michael N. 1974 Spontaneity and Tradition. A Study in the Oral Art of Homer. Berkeley, Los Angeles and London: California University Press. Nagy, Gregory 1990a Greek Mythology and Poetics. (Myth and Poetics.) Ithaca and New York: Cornell University Press. 1990b Pindar's Homer. The Lyric Possession of an Epic Past. Baltimore and London: Johns Hopkins University Press. 1996 Homeric Questions. Austin: University of Texas Press. Notopoulos, J. A. 1964 Studies in Early Greek Oral Poetry. Harvard Studies in Classical Philology 68:1-77. Parry, Milman 1971 [1928-35] The Making of Homeric Verse. The Collected Papers of Milman Parry. Edited by Adam Parry. Oxford: Clarendon Press. Revel, Nicole 1996 Kudaman: An Oral Epic in the Palawan Highlands. Oral Tradition 11:108-32. Roghair, Gene H. 1982 The Epic ofPalnädu. A Study and Translation ofPalnäti Vlrula Katha, a Telugu Oral Tradition from Andhra Pradesh, India. Oxford: Oxford University Press.

70

Minna Skafte Jensen

Rossi, Luigi Enrico 1978 I poemi omerici come testimonianza di poesia orale. In: R. B. Bandinelli (ed.), Storia e Civiltä dei Greci. I. 1: Origini e sviluppo della cittä, 73-147. Milano: Bompiani. Segal, Charles P. 1994 Singers, Heroes, and Cods in the Odyssey. (Myth and Poetics.) Ithaca and London: Cornell University Press. Senner, W. Μ. 1989 Theories and Myths on the Origins of Writing: a Historical Overview. In: W. M. Senner (ed.), The Origins of Writing, 1-26. Lincoln and London: University of Nebraska Press. Slyomovics, Susan 1987 The Merchant of Art. An Egyptian Hilali Oral Epic Poet in Performance. (University of California Publications in Modern Philology 120.) Berkeley, Los Angeles and London: California University Press. Taplin, Oliver 1992 Homeric Soundings. The Shaping of the Iliad. Oxford: Clarendon Press. Thomas, Rosalind 1989 Oral Tradition and Written Record in Classical Athens. Cambridge: Cambridge University Press. 1992 Literacy and Orality in Ancient Greece. Cambridge: Cambridge University Press. de Vet, Therese 1996 The Joint Role of Orality and Literacy in the Composition, Transmission, and Performance of the Homeric Texts: A Comparative View. Transactions and Proceedings of the American Philological Association 126: 43-76. Wolf, Friedrich August 1795 Prolegomena ad Homernm. Halis Saxonum e Libraria Orphanotrophei.

The textualization of South Slavic oral epic and its implications for oral-derived epic John Miles Foley In considering the textualization of oral epics from so many different areas, this conference addresses an important and complex area that has received inadequate attention in earlier research and scholarship. This "benign neglect" perhaps stems more from cultural predispositions about verbal art, technologies of communication, and our own deeply embedded dependence on texts than from any purposeful avoidance of the subject. Until relatively recently investigators have tended to overlook just how an oral epic reached textual form, preferring to deal with it as a readymade object that could be analyzed with available tools. Analysis and interpretation often began only after the transcription was made, after the multi-dimensional performance was reduced to a form that scholarly culture found more manageable.1 All of these assumptions mask an important part of the story. What we have come to Turku to tell, with Lauri Honko's inspiration, are the tales that lie behind epics-become-books. We are interested in the cultural matrices from which these performance-texts emerge, the roles of informants and folklorists who contributed to the making of these works, the intersemiotic translation of performance to manuscript and print, and the myriad other issues that are submerged when we interpose our concepts of object and stasis between a performed oral epic and its book-bound reception.2 For my part, I propose to tell the tale of how South Slavic oral epic was textualized by the two most prominent teams of investigators involved in its collection: first, the Serbian scholar Vuk Karadzic and his network of amanuenses; and second, the North American scholars Milman Parry and Albert Lord, with the important participation of their native assistant Nikola Vujnovic, who was himself an epic singer. In many ways both groups were informed and limited by the historical periods in which they made their collections and by the theoretical predispositions that they brought to the enterprise of recording and publishing oral epic. Toward the end of the presentation I will hazard a few comparative remarks that bear on two epic traditions that survive only in manuscript form - the Homeric poems and the Anglo-Saxon Beowulf - and whose textualization can therefore be reconstructed only hypothetically. Organizationally, I plan to proceed by asking three basic questions: What gets recorded? What gets published? What gets received? In simplest terms, what gets recorded is a performance, but the nature of that performance

72

John Miles Foley

necessarily varies from one instance to another, one subgenre to another, one investigator to another, and so forth. Both the Karadzic and Parry-Lord projects illustrate, and quite differently, how certain kinds of choices affect the outcome of the recording process. What gets published is usually a print transcription of a performance, with or without cues prompting the readership to appreciate other dimensions of the event, but not all transcriptions are created equal. The South Slavic case shows how unequal they can be, and how predispositions play a large part in what we might call "textual parallax". What gets received is essentially the reader's "take" on what he or she is offered in the published transcription. This part of the process cannot be detached from textualization because the very act of textualization always makes assumptions about the readership. Karadzic was creating a resource for a target audience, and so were Parry and Lord; the mere fact that these were manifestly different resources proves the importance of the projected audience's role in converting oral epic to a book.

Vuk Stefanovic Karadzic 1. What gets recorded Vuk Karadzic was, like all other collectors, a citizen of his time and place in history. Ground-breaking ethnographer and eminent lexicographer as well as what we would now call a folklorist, he responded to the nineteenthcentury European search for ethnic roots with benchmark accomplishments in many different fields. In addition to his massive Srpski rjecnik (A Serbian Dictionary) of 1852, which included many words and meanings from the epic and non-epic oral traditions of Serbia that were found nowhere else in contemporary discourse, he published numerous volumes and essays on the customs and rituals of his people, and, not least, modified the Latinbased and Cyrillic alphabets to make them both entirely phonetic. But the accomplishment that was to have the longest-term importance for the widest cross-section of people (and this includes an international scholarly constituency as well as generations of schoolchildren) was his unparalleled collection entitled Srpske narodne pjesme (Serbian Folk Songs). Although he began publishing narratives from oral tradition as early as 1814, Karadzic's standard edition of Srpske narodne pjesme in four volumes appeared from 1841-1862. Numbers 5 through 9 were posthumous publications, and four additional volumes, entirely unauthorized by Karadzic himself, were culled from his unpublished manuscript remains by 2ivomir Mladenovic and Vladan Nedic and issued in 1973-74. These anthologies run the gamut from lyric to epic, and within epic from medieval to modern. For our purposes the most significant part of the collection is volume 2 in the

The textnalization of South Slavic oral epics

73

original standard series, in which, its editor specifies, "su pjesme junacke najstarije" ('are the oldest heroic songs'). For this as for other volumes Karadzic employed two strategies for collection, both of which were informed by his personal history. For not only did he come from a family of traditional poets himself, but having grown u p in a village environment he was also very much aware of the realia of fieldwork: the morphology of performance, dealing with informants on an equal basis, and so forth. Early in his career he recorded most performances himself, by dictation of course, and from many different poets. I say "poets" and not "singers" because in the Christian tradition of the early and middle nineteenth century many bards did not use the gusle, tambura, or other musical accompaniment, and recited rather than sang their poetry. As Karadzic grew older and his project expanded and matured, he more and more m a d e use of a country-wide network of amanuenses who sent him transcriptions they h a d made, presumably but not absolutely dependably from oral performances. Even in this later period, however, he continued to deal personally with those poets he valued most highly, such as Old Man Milija and especially Tesan Podrugovic, the highwayman (hajduk) who recited m a n y of Karadzic's most memorable poems. What got recorded, then, was an enormous, geographically diverse sample of what has come to be known in the native scholarship as the "classic Serbian literature" of its time. Most significantly, it was first and foremost a collection of texts. It was written d o w n by many different hands, and largely without the kind of context we modern investigators require. Because the focus was on unearthing a treasure that could be portrayed as an ethnic heritage, and also because this was after all the mid-nineteenth century, very little attention was paid to those process-oriented dimensions that distinguish an oral traditional work from a literary text. In fact, one could fairly describe the act of collection itself as the initial step in translating an oral tradition into a text, with all of the linguistic and political complexities entailed in such a remaking. One additional dimension of what got recorded should be emphasized. Karadzic's quarry was the Christian epic tradition of the Serbs, a type of narrative that m a n y scholars of Western European and Asian epics might choose to deny admission to the genre. The poems from this tradition are seldom more than 300-400 decasyllabic lines in length, and, although they may constitute essentially the same story as a Moslem South Slavic epic of 2000 lines or more, still lack the elaborate descriptions of people, places, and events that we associate with the term and category of "epic". Nonetheless, Serbian scholars have always treated these p o e m s as epic, and they d o represent the most extensive and developed narrative form within their ethnic orbit. We may add one further note in passing. While the portrayal of any work from oral tradition as a separate, free-standing item is inherently

74

John Miles Foley

misleading, in that it epitomizes the instance at the expense of the tradition, in the case of the Christian epics from South Slavic it is especially problematic. Individual poems require a very substantial background to be merely intelligible, and the brief Christian narratives leave a great many "gaps" for the uninitiated reader. We would do better to think of Karadzic's poems as parts of a cycle, a concept that, as we shall see, he adopted himself, albeit unconsciously or at least tacitly.

2. What gets published From the many texts in his possession, both those he wrote down himself and those submitted via his network of collectors, Karadzic selected those which he thought would best represent the Serbian tradition. The first cut was simply on the grounds of perceived quality. Karadzic included in his notes to the standard edition and elsewhere a realistic appraisal of his personal archive: a fair number of texts were confused or otherwise flawed, many were mediocre, and a relatively small group he considered up to the demands of his program - to reflect the precious heritage of Serbian poetry he had heard since he was a child. Let me add that this procedure of gross selection, based on subjective, qualitative evaluation, may have been prescriptive by our modern standards. Nonetheless, it is also refreshing in another respect, namely, that Karadzic was willing to accept the fact that South Slavic oral poetry - like any contemporary poetry, for example - was inevitably uneven. Some performance-texts were poor, others average, and a few outstanding. The second stage of selection was the much-discussed matter of Karadzic's editing, often unfortunately compared to the outright rewriting undertaken by the Grimms or some of the other nineteenth-century collectors and editors. But within his own period, in fact, Karadzic was criticized for his perceived "failure" to sufficiently edit the texts he published, to bring them into line, a reaction that serves as a good index of what the assumed editorial policy really was. Since that time, the tendency has usually been to overestimate the extent to which he interfered in the texts he wrote down or received from others, probably under the influence of foreign exemplars. Discovering the true story on this score is no easy task, since Karadzic followed the practice of destroying fair copies of texts as soon as they were sent to the publisher. The result is that the Archive of the Serbian Academy of Sciences holds very few manuscripts of the poems that appear in the second of his original, personally supervised volumes ("the oldest heroic songs"). The only survivors, then, are those published poems that by sheer chance occupied the other side of a text not selected for publication. I have personally examined these exceptional manuscripts in the Academy archive and must report that I disagree with Mladenovic (1973, espec. xxiv-lv) and Duncan

The textnalization

of South Slavic oral epics

75

Wilson (1970: 317-18) that Karadzic's editing was heavy, frequent, and noteworthy. In fact, the editor interferes only seldom, and then very lightly and expertly. Specifically, he averages an added or substituted hemistich about every 75 lines, and a minor adjustment for dialect or incomplete syntax about every 30-40 lines. These changes, along with orthographical corrections, address inconsistencies that may well have stemmed from the process of recording via dictation: certainly the spelling and minor errors fall into that category, and the larger ones may as well. To impute absolute accuracy to the original transcription process would be naive even for a modern-day amanuensis; with a transcriber who may speak a different dialect and who may know a different version of the poem being performed, the process gains in complexity. Two more aspects deserve mention in answering the question of "What got published?" One is the matter of the editor's handling of variant versions of the same poem. In this regard Karadzic's practice was to print only what he considered the finest version of a given tale, the one that he judged the most finished poem. The sole exclusion to this rule was the occasional second, or even third, version of a particular narrative, labelled "Opet to isto" ('Again the same') and placed in sequence immediately after the initial text. It appears, although to my knowledge Karadzic never articulated it as a policy, that he included alternate versions only if they satisfied two conditions: (1) they had to meet his general criteria for quality in their own right, and (2) they had to offer his reader a significant variation of some sort (in overall story-pattern or some smaller narrative action). These rare windows into the morphology of the Christian oral epic in nineteenth-century Serbia, rare because Karadzic was committed to texts and a virtually literary heritage that would help his countrymen stand shoulder to shoulder with other European nations, can be supplemented by perusal of his posthumous volumes, and especially by consultation of those many alternate versions he chose not to publish (as preserved in the Mladenovic and Nedic edition of his manuscript legacy). Last on our list of what got published is Karadzic's actual configuration of his anthologies. In the second volume of his original standard collection he did not simply print the choicest available bits of the tradition, leaving each of them to make its own artistic statement as a singular poetic creation. Rather he provided an organic kind of structure to the presentation, organizing and sequencing his collection according to groups and lineages of principal characters and interrelationships among semi-historical and highly mythologized events. Two primary examples of this textual strategy are the groupings of stories surrounding the Battle of Kosovo in 1389, the defining moment for Serbian heroism and identity vis-ä-vis the Ottoman Empire, and the redoubtable Prince Marko (or Kraljevic Marko), the mostly legendary figure who served the Turks as a forced mercenary and who took every op-

76

John Miles Foley

portunity to outwit his oppressors in grand style. In speaking of the narratives that collectively present Kosovo and Marko in volume two, we would do better to recognize them as cycles of epics, incomplete in themselves but highly resonant as a group. By molding these individual poems into a cycle, Karadzic was in effect imitating the traditional process from which they derived, a process that never isolates but always contextualizes, situating the singular in the plural, the one performance in the implied network of the epic tradition. Although they were artificial facsimiles, the cycles in his anthology revealed Karadzic's own personal acquaintance with South Slavic epic as an ongoing presence and tradition that could not be reduced to a sampler of individual items.

3. What gets received

In the interest of saving enough space to do some justice to Parry and Lord and the Moslem tradition, let me give the very important question of "what gets received" short shrift here. Elsewhere I have written of how traditional structures imply traditional referentiality, how composition implies and drives reception, how the "signs" of oral epic open onto broad vistas of idiomatic meaning that, strictly speaking, are not a part of the literal performance or text. Using a homemade proverb - "Oral traditions work like languages, only more so" - I have tried to understand the depth and richness of the coding inherent in the special languages of oral tradition, especially but not exclusively in epic (see Foley 1990,1991,1995b, 1996b, 1996c, 1999). Against that background it becomes very obvious that the history and methods involved in the textualization of oral epic are absolutely crucial to the process of reception. Editors of texts that derive from centuries of entrenched literary tradition are beginning to explain how any act of edition is also necessarily an act of interpretation, that there is no such thing as "neutral" editing. How much more true is this for the oral or oral-derived epic performance, whose signals must undergo a semiotic translation even to appear in the "finished product" of the epic-as-book? Without further discussion, then, let us simply take note of the determinative role of the real and projected audience(s), or readership(s), for that product, as well as the fact that making an experience into a text has a profound effect on what gets received.

The textualization of South Slavic oral epics

77

Milman Parry and Albert Lord 1. What gets recorded The very first aspect that must be addressed herein is the question of Parry's and Lord's core motivation in mounting such an ambitious expedition to the former Yugoslavia for the purpose of collecting oral epic from the South Slavs. 3 Why did they spend the better part of two years in the hinterlands of Bosnia and Hercegovina, acoustically recording more than 700,000 decasyllabic lines on aluminum disks - a "half-ton of epic", as Albert Lord characterized their yield on more than one occasion? Why did they take the trouble to have a Connecticut-based electronics firm design and produce a state-of-the-art device that allowed inscription of those disks, and what was the impetus behind their hiring a native assistant to accompany them during fieldwork, the same person who would later spend more than a year in Cambridge, Massachusetts writing out transcriptions of many of the audio records? We will be unfair to their methodology - both its numerous positive aspects and its inevitable shortcomings - if we do not quote Parry's own stipulation that "it was least of all for the material itself" that he and Lord worked so arduously and creatively in the field. Parry was always primarily a classicist, and he was interested in one thing: a living analogy for the manuscript poems of Homer, a laboratory proof of the phenomenon of composition in performance that he had posited on the basis of the long-silent Iliad and Odyssey texts. If the formulaic method he had so carefully demonstrated in Homer could also be shown to be the root compositional strategy in the oral epics from then-Yugoslavia, then the circle, he felt, would be closed. Homer the ancient Greek aoidos would be mirrored in the South Slavic guslar. This policy and priority had far-reaching consequences. Most centrally, it meant that Parry and Lord limited their sample of oral traditions to the single subgenre that most closely resembled the Homeric poems - namely, Moslem epic. In this restriction they were very wise, in that only Moslem epic compares closely to the Iliad and Odyssey in length, elaborateness, and so forth. (An undesirable by-product of this focus emerged in later years, as many investigations used the specifics of this single subgenre of oral tradition as a universal model for all genres, but thereby hangs another tale.) Thus, "what got recorded" as part of their fieldwork project also determined what didn't get recorded: the shorter Christian epic anthologized by Karadzic was largely avoided, lyric or women's songs were collected but assigned a different status, and many other forms never were considered. This is certainly not to say that Parry and Lord should have conducted their fieldwork differently; on the contrary, they kept assiduously to a predetermined plan,

78

John Miles Foley

and their archive is the richer for it. I observe merely that any decision about areas to be examined or avoided has implications, immediate and long-term, for edition and scholarship in those areas. To their great credit Parry and Lord sought not texts but performances, and this was a significant step beyond any other collectors who had worked in the former Yugoslavia. And, in an important departure from Karadzic, they of course did all of the fieldwork and recorded all of the performances themselves, with the invaluable cooperation of Nikola Vujnovic. Whenever possible, they used their unique apparatus to encode the guslar's full-speed performance acoustically; these recordings were referred to as "sung", because the poet employed vocal and instrumental melodies as defining dimensions of performance. Songs that were voiced without melodic accompaniment, yet still recorded on the aluminum disks, were called "recited", and those taken down in the ages-old way by an amanuensis listening to a composing bard were termed "oral-dictated". Lord came to the conclusion that this last category of performances, which he viewed as an opportunity for the singer to show his best because of the relaxed pace of the song, was the finest, and it was these he chose to compare most closely to the supposedly oral-dictated texts of Homer, as he saw it. In my work with the guslari from the region of Stolac, I have not found the oral-dictated text such a dependable benchmark. Some of the singers adapted well to what was for all of them a new method of proceeding - speaking a line and waiting for the amanuensis to catch u p - but most did not. This may be one place where the tyranny of a general rule runs roughshod over the specifics of individual talent, preference, and adaptability. Recording actual performances insured that those who later examined them would be working with the "real thing". Another pioneering aspect of their investigation was Parry's insistence on multiple versions by the same singer, by different singers, in different regions, and so on. In this way he hoped to show dramatically how the principle of multiformity, or variation within limits, informed alternate performances. The South Slavic material illustrated beyond doubt that at least the junacke pjesme ('heroic songs') and, as Parry saw it, by analogy the Homeric poems as well - followed this pattern of rule-governed variation rather than memorization or wholesale reinvention. Although it is not widely known, they even conceived of a test, nicknamed the "proba" by Vujnovic, to procure more examples of alternate versions. Via this ruse they told the singer, perhaps 30-70 lines into the song, that it would be necessary to start over, thus generating two, three, or more successive versions of a song opening. The Stolac songs are full of such alternate beginnings, and they are very instructive. Having already mentioned Nikola Vujnovic a few times, we should add a few words about his crucial role in determining what got recorded by Parry and Lord. As a guslar himself, Vujnovic communicated easily and produc-

The textualization of South Slavic oral epics

79

tively with the singers they encountered. The tenor of the conversations makes it clear that he was both an insider and an outsider and that he used both relationships to good advantage. That is, he was able to bring to the table an insider's knowledge of both the tradition and its performance, and at the same time he could urge the research agenda of his academic employers, framing their inquiries in digestible terms. He was willing and able to cajole and flatter, but also to criticize and hold his singer-colleagues to a high standard in their responses. As every fieldworker knows, the answers one receives seldom match the questions posed, at least literally, given discrepancies in cultural and cognitive categories. But Vujnovic deftly and consistently uncovered much of great interest during the conversations, whether Parry and Lord actually intended the particular result or not, and he was their mainstay in discovering a context for the performances they were in the process of recording. A brief example will illustrate the value of his contribution beyond the more general functions of obtaining singers' repertoires and coaching them to perform certain songs. For one thing, Vujnovic, inspired by Parry, was especially tenacious about determining the guslar's notion of a rec u pjesmi (a 'word in a song'), a much misunderstood concept. The singers were often able to distinguish between our contemporary, print-oriented sense of "word" as an entry in the dictionary on the one hand (a definition that lacked any real meaning for them) and a traditional phrase or scene or story-pattern on the other. Thus within the conversational register Vujnovic and Salko Moric, an epic bard from Stolac, were able to agree that "Salko", the singer's name, was a word. And yet Moric also insisted that, when he was singing epic, u pjanoj mehani ('in a drinking tavern') was a single "word". The difference is something G. S. Kirk, author of The Songs of Homer and general editor of the recent Iliad commentary, was unable to see: a difference in two registers, two "ways of speaking". And the person most immediately in charge of what got recorded - both acoustically and via dictation - was fluent in both tongues. 2. What gets published From the "half-ton of epic" recorded in 1933-35, and to a lesser extent from later trips in the 1950's, it was left to Lord to decide what would get published. One strategy would have been to present an anthology of what he considered the best performances from all over the country with very few alternate versions - essentially the Karadzic approach. But Parry and Lord had collected regionally, and had become convinced that the basic organization of the epic tradition was regional (in later work I have identified the organization of South Slavic Moslem epic with the linguistic terms "idiolect", "dialect", and "language"; see Foley 1990: chs. 5,8). For this reason

80

John Miles Foley

Lord chose first to publish a carefully selected group of performances from the Novi Pazar district, featuring five singers, with different songs by the same guslar and alternate performances of the same song by the same singer. In this fashion he hoped to sketch the tradition surrounding any given performance by placing it in its natural context. With these volumes of SerboCroatian Heroic Songs in hand, he and David Bynum could then go on in volumes 3-6 to sample the district of Bijelo Polje, with the master bard Avdo Medjedovic and his performances rivalling the Iliad and Odyssey in length and elaborateness, and in volume 14 to the region of Bihac:, where the presentational form of epic, here accompanied by the tambura rather than the gusle, differs in some interesting respects. For my part, I am working toward original-language and translation volumes of performances from the Stolac area, a part of central Hercegovina that especially favors the Return Song, the story of the Odyssey. Lord's strategy for textual representation also differed significantly from Karadzic's. Instead of interfering occasionally between bard and reader, he aimed at printing precisely what the singer said, "errors" and all, within existing editorial conventions. Only in this way, Lord reasoned, could Slavic specialists appreciate a living oral epic tradition without textual distortion. Only in this way could comparative scholarship, and especially the classical scholar, find the South Slavic epic a useful analogy, a parallel that could allow for "filling in" the necessarily fragmentary picture of Homer as oral poet. 4 Of course, the nonverbal aspects of the guslar's epic performance were, as with Karadzic's Srpske narodne pjesme, silently deleted; the consideration of the role of vocal and instrumental melody, with rare exceptions like Bela Bartok (1953) and most recently Stephen Erdely (1995), was to remain an unstudied dimension. But Lord's insistence on an absolutely "literal" transcription, reflecting the actual performance as well as a transcription could, was certainly a major step forward in the study of South Slavic epic both in itself and as an analogy for other traditions. What got published was now much closer to what got recorded and, in turn, what got performed. I have already mentioned that the publication plan of Serbo-Croatian Heroic Songs (hereafter SCHS) called for proceeding region by region through the six major districts in which Parry and Lord collected, beginning with Novi Pazar and Bijelo Polje and continuing with Bihac in the initial volumes. Within each of these areas the internal strategy was to be the same: to present a selection of guslari, each with multiple versions of the same epic and, where possible, including instances of the same epic sung by more than one singer. Like Karadzic, they sought to re-create a traditional context for the performances they were making available to a highly literate world. Unlike Karadzic, however, they did so not by grouping song-texts around historically and politically important figures or events, but by attempting to simulate traditional multiformity, placing any one traditional instance in its natural,

The textnalization of South Slavic oral epics

81

resonant frame of reference. (I would in fact go further, understanding this frame of reference as the crucial dimension not only of structure but also of art, not only of composition but of reception, but that concern lies largely beyond the present discussion.) The theory driving this plan for "what gets published" was of course the Oral-Formulaic Theory. By presenting a digest of performances from a single area, with considerable overlap among singers and individual songs, Lord provided evidence for his tripartite scheme of compositional units: the formula (or traditional phrase), the theme (or typical scene), and the storypattern. And within its specified arena the demonstration works very well. The South Slavic epic abounds with formulaic expressions based on the heroic decasyllable, and this includes noun-epithet formulas (all six syllables long, so thrift is not an issue) as well as more loosely configured diction. Typical scenes are also very much in evidence: Arming the Hero, Assembly, Shouting in Prison, and Test of the Wife/ Fiancee are only a few that appear in the Novi Pazar songs (the first two volumes of SCHS) and should also be quite familiar from the Homeric poems. And the large-scale dimension of storypattern informs all of the Parry-Lord songs in one way or another - the Wedding Song and Return Song are particularly prominent in their remarkable sample of Moslem epic. The advantage in all three levels of structure is of course that the South Slavic epics actively and tangibly illustrate a morphology that could in many cases only be hypothesized for the Homeric epics. Do you want to study the deployment of the noun-epithet formula licki Mustajbeze, 'Mustajbeg of the Lika', through the phraseology? Then consider the literally hundreds of examples available in the work of numerous singers from one or more regions. Does the shapeshifting of the Arming the Hero scene interest you? Then have a look at the dozens of heroes armed hundreds of times in related songtexts. Are you embarked on the analysis of the Odyssey as a Return Song? Well, you have before you scores of Return Songs, with scores of different heroes in many superficially different but generically similar situations, with wives and suitors and homecomings. Although the Parry Collection has not yet been used to full advantage, its unique comprehensiveness and depth allow for an unmatched perspective on one tradition of oral epic - both its characteristic structure and its characteristic art.

3. What gets received

Again, my comments in this section will be brief, since relevant scholarship is available elsewhere (Foley 1991,1995b, 1996c, 1999). One way to describe what gets received is to ask for whom the SCHS volumes were created. Unlike Karadzic, who was assembling a historical and political as well as a "literary"

82

John Miles Foley

anthology for native speakers, many of whom had considerable familiarity with the Christian epics, Parry and Lord strove to simulate an oral epic tradition almost exclusively for non-native speakers who had no personal experience of the Moslem epic tradition. With that in mind, we can appreciate how important the literal transcriptions in the original language are for Slavic specialists, as well as how significant the English translations are for comparatists. Both groups can profit from the regional constellation of performance-texts in SCHS, with the resultant emphasis on the difference between an instance and a tradition of epic. And in the much understudied area of meaning and art, the Parry-Lord songs, both the published selections and especially the enormously larger archive of unpublished material, allow the serious evaluation of traditional referentiality, the ways in which the oral tradition informs each of its instances with a special, idiomatic force. This last aspect is the sense of the homemade proverb cited above: "Oral traditions work like languages, only more so."

Two applications to other epic traditions Let me close with two brief examples of how the textualization of South Slavic oral epic can have a bearing on editorial and interpretive challenges in other, oral-derived epic traditions. The first, having to do with Homeric epic, concerns a minor textual phenomenon with rather larger implications for formulaic structure as a whole; the second, a feature of the Anglo-Saxon epic Beowulf, is a fundamental and well known property that the South Slavic analogy helps us to understand.

2. Textualization of Homer One of Parry's earliest and most elegant ape^us was his explanation of metrical "flaws" in Homeric poetry as a function of formulaic combination and recombination. To take one general example, he argued that short vowel hiatus - one word ending with a short vowel and the next beginning in the same way - resulted from two phrases being joined together and creating hiatus at the point of juncture. This phenomenon resulted from the fact that each phrase had an independent life of its own in the tradition, he contended. While in some cases a final or initial short vowel would be paired with a consonant, and no hiatus would occur, in other cases two vowels would be juxtaposed and a metrical "flaw" would appear. Thus the apparent necessity for a glottal stop between two short vowels was very often a mark of formulaic structure: one "part" ended there and another "part" began.

The textualization of South Slavic oral epics

83

A look at the Parry and Lord versus the Karadzic song collections puts this phenomenon into a different, and I think clearer, perspective. Although Karadzic silently eliminated all traces of them in his Srpske narodne pjesme texts, performed epic in South Slavic makes generous use of hiatus bridges, consonantal sounds that are inserted in performance into positions where hiatus would otherwise occur. These non-lexical phonemes smooth the aural pathway for a guslar, obviating the need for a glottal stop and allowing him to maintain vocalization (and sung melody) throughout the line - whatever combinations of phrases he might use. The singer Halil Bajgoric from the Stolac district, for example, has an impressive array of such continuant sounds: he can and does deploy [h], [j], [m], [n], and [w] to solve the inevitable phonetic problems generated by the exigencies of composition. These "extra" sounds are audible throughout the Parry-Lord acoustic records, of course, and as part of the policy of exact transcription they have also been maintained in the published textual record, though not at all consistently in the transcriptions made by Nikola Vujnovic, himself a guslar, who perhaps did not "hear" them as part of the song or found it unnecessary or inappropriate to include what were strictly performance-based features in a textual record. Indeed, the phenomenon of hiatus bridges makes one think carefully about the (transcribed) nu-moveable and (untranscribed) digamma in Homeric epic. Might these - and other - sounds also have participated in the ancient Greek poet's negotiation of short-vowel hiatus? Might Homer's formulas also have been bounded - in performance - by such phonetic bridges rather than by hiatus per se? 2. "Completeness" and unity in Beowulf Students and scholars alike have long been frustrated by what seems the incompleteness of the textual record of this unique Anglo-Saxon epic poem. So many moments parallel to but not an integral part of the main story such as the Fight at Finnsburg and the Heathobard feud - are mentioned only telegraphically, leaving readers scurrying to the notes provided in Klaeber's standard edition or, better, to Adrien Bonjour's monograph on what he calls the episodes and digressions in Beowulf, written precisely to allay this difficulty. And the difficulty is, of course, that we modern readers of the oral-derived epic text are not "in on" the tradition from which it springs. We simply do not know the background on which the poet depended for economical communication with the audience. While he or she could confidently cite a shorthand version of a given story with full expectation that the audience would fill in the background and make the connection to the main story, that strategy leaves the uninitiated outside the arena, struggling to make sense of what they perceive as a partial exposition. This is a situation

84

John Miles Foley

not confined to epics but common to most traditional works to some degree: the very fact that the individual item emerges from a natural frame of reference makes interpretation more challenging for scholars of ballads, for example, or Native American tales, or even the homelier genres from the former Yugoslavia like laments and magical charms. The textualization of South Slavic epics illustrates the endemic importance of such a frame of reference, and of the varying need to simulate it for different readerships. Karadzic had much the easier task, in that the audience he was addressing was likely to be broadly au courant with the background of Srpske narodne pjesme, particularly the songs celebrating the Battle of Kosovo and Prince Marko. But Parry and Lord's SCHS project could not assume such a well prepared audience. Especially because we Western scholars are usually very narrowly trained, if at all, in the verbal art of Eastern Europe, most of the information a guslar could take for granted within his tradition is simply not a part of his new audience's working knowledge. Not only do they entirely lack fluency in the traditional way of speaking (as distinct from simple language competence), but they are also wholly unfamiliar with what that idiom points toward. Who is Djerdjelez Alija? What does a Turkish hero's armor look like (especially the feathered headpiece)? What is the conventional role of women in Moslem epic? The SCHS audience is in much the same boat as the modern-day readership of Beowulf, needing extensive editorial apparatus and multiple songs and versions (or their equivalent) to gain even a glimpse of what Karadzic's readership already had available to them as a working interpretive "kit". While in the present state of knowledge we cannot ameliorate the situation beyond a certain point, since no real analogues to Beowulf have survived, we can at least appreciate that the apparently "partial" nature of the text is not a paleographical or artistic blemish but betrays another way of accessing background information. The episodes and digressions are symptomatic not of a problematic text, but rather of the operation of traditional referentiality.

Conclusion The textualization of South Slavic oral epic is a complex and fascinating area of inquiry, and I have had time to focus on only a limited number of issues here. Left aside were the myriad other collections and their common and idiosyncratic features; I especially regret not having the opportunity to treat the published anthology of Luka Marjanovic (1898), whose two volumes of Moslem epic provide many analogues for the Parry-Lord songs, and more generally the published and unpublished holdings of Christian songs in the Zagreb archive. Then, too, there are the liminal cases of Bishop Njego§ and Andrija Kacic-Miosic (see Lord 1986), who seem to have been able to

The textualization of South Slavic oral epics

85

"textualize" their own personal epics, written in the special language of South Slavic oral epic by thoroughly literate, even scholarly authors. This latter phenomenon leads us into the nature of the traditional register, its accessibility to composer and receiver, and related matters that I have no opportunity to address here (on this point, see Foley 1995b: 60-98). What we can say in closing is that oral epics become books in South Slavic via a variety of routes, the two most prominent of which I have chosen for examination here. What emerges at the far end of the process as a volume on the library shelf must pass through many stages, from the initial decision about what gets recorded - effectively the root determinant of everything else - to the secondary decision of what gets published - not, as we have seen, an enterprise without choices - and on to what gets received. At every stage, theory and policy narrow the range of options, productively giving necessary definition and identity to whatever is collected, published, or read. It is this often unexamined process of manufacturing definition and identity that we must closely examine; rather than starting with the received text, in other words, we should begin by asking how this particular facsimile of an oral epic performance reached us in this particular form. South Slavic epic provides one opportunity to shed light on the evolution from performed epic to printed book, and I look forward to hearing about many more case histories during the course of this conference.

Notes 1. On the history of textual editing of traditional oral works, see espec. Fine 1984 and Foley 1995a. On the important approach through ethnopoetics, see espec. Hymes 1981. 2. I regret not having the opportunity to take into account in this paper the fascinating and potentially highly fruitful efforts now underway to provide electronic, hypertext editions of oral and oral-derived epics (I think especially of the HEURO project on South Slavic epics and the Hyper-Kalevala project). It may well be that electronic media, and particularly the Internet and multimedia computing, will make possible new levels of fidelity in the edition of traditional oral works. 3. For an account of Parry and Lord's fieldwork and scholarship, see Foley 1988:1956. 4. On the philological dimensions of the comparison, see Foley 1990,1996a.

86

John Miles Foley

References Bartok, B£la 1953

[Transcription of selections from The Captivity ofDjulic Ibrahim with Notes.] In: Albert B. Lord (ed. and trans.), Serbocroatian Heroic Songs (Srpskohrvatske junacke pjesme), vol. 1: Novi Pazar: English Translations, 435-67. Cambridge, Mass. and Belgrade: Harvard University Press and the Serbian Academy of Sciences.

Bonjour, Adrien 1950 The Digressions in Beowulf. Oxford: Basil Blackwell. Erdely, Stephen 1995 Music of Southslavic Epics from the Bihac Region of Bosnia. New York: Garland. Fine, Elizabeth 1984 The Folklore Text. Bloomington: Indiana University Press. Reprinted in 1994. Foley, John Miles 1988 The Theory of Oral Composition: History and Methodology. Bloomington: Indiana University Press. Reprinted in 1992. 1990 Traditional Oral Epic: The Odyssey, Beowulf, and the Serbo-Croatian Return Song. Berkeley: University of California Press. Reprinted in 1993. 1991 Immanent Art: From Structure to Meaning in Traditional Oral Epic. Bloomington: Indiana University Press. 1995a Folk Literature. In: David C. Greetham (ed.), Scholarly Editing, 60026. New York: Modern Language Association. 1995b The Singer of Tales in Performance. Bloomington: Indiana University Press. 1996a Guslar and Aoidos: Traditional Register in South Slavic and Homeric Epic. Transactions of the American Philological Association 106: 11-41. 1996b Oral Tradition and Its Implications. In: Barry B. Powell and Ian Morris (eds.), A New Companion to Homer, 146-73. Leiden: E. J. Brill. 1996c Traditional Signs and Homeric Art. In: Egbert Bakker and Ahuvia Kahane (eds.), Written Voices, Spoken Signs: Tradition, Performance, and the Epic Text, 56-82, 238-43. Cambridge, Mass.: Harvard University Press. 1999 Homer's Traditional Art. University Park: Pennsylvania State University Press. Hymes, Dell 1981 "In Vain I Tried to Tell You." Essays in Native American Ethnopoetics. Philadelphia: University of Pennsylvania Press. Karadzic, Vuk 1841-62 Srpske narodne pjesme. Vols. 1-4. Rpt. Beograd: Nolit, 1975. 1852 Srpski rjecnik. Rpt. Beograd: Nolit, 1975.

The textnalization of South Slavic oral epics Kirk, G. S. 1962

87

The Songs of Homer. Cambridge: Cambridge University Press. Rpt. in abridged form as Homer and the Epic, 1965. Kirk, G. S. (gen. ed.) 1985-93 The Iliad: A Commentary. Cambridge: Cambridge University Press. Klaeber, Frederick (ed.) 1950 Beowulf and the Fight at Finnsburg. 3rd ed. with 1st and 2nd suppls. Boston: D. C. Heath. Lord, Albert B. 1986 The Merging of Two Worlds: Oral and Written Poetry as Carriers of Ancient Values. In: John Miles Foley (ed.), Oral Tradition in Literature: Interpretation in Context, 19-64. Columbia: University of Missouri Press. Marjanovic, Luka (ed.) 1898 Hroatske narodne pjesme, vols. 3 and 4. Zagreb: Matica Hrvatska. Mladenovic, 2ivomir 1973 Rukopisi narodnih pesama Vukove zbirke i njihovo izdavanje. In Zivomir Mladenovic and Vladan Nedic (eds.), Srpske narodne pesme, vol 1, i-cclxxix. Beograd: Srpska Akademija Nauka. Mladenovic, 2ivomir and Vladan Nedic (eds.) 1973-74 Srpske narodne pesme. 4 vols. Beograd: Srpska Akademija Nauka. Wilson, Duncan 1970 The Life and Times of Vuk Stefanovic Karadzic. Oxford: Oxford University Press. Rpt. Ann Arbor: Michigan Slavic Publications, 1986.

Performance, textualization, and textuality of "elegy" in Old Norse Joseph Harris The major technical terms of our conference can be applied to my subject only after some special adjustments, first to the term "epic". Old Norse-Icelandic offers, of course, nothing similar to Homer, Virgil, or Milton - nor to Manas, Gesar, or Sunjara - though it is possible to see in a few late eddic poems, especially Atlamal and SigurdarkviÖa in meiri, an arrested development in that direction (especially relevant here are Andersson 1980, 1986). In the related Germanic language areas the development from Lied or lay to Epos, or epic in the usual English sense, seems to have happened twice, first in Old English and Old Saxon, then again in Middle High German - such at least is standard literary history (cf. Heusler 1905,1941; Haug 1975; Andersson 1987). The international sense of epic is much more tolerant, of course, and in that sense Iceland, in particular, is blessed with plenty of prose epic, in addition to a moderate amount of narrative eddic poetry. If the ultimately oral nature of both these types of "epic" is often taken for granted, a closer understanding of early Icelandic orality in general remains extremely elusive (relatively recent overviews in Harris 1996; Clover 1985; Quinn 1990). What we do have are many texts and a few anecdotes about saga-tellers and poets living in the pre-literate period. Oral forms, performances, and textualization itself are our constructs (or as we prefer to believe, reconstructions), and in the nature of the beast, they are controversial. (On performance, also cf. Harris 1985, and references there; Quinn 1990; Bauman 1986.) Even when w e have an author, written sources, and oral witnesses specified by name, as in the case of Snorri's Heimskringla, for example, the exact nature of authorship and textualization is not uncontroversial; and most of the passages in Old Norse-Icelandic literature that can throw light on oral stages, oral performance, and textualization have been battlegrounds for a century or more.

Oral performance and textualization The richest of such passages, the description of an Icelandic wedding at the farm Reykjaholar in the year 1119, is, however, still rewarding to consider for students of both prose epic and eddic verse. 1 Now there was revelling and much rejoicing, good entertainment and all kinds of games, both dancing, wrestling and saga-telling. .. .Something is told about

90

Joseph Harris this, though it is of little importance, about who the entertainers there were and what their entertainment was. What is in the account is now contradicted by many, who pretend not to have known it, for many are ignorant of the truth and think true what is invented and lies what is true. Hrolfr from Skälmarness told a saga about HrpngviÖr the viking and about Öläfr Warriors'-King and about the breakin in the funeral mound of prainn the berserk and about Hrömundr Gripsson, with many verses in it. King Sverrir was entertained by this saga, and he declared such "lying sagas" most entertaining; nonetheless, people can recount their genealogies back to Hromundr Gripsson. This saga was composed by Hrolfr himself. Ingimundr the priest told the saga of Ormr, skald of Barrey, including many verses and with a good flokkr [a genre term of skaldic poetry] at the end of the saga which Ingimundr had composed, but nonetheless, many well-informed people hold this saga to be true.

The most recent discussions of the familiar passage emphasize the performance context of the two named entertainers, who both performed prosimetrical narrative, prose sagas with copious verses; and recent students point out the connections between the entertainers, one of whom was a priest, and the group responsible for the first attested writing in Iceland, the writing down of the traditional oral laws in the winter of 1117-18, emphasizing the close involvement of writing and oral performance from as far back as we can reach with certainty.2 It seems, too, that, in certain obscure disagreements reported amidst the wedding festivities, we may perceive a clash of oral and literate mentalities - a point of view popular with humanists at least since Walter Ong's synthesis of 1982, but not undisputed among social scientists.3 No matter how hard we push the evidence, however, it is not possible to guess under what circumstances one of the two sagas recited on this occasion, but apparently not the other, was written down. The oral materials for textualization and the process itself remain a mystery.

Prosimetrum The prosimetrum, or saga with verse, as attested for 1119, may have been a major manifestation of the textuality of some types of performances. The oral, or at least the early literate, audiences were conscious of verse as a specific form, which could have an author, while saga was simply story, basically histoire, as opposed to discours; it could be true or false, appreciated or not, but the idea of a certain kind of prose does not seem to be attested. Yet most of the information on performances pertains more to stories preserved in prose than to verse. When we turn to narrative eddic poetry, far fewer such famous passages offer themselves to the late-coming gleaners. There is codicological evidence, of course. Inferences drawn from it suggest that Snorri's Prose Edda broke

Performance, textnalization and textaality of "elegy"

91

the ground, making possible the idea of a collection of the old heathen and heroic poetry, and some scholars regard the collectors of the poetry pamphlets that preceded the Codex Regius as having acted as cultural conservators in the spirit of the older folkloristics. (Cf. esp. Wessen 1947; and the survey in Harris 1985: 74-79.) On the other hand, the Codex Regius is not an archive or even fieldnotes, but a carefully arranged, and probably selected, anthology that seems to have its own import as a book. Its creator was not indiscriminately saving doomed oral poetry, but making a book that meant something, or many things, to his contemporaries, and like the Norwegian and Northumbrian Christian artists who employed pagan and heroic topoi in Christian settings, he probably intended to use the old poetry and not simply to preserve it. (Cf. Klingenberg 1974; Harris 1985: 77-78, 1996: 11820.)

Both these points of view are inferential approaches to textualization, but direct evidence is quite thin. The thickest description available is probably the picture of an itinerant, but non-professional entertainer who mixes eddic verse with sagas in the story Norna-Gests pättr, a fantasy set about 998 (Wilken 1877); another suggestive vignette reports how an ancient poem, the Bjarkamal, was declaimed on the morning of the battle of StiklastaÖir in 1030 (Harris 1985:118-20). Both reports are preserved at least two centuries after their events, and are generally taken cum grano salts; yet their typical features must at least teach how the early literate authors understood the old poetry to have been performed and transmitted.

The testimony of warning Most such evidence has been extensively debated by eddic scholars, but I have recently been interested in a pair of passages in the poems themselves that have not been extensively milked for our subject. In the older Atli poem, AtlakviÖa, GuÖrun sends her Burgundian brothers a warning against accepting the invitation of the king of the Huns: the message is conveyed by the exquisitely delicate means of a single wolf's hair twined around her gift of an arm-ring.4 Hvat hyggr ρύ bruöi bendo, pä er hon okr baug sendi varinn väöom heiöingia? Hygg ek, at hon vprnuÖ byöi. Här fann ek heiöingia riöit ί hring rauöom. Ylfskr er vegr okkarr at riöa erindi.

What do you think the lady meant when she sent us a ring wrapped in the heath-ranger's coat? I think she offered us a warning. I found a hair of the heath-ranger twisted round the red ring. Our way is wolfish if we ride on this journey.

92

Joseph Harris

The single stanza that conveys this motif offers some peculiarities typical of this poem, such as a sort of doubling or variation between the two parts of the stanza, that have led to the thought that the surviving poem combines elements of two predecessors. It is not due to communicative redundancy, however, that Gunnarr and Hogni have no difficulty understanding that their sister had "offered us a warning": "our way is wolfish if we ride on this mission." Here a single wolf's hair acts as a natural "index" of a wolf, just as a wolf's ears do in the proverb that forms the basis of a line in the eddic poem Fäfnismäl: "I have every expectation of a wolf when I see one's ears" (Kuhn 1962:186 [st. 35]). It is true that the pre-literary sign system imagined in AtlakviÖa does depend on social construction, for its user had to have the cultural knowledge to interpret a wolf, and the juxtaposition of the natural hair with the cultural ring almost seems itself a symbol of social construction. This symbol, if I may indulge this fantasy one step further, is structured like the two-part native kenning with a base and a modifier. In any case, this older poem places a Gunnarr endowed with extreme sensitivity into an oral world of sensory unity. At least its consentience stands out by comparison to the younger Atli poem, Atlamal, where GuÖrün's warning is incised in runes, which have been changed by the messenger (st. 3-4). The equivalent passage struggles through a long series of stanzas (st. 9ff.). Here it is not the heroes themselves but the wife Kostbera who puzzles over a "purloined letter", and appropriately enough Atlamal itself is impenetrable at the point at which GuÖrün's garbled text never does quite render up its secret to Kostbera. Instead the runereader's troubled mind seems in sleep to divine the writer's original meaning. The Atlamäl poet - like another witness to early literacy, Plato - clearly distrusts writing, and the poem overdetermines the women's warning by adding dreams of disaster with traditional male misinterpretations. In deconstruction writing (or language) is inadequate to expression. Here in Atlamal the physicality of runic writing also makes it further alienable from intention. In the older poem there is no misunderstanding though the brothers do ignore the warning and commit themselves to the journey; here the "slippage" so dear to deconstructive thought does not inhere in the sign system but in GuÖrün's understanding of heroic psychology: it was precisely her warning that made the journey inevitable. Eddic elegies The comparison of these stanzas adds nothing to our knowledge of the textualization and oral prehistory of the older poem, but the younger Atlamäl would seem to stem from a milieu in which traditional oral poetics flourished

Performance,

textnalization

and textnality

of "elegy"

93

in contact with writing. I do not think, with Ursula Dronke (1969: 111-12), that Atlamäl was composed in writing, but if we honor the manuscript description of the poem as "from Greenland", the possible time, place, and social circumstances of textualization seem relatively restricted. This late, rather modern-sounding poem - the domestic recriminations remind one of Ingmar Bergman - with its melancholy strain, is generally associated with the group of eddic elegies, so-called: the inner group of six poems concern women and focus on a grieving retrospective monolog; in addition most scholars will associate two longer poems, SigurÖarkviÖa in skamma, Atlamäl, and parts of a third, HelgakviÖa Hundingsbana II, with the group. The reasons for dating these poems late relative to the earliest heroic poems need not be questioned here, but a review of scholarship on them is relevant to the question of their textualization. It was Andreas Heusler (1906) who most memorably formulated the classic view that the softer ethos of these poems argued that they were products of a purely Icelandic renaissance, which he put in the late eleventh century and in the twelfth. Heusler imagined con texts of secular learning for their composition but did not specifically deal with the question of orality and literacy, probably because, in keeping with the older eddic scholarship in general, he did not see that the difference was important. In fact there has been only one really thorough-going attempt to show that some eddic poems came "aus der Schreibstube" (Gutenbrunner 1955), and in general the older eddic scholars assumed an oral poetics which is indistinguishable from writing. In such a context textualization would also present no special problem. Heusler's model was challenged by Wolfgang Mohr in two great and frustrating articles from 1938 and 1939 in which he argued that, like the older "Fremdstofflieder" based on continental story-materials, these younger poems had a prehistory that could be reconstructed through the internal references, especially to Denmark, and through external comparisons, especially with the later Danish ballads. Sometimes Möhr notices parallels in early medieval West Germanic poetry, and toward the end of the second article he weighs sympathetically the possibility of an old, Common Germanic origin of the heroic elegy; but on the whole his prehistory is an eleventhcentury continental one, and the twelfth-century Icelandic poems mainly reproduce this vanished stratum. Mohr's model was too complex and selfcontradictory to dominate post-war literary history, but it forms the background for the most recent contribution, Ulrike Sprenger's 350-page Die altnordische Heroische Elegie (1992).6 Sprenger reasserts Heusler's position, with the addition that the eddic elegies are strongly influenced by Christian religious language and concepts, which she attempts to document chiefly in the lexicon. Some of Mohr's continental comparanda are taken over as influences, but Sprenger's new and almost purely Icelandic genre came into being in the last quarter of the twelfth century and the first quarter of the

94

Joseph Harris

thirteenth. Thus Sprenger conveys her sense of the textuality and intertextuality of the heroic elegies, but no sense of how they were composed, performed, and textualized, and scarcely a word is wasted on writing as a thing apart from the milieu of an oral poetry.

A wider perspective My own work on the subject, beginning in the late 1970's, cast a wider net and argued that the several forms of elegy in Old Norse - the heroic elegies of the Edda, the death songs and other elegy-like poetry of the Eddica minora, and the various skaldic elegies, especially erfikvsedi - were all not only thematically but also genetically related, and further related to the manifestations of elegy in Old English (Harris 1982,1988). I postulate not only a typological relationship - the kind of relationship one might expect to find in comparing laments in only distantly related cultures such as the Balto-Finnic lament poetry studied by Lauri Honko (1974) and Aili Nenola-Kallio (1982) - but also the kind of reconstructable family relationship that most scholars of comparative Germanic literatures will grant for the heroic lay, gnomic poetry, eulogy, and of course for runes, meter, poetics, and language itself (Harris 1994). My conception of the ramifications and genre structure of an oral literature and of how to write its history is different from, but deeply influenced by Heusler's great map of Old Germanic literature, Die altgermanische Dichtung (1941), though on the specific subject of elegy the Swiss master had other ideas. Here my ideas are closer to some of Heusler's predecessors, especially Neckel (1908) and Sieper (1915), and not too far from Mohr in some of his moods. My notion of the ultimate origin of Germanic elegy combines Tsvetan Todorov (1976) on the origin of einfache Formen in speech acts with Levin Schücking's (1908) hypothesis that the roots of Old English elegy extend to funeral laments - a notion not unrelated in 1908 to what came to be known as the myth-ritual theory. None of this can throw direct light on the performance and composition of the eddic elegies in the period before they acquired their present place in the King's Book (always assuming they were not composed in writing for a specific place in this or another book). But the typologically and genetically related skaldic elegies, especially the erfikvaeöi (cf. Fidjest0l 1989), and the other related Germanic "elegies" are at least suggestive for the eddic problem. ErfikvseÖi means Erblied, and all the skaldic poems so designated, together with many others that lack the label, are formally and functionally related to one or more deaths. These poems are not "epic" in any sense; the "I" that speaks is the poet himself, though he may take on a persona as retainer, kinsman, or, especially father. The greatest example, Egill Skalla-Grimsson's Sonatorrek, shows numerous parallels in diction with the eddic group and

Performance, textnalization and textnality of "elegy"

95

makes clear that one function of such elegies was consolation and a substitution for revenge; Tacitus himself had already noted the gendered distinction in the demonstration of grief: women grieved publicly, men nursed grief in their hearts. But when revenge was not possible, a man, even a viking like Egill, would substitute words. The soteriological - or salvific - function of such words is also made plain in Egils saga, for example in the outcome of the composition of Sonatorrek (where Egill's renewed grip on life grew stronger as he composed) or in a masterfully written scene in which Arinbjprn encourages the grieving Egill to speak a lament poem and so to live: Madr skal eptir mann Ufa, "One man must live on after another's death." 5 Another connection between Sonatorrek and an "epic" or heroic elegy is the remarkable echo in Beowulf's "Old Man's Lament" of a proverb about filial loss - a proverb which, I have argued, is also quoted by Egill (Harris 1994). The eddic elegies too deal explicitly with death, loss, and consolation; each has as its core a woman's voice, distantly reminiscent of the Geatisc meowle whose lament for Beowulf includes her own fears. Sprenger insists that eddic elegies are "young", and "young" they may b e - i n fact, very little in Old Norse literature is not "young" according to her criteria - but their generic roots seem to be old. Instead of the men's hall, we should perhaps imagine performances of heroic elegy in the skemma or women's quarters with a predominantly female audience, perhaps engaged like the women of "The Second Lay of Gudrun" in weaving and embroidering narrative tapestry. (Lars Lönnroth's principle of the Double Scene [1979] encourages us to look into oral literature for reflections of its own performance milieu.) Unfortunately for my argument the one detailed vignette of the performance of an eddic lay happens to portray the recitation of one of these elegies by a male entertainer before a king's warband, but notice that there, in Norna-Gests pattr, the author shows his all-male audience as astounded by the poem. And their wild applause is squelched by the saintly king, who changes to a safer - more masculine and less un-Christian - subject. The author of Norna-Gests pattr used the same technique at another point in his fascinating short story, and I think the scene under discussion may be interpreted as showing such an elegy out of its usual context.

Conclusion But I must return to textualization proper for a closing frustration. The famous anecdote accompanying Sonatorrek refers to the textualization of this elegy; as you may remember, Egill's main audience for the oral performance of the poem is female - his wife and daughter - and it is precisely the daughter who offers to preserve the poem by carving it in rimes on a rune-stave. This

96

Joseph Harris

is fiction, and there was a time when most students took it to be fantastic as well, but that was before the Bergen rune finds made such preservation seem at least a distant possibility. On textualization in a strict and narrow sense, however, comparison of the eddic elegies with elegies elsewhere sheds little or no light. A resigned note on which to close, but at least an accurate one.

Notes 1. In Ρorgils saga ok Hafliöa (Brown 1952:17-18); the basic interpretation is still Foote 1955-56 [1984]; the translation is from a paper in progress by Judy Quinn. 2. Meulengracht S0rensen 1993:42-51,117; on the prosimetric nature of these sagas, Harris 1997:131-63. 3. Ong 1982; Finnegan 1988 (for example); drawing the same conclusion, Meulengracht Serensen 1993. 4. St. 8; text and translation from Dronke 1969: 4. 5. The classical edition of Egils saga is Nordal 1933; the episode with Arinbjprn is on p. 148 (ch. 56); Sonatorrek and its narrative context are pp. 242-57 (ch. 78). 6. Since this article was written in 1996, at least one important addition to the discussion of "elegy" and eddic dating has appeared: Sävborg 1997.

References Andersson, Theodore M. 1980 The Legend ofBrynhild. (Islandica XLIII.) Ithaca and London: Cornell University Press. 1986 Beyond Epic and Romance: Sigurdarkvida in mein. In: Rudolf Simek, Jonas Kristjänsson, and Hans Bekker-Nielsen (eds.), Sagnaskemmtan: Studies in Honor of Hermann Pälsson on his 65th Birthday, 26th May 1986,1-18. Vienna, Cologne, Graz: Η. Böhlaus Nachf. 1987 A Preface to the Nibelungenlied. Stanford: Stanford University Press. Bauman, Richard 1986 Performance and Honor in 13th-Century Iceland. Journal of American Folklore 99:131-50. Brown, Ursula (ed.) 1952 Ρ orgils saga ok Hafliöa. Oxford: Oxford University Press. Clover, Carol J. 1985 Icelandic Family Sagas (tslendingasögur). In: Carol J. Clover and John Lindow (eds.), Old Norse-Icelandic Literature: A Critical Guide, 239-315. (Islandica XLV.) Ithaca and London: Cornell University Press.

Performance, textnalization and textaality of "elegy"

97

Dronke, Ursula (ed./trans.) 1969 The Poetic Edda. Vol. I: Heroic Poems. Oxford: Clarendon Press. Fidjestal, Bjarne 1989 Erfidräpa (Erblied). In: Heinrich'Beck et al. (eds.), Reallexikon der Germanischen Altertumskunde 7: 482-86. Finnegan, Ruth 1988 Literacy and Orality: Studies in the Technology of Communication. Oxford: Blackwell. Foote, Peter 1955-56 Sagnaskemtan: Reykjahölar 1119. Saga-Book 14:226-39. (Rpt. in Aurvandilstä: Norse Studies. Odense: University Press, 1984.) Gutenbrunner, Siegfried 1955 Eddalieder aus der Schreibstube. Zeitschrift filr deutsche Philologie 74: 250-63. Harris, Joseph 1979 The senna: From Description to Literary Theory. Michigan Germanic Studies 5: 66-74. 1982 Elegy in Old English and Old Norse: A Problem in Literary History. In: R. T. Farrell (ed.), The Vikings, 157-64. Chichester: Phillimore. Rpt. in: Martin Green (ed.) 1983, The Old English Elegies: New Essays in Criticism and Research, 46-56. Rutherford, N. J.: Fairleigh Dickinson University Press and London and Toronto: Associated University Presses. 1985 Eddie Poetry. In: Carol J. Clover and John Lindow (eds.), Old NorseIcelandic Literature: A Critical Guide, 68-156. (Islandica XLV.) Ithaca and London: Cornell University Press. 1988 Hadubrand's Lament: On the Origin and Age of Elegy in Germanic. In: Heinrich Beck (ed.), Heldensage und Heldendichtung im Germanischen, 81-114. Ergänzungsbände zum Reallexikon der Germanischen Altertumskunde 2. Berlin and New York: de Gruyter. 1994 A Nativist Approach to Beowulf. The Case of Germanic Elegy. In: Henk Aertsen and Rolf H. Bremmer, Jr. (eds.), Companion to Old English Poetry, 45-62. Amsterdam: VU University Press. 1996 Romancing the Rune: Aspects of Literacy in Early Scandinavian Orality. Accademia Peloritana dei Pericolanti (Messina), Atti, classe di lettere filosofia e belle arti, vol. LXX, anno accademico CCLXV (1994): 109-40. 1997 The Prosimetrum of Icelandic Saga and Some Relatives. In: Karl Reichl and Joseph Harris (eds.), Prosimetrum: Cross-Cultural Perspectives on Narrative in Prose and Verse, 131-63. Woodbridge, Suffolk, U. K. and Rochester, Ν. Y.: D. S. Brewer. Haug, Walter 1975 Andreas Heuslers Heldensagenmodell: Prämissen, Kritik und Gegenentwurf. Zeitschrift für deutsches Altertum und deutsche Literatur 104: 273-92.

98

Joseph Harris

Heusler, Andreas Lied und Epos in germanischer Sagendichtung. Dortmund: Ruhfus. 1905 Rpt. Darmstadt: Wissenschaftliche Buchgesellschaft, 1956. Heimat und Alter der eddischen Gedichte. Das isländische Sonder1906 gut. Archiv für das Studium der neueren Sprachen und Literaturen 116: 249-81. (Rpt. in: Kleine Schriften, II, 165-94. Berlin: de Gruyter, 1969.) Die altgermanische Dichtung. 2nd ed. rev. Potsdam: Athenaion. Rpt. 1941 Darmstadt: Gentner, 1957. Honko, Lauri Balto-Finnic Lament Poetry. In: Pentti Leino et al. (eds.), Finnish 1974 Folkloristics I, 9-61. (Studia Fennica 17.) Helsinki: Finnish Literature Society. Klingenberg, Heinz 1974 Edda-Sammlung und Dichtung. Beiträge zur nordischen Philologie 3. Basel: Helbing und Lichtenhahn. Kuhn, Hans 1962 Edda. Die Lieder des Codex Regius nebst verwandten Denkmälern. I: Text. 4th ed. rev. Heidelberg: Winter. Lönnroth, Lars 1979 The double scene of Arrow-Odd's drinking contest. In: Hans Bekker-Nielsen et al. (eds.), Medieval Narrative: A Symposium, 8293. Odense: Odense University Press. Meulengracht Serensen, Preben 1993 Fortxlling og sere. Studier i islsendingesagaerne. Aarhus: Aarhus Universitetsforlag. Möhr, Wolfgang 1938-39 Entstehungsgeschichte und Heimat der jüngeren Eddalieder südgermanischen Stoffes. Zeitschrift für deutsches Altertum und deutsche Literatur 75:217-80. 1939-40 Wortschatz und Motive der jüngeren Eddalieder mit südgermanischem Stoff. Zeitschrift für deutsches Altertum und deutsche Literatur 76:149-217. Neckel, Gustav 1908 Beiträge zur Eddaforschung, mit Exkursen zur Heldensage, 367-89,49596. Dortmund: Ruhfus. Nenola-Kallio, Aili 1982 Studies in Ingrian Laments. (FF Communications 234.) Helsinki: Academia Scientiarum Fennica. Nordal, Siguröur (ed.) 1933 Egils saga Skalla-Grimssonar. (islenzk fornrit 2.) Reykjavik: Hiö islenzka fornritatelag. Ong, Walter 1982 Orality and Literacy: The Technologizing of the Word. London: Methuen. Quirin, Judy 1990 The Eddie Tradition: Α Study of the Mode of Transmission of Eddie

Performance, textualization and textuality of "elegy"

99

Mythological Poetry in the Middle Ages. Diss. Sydney: University of Sydney.

Schiicking, Levin 1908 Das angelsächsische Totenklagelied. Englische Studien 39:1-13. Sieper, Ernst 1915 Die altenglische Elegie. Strassburg: Trübner. Sprenger, Ulrike 1992 Die altnordische Heroische Elegie. Ergänzungsbände zum Reallexikon der Germanischen Altertumskunde 6. Berlin and New York: de Gruyter. Sävborg, Daniel 1997 Sorg och elegi i Eddans hjältediktning. (Stockholm Studies in History of Literature 36.) Stockholm: Almqvist & Wikseil. Todorov, Tsvetan 1976 The Origin of Genres. New Literary History 8:159-70. Wessen, Elias 1947 Den isländska eddadiktningen: dess uppteckning och redigering. Saga och sed 1947:1-31. Wilken, Ernst (ed.) 1877 Die prosaische Edda im Ausziige nebst Vplsunga-saga und Nornageststhättr, Theil I: Text, 235-61. Paderborn: Schöningh. (2nd ed. rev. of 1912 omits the introductory discussions.)

Turkic and Siberian epics

Silencing the voice of the singer: problems and strategies in the editing of Turkic oral epics Karl Reichl Although an enormous amount of Turkic oral epic poetry has been recorded, there is surprisingly little information on the actual performance of oral epics, and documentation is particularly scarce when it comes to non-textual matters. While texts of oral epics have been written down in full since the early 19th century, descriptions of the performance are only occasionally given. Here is Hermann Vambery's impression of the performance of a Turkmen singer, gained during his stay in Central Asia in 1863: In his home, the Turkmen man presents a picture of complete idleness. He considers it the greatest shame if a man puts his hand to any household task. He is only busy with his horse, and when he has finished with it he visits his neighbour or joins a group of men sitting in a circle in front of a yurt and he takes part in their conversation, which is either about politics, new raids or horses. Meanwhile the inevitable cilim, a kind of Persian pipe, which is however smoked with moistened tobacco, is handed round. Only in the evening hours, in particular in winter, they like to listen to enchanting fairy-tales and stories, and it is considered a higher form of enjoyment when a baxsi (troubadour) arrives, who - accompanying himself on his dutar, a two-stringed instrument - recites some songs of Köroglu, Amanmolla or the almost idolized national poet Makhtumquli... It was in Etrek that one of these troubadours had a tent next to ours, and as he visited us in the evening with his instrument, it was not long before some young people gathered round him and he was asked to recite some heroic songs. His song consisted of certain hoarse guttural sounds, which we might consider more a rattling noise than singing and which he accompanied at first by a soft, and later, when he became agitated, by a wild strumming of his dutar. To the degree that the battle became fiercer, the excitement of the singer and the enthusiasm of his young listeners increased, and it was a truly romantic sight when the young nomads, heaving deep sighs, threw their hats to the ground and tore their hair out in a veritable fit of rage, as if they wanted to do battle with themselves. (Vambery 1865:257-58, my translation.)

Vambery's description of the musical side of this performance is typical of the layman's account of non-western music in general: to Vambery's ears the singer utters "hoarse guttural sounds" ('rauhe Kehllaute') and his singing, rather than being musical, resembles the rattling sound of a sick man ('Geröchel'). Having only recently recorded a complete branch of the Köroglu epic from a Turkmen baxSi in Khorezm (Uzbekistan), I can testify that I heard neither hoarse guttural sounds nor any other unpleasant noises. The singer

104

Karl Reich!

I recorded, Gulum-baxäi Ilmetop, accompanied himself on the dutar, a twostringed lute, and was furthermore accompanied by a musician playing the gididiak, a spike-fiddle. While textual variants of the epic he performed are available in print, the music of Turkmen epics is only poorly documented. A reliable scholarly edition (and Russian translation) of one Turkmen version of the Köroglu-cycle was published in 1983 in the series "fipos narodov SSSR", renamed in 1993 "fipos narodov Evrazii". As is customary for this series, there is an appendix on the music of the epic, in this case consisting of short extracts from V. Uspenskij's and V. Beljaev's book on Turkmen music (which appeared in 1928) and musical transcriptions of three verse-passages, also taken from the same book (Karryev 1983: 778-84).1

Textualization of the living sound in performance Although there is some musicological literature on the music of Turkic oral epic poetry, editions (and translations) of oral epics are entirely focussed on the text. A notable exception are the more recent editions of Siberian folklore in the series "Pamjatniki fol'klora narodov Sibiri i Dal'nego Vostoka", which include not only somewhat more detailed chapters on musical matters but also samples of performances on a small record. Thus the edition and translation of the Jakut heroic epic Qiis Debeliye contains, apart from a record, a most enlightening (and thorough) chapter on the music of the Jakut olonxo (epic) by the musicologist A. P. Resetnikova, with ample musical illustrations (Resetnikova 1993). It is probably fair to say that on the whole collectors of Turkic epics have ignored the living sound of the epic. This is reflected in the fact that in the archives of the Turkic-speaking areas of the former Soviet Union and of China which I have visited, I have noted a certain scarcity of tape-recordings and musical transcriptions. Several years ago, when I asked a group of Kazakh scholars working in the Kazakh department of the Sinkiang Academy of Social Sciences in Ürümchi why no one had bothered to write down the music of the epics, I was told: "Why should they? We know what an epic sounds like!" By contrast, I was impressed to find in the manuscript department of the Kazakh Academy of Sciences in Alma-Ata that the manuscripts of oral epics which I was able to consult were as a rule accompanied by musical transcriptions of selected passages. There are, of course, records of oral epic poetry, and the performance of singers (both traditional and conservatory-trained) for television has helped to preserve the sound and visual appearance of an epic recitation, albeit in a somewhat staged and folkloristically polished form. There are also films, both of a documentary and an artistic bias; the Kirghiz film Manasci which incorporates scenes with the famous Kirghiz bard Sayaqbay Qaralayev is a well-known example.

Silencing the voice of the singer

105

But maybe the Kazakh scholars from Sinkiang were right: Why should anyone bother to record the music (and other performance aspects) of the epic? Must not our main concern lie with the recording and preservation of the text? Isn't it true that once the text has been written down, the principal task of the collector has been achieved? This is certainly the general practice, although it has also been common knowledge for a number of years that the edition of an orally performed poem, song, narrative, sermon or speech is by no means the simple conversion of the spoken word into the written word. Scholars working in the field of ethnolinguistics and the ethnography of communication have convincingly shown that orally performed poetry (taking "poetry" in the wider sense of the word) must be understood as a particular type of speech-event and that the transcription of oral poetry should hence be replaced by a transcription of the speech-event, or failing this, that the edition of an oral text must be supplemented by at least a partial documentation of the elements making up the speech-event. What this entails for the edited text has been amply illustrated in the work of Dell Hymes, Dennis Tedlock and other ethnopoetically oriented linguists and folklorists and summarized in Elizabeth Fine's book on the Folklore Text (1984), so that there is no need for me to pursue this question here. (See in particular Hymes 1981 and Tedlock 1983; compare also Bauman 1986 and the survey by SavilleTroike 1989. On the performance aspects of Turkic oral epic poetry see also Reichl 1995.) Although the importance of the insights into the structure of speech-events gained by ethnolinguistic research can hardly be overrated, the utilization of these insights for the actual textualization of oral epic poetry is far from being a settled affair. There are two basic questions one will have to confront when editing oral poetry: (1) Is it at all possible to effect a "translation" from performance to text? In other words, are not the two entities - the speech-event and all it entails on the one hand and the written representation of this event on the other (with "written" including possibly other media than the written word, such as sound and visual recordings) - of such a different nature that a one-to-one translation or "transformation" is categorically excluded? (2) Granted that a translation from performance to text is at all possible, should the written representation of the speech-event - the "text" (in its wider sense) - be a transcription ("documentation") of the event or should it rather be an edition of the event? Put differently, should the written text "contain" everything that made up the performance? An attempt to answer the first question would lead us so far afield that there would be no space left for the discussion of a specific case such as the tex-

106 Karl Reichl

tualization of Turkic oral epic. It entails also the discussion of the ontological status of a literary text, in other words the deconstructionist debate that characterizes much of contemporary philosophical and critical discourse. I would just like to state my own position briefly. For one thing it seems to me that even the most complete replica of the event, catching and preserving sound and sight to the fullest technologically possible degree, is radically different from the event itself. While the event itself is open - it consists of a sequence of choices from various possibilities - the replica is fixed and in its replay pre-determined. Furthermore, the audience of a speech-event takes part in the synchronous event - it is part of the event, even in the case of an entirely passive audience - while the audience of a replay watches a past event, without being a part of it nor taking an active part in it (which does not, of course, exclude emotional involvement, aesthetic enjoyment, etc.).

The singer's text in performance and beyond Although we can conceive of an exact replica of the event (whatever its relationship to the event itself), it can be argued that such a replica, though useful and instructive, is not the ultimate goal of scholarship in the field of epic studies. As the issue I want to raise here is so complex and involved, I will try to shortcut discussion by focussing on a specific example. My example comes from the repertoire of the Karakalpak singer 2umabay-zi'raw, with w h o m I have been acquainted for fifteen years. (For a general discussion of the Karakalpak tradition in the context of Turkic oral epic poetry see Reichl 1992.) I have recorded his complete repertoire, that is to say I have recorded (and transcribed) one or more performances of the various poems and epics making up his repertoire. One of the epics I have recorded (both on audiotape and on videotape) is the epic Qoblan. What I am interested in as an editor of oral epic poetry is 2umabay's Karakalpak version of the epic Qoblan. (I am also interested in the relationship of his version to other Karakalpak and Turkic versions, but this is a different matter.) In a purely nominalistic vein I could maintain that Zumabay's version consists of his various performances of Qoblan, of which only some have been recorded. All I can do is edit the recorded versions, admittedly only a tiny fragment of the singer's "complete" version. But I could also maintain that in order for the singer to be able to perform Qoblan there is an exemplar of the epic somewhere stored in his mind (what Lauri Honko calls "mental text" [Honko 1996]). This exemplar is the product of his training and clearly fashioned on the model of his teacher's version. (One performance of his teacher's version has incidentally been recorded.) When we talk of "the singer's version of Qoblan", we could hence denote by this term his exemplar, which is, of course, only indirectly and partially accessible to us by means of the various performances that

Silencing the voice of the singer

107

have been recorded. Going one step further, and we are now treading on dangerous ground, we could cut the psychological moorings of the singer's exemplar and switch from nominalism to Platonism by assuming a kind of Popperian World Three existence of "the singer's version of Qoblan". This brings us back to deconstruction. Is there a "text" of say "The Ode to the Nightingale" or are there only various readings of this poem? (Of course, there is Keats' manuscript, preserved in the Fitzwilliam Museum in Cambridge, and there are innumerable printed replicas of the words of this manuscript. But what is meant by "text" is really a Platonic entity.) If one conceives of the singer's version as an object on its own (mental or Platonic) which is not equivalent to the sum of its various realizations in performance, then, of course, the editor's task involves some sort of critical and not only documentary activity. Before there is a misunderstanding, it is perhaps appropriate to underline at this point that, while believing in the existence of "texts", I do not advocate the search for some kind of phantom version, constructed from specific texts by methods of textual criticism which seem hardly suitable to orally composed and orally transmitted poetry. In medieval studies, too, the method of establishing critical editions along the lines of Lachmannian textual criticism has been viewed with scepticism for a number of years (cf. Cerquiglini 1989). But I would maintain that despite the importance of complete documentation, the editor's goal is in the end the edition of a singer's version, as represented in its various performances. Editing a "text" means that critical activity is called for, an evaluation of performance for what it is, namely an individual realization of a work of poetry (as mental or Platonic "text"), controlled and limited by time, place and circumstance. Although it is the individual performance on which our edition must be based - and it must be based on this performance according to strict linguistic, ethnopoetic, musicological and philological principles - we must also look beyond performance. Looking beyond performance implies among other things that we define the objective of an adequate edition not as complete documentation but as critical selection.

Transcribing a Karakalpak singer's performance I will illustrate this contrast between documentation and edition by an excerpt from Qoblan. Prior to the discussion and presentation of my excerpt, a few remarks on 2umabay-zi'raw and the tradition in which he stands might be useful. The majority of the Karakalpaks live in Karakalpakistan on the shores and to the south of the Aral Sea; they are culturally and linguistically close to the Kazakhs, with whom they also share a number of epics. There are two types of Karakalpak singers who perform epics - or dastans in the

108

Karl Reich!

native terminology - the baqsi and the ziraw. The Karakalpak baqsi sings mostly love dastans, such as Garip-Asiq Sayatxan-Hämre, Näzeb-oglan and others. The baqsi plays the dutar like the Turkmenian baxs'i of Khorezm and is also often accompanied by a further musician, who plays the girzek (gidzdzak). The dastans performed by the Karakalpak baqsi' are closely connected to the repertoire of Turkmenian and Khorezmian (Uzbek) baxsis; in their musical style, too, a number of similarities in particular to Turkmenian singers exist. The strong Turkmenian influence on the Karakalpak baqs'i is also underlined by the fact that the poems of the Turkmenian poet Makhtumquli (c. 173080) are generally part of his repertoire. The Karakalpak ziraw, on the other hand, plays the qob'iz, a bowed instrument with two horse-hair strings. The word qob'iz as a term for a musical instrument is widely spread among Turkic and non-Turkic peoples (Kazakh qil-qobiz, Kirghiz qomuz, Ukrainian kobza etc.); the instruments denoted by this term are not always identical. While the Karakalpak and Kazakh qobi'z is a bowed instrument, the Kirghiz qomuz and the Ukrainian qobza are plucked instruments. The term is already found in the Old Turkish Book ofDede Korkut, where Dede Korkut himself plays the qopuz, here almost certainly a plucked instrument. Among the Kazakhs the qob'iz was the instrument of the baqs'i. The word baqs'i in Kazakh does not, as in other Turkic languages, denote the singer but the descendant of the pre-Islamic shaman, the faith healer. This points to an early association of the qob'iz with shamanism, one of many links between the singer of epics and the shaman in the Turko-Mongolian world (Reichl 1992: 57ff., lOOff.). The Karakalpak qob'iz is tuned to the voice of the singer, the higher string in the vicinity of c' (in the following example b flat), the lower string a fourth down (in the example f). The strings are used individually for playing the melody as well as both together for double stops. Due to extensive fieldwork, mostly in this century, a great number of epics have been recorded from Karakalpak ziraws. Today only one Karakalpak ziraw who has acquired his art in a purely oral tradition is still alive, £umabayzi'raw Bazarov from Shomanay in Karakalpakistan. 2umabay-zi'raw was born in 1927 and learned the art of performing from the singer Esemurat-zi'raw Nurabullaev (1893-1979) in Kongrat in the forties. He stayed with Esemurat for a period of three years and acquired a repertoire of three epics, Edige, Saryar and Qoblan.The transmission of these epics can be traced back for two further generations of singers to the early 19th century. Edige is centred on two historical personages, Khan Tokhtamysh and Idiku (Idige, Edige) of the Golden Horde, who reigned at the end of the 14th century. The various versions among different Turkic-speaking peoples (some verse epics, some narratives in a mixture of prose and verse, some prose tales) allow us to postulate an origin of the epic some time in the 15th/16th centuries. Saryar might be termed a fairy-tale epic; its plot is very similar to Pushkin's Tale of Tsar Saltan and the dastan is, like Pushkin's tale, a representative of the inter-

Silencing the voice of the singer

109

national folktale of the Calumniated Wife, No. 707 in Aarne and Thompson's type index (1961:242). Qoblan is like Edige a heroic epic; the epic is extant in a number of Karakalpak and Kazakh versions, but there are also traces of this epic in other Turkic languages. Qoblan belongs to the cycle of Noghay heroes; the time of its first composition is difficult to determine; conservative estimates place the origin of the epic in the 15th/16th centuries. Besides these epics, the singer has also shorter songs (terme) in his repertoire, among them a praise song on his native town Shomanay (Somanay täribi), a poem of advice (nasihat), and a historical song (tolgaw) on Ormanbet biy.2 When transcribing the spoken and sung words of the singer, the editor is faced with a number of choices. The basic choice is probably that between a transcription which is essentially a rendering of the spoken (and sung) word into a sequence of lexemes and grammatical particles and a transcription which incorporates paralinguistic features, of a verbal and non-verbal kind, as well as various prosodic features, such as stress or intonation. And if we limit ourselves to the purely textual (lexical) level, we are still faced with various alternatives: Should the transcription of the text be phonetic or rather phonemic? Should it contain idiosyncrasies of the singer's pronunciation? Should dialect features be included or should the written form of the epic conform basically to the standard language? As to the actual editions of Turkic oral epics from Central Asia (Uzbek, Karakalpak, Kazakh, Kirghiz), one can say that even in cases where the recorded text has not been tampered with, paralinguistic and prosodic features are not taken into account. The text is generally converted into the standard orthography of the standard language in question (see Petrosjan et al. 1977). This means that the written representation of the singer's words is basically phonemic. Individual idiosyncrasies of pronunciation or dialect features are sometimes recorded, but as a rule only when their recording is necessary from the point of view of metre. In the Uzbek edition of Fäzil Yoldäs-ogli's version of Alpämis (a highly edited text), for instance, the singer's pronunciation is given in the following lines for reasons of rhyme: Äq kiräwka, ältin säwut kiyaman, Bär kucimni bilagima dziyaman.3 I will put on the white kiräwka [a kind of shirt], the golden coat of mail, All my strength I will unite in my arm. The form dziyaman is typical for the singer's dialect, which is actually Kiptchak Uzbek, a dialect which from the linguistic point of view is closer to Kazakh and Karakalpak than to Standard Uzbek. In this dialect Old Turkic / y / in initial position becomes / z / or / d z / as in Kazakh and Karakalpak. Although the manuscript in which Fäzil's version is transmitted has / d z / in these contexts (the manuscript dates from 1927/28 and is written in Arabic

110 Karl Reichl letters), the edition cites these forms only when this is necessary for the metre. In this case the Uzbek form is yigaman, which would not rhyme with kiyaman. In Karakalpak epic poetry, we find the opposite phenomenon, i.e. the pronunciation / y / instead of / ζ / ~ / d z / . Sometimes this pronunciation is a Turkmenism in poetry which has been heavily influenced by the Turkmen tradition. These Turkmen forms are generally left in the edited texts if they affect metre (rhyme, number of syllables). Sometimes, as in the case of £umabay-zi'raw, the variation / y / - / z / ~ / d z / is a dialectal feature (Nasyrov 1983). How should one deal with these and similar phenomena? There is no doubt that a detailed phonetic transcription of the singer's words is interesting from the linguistic point of view. For the dialectologist, for instance, such a narrow transcription, using the international phonetic alphabet, is invaluable. The sound of poetry is, of course, also important from the literary point of view. There is no need to invoke Roman Jakobson's insistence on sound and self-reference in poetry to substantiate this. But, however paradoxical it might seem, the sound of poetry is the sound of language, not of speech. Patterns and oppositions on the level of sound are those embedded in the sound-structure of the language (and that includes dialects) and not those dependent on an individual's voice qualities or idiosyncratic pronunciations. To give just one example, the very first line of Qoblan reads in "broad transcription", modelled on the orthographic system of Standard Karakalpak (and using the new Latin alphabet): Äyyemgi ötken zamanda In former times long past

In phonetic transcription, taking account of the singer's pronunciation, the line reads: ajjemgi Miutkiun zomondo

The actual pronunciation is hence marked by velarization: [ä] > [a], [a] > [o]; vowel-harmony is stronger in the spoken language (depending on dialect) than in the written language; initial [o] and [ö] are labialized; rounded [o] becomes unrounded, velar [ui]. But despite these differences, the oppositions are the same in both lines: the phonemic representation is just as indicative of the sound structure as the phonetic representation. Additional information on the pronunciation of Karakalpak and the dialect of the singer flesh out the phonemic representation of his words to give an idea of their actual pronunciation, but they do not, on the whole, significantly change the oppositions inherent in a phonological rendering of the spoken words. This is not to say that dialect features should not be incorporated into the written form of the spoken word. There is, for instance, a certain amount of

Silencing the voice of the singer

111

variation between / y / and / d z / ~ / z / . As this affects the sound of the poetry (irrespective of the question whether these two representations are allophonic or not), I indicate this change (cf. Kirchner 1992,1:133ff.). In order, however, to show that the initial sound of words like zer and yer is a free variant I write the standard form zer for both forms and indicate by italicizing (or underlining) the initial / z / when it is actually pronounced / y / . Notation of prosodic and paralinguistic features Far more difficult (and problematic) is the notation of paralinguistic and prosodic features. Paralinguistic elements include non-verbal features like gestures and facial expressions and verbal features like Trager's vocalizations. Prosodic elements comprise stress (suprasegmentale over lexical items), pitch variation (intonation, i.e. suprasegmentale over utterances) and phenomena like rhythm, speed of utterance, degree of loudness or pausing.4 There are two problems here, one the technical problem of how to render these features in print - a problem for which several solutions have been found in the linguistic and ethnolinguistic/ethnopoetic literature - and the more fundamental problem of which of these features should be represented at all. As to the technical problem, given a satisfactory recording of the singer's performance, there is no reason why it shouldn't be possible to transcribe the various prosodic and paralinguistic features present. David Crystal, for instance, has devised an elaborate notational system for English (1969), which could be used also for Karakalpak (and other Turkic languages). He lists among prosodic systems tone (pitch direction), pitch-range, pause, loudness, tempo, rhythmicality, and tension; among paralinguistic systems voice qualifications (laugh, giggle, tremulous etc.), voice qualifiers (whisper, breathy, husky etc.) and tension (slurred, lax, tense, precise), the latter being also an element of prosody. But the question is once again, how much of this information is necessary when the goal is not total documentation but critical edition. When we look for these features in the performance of the Karakalpak ziraw, we can observe that they are intimately linked to the various modes of performance. These modes are in turn dependent on whether the text is in prose or in verse (or, as a third alternative, in rhythmical prose). The general rule (which covers other Turkic traditions as well) is that the prose parts of the epic are declaimed, while the verse parts are sung. Although there is considerable variation among traditions, singers and epics, the verse parts tend to be monologues or dialogues of the various figures in the epic, while in the prose portions third-person narrative predominates. Notable exceptions are set descriptive passages, "themes", such as the hero's ride through the desert or battle scenes, which are generally in verse. On the other hand,

112

Karl Reichl

the prose portions can be quite elaborate, with lively dialogues and descriptions in rhymed prose (saj c ), in particular in the Uzbek dästäns. For the Karakalpak ziraw's performance of verse and prose the opposition between a tense and a lax voice quality is crucial; it characterizes the two main modes of performance, singing and speaking. The origin of the tense voice quality of singing is obscure; it is possible that there is a connection to shamanism, a connection for which there are also other pointers. It is in any case obligatory for the singer to sing in a tensed, glottalized style. When I asked 2umabay for a reason, his answer was simply that this is the way he learned to sing and that this is the way the epics have to be sung. There is no doubt that "tension" must be incorporated into an edition of the text, either in some notational form or at least in a descriptive gloss on the manner of the singer's style. For the narration of the prose portions of the epics, on the other hand, 2umabay uses his natural voice; he speaks clearly and with a loud voice, his vocal expression being neither strained nor distorted. There is a certain relaxed air about his narrating; passages in direct speech are spoken in the intonation one would expect of the speaker in the circumstance in question, without, however, exaggeration or dramatization; the events are narrated in a way that makes what is happening clear to the audience; there is a certain challenge in his voice and facial expression at times, as if to say: "Have you understood? Isn't it stupendous?" There are pauses at particularly important moments, and the singer expects his audience to react to surprising or outrageous events. While narrating, the singer might take up his bowl to sip tea, stroke or twirl his moustache, readjust his headgear or wipe his forehead with a handkerchief. Pauses are filled by re-tuning his qobi'z or by playing one or two motifs. The speed of narration is steady and fairly rapid and the singer never hesitates while he is engaged in relating the story. To illustrate at least some of these prosodic features, I will give a highly simplified transcription of a short prose passage from the version recorded on 19 and 20 June 1994 in Nukus: Ex. [1] A...qsaxan patsa qulaq asi'yi'p turfp | häT bala/m || öziq nawzuwan zassa/q || [tea] aw 0 onda 0 sennen 0 ülken 0 qatta 0 palwannan ° bir nesse miqi zatfr goy sira/gi\m || sagan qalay_zol_beredi_ol_zatqa/n palwan | bira/q safari^ o//q_bolup ba/ra goysaq || soga/n oqtiy ata I goysaq || Aqsaxa/n ata/m usfrn attif de/p aytqayisaq || talabit/q särti/q aliaJ igbatiqa beri/p agasti'_qulata goy/saq ||

Silencing the voice of the singer

113

Qurtqani' uzati'p elige_ali'p qayta goysa//q | özime_äkelip tapsi'rgaysaq deldi\ || [tea] The padishah Aqäaxan listened carefully [and said]: "O my boy, you are an inexperienced youth. Hey, at that place much greater wrestlers than you, more than a thousand, are assembled, my friend. How will these assembled wrestlers greet you? But, may you go on a successful journey! May you shoot the arrow there [into the goal]! May you say that you shot for your father AqSaxan! God giving you luck for your wish and your endeavour, may you fell the tree! Taking Qurtqa as a bride, may you come back! May you get her and bring her to me!" he said. Context: Aqsaxan is the ruler of Samarkand. He wants to win the fair Qurtqa's hand but has no success in the suitor contest (which consists in felling a tree by an arrow shot). Qoblan, the hero of the epic, rides out to win Qurtqa for himself; on his way he is met by AqSaxan, who is returning from his failed bride-winning expedition and charges Qoblan to woo Qurtqa for him. Gloss: This transcription is on the segmental level basically phonemic. Sandhi phenomena (such as the pronunciation qulag instead of qnlaq in 1. 1), or labial vowel harmony (such as the pronunciation turup instead of tur'ip in 1. 1) are not indicated. Also, the stress on syllables, as far as it is part of the linguistic system of Karakalpak, is not marked. As to the various other features transcribed, they conform to a fairly reduced system of basic oppositions and distinctions: pause is marked by | and ||; if there are short musical interludes, these are marked by ./. and .11., respectively; a rise in pitch (generally accompanied by an increase in loudness) is symbolized by I or II after the vowel; a fall of pitch (generally accompanied by a decrease in loudness) is symbolized by \ and \\. Double marks signify an increase of the feature in question (higher, lower, longer). Very strongly marked syllables are indicated by T, very weakly marked syllables (approaching murmuring) by i . A staccato enunciation, generally slower than the preceding section, is symbolized by a legato enunciation, generally in a quicker tempo, by _. Finally, lengthened vowels are followed by dots; gestures are enclosed in square brackets. This transcription could be refined in various ways. But rather than discuss the possible elaborations of the above transcription, I would like to discuss the relevance of the features noted for an edition of the text. They fall into three groups. There are first features which are part of the language system and hence embedded in the rule system of the language. In lines 7 , 8 , 1 0 and 11 the particle goy, an emphatic particle, is accompanied by a pitch rise at the end of the line; the same applies to line 9 and 12 with the optative morpheme gay. In an edited text these lines would be punctuated with an exclama-

114

Karl Reicht

tion mark, a sign that, given the intonation rules of Karakalpak, symbolizes precisely the given intonation in its general form. Similarly, the murmured pronunciation of dedi at the end of this passage corresponds to the intonational rules of Karakalpak. These rules govern also the general intonational pattern of the other lines. What the rules of grammar do not specify, of course, is the exact intensity of pitch and loudness, also the exact relationship between various degrees of pitch and loudness in the text. A second group of prosodic features is only indirectly, if at all, rule-governed and encoded in the language system. These are the various features related to pause, tempo, and rhythm. Some of the pause features can be expressed by punctuation; indeed one might argue that punctuation has to be based mainly on pausing rather than other criteria such as grammatical construction. As it turns out, however, the pauses often agree with syntactic segmentation. The staccato or legato enunciation of a group of words as well as the slower or faster pace or utterance belong to parole rather than langue and often have a foregrounding or backgrounding function (emphasis = slower, staccato; background information = faster, legato). But they can also depend on either the singer's need to reflect or his wish to proceed with his story faster. Finally, there are paralinguistic features like the singer's drinking tea which lie completely outside the rules of linguistic behaviour. First of all, there is no denying the fact that all the features mentioned are important for a correct transcription of the singer's performance as a communicative event. It is, to start with tea-drinking, part of the ritual that a singer performs in a context of feasting. Even when recording in the House of Culture in his native town (where lighting and acoustics were better than in a private home), it was felt necessary that there be a table with bread, tomatoes, and fruit for the singer. The tea is actually necessary both for his voice and more generally as a means of refreshment. But to come back to the edition of the text: is it necessary to indicate at what stages the singer drank tea? A parallel might be helpful. Take a performance of Bach's A minor violin concerto by David Oistrakh. If my main interest is the music itself, an "edition" of Oistrakh's performance in the form of a CD will do. If I'm interested in violin technique, I want to see his bowing, the way he holds his violin, possibly the way he moves his body, and maybe even at what stages he wipes his brow (if at all). I need an "edition" of his performance in the form of a video tape. The same applies to our singer. If my concern is the communicative event, I will want to record at what stages the singer twirls his moustache, wipes his forehead, drinks tea etc. If my concern is with the epic as a poetic, aesthetic entity, the drinking of tea seems to play only a marginal role, if it plays any role at all.

Silencing the voice of the singer

115

While it might be clear in this particular instance of drinking tea where the dividing line between documentation and edition lies, the same clarity is unfortunately not possible for all the features mentioned above. If we conceive of the oral epic as a work of literature (apart from other non-literary dimensions such as ritual), an edition must preserve the fullness of its linguistic and poetic/aesthetic structure. And it is the latter which comprises all the various manifestations of sound: the phonological and phonetic level of language, the patterning of sound, the rhythm and melody of speech, and the music of poetry. Some of the prosodic features discussed above will be retrievable from any sensitively punctuated and graphically displayed edition of the text, while others will be lost in a more traditional text edition. I would argue for the inclusion of more features than generally found in the edition of oral epics (such as rhythmic patterning and in particular the musical side of epic). But I find it difficult to draw a line between "essential" and "nonessential" prosodic (and similar) features, while feeling nevertheless that such a line exists. The dilemma is similar to that found in medieval philology. If we have a medieval text preserved in a manuscript we can either produce a so-called diplomatic edition, which reproduces the text exactly as it is found in the manuscript, abbreviations, punctuation, line-division and all (nowadays largely superseded by photographic facsimile editions), or we can produce an edition of the text in which the text has been "re-designed" to conform to the kind of printed text we are familiar with. Today we have our doubts about heavily edited texts: modern punctuation is nowadays sparingly applied, and in some editions the medieval system of punctuating the text has even been re-introduced; we do not normalize spellings any more (unless in editions specially designed for the general reader or the undergraduate); and we are generally careful to preserve as much of the original appearance of the text as possible. There is therefore a movement towards the diplomatic text, but even so the most faithful edition is marked by the hand of the editor and differs radically from a diplomatic edition (or facsimile), if not in appearance then at least in intention. Inclusion of performance features Although I have no safe and sound rules to offer for the textualization of prose, I would nevertheless like to underline the importance of performance aspects also for what I have been calling a critical edition. This is even clearer in the case of verse. One important paralinguistic (and prosodic) feature of both prose and verse is the playing of the qobtz. While (among the Karakalpaks) there are no gestures which might be specifically related to the performance of epic, the qobi'z marks the rhythm and speed of performance. Just as

116 Karl Reichl pauses are important for punctuation, so are the bow strokes the singer inserts into his text. As punctuation marks and structural signals I indicate them (in verse and prose) by various slashes, depending on the type of "interlude" (one short stroke, several short strokes, a short musical motif, repeated motifs etc.; see above). When it comes to the verse passages, their performance follows a fixed pattern. The following extract from Qoblan will illustrate this; it is the verse passage immediately preceding the prose passage quoted above: 5 Ex. [2] [S] Hä...ä, hä...ä bala-awIII Sarfala atqa er sali'p, Ayildi' mäkkem men sali'p, At beline minedi-ay. Qabarlaspay hes kimge 5 2eke basi', tän özi Bir allaga siy'ini'p, Äyni sähär waqtfnda, Qoblan zolga kiredi-ay .11 Äne, zolga enedi. 10 Asti'ndagi' sariala at "Alle maydan Söllerde Tulparli'qtf bilsin!" dep, "Palwanni'q keypi tfnsi'n!" dep, Här bir maydan zerlerde 15 "2abi'rayil!" dep angi'ydi", Atqan da oqtan kem ernes, äy.// Atqan oqtan kem ernes, Hes hay wanni'q terji ernes. Alle maydan söllerde 20 "Zabi'rayil!" dep Sirqi'rap, At iistinde Qoblanli' Atlarga qamsi' beredi-äy.// Bir ne§se maydan ötkende, Asti'ndagi sariala at 25 Taqi'rqani'p qaradi'-aw, Qulagi'n tigip turadi'-ay. Arti'q tuwgan Qoblanli" "Ne äzayi'p boldi?" dep, Aldi' endi arti'na 30 Közin sali'p bayqasa, Bir bölek zaqgi't kiyati'r, Aldi, betten qusagan, häy .11 Ilgi'y quwla qarasa, Dawislarf elatqa 35 Uqsaydi' endi bätsagar.

Silencing the voice of the singer "Kim eken?" dep arjlasa, Sapan oqbay kiyatqan Aq§axandi'-aw köredi-ay .11 At aldina keledi: 40 "Assalawma äleykum!" Haq sälemin beredi. "Wa äleykuma assalam!" Sälemin älik aladi. "Qaydan qaytqan balasaq? 45 Negi'p ziirgen zan barsaq? Sapariri qayda, äiragi'm, äy?"// [R] Anda Qoblan söyledi,// Anda Qoblan söyledi: "Sapar tarti'p si'qqanman, 50 Azgana üyli Qi'psaqtan./ Qorgansaday qalada Qidi'rbay Saldi'q yaratqan, Bir alladan tilep algan Balasi' bolaman./ 55 Talaban bar Qatagan Nogay xalqi'na ki'z Qurtqaga. Talaban boli'p endi baratqanI Men de bala/ häw bolarman-aw."/// He put the saddle on the yellow-dappled horse, made the saddle-girth fast and mounted on horse-back, ay. Without talking to anybody, 5 all by himself and alone, Qoblan bowed to the one God and started in the morning time on his way. He started on his way. 10 The yellow-dappled horse under him said: "In every field and steppe the fact that I am a winged horse shall be known!" It said: "The hero's heart may be calm!" In every field 15 it said "2abi.'rayil" [the Archangel Gabriel] and was no slower than a flying arrow. It was no slower than a flying arrow and no animal was its equal. In every field and steppe 20 it said "2abi'rayil" and whizzed along. Qoblanli on the horse gave the horse the whip. When it had passed many fields,

117

118

Karl Reichl

the yellow-dappled horse 25 looked in surprise, raised up its ears. The noble Qoblanli' said: "What strange thing has happened?" When he looked in front and behind, 30 he saw a heap of dust coming, as if coining from ahead. As he was looking straight towards the south, their voices resembled those of people, 35 the cursed fellow. As he realized who it was, he saw Aqsah-khan returning from his unsuccessful journey. He rode up to him 40 and greeted him respectfully: "Assalawma äleykum!" "Wa äyleykum assalam!" answered the other. "Where do you come from, child, 45 with what purpose are you riding along, where does your journey lead you, my dear?" Then Qoblan said, then Qoblan said: "I have left on a journey 50 in order to get married. I am the son whom Qi'di'rbay the old man has begot, has received as his wish from the one God in the town of Qorgansa. 55 I am a suitor for the girl Qurtqa from Qatagan of the Nogay people. I am a young man who is now on his way as a suitor."

The actual performance of this verse passage comprises six elements: (1) The singer begins by tuning his instrument and playing various melodic snatches, consisting of the basic melody of the passage and slight variations [Introductory melody]. (2) He then begins the actual singing [S], starting with a melodic line (g-f-b flat) which consists basically of a sustained note ("ä" on b flat). (3) The first seven lines of the actual text are then sung to one and the same melody. The following transcriptions are "emic" rather than "etic", i.e. they capture only the distinctive features of the melodic line and not all the acoustic details of the actual performance (for this distinction see Nettl 1964:103ff.):

Silencing the voice of the singer

119

Ex. [3]

S 1. 2. 3. 4. 5. 6. 7.

Sa Ayil At CJa Ze Bir Äy

4-

)

4-

ral di" be bar ke al ni

s

4

at mäk Ii las ba la sä

)

J*

^ er men mi hes tän si waq

qa kem ne pay si', ga här

*

^ sa V sa ne kirn ö yi tfn

β

1

li'P, lip, day. gc zi nip, da,

The melody of the text is stichic and not strophic. This accords with the metrical structure of the verse passage. The verse lines (of generally seven or eight syllables) are not grouped into regular stanzas by rhyme or assonance. They are rather arranged into an irregular pattern of non-rhyming and rhyming lines (e.g. saltp: Saltp 1-2; bilsin dep: ttns'in dep 12-13, kem emes: terji emes 1 18 etc.). The melody is syllabic, i. e. there are no melismas. It falls rhythmically into two parts, corresponding to the two parts of each verse line, which are separated by a caesura (generally after the fourth syllable). These two parts have a clearly stressed first beat; they can be interpreted as a sequence of a 4-beat and a 3-beat bar. Extra syllables are incorporated into the two bars by shortening the notes. The tonal range is comparatively small: it does not exceed a fourth (f-b flat). The melody begins on g, which is repeated several times, descends to f and ends on b flat; it is accompanied in unison on the qob'iz. Line 8 has a slightly varied melodic form: Ex. [4]

5 8 Qo-blan

zol-ga ki-re - di-ay

(4) At the end of this line there is a short instrumental "interlude" (marked in the text by //): Ex. [5]

Β

\

Π

Π

\

ι

\

at

jyjffu *

n

n

Π

\

Π

π

120 Karl Reichl Similar instrumental breaks occur also after lines 16, 22, 32, 38 and 46. In each case the line preceding the instrumental "interlude" is sung to a slightly altered version of the basic melody, similar to the melody given for line 8 (example [4]). (5) The last part of the passage, beginning with line 47, is spoken in a high declaiming voice [R]. In the present example the singer inserts an instrumental break after the first line of the declaimed passage (1. 47) which leads him to repeat this line when he re-starts declamation (1. 48ff.). The spoken part of the passage is several times punctuated by a short musical phrase on the qobtz (marked by / in the text):

Ρm

Ex. [6]

\

m

m

Ψ*

glissando

The passage ends with (6) a sequence of sustained notes on the last syllables of the text (marked by ///): Ex. [7] S— Jf

f

[

* ' ß

^

a

häw bo-larmanaw

The use of a basically stichic, syllabic melody of comparatively small range is fairly typical of the singing of epic poetry in a great number of traditions. This type of melody is generally considered archaic and is contrasted with melodies of a more song-like, often stanzaic or at least poly-motif structure, with melismatic elements and a wider melodic extent. Robert Lach, for instance, remarks on the Finnish runo melodies that they "bear all the symptoms of archaic melody in the highest degree: they are characterized by an utmost monotonousness, uniformity and obstinate repetition of the same motif" (Lach 1913:328, cf. also Kondrat'eva 1975). Among the Turkic-speaking peoples this type of melodic organization is found in various traditions. B. G. Erzakovic distinguishes three types of musical performance of Kazakh epic poetry, the "singing-recitative style" (pesenno-reöitativnyj vid), the "musical-illustrative style" (muzykaV no-illjustrativnyj vid) and the "programmatic style" (muzykaVno-soversennyj, programmnyj vid). It is the first type which is typical of the performance of the heroic epic (zir), and the transcriptions Erzakovic gives for the Kazakh versions of Qoblan, in particular the ones of the Kazakh singer Amanzolov Erezep-zi'raw from Karakalpakistan, show a close resemblance to 2umabay-zi'raw's melodic style (see firzakovic 1966: 158ff.; see also firzakovic 1975). A similar melodic organization is found in

Silencing the voice of the singer

121

the musical performance of Uzbek dästäns, in particular of verse-passages in lines of seven or eight syllables which describe some typical scene such as a battle or the hero's ride through the desert. It is interesting to note that there are also similarities in the quality of the voice: Uzbek singers from Southern Uzbekistan generally start in a lower register and sing in a pressed, strained voice (on the performance of Uzbek dästäns see Dzabbarov 1971; on the style of an Uzbek baxsi from Southern Uzbekistan see Reichl 1985: 616ff.). The use of a stichic, syllabic melody for the performance of epic poetry is also typical of the Kirghiz tradition. V. S. Vinogradov distinguishes four types of melodies in his description of the musical performance of Manas (see Vinogradov 1984; compare also DyuSaliev 1993:159ff.). By far the most common of these is the so-called dzorgo söz 'flowing speech', which is characterized by its final cadence, a "triadic" descent from g or a to c', either in a succession of minor and major third (g-e-c) or major third and fourth (a-fc). In its melodiousness this melody is quite different from Zumabay-zi'raw's tolgaw melody. The musical patterning of verse sections (tirades or laisses) in Kirghiz is, however, very similar to that found in the performance of the Karakalpak ziraw: besides the melodious dzorgo söz more recitative-like melodies are also used; there is a change from singing to declaiming at the end of each tirade, and the verse section is generally rounded off by a melismatic flourish corresponding to the final sustained notes of the example transcribed above. The change from singing to declaiming is even more noticeable in the Uzbek musical style, where the singer is said to "boil" when his singing switches into a high, exalted declaiming manner at the end of a verse passage. There is no space here for further comparisons - for instance to the musical style of the Jakut οίοηχο or to the Altai qay - or for further elaborations of my musical analysis. It should be stressed again that the performance style of the Karakalpak ziraw is only one of several styles found among the Karakalpaks. It stands at one end of a scale, at the other end of which we find the musically elaborate style of the baqsi. Both, in their different ways, are representatives of the musical and poetic heritage of the Turkic-speaking peoples. But while in the case of the ziraw the roots of his art in an archaic layer of Turkic oral tradition are still clearly discernible, in the case of the baqsi the singer of tales has been transformed into a singer of songs and skilful instrumentalist. We are reminded of a similar development in Europe in the Middle Ages, where the singer of epics gave way, in the later Middle Ages, to the minstrel, the entertainer, musician and performer of romances. There is a third mode in Zumabay-zi'raw's performance of epic. He sometimes recites rather than sings passages which are so highly regular in their rhythmic arrangement that they could be printed as verse. Here his voice loses its natural inflections and takes on a more stylized reciting tone, similar to a particular Western way of reciting poetry (with a strong marking of the verse rhythm and a certain sing-song intonation). The border-line between

122

Karl Reichl

verse and prose is sometimes blurred, also in other Turkic traditions. One characteristic trait of Turkic oral epic poetry is the marked preference for parallelism, and it has been argued that the development of verse in Turkic oral poetry takes syntactic parallelism as its starting point (cf. 2irmunskij [Zhirmunsky] 1985 [1965]). It is certainly the case that the prose passages can be highly patterned, with "purple passages" in rhymed prose (saj% especially in the Uzbek dästäns (cf. Sarimsäqov 1978). Generally, the difference between verse and prose is one of singing and saying, but intermediate positions, like rhythmic recitation (but not singing), are possible. In some Uzbek traditions the prose passages are incidentally also performed to the accompaniment of a musical instrument: When performing the dästäns the prose parts are narrated, but in a number of regions they are performed as recitative to the accompaniment of the dombira. We have observed such a way of performance from Razzaq-baxäi Qazaqbaev of the village Pe§qärän in the Namangan district and from Ergas-säzanda Zakirov of the village of t a r w a q in the Osh district, where the prose as well as the verse of the dästän is performed from beginning to end in the manner of a recitative. (Abdullaev 1989:114, my translation.)

As a final point I would like to mention that the transcription of the music is riddled with the same problems that we encountered in relationship to prosodic (and paralinguistic) features. There is the technical problem of adequately transcribing the music and the theoretical problem of knowing how much of the musical aspect of the epic should be incorporated into an edition. Gilbert Rouget, in an important article on the problem of transcribing non-Western music (1970), has suggested that a musical transcription must contain both a notational rendering and a descriptive statement on the music (enhanced/ supported by spectograms and other technical devices). But even if the music is transcribed in the traditional way, there is still a choice between a detailed, "phonetic" transcription like that given by Bela Bartok for samples of South Slavic heroic song, and a broader, "phonemic" transcription like that recently given by Stephen Erdely (1995), also for South Slavic heroic songs from the Parry corpus. Recapturing the voice Coming back to the two questions I posed earlier - the question of "speechevent" vs. "(mental/Platonic) text" and of "total documentation" vs. "critical edition" - I would like to emphasize once again that the distinctions I have been trying to draw are not quite as clear-cut as I would like them to be. The "text" is only accessible (and understandable for what it is) in the context of the communicative event, and the textualization of performance

Silencing the voice of the singer

123

must try to do justice to the fullness of sound and rhythm and hence reach beyond the limitations of conventional editorial techniques. In fact, presentday computer technology would allow us to create a multi-media edition where the written text is supplemented by a visual and acoustic "text". As long as we are dealing with a living tradition, the sound of epic can still be illustrated and documented. But it has to be admitted that an "edition" however close it is to "documentation" - is no substitute for the living event. It transforms the event, and in doing so inevitably silences the voice of the singer. This makes it all the more important to recapture as much as possible of the singer's voice, at least through an appeal to our imagination as readers of epic texts, rather than listeners to the spoken, declaimed, recited, chanted and sung word. As editors of oral poetry we still have a long way to go; but neither the length of the way nor its difficulty are discouraging; what is discouraging is that oral epic poetry is quickly disappearing, at least in some regions of the world, and we must therefore speed up our pace on the way.

Notes 1. On the music of Turkmen epic poetry see also Gullyev 1989. Slawomira ZeraAskaKominek has edited a number of CDs with the music of Turkmen baxSi's; cf. Turkmen Epic Singing/Körogln (Paris, AUVIDS/UNESCO 1994); Turkmenistan: La mitsique des "bakhshy"'(Geneva, ΑΙΜΟ & VDE GALLO 1991). 2. On the legend of Edige and the various epics on Edige see 2irmunskij 1974; two Karakalpak versions, one by Erpolat-zi'raw, the other by Qi'yas-zi'raw, are published in Bayniyazov & Maqsetov et al. 1990. Two versions of Saryar (by Oteniyaz-z'iraw and Qulamet-z'iraw) are edited in Maqsetov 1984; on the relationship of the plot of the epic to Pushkin's tale, see Nurmuxamedov 1983. On Qoblan see Nurmagambetova 1988; Esermurat's version is edited in Zapaqov et al. 1981. 3. Zarif, Mirzaev & Hakimdzänova 1992-93:2:23. All translations of Turkic texts are mine; the transcription system employed is that of Reichl 1992: 7-11. 4. The question of whether prosodic features like stress and pitch have phonemic status is a tricky one and depends on the language in question. For a succinct summary see the article "Phonemics" by Tsutomu Akamatsu in Malmkjaer 1991: 339-45. On the analysis and notation of "paralanguage", see the seminal article by Trager 1964; for a discussion of non-verbal communication see also Hinde 1972, and on the "body-language" in Afro-American discourse see Cooke 1972. The main lines of the linguistic study of prosody from the Renaissance to the present day (applied mainly to English) are traced in Crystal 1969: 20-96. For a summary of the systems of prosodic and paralinguistic features in English according to Crystal see ibid., 177. 5.1 have used the following example also in Reichl 2000:133-38.

124

Karl Reich!

References Aarne, Antti and Stith Thompson 1961 The Types of the Folktale: A Classification and Bibliography. (2nd rev. ed. FF Communications No. 184.) Helsinki: Academia Scientiarum Fennica. Abdullaev, R. S. 1989 Bytovanie dastanov ν Uzbekistane. In: Zemcovskij 1989:113-17. Bauman, Richard 1986 Story, Performance, and Event. Contextual Studies of Oral Narrative. Cambridge Studies in Oral and Literate Culture, 10. Cambridge: Cambridge University Press. Bayniyazov, Q. and Q. Maqsetov et al. (eds.) 1990 Edige. Qaraqalpaq xaliq dästanü Nukus: Qaraqalpaqstan. Cerquiglini, Bernard 1989

Cooke, B. G. 1972

Crystal, David 1969 DyuSaliev, K. 1993 Dzabbarov, A. 1971

tloge de la Variante. Histoire critique de la philologie. Paris: fid. du

Seuil.

Nonverbal Communication among Afro-Americans: An Initial Classification. In: T. Kochmann (ed.), Rappin'and Stylin' Out: Communication in Urban Black America, 32-64. Urbana: University of Illinois Press. Prosodic Systems and Intonation in English. Cambridge Studies in Linguistics 1. Cambridge: Cambridge University Press. Pesennaja kul'tura kyrgyzskogo naroda. Bishkek: Nacional'naja AN Resp. Kyrgyzstan, Institut Literaturovedenija i Iskusstvovedenija.

Uzbekskij dastan (epos). Voprosy muzykoznanija, Vypusk 2,13-35. Tashkent: Izd. Fan. Erdely, Stephen (transcr.) 1995 Music of Southslavic Epics from the Bihac Region of Bosnia. (Milman Parry Studies in Oral Tradition.) New York: Garland. firzakovic, B. G. 1966 Pesennaja kul'tura kazaxskogo naroda. Muzykal'no-istoriieskoe issledovanie AN Kaz. SSR, Inst. lit. i iskusstva. Alma-Ata: Izd. Nauka. 1975 Notnye zapisi melodij iz eposa "Koblandy-batyr". In: Koblandybatyr. Kazaxskij geroiceskij epos, 420-37. 6pos narodov SSSR. Moscow: Izd. Nauka. Fine, Elizabeth C. 1984 The Folklore Text: From Performance to Print. Bloomington: Indiana University Press. Gullyev, S. 1989 Destannyj epos i iskusstvo turkmenskix baxsi. In: Zemcovskij 1989: 125-32.

Silencing the voice of the singer

125

Hinde, R. A. (ed.) 1972 Non-Verbal Communication. Cambridge: Cambridge University Press. Honko, Lauri 1996 Epics along the Silk Roads: Mental Text, Performance, and Written Codification. Oral Tradition 11/1:1-17. Hymes, Dell 1981 "In Vain I Tried to Tell You." Essays in Native American Ethnopoetics. Philadelphia: University of Pennsylvania Press. Karryev, B. A. (ed. and transl.) 1983 Gjor-Ogly. Turkmenskij geroiceskij epos. Epos narodov SSSR. Moscow: Izd. Nauka. Kirchner, Mark 1992 Phonologie des Kasachischen. Untersuchungen anhand von Sprachaufnahmen aus der kasachischen Exilgruppe in Istanbul. (Turcologica 10.) 2 vols. Wiesbaden: Harrassowitz. Kondrat'eva, S. N. 1975 Κ istoriceskoj tipologii epiceskix napevov. Tipologija narodnogo eposa, 152-63. AN SSR, Inst. Mirovoj Lit. Moscow: Izd. Nauka. Lach, Robert 1913 Studien zur Entwicklungsgeschichte der ornamentalen Melopöie. Leipzig: C. F. Kahnt. Malmkjaer, Kirsten (ed.) 1991 The Linguistics Encyclopedia. London: Routledge. Maqsetov, Q. 1984 Saryar. Qaraqalpaq xal'iq dästan'i. Qaraqalpaq fol'klori 13. Nukus: Qaraqalpaqstan. Nasyrov, D. S. 1983 Dialekty karakalpakskogo jazyka. In: D. S. Nasyrov et al. (eds.), Voprosy karakalpakskogo jazykoznanija. Qaraqalpaq til bilimi mäseleleri, 56-113. Nukus: Qaraqalpaqstan. Nettl, Bruno 1964 Theory and Method in Ethnomusicology. New York: Free Press of Glencoe cop. Nurmagambetova, O. 1988 Kazaxskij geroiceskij epos Koblandy batyr. Alma-Ata: Izd. Nauka. Nurmuxamedov, Μ. K. 1983 Skazki A. S. Puskina ifol'klor narodov Srednej Azii. Tashkent: Izd. Fan. Petrosjan, Α. A. et al. (eds.) 1977 Fol'klor. Izdanie eposa. Moscow: AN SSR, Inst. Mirovoj Literatury. Reichl, Karl 1985 Oral Tradition and Performance of the Uzbek and Karakalpak Epic Singers. In: Walther Heissig (ed.), Fragen der mongolischen Heldendichtung. III, 613-43. (Asiatische Forschungen 91.) Wiesbaden: Harrassowitz.

126

Karl Reich! 1992

Turkic Oral Epic Poetry. Traditions, Forms, Poetic Structure. (The Albert Bates Lord Studies in Oral Tradition 7.) New York: Garland Publishing. Epos als Ereignis: Bemerkungen zum Vortrag zentral-asiatischer 1995 Turkepen. In: Walther Heissig (ed.), Formen und Funktion mündlicher Tradition. Vorträge eines Akademiesymposiums in Bonn, Juli 1993, 156-82. (Abhandlungen der Nordrhein-Westfälischen Akademie der Wissenschaften. Band 95.) Westdeutscher Verlag, Düsseldorf: Nordrhein-Westfälische Akademie der Wissenscaften. The Performance of the Karakalpak Zhyrau. In: Karl Reichl (ed.), 2000 The Oral Epic. Performance and Music, 129-50. (Intercultural Musical Studies 12.) Berlin: Verlag für Wissenschaft und Bildung. Resetnikova, Α. P. 1993 Muzyka jakutskix olonxo. In: P. N. Dmitriev and Ν. V. Emel'janov et al. (ed. and transl.), Jakutskij geroiceskij έροε Kyys Oebilije, 26-69. (Pamjatniki fol'klora narodov Sibiri i Dal'nego Vostoka.) Novosibirsk: Izd. Nauka. Rouget, G., avec la coli, de J. Schwarz 1970 Transcrire ou Decrire? Chant Soudanais et Chant Fuegien. In: fichanges et communications. Melanges offerts ä Claude LeviStrauss. Studies in General Anthropology 6: 677-704. Sarimsäqov, B. 1978 Ozbek adabiyätida sadz\ Tashkent: Izd. Fan. Saville-Troike, Muriel 1989 The Ethnography of Communication. An Introduction. Language in Society 3. 2nd ed. Oxford: Basil Blackwell. Tedlock, Dennis 1983 The Spoken Word and the Work of Interpretation. Philadelphia: University of Pennsylvania Press. Trager, George L. 1964 Paralanguage: A First Approximation. In: Dell Hymes (ed.), Language in Culture and Society: A Reader in Linguistics and Anthropology, 274-79. New York: Harper and Row. Uspenskij, V. S. and V. Beljaev 1928 Turkmenskaja muzyka. Moscow: Gos. Izd. Muzykal'nyj sektor. Vämbery, Hermann 1865 Reise in Mittelasien von Teheran durch die Turkmanische Wüste an der Ostkiiste des Kaspischen Meeres nach Chiwa, Bochara und Samarkand, ausgeführt im Jahr 1863. Leipzig: F. Α. Brockhaus. Vinogradov, V. S. 1984 Napevy "Manasa". In: Manas. Kirgizskij geroiceskij epos, 492-509. (Kniga 1. £pos narodov SSSR.) Moscow: Izd. Nauka. 2apaqov, N. et al. (eds.) 1981 Qoblan. Dästan. Esemurat zi'raw Nurabullaevtan zazi'p algan. (Qaraqalpaq fol'klorf 8.) Nukus: Qaraqalpaqstan.

Silencing the voice of the singer

127

Zanf, Η., T. Mirzaev and A. Hakimdzänova (eds.) 1992-93 AlpömiS. Ozbek xalq dästänlari. (Ozbekistän Respublikasi Fanlar Akademiyasi Aliäer Nawäiy nämidagi Adabiyat Instituti.) Tashkent: Ozbekistän. Zemcovskij, 1.1, (ed.) 1989 Muzyka eposa. Stat'i i materialy. Yoshkar-Ola: Komissija muzykovedenija i fol'klora Sojuza kompozitorov RSFSR. Zirmunskij, V. M. 1974 Skazanie ob Idige. Tjurkskij geroiceskij epos, 349-86. Leningrad: Izd. Nauka. 1985 [1965] Rhythmico Syntactic Parallelism as the Basis of Old Turkic Folk Epic Verse. In: V. Zhirmunsky, Selected Writings. Linguistics, Poetics, 320-52. Moscow: Progress. [Originally published in 1965.]

Textology and epic texts from Siberia and beyond Arthur T. Hatto The archaic cultures of Siberia, varied though they were, are susceptible of some generalization for the purpose in hand for both the Czarist and the Soviet periods. Since the middle of the last century and even earlier, farsighted scholars attempted to make records of Siberian oral traditions, often as exemplified in heroic and epic songs, in the spirit directly or indirectly of a salvage operation: directly by recording such songs for their own sakes; or indirectly to illustrate their works on languages deemed moribund. Here, thanks to the development of native schools and universities under the Soviets, albeit on the latters' terms, incomparably more natives became involved in the recording and study of their own cultures during that time than under the Czars, yet on foundations created by wellnigh legendary Czarist - and anti-Czarist! - scholars under the patronage of the Imperial Academy of Sciences, a patronage continued in the same lordly fashion by the Soviets, and with commendable results so far as oral studies are concerned. In the present age of high technology it would be facile to judge by our so recent standards the failures and near-failures in the textualization of Siberian and other epics. We ourselves do not always achieve the ideal, prior to textualization, of a complete audio-visual record of an Epic Event in a customary environmental setting, spontaneously/ritually attended by the habitues, with a minimum of three different angles of vision to sample audience reaction. Thus it would be misplaced to take a haughty view of the fact that there have fallen into my poor net only three reports of phonograph cylinders with heroic narratives: one with an excerpt from the Kirghiz Semetej (Manas cycle) of 1903; the second from an unknown narrative from an ExiritBulagat (West Burjat) singer in 1912; the third, several cylinders containing Nenec (Samojed) hero-songs of three different subgenres, recorded by Toivo Lehtisalo in 1911-12. Later - as late as December 1959 - a conference of Sibirologists was held in Ulan Ude, Burjatia, at which the musicologist Β. M. Dobrovol'skij, in the spirit of a rescue-operation, asked for the use of the tape-recorder in ethnographic fieldwork, and to that end promised to bring pressure on both the Academy of Sciences and the Ministry of Culture to speed up production of portable apparatus (Sidel'nikov et al. 1961:133). While he was speaking, Ζ. N. Kuprijanova may still have been recording the exordia of Nenec hero-songs on her (imported?) tape-machine. Even in 1966, Erika

130 Arthur Τ. Hatto Taube had to plug into an external source of power for her "Smaragd" to record Altai-Tuvinian hero-tales (Taube 1994:19). Even with the aid of the most sophisticated sound-recording apparatus, the task which exercised the 19th-century field workers remains in the making of a verbal text: to achieve a valid broad transcription by phonemic analysis, and that done, with peculiarities safely preserved in the apparatus criticus, how to relate the transcription to the musico-histrionic dimension. In his musicological Appendix to Ζ. N. Kuprijanova's edition of Nenec hero-songs (1965), the afore-mentioned Β. N. Dobrovol'skij argued strongly that in sung/recited narrative the musicological analysis must precede the verbal on prosodic grounds and that tape-recording was thus a minimal requirement. It is further to Dobrovol'skij's credit that in his paper at the Ulan Ude conference in 1959 he had recommended the use of cinematographic apparatus to record movement, and even movement of the singer's mouth when producing "a series of phonemes" (sicfonem, for which read "sounds"; Sidel'nikov et al.: 134). Recorders of epic song, perfectly well able to produce a phonematically based transcription for their texts from the dialect/speech of a given singer will nevertheless have found themselves before the dire choice as to whether their texts were to be purely scholarly and to be published as they stood, or whether they were to be part of an emergent National Literature borne by a National Grammar and Orthography based on a painful compromise between dialects. Where such tensions occur, as for example in Nenec, they are indicated below.

Ainu Writing in 1980, the late C. J. Dunn stated that a considerable amount of the rich Ainu oral literature had been preserved either by transcription into the Japanese syllabary or more satisfactorily into Roman letters, or in more recent years by recording on tape (Dunn 1980: 330f.). Dunn further mentions 20 kamui yukar (songs of gods, mostly zoomorph) on disc as supplied by the NHK (Japanese Broadcasting Corporation), dated 1965 and 1967 by the late B. Hickman, who supplied C. J. Dunn with an expert Bibliography (Dunn 1980: 342). The Revd. J. Batchelor published the first notable edition of an Ainu epic text in 1890 in the Transactions of the Asiatic Society of Japan, with the Ainu original in the left and his English prose translation in the right column of one and the same page, with copious footnotes where needed (Batchelor 1890). The Ainu text is in the Italic form of the Roman, is devoid of diacritics, including length-marks, and of any non-phonetic renditions other than sh and ch. The apparent simplicity of this transcription may be due to Batchelor

Textology and epic texts from Siberia and beyond

131

accepting a Japanese version of Ainu sounds, compare below how N. A. Nevskij's field-records were textualized. Working in Hokkaido in 1919-22, Nevskij wrote down his texts, among them epics, phonetically in a Cyrillic script with additional Roman letters and some diacritics, but without standardization (i.e. in a narrow transcription). In 1972, L. L. Gromkovskaja published some of Nevskij's epics together with other pieces from his field-texts preserved in the archives, following the convention in her day of "the canonical transcription of Soviet Japanology for the Japanese language" (Nevskij 1972:5). Thus the epics heard by Nevskij in Ainu were filtered through two other languages as remote from each other as Japanese and Russian. As the translator, one might say donor, of an abundance of Ainu Songs of Gods, Songs of Humans, Donald L. Philippi was not bound to say how his pieces had been recorded. Thus beyond the formula "recorded in writing" by this or that collector or by the singer in person, we learn nothing about either the material or "epical" circumstances of the recordings (Philippi 1979: passim). Batchelor had taken down the above-mentioned epic (now known as "The epic of Kotan Utunnai", retranslated and set stichically on the basis of the revised original, by Philippi as his No. 33) prae 1889, and he delivered it to the Asiatic Society of Japan in December 1889. His convert Imekana wrote down her whole epic and other repertoire as from 1926. Batchelor introduced her to K. Kindaichi, the future editor of 7 volumes of materials from Imekana (ed. 1961-72). I. Kubodera continued Kindaichi's work into the age of tape, by which time, as was seen, the discs of the Japanese Broadcasting Corporation were also available. Altaian In Altaian heroic epic, the ever-notorious impediments to recording to dictation were augmented by the nature of the epic-specific style of voice-production known as kay. Subjectively, A. Anochin had likened the low laryngeal kay to the sound of a beetle in flight. The musicologist Β. M. Sul'gin, however, states that kay was not a gamut of squeaks but a type of true singing through the somewhat compressed larynx, subject to the usual laws of music. Moreover, he asserts, familiarity with this mode of singing leads to liking, with a soothing effect on the nerves. Nevertheless, the effect of this athletic vocal style on the singer was one of exhaustion. The path to fame of the celebrated bard A. G. Kalkin was from narrative prose through recitative to kay. When in performance he tired from kay, he retreated into recitative, laying aside his stringed topsur as he always did for dictation. (Sul'gin 1973: 454-59.)

132 Arthur Τ. Hatto It follows from this last that dictated Altaian epic texts are in one sense even less authentic than the common run of dictated epic texts in that they lack the pervading influence of their specific mode of utterance, the kay. Assuming that the kay once rendered the voices of spirits through the mouth of the bardic medium, dictated epic texts in Altaian will have lost a mysterious vital semantic link with the past. (Cf. Sagalaev and Oktjabrskaja 1990:179ff.) The circumstances of the recording of the Altaian epic Maadaj kara are so revealing that they must be given in some detail. A. G. Kalkin (b. 1925) performed this epic before an audience of Altaians in February-March 1964 in the Institute of Pedagogy, Gorno-Altaisk, and it was exceptional that it was recorded on tape. The brilliant, prolific native fieldworker S. S. Surazakov deciphered the tape within 20 days, during which time the bard was at his full disposal for consultation. Surazakov also disposed of two earlier variants of Maadaj kara from Kalkin in manuscript to his dictation. The first, taken down by Surazakov himself in 1948 was edited and published by him in a shortened form in 1957. The second variant, taken down by another in 1951, remained in the archive owing to its prose form, violating the verse-structure, and also because of its inaccuracies. (Surazakov et al. 1973: 443.) Surazakov reports that Kalkin could perform for eight hours together. He sang seven of the stock shortish lines of Turkic epic with one breath. When he paused to rest his voice from kay, his listeners were free to interpose their comments (Surazakov et al. 1973: 439-42). For all we know, the upshot of such responses by the audience may have been incorporated in the bard's utterance, yet there is no indication of them in the published text of 1973 (7738 lines), nor in the Commentary. Surazakov was exceptionally well-versed in the process of taking down epics from singers and preparing them as texts for publication, it is therefore a painful loss that his untimely death in 1980 frustrated the completion of an essay by him, the Russian title of which I render as "The speech-structure of Altaian epic poetry and the means of establishing its texts" (Recevaja struktura altajskogo eposa: sposoby ee tekstovogo oformlenija) (V. M. Gacak, in Surazakov 1985: 9). Detached theoretical discussion of his consultations with the bard Kalkin on the basis of his prompt transcript from tape, together with the two earlier variants, must have engaged Surazakov at the level of "mental text" (on the concept of "mental text", see Honko 1996). This in turn would have thrown light on the bard's and the scholar's joint endeavour, whether conscious or unconscious, to furnish an optimal realisation of the "mental text" for display in an inter-Soviet series such as that of Epos narodov SSSR, with all its latent competitiveness. Those who are interested in the typographical manipulation of the printed page as a means of at least signalling the non-verbal semantics of epic performances, for example, histrionics and voice-production, will find a mighty

Textology and epic texts from Siberia and beyond

133

challenge in the Altaian singing-mode of kay, not excluding the vocalized intake and release of breath! In an introduction to Maadaj kara 1973,1. V. Puxov appraises the pioneer Turkologist W. Radioff's recording of Altaian hero-tales in the early 1860s (published 1866) in these terms: "Although Radloff effected his recordings of the texts to the dictation of the performer, he was unable fully to preserve the verse-structure, since he wrote down not the performance but a paraphrase by the Narrator" (Surazakov et al. 1973: 10). This perfectionist assessment by so high an authority is not to be gainsaid, yet recourse to Radloff's Altaian texts (Radloff 1,1866:10-112) shows many blocks of genuine prosodic speech-patterning such as will have been shaped for, and in, narrative song. In his Introduction (Radloff 1,1866: xvii-xxiv), Radloff sets out 10 voweland 30 consonant-symbols and discusses their application in detail to the varying Turkic tongues of South Siberia illustrated by his texts. In other words, he used a narrow transcription which he could tune into the case in hand. This alphabet is based mostly on Cyrillic, but it also uses a few Roman symbols, and further uses diacritics. Radloff does not refer to Böhtlingk's specifically Jakut alphabet of 15 years earlier (see infra), which employed 9 vowel- and 21 consonant-symbols, although there is some overlap with it in Radloff's system. Nor does Radloff refer to any "Academic Transcription" his ambition had already soared to embracing all Turkic tongues comparativistically, he was already shaping a transcription which he could adapt to the needs of his great Opyt slovarja tjurkskix narecij (1893-1911).

Altai-Tuvinian Under strong Mongolian cultural influence as the Altai-Tuvinians were, it is not surprising that they favoured hero-tales (Heldenmärchen) rather than epics. Though technically, these narratives could be passed over in this paper, Dr. Erika Taube's account of their collection and initial textualization is highly relevant to our enquiry. I translate: C. Suxbaatar, assigned to assist me and unacquainted with the Tuvinian language, took charge first of all of organizational work: his responsibility was to guarantee us electrical current, on which we still depended in 1966, when I had to work with an inconvenient and under those conditions unpractical tape-recorder of "Smaragd" brand, having no other. C. Suxbaatar procured a mount for this or that narrator to ride into our centre, he bustled about on such occasions to free the narrator from chores... (Taube 1994: 19.) ... In our Altai-Tuvinian tales, the text is in most cases based on tape recordings. But of course it also happened that we heard some performance or other where it was not possible to use the taperecorder, working from the electrical grid as we were, or where, not anticipating a "catch", we did not have it with us. [Once] in the yurt of a relation of Galsan's,

134 Arthur Τ. Hatto a woman, A. Baajaa by name, met us, who in response to our enquiry regarding tales gave the unexpected reply "My heart is full of them!" and began to narrate forthwith. In such cases we asked them to narrate more slowly so that Galsan could write it down hard after the performer. But that inhibited the free flow of the narrator's utterance. So we let him [sic!] tell on tranquilly, with [Galsan] taking down only the "strongpoints" [opornye mesta], but after the performance of the tale, Galsan recorded everything and only in individual cases - on the next day, basing himself on notes made at the time of the narration, and fixing the names, special concepts, felicitous or original thematic turns and expressions, picturesque comparisons, formulae etc., as on a frame the scarlet thread of the action's unwinding. For tales written down in this fashion the narrators' names are given in the Commentary. (Taube 1994: 27.)

Burj at Burjat epics, named üliger, can be roughly divided into three groups: i) epics without the hero-name "Geser"; ii) epics with a super-hero "Geser" yet entirely free of subject-matter from the Tibeto-Mongol Geser-cycle; iii) full derivates of that same Geser-cycle. Although there were some ten or so important Burjat dialects, there were three main dialect-groupings: Xori, Ungin, Exirit-Bulagat. It was only in the last decade of the 19th century that the contents of some üliger came to the notice of scholars in Russian re-narration as published by G. N. Potanin and Μ. N. Xangalov (Ulanov 1963: 9), and indeed Xangalov's re-narration of Abai Geser Bogdo Xan is said to be very detailed (Burcina 1990: 70). The first recordings deemed scholarly appeared only in the present century. This comparatively very late emergence to notice of the üliger in an area not remote from Irkutsk must be due to the known reluctance of Burjat singers to let their words be fixed in writing. Where clan customs remained strong, to divulge the contents of epic performance to outsiders was tabued. This ties up with the fact that the earliest translations, those of üliger of the third group, come from the East, where Mongolian influence was strongest; whereas those from the first group, from the archaic Exirit-Bulagat tribes, came to notice only later, when clan-structure was breaking down. The zealots for clan traditions demanded faithful transmission of essential content from the bards. The hallowed time and season for epic performance was by night, when the Pleiades had risen. Disclosure of epic content, it was feared, would expose the clan to further Russification through the transfer of the epic's power. The reality of such fears is underscored, for example, by the refusal of the bard P. Dmitriev to allow "The Battle with the Monster Sire Ninata" to be written down - a figure about whom many iiligersin would not even sing (Ulanov 1960: 9b).

Textology and epic texts from Siberia and beyond

135

This last raises a question of some theoretical interest for textologists. What is the status of a personage of a "mental text" who remains locked up in the bard's mind? Is the Monster in question in the "text" or out of it? I leave it to you. It is highly significant that an experienced fieldworker, editor and general authority on the üliger, A. Ulanov, includes the clan arbiters on tradition among the text-shapers of epic evolution (Xomonov 1961:6a), a view which doubtless many of us here have long held with regard to epic poetry in general (Hatto 1989:150f.). Ulanov here speaks of an oral epic text establishing itself, of its clearing and settling (otstojavSegosja teksta). Would we say that the arbiters, over time, force their "mental texts" on the bards, or, if not, refuse to let the bards impose novel "mental texts" on them? Under such surveillance, Burjat epic texts are judged to have been very stable, with only exordia and epilogues in a freer style (Ulanov, in Xomonov 1961: 6b). When at last fieldworkers were given access to Burjat epics, they encountered new and intricate problems. Which means were they to choose for their records? How were they to deal with the all-pervading phenomenon of dialect? As to means. With regard to mechanical recording, in 1912 two young scholars, V. A. Mixajlov and Β. E. Petri, recorded a nameless narrative from an unknown bard on a wax cylinder. The text was deciphered by G. O. Tudenov and the melody transcribed by V. V. Korguzalov and published in 1958 (Tuloxonov 1991: 39). It emerges from Tuloxonov's Bibliography, item 62, that the Tudenov-Korguzalov transcription forms part of a treatise on Burjat poetics: its length is not indicated, nor is it designated an üliger. In 1962, the exordia of four üliger were recorded on tape by S. P. Baldaev and M. P. Xomonov from the Exirit-Bulagat bard B. Barnakov. A seemingly entire üliger was recorded on tape in 1966 from the singer S. S. Sontoxonov. All six mechanical recordings were made in the Exirit-Bulagat dialect. Apart from perhaps the last, the aim of the mechanical recordings was musicological. The authors of the section on the music of the Exirit-Bulagat epic in Tuloxonov's edition of Alamji-mergen (1991), A. S. Dugarov and Ju. I. Sejkin, observe on the difficulties of associating melodies recorded so late (mostly in the 1960s) with texts recorded at the beginning of the century, and they draw attention to C. Jamcarano's insistence on taking down his texts from live performance among a bard's habitues, whereas tape-recording of the musical exordia was in isolation (Tuloxonov 1991: 52). Since not a single musical exordium of Alamji was available to them, the two above-named musicologists had to work by extrapolation from Exirit-Bulagat exordia from other epics.

136

Arthur Τ. Hatto

Jamcarano took down his first Burjat texts in "Mongol writing" (pis'mo), but he did not persevere with it because it did not reflect the nuances and lilt of Burjat speech, while the marvellous epic style, in particular that of ExiritBulagat iiliger, was almost entirely lost (Xomonov 1961: 229b). The Russian Academic Script was available for purist fieldworkers, and we shall see that Jamcarano later availed himself of it. But it, too, was open to criticism. Among a people hitherto without a national literary language, the advent first of a Roman alphabet in the early 1920s, then of a Cyrillic alphabet in 1939, could not fail to exert a pull on the textual presentation of any scholar or writer concerned to bring his hard-won epic to the eyes and hearts of a wider audience. The changing preference for one or other of ten or more significant dialects as a basis for the national language, first the Songolo-Sartul', rendered in Roman up to 1939, then the well-tempered Xori in a Cyrillic adaptation, had inevitable repercussions in the sphere of textualization. Three brief histories of editions of iiliger will illustrate these pressures in general terms. (A) The 1991 edition of Alamji-m0rg0n, a major Exirit-Bulagat epic, has the following history. 1) Jamcarano took it down in the field in 1903 in an evidently narrow phonetic transcription from the bard Elbon Salbykov as he performed to his habitues. 2) A copy of this field-record was made by Jamcarano himself in 1912, evidently for 3) the St. Petersburg edition of 1913 by A. D. Rudnev and two assistants. While preparing his edition, Rudnev received advice from Jamcarano, who was on expedition in Urga, Mongolia. Rudnev held the MS copy (our 2, above) inviolate as far as possible, but when he had to emend (as for example with variant spellings of one and the same word) he retained the rejected forms in footnotes. Rudnev also admits to silent emendation of e to ε in monosyllables and diphthongs. 4) M. Tuloxonov based his 1991 edition of Alamji on 3), above, as the combined work of outstanding documentary reliability of Jamcarano and Rudnev. Tuloxonov's text is no longer in the narrow Academic Transcription but in the Burjat National Orthography, which is also based on Cyrillic but avoids diacritic signs. Here Tuloxonov lists dialect features of the original which his new text follows, yet without recourse to diacritics. (B) In 1906, Jamcarano recorded the üligersin Mansut's variant of Abai Geser Xubun in Mongol script. In a sense this was from a public performance, since Jamcarano was acutely aware of the disruptive effect on the bard of dictation-to-hand, so that (as we saw) he insisted that others be present. At Jamcarano's side was V. A. Mixajlov, a master of Mansut's dialect, to explain baffling words or to interpret what Jamcarano had taken down, in terms of the Exirit-Bulagat dialect. Later, bypassing the newly introduced Roman alphabet, Jamcarano transcribed his old record in the Mongol script into the Russian Academic Script for an edition of this epic by Rudnev in 1930. In 1961, M. R Xomonov republished the 1930 edition of the epic in a version of

Textology and epic texts from Siberia and beyond

137

the National Burjat Cyrillic sensitized to the peculiarities of the Exirit-Bulagat dialect, which had been caught and / o r reconstructed in Jamcarano's use of the Russian Academic Script, the drawbacks of which Xomonov specifies as "difficulty in reading and fuss" (and doubtless expense) with diacritics... (Xomonov 1961: 229f.) These assessments should, however, be set beside an assessment of thirty years later by Tuloxonov in the textological section of his edition of Alamji-mirgen (1991: 283): Thanks to special training and acquired routine, C. Jamcarano achieved as nearly as ever possible a precise and complete fixation of the narrative text of an iiliger. It is no accident in this connection that he writes: "...a sort of shorthand record was obtained which had to be checked there and then and any obscurities rectified lest one forget them."

(C) In 1924 in Kijing, the writer X. Namsaraev took down the iiliger Xaratuur xaan from 60-year-old Mensxen Zajatuev, apparently in vertical Mongol script, at a time when a Roman script was gaining favour for the literary language. The current literary language was based on the SongoloSartul' dialect. For his edition of 1933, however, Namsaraev transcribed his Mongol field-record into Roman script. In this Romanized edition, nevertheless, one and the same word might appear in different guises. In 1938, one year before Roman script ceded to Cyrillic in Burjatia, Namsaraev republished Xaratuur xaan in a "re-processed" (pererabotan) form, with a phonetic structure based on the soon-to-be standard Xori dialect and with the addition of non-textual words for greater precision (Xomonov 1964:107).

Even (Lamut) In 1937-38, the schoolmaster N. R Tkacik (1905-44), son of an immigrant Ukrainian family, and an idealist, recorded by hand the epics Delg3ni and Civd3O3l and the greater part of Geakcaval from the 98-year-old narrator (nimkalan) N. G. Mokrousa (b. 1830), a speaker of the Oxotsk dialect. The aged singer having died before reaching the end of Geakcaval, Tkacik enlisted the help of the former's foster-son A. M. Gromov, himself a nimkalan and well-versed in the foster-father's repertoire. In the process of preparing the three epics for the press, Tkacik replaced dialect peculiarities with elements of the new literary norm. These authorially revised texts rest unpublished in the archives. But in 1969, Tkacik's widow, Ε. V. Bulatova, placed the original field-records (phonetic system not stated) at the disposal of the celebrated Tungusologists G. M. Vasilevic, V. I. Cincius and V. D. Lebedev. Tkacik's touching devotion to the Even proved not to be fully matched by his use of their tongue: errors in the forms of Even words and violations of stichic

138

Arthur Τ. Hatto

structure were found. During expeditions to the Even of the Oxotsk Region in 1969-73, Lebedev achieved a substantial textological work on the epics. With the cooperation of Gromov, he clarified obscure words and phrases and restored the jingling "leitmotifs", so characteristic of Tungusic epic, which indicate which person is uttering (zapevy). Lebedev's laborious work on these texts and his own field-materials on tape (magnitofon, no date) gave him a foundation for restoring shattered stichic direct speech. Lebedev did not live to see his work in print (1986) but before he died it was his endeavour to produce "national texts", seemingly in the national Even language with some broader phonetic and morphological dialectisms permitted but with narrower dialectisms committed to the Notes. A national language for the tiny Even people had begun to emerge in the early 1930s, based not on the speech of Oxotsk but on another Eastern dialect, that of Ol'sk (Lebedev 1986:3-8,296ff.; Lebedeva 1981:13; Lebedev 1982:14).

Evenk (Tungus) From the angle of later, longer recordings of nimngakan or Evenk heroic narrative the specimens published by G. M. Vasilevic in 1966 with translation and commentary have been deemed fragments. An exception is her No. 2 Ködäkcen, which was recorded on tape in Leningrad before 1966. The singer was Semen Savin, born in 1894 of an Evenk mother and a Jakut father (Vasilevic 1966:341; Myreeva 1990:75; Dobrovol'skij, in Vasilevic 1966:385). Dobrovol'skij hails this tape-recording as enabling him for the first time to adduce, from tape, examples in musical notation on the page, of the famous Evenk musico-verbal "leitmotifs" initiating sung monologues or arias. A further advance was marked in 1990, when A. N. Myreeva's musicological collaborators appended a disc derived from taped leitmotifs from Irkismond' in support of their comparative analysis of this phenomenon (Myreeva 1990: 89-124). Irkismond', taped from the great singer N. G. Trofimov (1915-71) is a much longer epic than Ködäkcen. Some say the recording was made in 1964, others in 1966. In either case, the sound recording was preceded by A. N. Myreeva's recording of Irkismond' by hand to dictation from the same singer, who himself suggested from past experience with the Jakut οίοηχο that performance before a tape-recorder would be preferable. Myreeva had to quit the field before her written record of Irkismond' was completed, and it was characteristic of Trofimov that he sat down and finished the epic on paper for her with all the long vowels indicated by length-marks. Myreeva proudly records that "in a year's time" Irkismond' was taken down on tape from Trofimov, but whether this was in its entirety and by whom is not stated (Romanova and Myreeva 1971: 97). It would have been of even greater interest to us to know whether

Textology and epic texts from Siberia and beyond

139

A. N. Myreeva, to whom we owe so much, based her edition Irkismond' esöniq on her dictated text or on the tape or on both eclectically, and according to which rules, a matter on which she is silent. In the Even section we have seen how N. P. Tkacik employed the fosterson and pupil of an aged bard to finish the epic the old man had died over. This may be called "extrapolation" from within the same tradition. In the musicological study in Myreeva 1990, the two authors, M. Ayzanstadt and Ju. I. Sejkin, actually use the word "extrapolation" when applying knowledge gleaned from taped or otherwise musically illustrated epics to the epics in hand, Sodani and ΌένέΙεέη (see Myreeva 1990: 89). Here the extrapolation remains extra-textual, yet it sits at the forefront of commentary and might yet intrude into published texts. Despite the existence of a tape-recording of Irkismond', A. N. Myreeva's edition of it is set in prose (1971). In her edition of the two other long epics from Trofimov, Oevelcen (3,143 vv.) and Sodani (2,632 vv.), the texts are given in stichic form, although no tape-recordings were said to be available. ΌένέΙύέη was taken down from Trofimov in 1960 with Myreeva using a phonetic transcription as befits an authority on Evenk linguistics. Sodani was written down by Trofimov himself under canvass and by candlelight when each hard day's work herding the reindeer was done. The great bard was ailing, and a day came in 1971 when he had to be flown to hospital. He died in the operating-chair, but in his hold-all his record of Sodani was found in an exercise-book addressed to Madame Myreeva. What had kept him going was his sense of responsibility for a tradition to which very few had access. Elders who had dropped in in earlier days to listen to his dictations asked: "When are you going to stop writing?", and soon went out again. The young people said: "Translate it into Russian or Jakut, and publish it soon!" (Myreeva 1990: 371,380; Romanova and Myreeva 1971: 97). Thus each of the great Evenk bard's three substantial epics was recorded by a different means: by hand in phonetics to dictation; by the bard's own hand from a combination of memory and re-creation at which we can only guess; and by tape, probably in camera, for had it been before an audience, A. N. Myreeva would surely have noted it. Will it one day be a task for textology to edit all-three epics of N. G. Trofimov to a unified textological convention? Jakut The various stages in the editing of Jakut epics (οίοηχο) can usefully be indicated by the names of the leading editors of their day, O. Böhtlingk (1851); Ε. Κ. Pekarskij (Obrazcy 1907-18); G. U. fergis (1947); I. V. Puxov (archival,

140

Arthur Τ. Hatto

1948-49); with Ν. V. Emel'janov combing the archives for well-planned thematic studies (1980,1983,1990). The first complete olonxo published in Jakut was a short variant of £r Soyotox "The Solitary Hero", not for its own sake but as a specimen to illustrate the monumental Jakut grammar and dictionary of the great Sanskritist Ο. Böhtlingk (1851), written to keep a promise to his friend A. T. von Middendorf that he would see to it that the latter's Jakut materials were published. The original text of tr Soyotox was written down "from memory" by A. Ju. Uvarovskij, the bilingual son of a Cossack, and a Jakut by heart, when Böhtlingk to his delight found and took him as tutor for Jakut in St. Petersburg. On the basis of οίοηχο fragments among Middendorfs materials, Böhtlingk judged Uvarovskij's "Märchen" (as Böhtlingk called it) altogether in the Jakut style (1851: xlix-1), while over a century later οίοηχο specialists, noting its artificial/artistic source and unusual brevity, accept it nevertheless as reflecting genuinely traditional οίοηχο style and content (e. g. Puxov 1962: 9; Ergis 1974: 30); and, as it chances, the present writer quoted Uvarovskij's tr Soyotox in support of his contention that the colourful stock figure of the Crone Simexsin was derived in part from the Guardian Spirit of the Homestead, an explicit trait lost in later οίοηχο (Hatto 1994: 8ff.). Uvarovskij had developed his own inaccurate use of Cyrillic to write in Jakut, but Böhtlingk took him patiently through the texts of £r Soyotox and Uvarovskij's autobiographical "Memories" in terms of his own Cyrillic-based system, never obtruding his private judgements but always leaving Uvarovskij to supply his own responses until they were acceptable. Thus this specimen of textualization was the product of shrewd, self-effacing investigation and deep trust between a great scholar and an informant who became a friend. Böhtlingk's phonological system was made up of 29 symbols - 8 vowels and 21 consonants. 10 of his symbols do not occur in the Cyrillic of c. 1850, though 6 of them are adaptations of Cyrillic characters and only 3 are from Roman (Böhtlingk 1851:97, § 1). The profundity of his analysis is underlined by the fact that Pekarskij in his glorious Slovar' in three mighty volumes adopts Böhtlingk's system with only one minor deviation, though not entirely the shape of his symbols (Obrazcy 1,1907: 3, 7ff.). Pekarskij came to his "irrevocable decision" to apply the Böhtlingk system to his own lexicon and his equally bulky editions of οίοηχο just in time, and he imposed it on texts collected by others too (Obrazcy 1907-18). The instance of the bard Timofeev's Kulun Kullustuur, taken down by V. N. Vasil'ev in 1905 is highly illustrative of the process of textualization accomplished by the great editor, his staff and his fieldworkers. The facts are given at the head of the edition under the titles "From the Collector" (Vasil'ev) and "From the Editorial Office" (Pekarskij). Vasil'ev writes that he was "commanded" into the field by the Museum of Anthropology and Ethnography under the auspices of the Imperial Acad-

Textology and epic texts from Siberia and beyond

141

emy of Sciences (where Pekarskij already enjoyed considerable influence). Whether it was on his own initiative or the result of Pekarskij's insight, Vasil'ev's method in his fieldwork has an original touch such as appears all too rarely among the epic recordings of Siberia reviewed here: if other collectors employed it, they did not say so. As a preliminary to writing the text down to dictation, Vasil'ev had his bard perform his οίοηχο in full in its normal rendering and in the appropriate surroundings (obstanovka) in the presence of the habitues. This done, Vasil'ev could proceed to recording it by hand with some confidence that under the bard's dictation he would at least notice major omissions, alterations, or deviations. Otherwise, he did his best to reproduce the pronunciation of the bard, presumably in the "Academic Transcription". Pekarskij's response reveals something of a textological nightmare. He states that for the published edition he had not used Vasil'ev's field-record, but a copy of it made by P. N. Malygin, 1.1. Govorov and G. Nosov - but also by V. N. Vasil'ev! This last was on the latter's "return" (in 1916 one imagines "from military service"). All four assistants used the "Academic System" for their copies, yet with some deviations, which could not fail to complicate preparations for the press. Here it was not always possible to determine which of the four transcribers' deviations had been caused by the peculiarities of the Narrator's pronunciation or by the Four's personal views regarding these or other phenomena of Yakut phonetics. To make the editorial position crystal clear, Pekarskij appends as Note 1 to the text proper: "The first pages of the story [οίοηχο] were not transcribed by V. N. Vasil'ev himself but by various other persons, thereby is explained the lack of consistency in the transcription of many Jakut words indicated in the notes." Pekarskij seems to complain when he writes that Vasil'ev was trying to convey the Narrator's pronunciation, but that together with his fellow transcribers he for example always replaced intervocalic h of the original with the sign s. Then, nota bene, the great man turns to the difficulty caused by attempts to render the lisping speech of Tungus and other characters, not consistently adhered to by the Narrator and so baffling to the recorder. Here Pekarskij concluded that the attempt to render such sounds must be abandoned in favour of ordinary forms... (iDbrazcy III, 1916: i-v, 7.) This last in effect negative solution of a histrionic problem of textualization, more or less adequate to the date of publication in an edition without commentary, poses a challenge to users of modern equipment directly they sit down to their word-processors with their tapes. And this challenge is left to them. Pursuing the textualization of Timofeev's Kulun Kullustuur, we come to another problem. In 1985, the Timofeev-Vasil'ev Kulun Kullustuur was reedited, translated into Russian and furnished with a commentary by a very distinguished team: G. U. firgis transmitted Pekarskij's Jakut text with its

142

Arthur Τ. Hatto

apparatus criticus intact; Α. Α. Popov's translation was taken from the archives under the care of Ι. V. Puxov; the commentary came from Popov and was supplemented by Puxov and Ergis (Ergis et al. 1985). This work was long in the making, since Puxov refers to it in 1962 (Puxov 1962: 251), and Ergis had died in 1968. In his notice "From the Collector" already referred to, Vasil'ev had noted that to enliven the interest of his listeners the Narrator would introduce coarse descriptions of a pornographic nature. These were ready-made and adaptable to the situation in hand. They were usually omitted in the houses of the well-to-do and eminent as unseemly, but in the yurt of a poor man and even of a person of middle means the Narrator would resort to such descriptions for comic effect. As ethnographer, Vasil'ev considered that such sallies were part of the text and he encouraged his singers to perform according to their wont. I. V. Puxov and G. U. firgis were also aware of their public! The famous bilingual series in which they were proud to publish the Jakut Kulun Kullustuur was one in which each ethnos of the Soviet Union was eager to display its epic (real or imagined) at its best in friendly rivalry. So, despite the very high standard of scholarship otherwise brought to bear on this edition, the bawdy passages so characteristic of the Old Jakut had to go. No attempt is made to hush up these cuts: they are claimed as "naturalistic" in both the Preface and in the Commentary (Ergis et al. 1985: 6,601 n. 4). More important and on the positive side in this edition is Srgis's setting of the text as stichic, where Pekarskij adhered to his regular practice of prosesettings; and the Russian translation follows suit. Such work is best left to native ears like his when there is no audible support, and he had had the experience in 1947 of re-publishing the inspired amateur bard Orosin's N'urgun, recorded by Orosin himself in 1895 "from memory" and first published by Pekarskij (Obrazcy 1,1907). In their pithy Preface to the new edition of Kulun Kullustuur in 1985,1. V. Puxov and the then already deceased G. U. fergis discuss Vasil'ev's transcription. Here they say he took down this οίοηχο "in the so-called Academic Transcription based on Russian", since the Jakuts had no writing then. In it, sounds not covered by Russian letters were rendered by letters from other alphabets, mainly Roman, or special signs were added to [Russian A. Τ. H.] letters. In the new edition of 1985, the late lamented G. U. Ergis had transposed the Vasil'ev-Pekarskij text into the current [Cyrillic-based A. T. H.] Jakut alphabet. "Therefore we omitted everything from Ε. K. Pekarskij's Preface ["From the Editorial Office" Α. Τ. H.] concerning the correction of V. N. Vasil'ev's errors in the Academic Transcription" (firgis et al 1985: 5). These two outstanding scholars, I. V. Puxov and G. U. Ergis, commend Pekarskij highly for his great textological work on Timofeev's Kulun Kullustuur. They preserve his every comment, changing only the numera-

Textology and epic texts from Siberia and beyond 143 tion to match their new stichic arrangement, whilst clearly marking any changes they make in his text. The Orosin-Pekarskij-Ergis N'urgun must be treated more briefly. In 1895, Pekarskij asked Orosin, the gifted, well-known connoisseur-performer of οίοηχο and man of independent means, to write down "his" N'urgun for him. This Orosin did "in Russian letters". Pekarskij then transposed this text into the "Academic Orthography". "I tried to preserve in the text all the orthographic peculiarities of the original (written in everyday Russian letters) in as far as they did not contradict the laws of Jakut phonetics" (Obrazcy I, 1907:1, n. 1). Orosin died in 1903, and Pekarskij's edition appeared in 1907. Since Orosin belonged to Pekarskij's circle it seems fair to assume that Pekarskij will have discussed every crux with Orosin, not only for the benefit of his οίοηχο but also for that of his lexicon. The designation "the Academic Transcription" occurs in the textology of other than Jakut traditions, where it probably refers to a system favoured by the Imperial Academy of Sciences. Here it refers to Böhtlingk's alphabet of 1851 (see above, and Ergis 1947: 365). Taking up the torch in his N'urgun edition of 1947, Ergis passed on to his readers the entire apparatus criticus of Pekarskij's edition lest any detail of Orosin's local speech be lost, even if embedded in an error (Ergis 1947: 365). As already mentioned, Ergis sets his text as stichic. New criteria have emerged, he says, which have to do with verse-structure, such as short pauses for breath, longer pauses for reflection... In the last analysis, for example, style determines whether what seems like two verses are in fact one or two... Well-versed in live performances, this Jakut scholar was glad to apply his insight to texts inherited from others. But he would obviously have been happier making his own texts from tapes of live performances. At the conference of Siberian folklorists in Ulan Ude in December 1959, Ergis was to represent the Jakut Institute of Folklore and speak of the very urgent need for action to record Jakut folklore. "It is time to apply recording by taperecorder and film to obtain maximal precision in the reproduction of all the circumstances of performance of and listening to oral productions" (Sidel'nikov et al. 1961:55). Erg is warns that after the great burst of recording (in 1935^49, Α. Τ. H.) no οίοηχο had been recorded in the 1950s. But, curiously, one gleans from A. N. Myreeva's account of the bilingual bard Trofimov that having been taped for Jakut epic (date vague!) he would like to be taped for Evenk (cf. supra). Though Myreeva gives a full reference as to where Trofimov's οΐοηχο-tape resides in the Jakut archives, Puxov's inventory does not name a single οίοηχο as having been recorded mechanically (1962: 24352), and the same is true of Ν. V. Emel'janov's list of 75 items (1980:295-300). What might have been the normal Siberian sequence in the creation of a National Jakut Alphabet, from Orthodox Ecclesiastical usage to gifted native

144 Arthur Τ. Hatto

writers seeking a mode of self-expression, was heavily influenced by the impact of the great scholars Böhtlingk (who achieved an all but perfect system) and Pekarskij (who, led by V. M. Ionov, perfected it [Böhtlingk 1851; Pekarskij Obrazcy 1,1907 etc.; Ubrjatova et al. 1987:21ff.; Okonesnikov 1982: 39ff.])· Looking back, it seems like light relief that after the Revolution, S. A. Novgorodov introduced a Roman-based alphabet of a phonetic nature with which one spelt democratically as one pleased, with some order introduced later by the admission of full-stops and capital letters! When, as elsewhere in 1939, a Cyrillic-based alphabet was imposed, the work for a broad transcription was of course already done. Despite the relatively large population scattered over a vast territory, the emergence of a unified National Orthography did not carve the people up, as for example among the Nenec (infra), since dialect variation was surprisingly weak. Thus the Literary Alphabet should exert minimal pressure on the rich harvest of οίοηχο after 1959 when they come to be textualized.

Nanaj (Gold) Nanaj epic material comes from both sides of the River Amur, the Chinese and the Russian. From the southern region we are fortunate to have a captivating epic of considerable length (some forty book-pages) entitled Moculin. Taken down in the Hezhen dialect (I am indebted to Dr. Jörg Becker for this information) during a three-month expedition by the Western-trained ethnographer Ling Chun-sheng in 1930, it comes to us re-translated into German by J. Hefter (1939) from Ling's Chinese translation (1934). The Hezhen field-record remained unpublished. Hefter's German text displays umbilical links with the original, however, in the form of Nanaj keywords in an Italic script apparently broadened by Hefter to accord with the then more recent scholarly norms, such Nanaj keywords, according to Hefter (1939:108) being embedded in Ling's Chinese translation in Roman script. Like the structure of some other riverine epics, that of Moculin is strong. Having lost his people from plague, Prince Moculin moves from town to town up the River Sungari, winning them, their rulers and rulers' daughters. His end gained, he returns to his capital downriver with his new people in thousands of ships. A fascinating aspect of this epic is that its Hero is a seven or protector-spirit (outwardly a Tree-stem) whose dazzling beauty is revealed at first only to his brides at their nuptials. In the northern region, the term ningmä embraces narratives of varying character and length and includes some specimens that we would style "heroic epics". In the impeccably presented Materialy po nanajskomu jazyku i fol'kloru (1986), V. A. Avrorin and, after his death, E. P. Lebedeva list 25 ningmä

Textology and epic texts from Siberia and beyond

145

from the Najxin Nanaj, including several of epic stature. They were taken down by V. A. Avrorin, a distinguished Tungusologist and linguist, during an expedition in 1948. Of the published form of his field-records (Materialy 1986), E. P. Lebedeva states that the rendering of the sounds is the same as in Avrorin's Grammatika nanajskogo jazyka of 1959, with the sole exception of the symbol 3. The script used by Avrorin and Lebedeva is an adaptation of Cyrillic with a tolerable number of diacritics. Rather than "work up" Avrorin's field-record, E. P. Lebedeva preferred to leave it as it was. This probably means that she did not wish to accommodate the field-record to the emergent Nanaj literary language, for which Avrorin himself had prepared the way in 1932 with the first Nanaj ABC-book. The old Tungusic epic style of recitative laced with arias introduced by musico-verbal "leitmotifs" (cf. supra) was in decline when Avrorin came to the field, which probably accounts for the setting of his narrative texts as prose.

Nenec (Samojed) Nenec heroic songs rising to longish heroic lays can be approached through four main collections: 1) that of Toivo Lehtisalo (1940), publishing specimens taken down by M. A. Castren during his expeditions in 1842- 43 and 184549; followed by 2) Lehtisalo's own recordings of 1911-12 and 1914, published in 1947; 3) that of Ζ. N. Kuprijanova (1965) from expeditions by her (in the main) and others (notably A. N. Scerbakova) in 1939, 1946, 1949, 1959; 4) that of Ν. M. Terescenko (1990), mostly from 1934 but also from 1956 and 1958, among them a recording "from memory" by A. P. Pyrerka, the first Nenec scholar and writer. Strangely, although Terescenko names Kuprijanova's essay on Nenec genres in her Bibliography she fails to list there what others regard as Kuprijanova's monumental work of 1965 (Editor-in-chief E. M. Meletinskij), yet refers to it in footnotes without mentioning her name. In 1940, in celebration of the 300th Anniversary of the Foundation of the University of Helsinki, on behalf of the Finno-Ugrian Society, Lehtisalo edited and published a volume of Nenec heroic and other songs whose texts had been taken down by Castren (1813-52) on his expeditions to the Samojeds and preserved in the Library of the University of Helsinki, where Castren had held a Chair. Lehtisalo regarded his commission from the Finno-Ugrian Society to edit specimens of Castren's recordings of Samojed oral poetry as an act of piety, so that his aim was to preserve as much of Castren's original texts as ever possible. Lehtisalo therefore operated in the critical manner as would an editor of a unique post-oral epic in manuscript, that is, philologically, with the advantage of being at only one remove from the original. This fact is

146 Arthur Τ. Hatto brought home to one when Lehtisalo notes that he has preserved in his apparatus forms subsequently rejected and deleted by Castren, sometimes twice over, and that by attending to variations in the colour of Castren's ink he could distinguish between corrections made in the field and those made later, presumably during exchanges with native speakers. On the few occasions when Castren gives a second or third version of a hero-song in other dialects, where one surmises that linguistic not ethnopoetic considerations were uppermost in Castren's thoughts, Lehtisalo publishes these too, even though he deduced that Castren and his native informants had themselves translated them into other dialects. In other such cases he allows for the possibility that a [true] variant of a song represented by the original might have been current among speakers of other dialects. Lehtisalo sets Castren's hero-songs as stichic without indicating whether he is following Castren in this particular. Later, in 1947, editing his own fieldrecordings of Tundra and Forest Nenec oral material, Lehtisalo again sets the hero-songs as stichic without discussion, but, then, this time he has heard them performed to melodies of strange beauty that moved him deeply. Cursory inspection of the sound-patterns in Castren's texts shows that a stichic setting might have some slight but not systematic support in alliteration and assonance. It is interesting that Ζ. N. Kuprijanova and Ν. M. Terescenko agree that Nenec hero-songs are stichic, yet set their texts as prose. On the basis of his musicological analysis of some tapes in his Appendix to Kuprijanova's book of 1965, Β. M. Dobrovol'skij sets stichic texts firmly founded on melodic/prosodic structure, and moreover, he subjects his Russian stichic translations to the same musicological discipline with an eye to obtaining truly poetic renderings from within, adding a warning to the reader that his Russian texts are unfinished and designed merely to point the way (Kuprijanova 1965: 7 5 7 - 80). Dobrovol'skij was of the opinion that decipherment of the text from tape can only be carried out when analysis of the musical component has been accomplished, since that component integrates the text prosodically and hence semantically. Nevertheless, in a generally excellent contribution to a symposium in Hamburg in 1965, W. Veenker notes that Ζ. N. Kuprijanova offers no evidence on verse-structure and metrics, adding that, in the texts he had studied he had failed to notice the forms of rhyme and alliteration and judged that the JurakNenec hero-songs had probably best be characterized as in rhythmic prose, after which denial he refers the reader to Dobrovol'skij's aforementioned Appendix (Veenker 1968: 51). For textologists it follows in any event that the recording of Nenec sung heroic narrative by tape is a minimal requirement. As has already been noted, Dobrovol'skij addressed the Sibirologists foregathered in Ulan Ude in December 1959 and encouraged them to use tape for recording oral texts. Z.

Textology and epic texts from Siberia and beyond

147

N. Kuprijanova is not named as present there. She was or had been among the Nenec in the field, and had a tape-recorder at her disposal (1965: 756), so that it is likely that it was she who supplied Β. M. Dobrovol'skij with the taped exordia which he was to examine musicologically in his Appendix to her book of 1965. In the absence of evidence to the contrary, however, it seems that Kuprijanova based her own texts of 1959 as published in 1965 entirely on writing to the bard's dictation, not on tape, for her texts cohere only very broadly with Dobrovol'skij's from tape. During the 19th century, the Orthodox Church had used a transcription for rendering the Scriptures in Nenec, combining Cyrillic, Church Slavonic and Greek characters, but it was eventually withdrawn as unsuccessful. After the Revolution an alphabet based on Roman script came into use, and it was presumably in this script that the first published Schoolbook in Nenec was couched in 1932. In 1937, however, a new Cyrillic-based script was promulgated for use with the Literary Language. The Nenec Literary Language was based on the Central Tundra, i.e. Bol'sezemlja dialect, and has probably exerted an influence on the textualization of Nenec hero-songs. For most of the specimens recorded by M. A. Kuprijanova and published by herself were in the (Western) Malozemlja dialect, while those taken down by others were in the (Central) Bol'sezemlja dialect. Most of the heroic narratives published by Terescenko on the other hand were taken down in the (Eastern Tundra) Jamal dialect, with a minority in the (Central) Bol'sezemlja dialect. Both Kupriyanova and Terescenko claim that they publish in a "practical transcription". Kuprijanova adds: "We strove to preserve the fundamental phonetic and morphological peculiarities of the performer's speech" (1965: 5). For her part, Terescenko continues: ...practical transcription constructed on the basis of the Russian system of writing ([as] it applies to the Nenec Literary Language) [and] it has as its aim to render the texts accessible to the Nenec intelligentsia. In them, fundamentally, are preserved the most distinctive peculiarities of the Jamal and Bol'sezemlja dialects. (Terescenko 1990: 3.)

Here it is perhaps relevant to record that at least one of Terescenko's narratives, No. 2, taken down in 1934 from the Party activist Ν. I. Pugurcin, is a Progressive-Patriotic pastiche. With her prominent role in the movement for a Nenec National Language as a vehicle for an emergent Nenec Literature, Terescenko seems, understandably, to be wishing to have the best of both worlds, that of the past and of the future, whereas Kuprijanova could well afford to come from her university with an academic education and outlook, together with an academic alphabet, and finally in 1959 with a tape-recorder. In a scan of their opening pages, Kuprijanova's texts make a very clear Cyrillic impression, with the addition of no more than three extra-Cyrillic

148 Arthur Τ. Hatto symbols, implying a high degree of phonemic analysis, while Terescenko's show four non-Cyrillic usages. There thus seems little at this level of textualization to justify Terescenko's exclusion of Kuprijanova's highly commended volume of 1965 from her Bibliography. On the other hand it is clear that the emergent Nenec Literary Language has exerted some pull on the presentation of Terescenko's hero-narratives. One wonders what the Forest Nenec think of all this, having been entirely left out of the picture owing to the remoteness of their dialects. During his expedition of 1911-12, T. Lehtisalo made recordings on phonograph of Nenec songs of various genres. With his main interest in linguistics, he omitted to document their background systematically (Niemi 1995: 172). No identity has been found between the exordia of the hero-songs he published in 1947 and those deciphered from his cylinders (Niemi, personal communication), compare the similar situation as between Dobrovol'skij's musical notations from tape and Kuprijanova's texts from dictation (supra). In collaboration with the Nenec journalist Ms. Anastasia Lapsui, J. Niemi recognizes in Lehtisalo's cylinder (mostly of poor quality) "epics" [here "herosongs"] among other genres, in particular sjudbabc [3rd-personal hero-songs]; jarabc [lst-personal hero-songs, in latter years moving to more general, topical themes]; and hinabc [a narrative genre deserving closer study and definition] (see Niemi's detailed and complex, yet clear, catalogue [1995:134-37]). On p. 130 of his current article, J. Niemi signals prospects of acquiring insights into cylinders, housed in the Institute of Ethnography in St. Petersburg. It is much to be hoped that he and Ms. Lapsui will, among other productions, be able to place a corpus of Nenec hero-songs on record in a modern holistic edition, naturally with renderings of the verbal text into a language of international scholarship. So-called "filler-" or "nonsense-syllables" proliferate in Nenec hero and other songs to compensate, so it is said, for a lack of Nenec metre or metrics, as a means of marrying "basic texts" to their melodies. They pose an acute challenge to the heroic breed of conscientious textualizers. Peter Hajdü, for example, insists that the musically borne "text" of actual performance must be published as well as the unauthentic "basic text" (1978: 356). Toivo Lehtisalo's work with the phonograph in the field shows him fully aware of the need to avail oneself of the best contemporary means of recording oral utterance. It therefore comes as no surprise that as early as 1947 he writes (I translate): "It would be desirable to collect the folk-poetry of those peoples [the "Jenissei- and Tavgy-Samojed", the "Ostjak-Samojed" and "Jenissei-Ostjak"] at the first opportunity, and especially now that it can be accomplished speedily and relatively easily by tape-recording (Bandaufnahme)" (Lehtisalo 1947: xii).

Textology and epic texts from Siberia and beyond 149 Ob-Ugrian: Khanty (Ostjak) - Mansi (Vogul) (only Khanty sampled) It is a cause for sorrow that the Ob-Ugrian tradition of heroic epics is at once the most arresting in the whole of Siberia and one of the most problematic as regards textualization. In pursuit of their lofty, inexorable code of manly honour, warriors, say, on a wooing expedition, find themselves on a fateful course. Arriving on alien ground they display a magnificent indifference to danger. It is as though they have known death all their lives. When battle erupts, a bloody mist, täran, bedews them from On High. One of the heroes is a god, the narrative from his lips is the epic, which, on his mounting to his high lodge to receive sacrifice, becomes his myth. So far as the Khanty are concerned, little of their epic traditions would be known but for A. Reguly's (1819-58) and J. Päpay's (1873-1931) ardent preoccupation with their texts. Reguly was inspired to go on an expedition to the north-eastern Khanty in 1844-45. There he took down 12 complete heroic epic songs. In 1898-99, in order to gain the fullest possible background to Reguly's texts, Päpay went to the field to work with speakers of neighbouring dialects, for the texts were bare of all translation and commentary and in a dialect that had since died out. In 1905, Päpay published 4 of Reguly's epical texts, both in Reguly's original transcription and in his own Päpayan transcription, together with a Hungarian and a German translation. A century after Reguly's first recordings, Miklos Zsirai published 2 of Reguly's epics, with Reguly's text, Päpay's phonetic transcription, Päpay's Hungarian translation and a German re-translation (1944). Seven years later, Zsirai published 4 more of Reguly's epic texts, the first 2 of which gave Reguly's original followed by Päpay's transcription and Hungarian translation, then a German re-translation by another scholar, whereas the second pair gave Reguly's texts but lacked Päpay's phonetic transcription yet nevertheless had Päpay's Hungarian translation, followed by a German re-translation (1951). After Zsirai's death, D. R. Fokos came out in 1963 and 1965 with the 2 remaining epics recorded by Reguly. Of the first of these, Fokos gives the whole of Reguly's text and prints Päpay's fragmentary phonetic transcription of vv. 1-657, adding his own transcription of vv. 658-1346 (end) based directly on Reguly's text, after which he prints Päpay's complete Hungarian translation. For the whole of the second epic in this edition - Reguly's 12th text - Fokos again edits Reguly direct, prints Päpay's Hungarian translation up to v. 1775, where it breaks off, and completes it with his own Hungarian translation, ending at v. 2437. This is followed by a German re-translation of Fokos's own. On dealing directly with Reguly's texts in epics Nos. 11 and 12 with less and less help from Päpay's legacy, Fokos found ever greater need to com-

150 Arthur Τ. Hatto

mune with himself and fellow specialists in a commentary which eventually ran to 120 packed pages offering many improved readings to the editions which preceded his own. Weighing up whether he should postpone publication until certain Ob-Ugrian lexicons had appeared or publish at once so that the lexicographers could profit from his work, he, a scholar in his late seventies, opted for the latter. The piety and resolution of Reguly's successive editors, from Päpay to Fokos, evokes an image of a team of heroic runners taking up the torch as their predecessors were struck down, and indeed Fokos himself characterizes them as heroes, too. It was with humility that he strove to improve on their endeavours, as he identified the causes of their shortcomings. For example, he cited the Khanty bards as custodians of ancient songs, some of the words or expressions of which they no longer understood. The informants whom Reguly and in a later generation Päpay consulted knew even less than the specialized bard. Again, Reguly recorded in haste, left no translation or commentary, had no chance to check with the singers, and did not comprehend an appreciable proportion of his texts, for evidently he cut some words in twain and fused others together (Fokos 1965:268). Päpay had found good interpreters, but what could they do with archaisms in a dialect that had died out before their time? Whilst Päpay was on his expedition to the Northern Khanty seeking to trace the dialect in which Reguly's records were couched, he took down 9 complete epics and 12 fragments in other North Khanty dialects. It was from these that he worked his way into Reguly's texts. Four of Päpay's own recordings were edited in 1972 by D. Fokos's pupil I. Erdelyi. Taking up the torch, Erdelyi commends Päpay highly for his decipherment of Reguly's texts but warns that there is much to criticize in his textualization of Reguly's records. He took exception to Päpay's going back home and in the seclusion of his study overloading his field-text with such a plenitude of diacritics that he fell foul of "phonetic extremists", responded to their criticisms, and yet failed to achieve a proper text. Erdelyi judged that, as a result, a phonemic transcription was barely possible and that since the main purpose of the Päpay Library was to honour the scholar's memory, he had best print the Khanty texts as Päpay himself had prepared them (Päpay and Erdelyi 1972: 9). This account obviously implies that Päpay's field-record of 1898-99 did not survive, which in turn raises a problem for fieldworkers. It seems very odd that a scholar who valued every syllable in his predecessor's record should deny his successors access to his own - or were they simply mislaid or lost? From outside the heroic Hungarian tradition of Ob-Ugrian scholarship, W. Steinitz makes the following observations in an Excursus "On the transcription of Ostjak in A. Reguly's recordings" (Steinitz 1976:67ff.). With such

Textology and epic texts from Siberia and beyond

151

materials as those in Reguly's legacy of texts, Steinitz deemed it axiomatic that one must first establish the sound-values of the transcription used in them, a transcription understandably old-fashioned and less than perfect, and relate this transcription to that of the modern, phonetically trained researcher into Ostjak, as K. F. Karjalainen has done for Castren's transcription of Ostjak, but which Päpay failed to do for Reguly's after a first tentative step in that direction. Päpay was unable to cope with the difficulties of Ostjak phonetics. He did not understand the phonetic peculiarities of the Ostjak dialect represented in Reguly's recordings. (Steinitz 1976: 62-64.) Steinitz then reconstructs the dialect situation in the relevant area (Steinitz 1976:64), and follows with over nine pages of detailed analysis of Reguly's phonetic usage. On pp. 63f., in a footnote in smaller print (n. 4), Steinitz nevertheless asserts that his strictures in no way diminish Päpay's merit for the decipherment of Reguly's recordings, or his, Päpay's, own recordings in the field despite their defects they constitute a quite outstanding, epoch-making achievement! This bombshell went off twenty years ago. As far as this writer could ascertain, no one seems to be in a hurry to apply Steinitz's hard-won insights to even one of Reguly's Ostjak heroic epics. Let us hope, however, that someone is working away on refining these precious texts, even if some or all of us may not live to taste the fruits. On an edition of a mid-Nineteenth-century Kirghiz epic from Arabic script In 1977 I published a text of the Kirghiz heroic epic The Memorial Feast for Kökötöy-khan (Kökötöydün asi) - from a photocopy of the once "lost" MS Leningrad Oriental Institute of the Academy of Sciences of the U.S.S.R., Oriental Archive. Razrjad II, opis' No. 4, delo No. 36 in Arabic script, given me by the finder, the late lamented Professor Margulan Alkey, a distinguished Kazakh archaeologist (Hatto 1971, 1977: vii). My edition consists of circa 3,251 vv. ("circa" to allow for improved verse-divisions by later editors), a translation into English prose facing the Kirghiz text, a commentary of some 150 pages, and 11 appendices. I gather from my Preface of so long ago that the motive which inspired me to furnish this edition was to bring to the notice of comparativists a major work of a major tradition of heroic epic poetry, and indeed the first recorded text of that tradition securely dated at least to the Spring/Summer of 1856 by the diary of its bard's patron of the moment, Sultan Cokan tingisovic Valixanov (1835-65). A. Margulan's homing on the MS in the 1960s came as a release, since Valixanov's Russian prose rendering of vv. 1 - about 890 had haunted students of epic since its publication by N. I. Veselovskij in 1904. A. Margulan came out with a Kazakh paraphrase of vv. 1 - 885 in 1965 and a Kazakh paraphrase of the whole in 1973. His facsimile

152

Arthur Τ. Hatto

edition of the MS in reduced size in 1971 proved unusable for finer points of transcription, e.g. Arabic diacritics, but for me personally this was redeemed by the magnanimous gift of his own fine photocopy. By trade no Orientalist, but a scholar in the Medieval German and wider Germanic field, and a comparativist, with six wartime years in a "room" of the Foreign Office, with some acquaintance with Kirghiz epic diction and a copy of Κ. K. Judaxin's magnificent Kirghiz-Russian Lexicon of 1965, together with W. Radloff's edition of Kirghiz epics of 1862 and 1869 in text and German translation (1885, re-edited by me, 1990), I attacked the Arabic transcription of Kökötöy diin asi (henceforth Kökötöy).

Up to 1977 there had been no appraisal of the Kökötöy MS by a professional paleographer. Its bold hand had at times been corrected or overwritten by a smaller, exceptionally fine and neat hand which I took to be Valixanov's. I further deduced that the Kökötöy MS could not be the original field-record, but instead a fair copy, and gave my reasons in Appendix 1 "Demonstration that the manuscript was copied fair from a Field Record". This contention, based for example on philological minutiae such as the classic wandering of the scribe's eye from like to like, and the presence or absence of Arabic diacritics, was clinched by the discovery that the assumption of an interchange of two leaves, or of the reversal from recto to verso of one leaf, of the original field-record in the course of copying Bok-murun's famous Itinerary, would replace a wildly absurd journey with a realistic one, and as later study revealed, a highly significant because ecologically determined nomadizing. History has dictated that various Turkic tongues have had to be recorded in inappropriate transcriptions imposed from above, of which Arabic is one of the least appropriate. How deeply moving is the epic moment when Atatürk is shown in full morning dress, his topper on the table and a chalk and pointer in his hands before the blackboard, as he instructs his highest officials in the writing of Turkish in Roman script! What a liberation from the tyranny of Arabic script, which had been stretched from Egypt to Singkiang to render varying Turkic speech! As it moves from tongue to tongue its conventions, naturally, change. From the sporadic use of the Arabic symbol for s for that of s in my MS, I soon sensed the proximity of the Kazakh frontier, even when allowing for Valixanov's probably Kazakh scribe. And since Radloff's Kirghiz epic texts of the same period and same general locality display similar phenomena (see Hatto 1982:8ff.), and since Kazakh, Karakalpak and other more western traditions agree with Kirghiz in placing their heroic epics in a Nogay quasi-Heroic Age, I formed the opinion that the possibility of a post-Nogay supra-dialectal epic diction - or a tendency thereto - should be investigated for older times and not summarily dismissed today on nationalistic grounds. The Arabic script, rich in consonant-symbols, possesses fundamentally no more than three vowel-indicators, whereas Turkic tongues are richer in

Textology and epic texts from Siberia and beyond

153

vowels, with two parallel series determining vowel-harmony. The possibilities of ambiguity of Turkic in Arabic script, are however reduced, though not eliminated, by conventions which vary empirically from region to region. Fortunately for me, a further limiting factor, and a strong one, was the epic diction in which my text was couched, compare the special operational jargons in warfare that ease the cryptographer's task. With the speech-mode of the bard unknowable or at least unknown, I chose a Roman transcription with a few diacritics corresponding very closely to Judaxin's Standard Kirghiz Cyrillic, with the exception of the politically inspired, but phonetically silly use of internal Λ [ja] and e [jo], which must have irked Judaxin for all his medals. With a post-Nogay bardic diction tinged with an unassessed Kazakh component and filtered through a presumably Kazakh scribe who recorded in Arabic script using a post-Cayatay convention, all that can be hoped for in a Kökötöy-edition is a plausible narrative text. My philological task done, I sat back and got a great epic thrill as I perused it. No textologist - not I or another - will ever succeed in taking from me Ancient Kosoy's aristeia in the Wrestling Match with Joloy (vv. 1628-50), or the scene in which the Amazonian Lady Oroηgu, stripping with an athlete's innocence to untie a prize Camel with her teeth, silences Manas's ribaldry with the assertion of her superiority as Matriarch and Ancestress of Heroes (vv. 2136-86). Lt. Col. Xomentovskij's picked reconnaissance party, of which Valixanov was a member, will scarcely have constituted a traditional audience for a bard being taken down to dictation, but in any case what Valixanov saw is lost forever. It is our luck, however, that we can assert that the text of Kökötöy enshrines the "mental text" of Kökötöydün asi one hundred per cent, for our text runs exactly parallel with Radloff's Bok-murun, recorded from another North Kirghiz bard only six years later. Analysis showed that the two epics had a deep shared structure. On this I wrote in 1977: ...two bards improvised not only on the same major theme [the Memorial Feast and Games]... but also on the same ten constituents of the theme... the two bards also tended to select the same minor motifs at the same place for bravura treatment. .. It was as though the bards held switches in their hands whose buds they could charm into leaves, side-shoots, blossoms or whole sprays, or let sleep, at will. (Hatto 1977: 98.)

This raises the problem "Why are epics so hard to destroy?" I have shown, for example, in the Ob-Ugrian section how precarious the textology of A. Reguly's epics is after 150 years of minutest philological analysis, and why. All the same, something so very powerful and exciting comes

154 Arthur Τ. Hatto through that were I twenty years younger, I would be tempted to learn Hungarian and then Ob-Ugrian to get closer to them - there is something there of unique interest to students of heroic epic poetry regarding not only content, i.e. mainly action and ethos, but also strong echoes of elaborately constructed names, epithets and other formulae. In his contribution to a symposium in Bonn in 1996, Lauri Honko threw a searchlight on the shortcomings of W. Radioff's recordings of mid-19th-century Kirghiz epic to dictation even more emphatically than Radloff does himself. There is no need to make your flesh creep by going into detail here, for Radloff's field-records to dictation are likely to be better than average, as befits the great pioneer of Turkic studies. I will therefore turn to Lauri Honko's second example, with which by a curious chance I am as familiar as with the former. (See now Hatto 1999. See further Honko 1998:185-88 [Kroeber], 18890 [Radloff].) The Mohave Epic This Epic from the outset represents textology of a whole grade lower in desirability than Radloff's and his peers'. Nevertheless, at the level of "mental text", I for myself rate it as at least the equal of anything Radloff took down. We owe it to the acumen of another great pioneer, A. L. Kroeber, of the American Southwest. The (as I understand) "horrific" nature of Kroeber's textification of Inyokutavere's "great dream" is best given in Kroeber's own words, but not before something is said on the bare circumstances of his recording. Kroeber was in the field in search not of heroic epics but of traditions on the origins of clans. It was only in the course of what became a six-day session that he became aware that he had a weighty epic in hand, and he had to return to Berkeley before Inyo-kutavere said he had finished. When Kroeber came back for "the end", his old Narrator, alas, had died. My analysis of the strong, convoluted structure of The Mohave Epic shows that the traditional epic had in fact ended and that at the conclusion of the last session with Kroeber was embarked on an excrescent latterday yarn ("The Half-Walapai Boy"), which I guess Inyo-kutavere was ready to spin out as long as Kroeber's high-grade cigarettes lasted. Even with a phonograph, Kroeber would not have been equipped to record a six-day session. As to the recording, Kroeber's bilingual guide and friend Jack Jones allowed the old man to proceed - for perhaps five to ten minutes - until the interpreter had as much as he could remember, then Englished it to me. With omission of repetitions, condensation of verbiage, and some abbreviating of words, I nearly kept up writing in longhand. If Jack got too far ahead, I signaled him to wait. On

Textology and epic texts from Siberia and beyond

155

the other hand, if names of n e w places or persons came too thickly, Jack w o u l d stop translating and ask Inyo-kutavere to repeat the names slowly, directly to me. (Kroeber 1951: 72b.)

To this must be added that with the insight of the blind, the aged Narrator learned to give ethnographic comment before it was requested, and that some of this has entered the text. Obviously a translation-paraphrase obtained in such circumstances, where only names and toponyms survive in the original tongue, falls very far short of the standards of the Turku Conference on the Textualization of Oral Epics, 1996. And yet the far more than merely "mental text" conveyed in Kroeber's record captivated me, drawing from me willy-nilly a year-and-a-half of the brief time that remains to me.

Horror and joy of epic texts - how are we to rhyme them? First we mourn the irreparable loss of the Mohave original. Though it was thought by the collectors to have been in songless prose, the power of the narrative would have tempted many a linguist into learning Mohave. Then we go for the deep structure, embedded in this one-sentence resumee, always the test of an heroic epic: Later, in Profane Time, while the Mohave were still hunter-gatherers, they left their fertile Valley for no reason, wandered into the desert WSW, dispersed, wandered again, in the main SE to where they learnt farming; then, after two abortive attempts by self-reliant Super heroes to expel Intruders from the Valley, were reunited by the Supreme Hero Hipahipa; who, after a reverse, led them to Reconquest at the cost of his life, a Reconquest the Mohave upheld against counterattack (Hatto 1999: 9).

Further analysis reveals that the character of the Supreme Hero Hipahipa is one of the most lofty, versatile and highly integrated in epic poetry, though conveyed in a supremely laconic style. Moreover, Hipahipa is accompanied by a Gentle Hero, MaGkwem-kwapäive, gentle despite his rank as a Superhero, who dies cradled on Hipahipa's knees from wounds received in battle - a character most rare in our genre. Despite the wretched textology of The Mohave Epic, we are speaking here in the way we speak of such epics as the Fula Super-hero Silamaka's, of high-grade textology complete with floppy disc (e.g. Seydou 1972), or of Päbüji, ditto, with the colour-plate of his par (Smith 1991). I was going to say we are speaking about The Mohave Epic as we speak about the Iliad, and indeed Kroeber spoke so, often luminously, only the text of the Iliad has been obelized, athetized and at times anathematized for well over two millennia!

156

Arthur Τ. Hatto

It must be clear by now that what makes heroic epic poems indestructible is not a high degree of textual probity but these constituents: vivid action moved by an heroic ethos, and borne by firm traditional structure articulated by epic moments, where structure and moments have mnemonic function, all deeply embedded in the "mental text". The tale of the textualization of epics in Siberia and beyond is thus a tale of woe and of comfort. At first, collectors lacked the technical means and then were too poor or slow to apply an optimal textology with modern equipment. On the other hand, by using dictation to hand in various styles with all their imperfections and frustrations, collectors have nevertheless secured from oblivion a huge corpus of often fascinating, powerful and colourful near-texts, which posterity will esteem the more, the more homogenized and bored it becomes. "Near-texts", of course because the second and third elements of the Epic Triangle, the Patron and the Audience (Hatto 1989:149ff.), have no tangible reflection, let alone presence in them. How fortunate that true Epic Occasions with attendant ritual still occur in a few areas, during which the expressions on listeners' faces may determine the significance of a bardic phrase, or frowns on the brows of aged connoisseurs condemn too bold an innovation by the bard! How highly commendable, how enviable, that you field-workers, among them Lauri Honko and his Finnish-Tulu Team, go there to set up your headquarters, perhaps joining residents on the same mission, fully equipped at last to bequeath to the future explicit evidence of what Epic Occasions were like, with appropriate textology and texts! May the means you employ survive even longer than papyrus, parchment, or high-grade paper! References Avrorin, V. A. 1986 Batchelor, John 1890 Böhtlingk, Otto 1851

Burcina, D. A. 1990 Dunn, C. J. 1980

Materialy po nanajskomu jazyku ifol'kloru. Leningrad. Specimens of Ainu folklore, read Dec. 4. 1889. Transactions of the Asiatic Society of Japan, vol. 18, part 1: [25]—86. Über die Sprache der Jakuten. St. Petersburg. Reprint: J. R. Krueger (ed.) 1964: Indiana University Publications, Uralic and Altaic Series, Vol. 35. The Hague. Geseriada zapadnyx burjat. Novosibirsk. Ainu. In: A. T. Hatto (ed.), Traditions of Heroic and Epic poetry, I, 328-44. London: The Modern Humanities Research Association.

Textology and epic texts from Siberia and beyond

157

Emel'janov, Ν. V. 1980 Sjuzety jakutskix οίοηχο. Moscow. 1983 Sjuzety rannix tipov jakutskix οίοηχο. Moscow. 1990 Sjuzety οίοηχο ο rodonacal'nikax plemeni. Moscow, firgis, Georgii Ustinovic 1947 D'ulurujar N'urgun bootur. Njurgun bootur stremitel'nyj. Yakutsk. 1974 Ocerki po jakutskomu fol'kloru. Moscow: Izd. Nauka. firgis, G. U., A. A. Popov and I. V. Puxov Kuruubaj xaannaax Kulun Kullustuur. Saxa olof]xoto. Stroptivyj Ku1985 lun Kullustuur. Moscow. [Bilingual edition.] Fokos, D. (ed.) Osztjäk (Chanti) Hös^nekek. Reguly-Könyvtdr, III, 1,2. Budapest. 1963,1965 Hajdü, Päter 1978 The Nenets Shaman Song and Its Text. In: V. Diöszegi and M. Hoppäl (eds.), Shamanism in Siberia, 355-72. Budapest: Akadämiai Kiadö. Hatto, Arthur Τ. The Kirgiz original of Kukotay found. Bulletin of the School· of Orien1971 tal and African Studies XXXIV: 379-86. 1977 The Memorial Feast for Kökötöy-khan (Kökötöydiin ast). A Kirghiz epic poem. Edited for the first time from a photocopy of the unique manuscript with translation and commentary. (London Oriental Series Vol. 33.) Oxford: Oxford University Press. The Marriage, Death and Return to Life of Manas. A Kirghiz Epic 1982 Poem of the Mid-Nineteenth Century. II. Turcica XIV: 7-38. Towards an Anatomy of Heroic Epic Poetry. In: J. B. Hainsworth 1989 (ed.), Traditions of Heroic and Epic Poetry II: Characteristics and Techniques, 145-306. General editor A. T. Hatto. London: The Modern Humanity Research Association. The Manas of Wilhelm Radloff. Re-edited, newly translated and 1990 with a commentary. Asiatische Forschungen, Band 110. Herausgegeben von Walther Heissig. Wiesbaden: Otto Harrassowitz. On some Siberian and other Copper-crones. Journal de la Sociiti 1994 Finno-Ougrienne 85: 71-105. The Mohave Heroic Epic of Inyo-Kutavere. Re-appraised and further in1999 terpreted on the basis of the edition of A. L. Kroeber and consultation of his field record. (FF Communications 269.) Helsinki: Academia Scientiarum Fennica. Hefter, J. Moculin - Ein Heldenepos der Golden. Sinica XIV: 108-50. 1939 Honko, Lauri 1996 Epics along the Silk Roads: Mental Text, Performance, and Written Codification. Oral Tradition 11/1:1-17. 1998 On the Difficulty of Documenting Oral Epics. In: Walther Heissig and Rüdiger Schott (eds.), Die heutige Bedeutung oraler Traditionen. Ihre Archivierung, Publikation und Index-Erschließung. - The PresentDay Importance of Oral Traditions. Their Preservation, Publication and

158 Arthur Τ. Hatto

Judaxin, Κ. K. 1965 Kroeber, A. L. 1951

Indexing. Abhandlungen der Nordrhein-Westfälischen Akademie der Wissenschaften. Band 102. Opladen/Wiesbaden: Westdeutscher Verlag. Kirgizsko-russkij slovar'. Moscow.

A Mohave historical epic. (University of California Anthropological Records 11/2: i-iv, 71-149.) Berkeley and Los Angeles: University of California Press. Kuprijanova, Ζ. N. 1965 tpiceskie pesni nencev. Moscow: Izd. Nauka. Lebedev, V. D. 1982 Oxotskij dialekt bvenskogo jazyka. Leningrad. 1986 tpos oxotskix tvenov. Jakutsk. Lebedeva, Ζ. K. 1981 Arxaiceskij bpos ένβηον. Novosibirsk. Lehtisalo, Toivo (ed.) 1940 Samojedische Volksdichtung. Gesammelt von Μ. Α. Castr£n. (Μέmoires de la Soci£t£ Finno-Ougrienne LXXXIII.) Helsinki: Soci£t£ Finno-Ougrienne. 1947 Juraksamojedische Volksdichtung. Gesammelt und herausgegeben von Toivo Lehtisalo. (Mämoires de la Soctete Finno-Ougrienne XC.) Helsinki: Sociiti Finno-Ougrienne. Maadaj kara 1973 = Surazakov et al. 1973. Margulan, A. 1965 äokan jazip alyan "Manas" Jin. Kazak S. S. R. Gilim Akademijasimri Xabarstst. Vestnik AN Kazaxskoj S. S. R., No, 8, Aug. Alma Ata. 1973 Köketajdtη ertegisi. "Manas"]irinir\ Sokan jaztp alynn böligi. Kirgizsadan Alkej Margulan. Alma Ata. Myreeva, Α. N. 1990 Nimi)dkar. tvenkijskie geroiceskie skazanija. Novosibirsk. Nevskij, Nikolai Aleksandrovic 1972 Ajnskijföl'klor. Sostavitel' L. L. Gromkovskaja. Moscow: Izd. Nauka. Niemi, Jarkko 1995 Transcriptions of Nenets songs in Α. Ο. Väisänens "Samojedische Melodien" re-evaluated. Journal de la Societi Finno-Ougrienne 86: 129-146; Appendix 1 with A. Lapsui, pp. 147-61; A. Lapsui Appendix 2, pp. 162-66. Obrazcy I—III. See Pekarskij, Ε. K. Okonesnikov, Ε. I. 1982 £. K. Pekarskij kak leksikograf. Novosibirsk. Päpay, Jözsef (coli, and ed.) 1905 Osztjäk nipköltisi gyüjtemeny. Sammlung ostjakischer Volksdichtungen. Heldengesänge mythologischen Inhalts, Götterbeschwörungsformeln und Bärenlieder. Auf Grund des Regulyschen Nachlasses und eigener Sammlungen. Dritte asiatische Forschungsreise des Grafen Eugen Zichy. Budapest and Leipzig.

Textology and epic texts from Siberia and beyond

159

Päpay, Jözsef (coll.) and Istvän Erunir pel

pa-lel, p'üqanup'ulsl büha, boyum kurit, ?ilawinum, k'ete-mek-sum wi-tam, k'ete-m p'oqtam, tiqele-m kü-ra.

10

Tu-nku-ra wi-tam iälim. *>unibuha put elinile yala· Calimleluherem *>uni yica·, ^elinile yala·, ts'ilqoq&n direwilbuha yala·. Buha,

büha,

büha pite*rum ^usat ilawi q'ayasuke, läqalsuke ''ilawi. 20 c Pi p o»qta buha, p'oqta bohema, p'oqta lelna. ^unir po-qta k'ete hima me«mto n hara·, memtomestopiton hara ken-la- pi p'o-qta. pominwine, wine tomoi, fine-

15 [II] [Discovered and Desired] [i] (A)

(B)

(C)

25

326 Dell Hymes wine,

30

wine· tomoi, k'ete-m tomoi.

Tomoi niqa-'a wine: "Heketun tomoi?" fipna«skuya. Wine honda wine tomoi, kelas tomoi k'ete-m. P'o-qta pi yomoi, yomoi, ''Heketun tomoi?" Pi Cälimleluheres ko m hima harä· dumia ?eleu pur harardumCi heket fipnamina. Cälimleluherestot pi hara· memtonin, dumfa, wira qewel.

(D) 35

}

40 (E)

45

K'ete-m tomoiX'uqel, (F) xanX'uqel. c Pi p 6qta niqa· tomoi memq'epCikenhotoin b e s "T'ipna· skuda heketun tomoi", ?uni yomoi. Tomoito-n bima- pc0»qtat0't. } Wi«täm yömoyar, (A) (ab) "Pohima nom haräwira ['] ibida", 7uni pi p o-qta. "Hariliskuda hekem pattähuna·, "Harä wira'ibi da", ?uni punet leweqana·. "Hara wira'ibi da nom", ">uni, wi-täm yomoyar 7uni t i n . "Di-h", 7uni.

(cd)

"Hekem härilwiram", ?uni. *>eleu hekem ti-nmina.

(ef)

P5uqa5ilah, (B) (ab) "Haril ma n *?ewet k'etem küfemto· mat leik'utem", ?ut ρ'ό-qtatot 7üni, '">eleuda hariliskoyumina." "Hekem härilwiram?" ^uni. (cd)

50 [ii] 55

60

65

Sang epic and Native American ethnopoetics

327

"xä-lum häriliskuda." '">ewet keif an haril." '">eleuda", 7uni p'oqtato-t.

(ef)

^uni honda ^ünihara, k o m namia. Cälimleluheres piat'et hüra·, tälimleluheres P'uqa'ilah büha, seX'amahnabuha p'uqa'ilah

(C)

Pi ρ'ό-qtatot 9 ύηΐ tomoito· X'äpurma·. Pubanp'urum tomoi *>eleu X'apurimina. K'ete m pur neqito· tomoi kela.

(D)

70

75

}

P3uqa'ilah büha, seX'amahna· buha. ^ünihara-, " H o h , *>ewet m a n hari-1", *>ύηϊ, "Calimleluherem", pürkurut. "Ho-h", ?ύηί.

(Ε)

Ρ'ό-qta depel put hari lwira. Tomoi X'apurma·, tomoito-n X'apure·, "Ewetam hari'lwiraibida", ^ύηϊ.

(F)

80

85

}

Loon Woman (He who is made beautiful and She who becomes Loon) Told by Jo Bender They live there, many people came into being.

PART ONE [The Beautiful Boy] [I] [Hidden]

After that two, a pair a man and wife live there, they have many children,

5

328

Dell Hymes a lot of children nine boys, one girl, ten children.

10

The first born (is) a beautiful boy. And then they leave him put away inside. He-who-is-made-beautiful is what they name him, they leave him put away inside, they leave him to stay rolled up in a bear hide.

[II] [Discovered and Desired] [i] (A)

They live there, live there, live there, some of the children walk around children play around. That "girl" lives there, the girl grows bigger, turns into a woman.

15

20 (B) }

After that, one morning the woman goes to a certain stream, she goes to where they get water, she sits down, that woman. She looks at the ground, she sees a hair, she takes it up. She looks, she looks at the hair, one hair.

(C)

She looks at the hair she has found: "Whose hair?" she wants to know. She looks at it long, looks at the hair, one long hair. That woman thinks, she thinks, "Whose hair?"

(D)

25

30

35

}

That He-who-is-made-beautiful goes every morning to bathe. No one has any idea that he goes to bathe. That He-who-is-made-beautiful goes to that water, bathes, comes home.

40 (E)

45

Sung epic and Native American ethnopoetics One hair has come loose, comes off his head. The woman finds it, the hair, finds the hair at the flat where they dip up water. "I want to know whose hair it is", so she thinks. That woman keeps that hair. }

(F)

Thinking of the man, "This morning I shall go west", so says that woman. "I want someone to go along to guide me. "I am going to go", so she tells her mother.

(A) (ab)

"I am going to go west", so she says, thinking of the man, so she spoke. "Hm", she [her mother] says.

(cd)

"Whom do you want to go with you?" she says, She will not tell whom.

(ef)

329

50 [ii] 55

60

The old woman, (B) (ab) "Well, surely go with this little one, your younger brother." Now that woman says, "I don't want to go with him." "Who will you go with?" so she says. "I want to go with another one."

(cd)

"Go with this one then." "I won't", so said that woman.

(ef)

She does this for a long time, she goes through them all. He-who-is-made-beautiful alone is left, He-who-is-made-beautiful. The little old woman sits, sits pondering it all, the old woman.

(C)

That woman measures that hair like that. The hair of the rest does not match. The one hair she had found is long(er).

(D)

65

70

75

}

The old woman sits, sits pondering it all. Suddenly, "Yes, surely go with this one", so she says, "He-who-is-made-beautiful",

(E) 80

330

Dell Hymes

"Yes", so she said.

her son.

The woman is happy she is going to go with him. She measures the hair, that hair matches. "I'm going to go with this one", she says.}

85 (F)

Appendix 5 [He-who-is-made-(to-be)-beautiful and She-who-is-made-to-be Loon19] One. A Beautiful Boy. I. Hidden. A first-born boy, beautiful, is hidden. II. Discovered and Desired. His sister ponders a discovered hair, the mother ponders her insistence, she go west with that brother, consents. Two. Incest, Destruction. III. Incest. The sister goes with her brother, and causes him to sleep with her. IV. Destruction. The brother wakes to discover the fact, flees into the sky with the rest of the family, they fall and burn, the sister laments her parents, finds the hearts of them all but her brother, wears them around her neck. Three. Restorations, Retribution. V. Rediscovered, Restored. A woman discovers the brother's heart from its singing, and nourishes it; restored, he joins her and her sister as their husband. VI. Retribution and Family Restoration. Their two sons learn about the bird that wears the hearts, and how to shoot it, from another bird; they shoot it, then show their father, who takes the heart and restores them.

Sung epic and Native American ethnopoetics

331

Appendix 6: Profile Underlining indicates repetition or other connection between lines. Braces around a symbol for a unit indicate amplification within it. A closing brace at the end of a row of numbers indicates that they are paired with the preceding row. Square brackets indicate that words are inserted in English (165-66, 169-70). Boldface underline indicates partial repetition, analogous phrases. Acts Scenes/Stanzas Verses

Lines

PART ONE I

i

A Β C

ab abc abcde

II

i

A Β {C} {D} Ε F A

abc de 16,17,18 + 19-20 abc 21,22,23 } abc abc abc 24,25,26; 27,28,29; 30, 31,32 ab abc ab 33-35, 36-38,39-10 abcde 41, 42, 43-15 ab cd ef 46, 47; 48, 49; 50, 51 } ab 52-54,55 cd 56-57,58 ef 59,60 ab 61-62,63-64 cd 65,66 ef 67,68 ab cd ef 69-70, 71-72, 73-74 abc 75, 76, 77 abc 78-79, 80-84, 85 abc 86, 87-88, 89 ab cd ef 90-91, 92-93, 94-96; g hij 97; 98, 99,100

11

Β

111

C D Ε F {A}

1-2 3-5,6-7,8-10 11,12,13,14,15

PART TWO i

ii

A Β C D Ε A Β C D Ε

abc abc abc abc d cd a a abcde abcde a

101,102,103 104,105,106-07 108,109,110 111-13; 114 115-17,118 119 120-21 122-26 127-31 132-34

332

Dell Hymes

iii

i ii iii

iv

V

F A Β C

a abc abc abc

135-37 138,139-41,142 143-44,145a-b, 146-47 148-50,151-53,154-55

}

A Β C A Β C A Β C D Ε A Β C D Ε F A Β C D Ε

ab ab[c]d[e] ab cd efg abc abc abcde abcde ab cde ab c de ab cde ab cd ab cd ab cde ab c def ab c d e abc abc ab cd ef gh ij

156-157,158-59-60 161-62,163-64,165-66,167-68,169171-72,173-7£ 17S-77 178,179-80,1S1-M 184-85,186-87,188-89 190.191,122-94,125,196-97 198,199-200, 201-02. 203, 204-Q2 210, 211; 212, 213, 214 215,216; 217; 21S, 219 220.221; 222. 223. 224 225.226:227. 228 229,220; 231,222 233.234:235-36-37 1 238. 239:240-242: 243-44-15 246.247: 248. 249:250 } 251.252. 253 254, 255, 256 } 257-59,260-62 263-65, 266 267-69, 270 271-73, 274 275-76, 277-79

abcde abc abc abc

280, 281-82,283, 284, 285 286-88,289-90,291-92 293-94,295-96, 297 298, 299, 300 301-02 } 303, 304, 305 306-07, 308-09, 310 } 311-12, 313, 314 315, 316, 317 } 318, 319, 320-23 324-29 } 330, 331-33, 334 335, 336-37, 338-40 } 341-43,344-45, 346-48, 349, 350-51 352-54 355, 356, 357-59, 360, 361-64 365-67, 368-69, 370-73, 374-77, 378-80

PART THREE A Β C A Β C D Ε F G Η I

J

A Β C D

abc ebcde abc abc abc abc abc abc abcde abc abcde abcde

Sung epic and Native American ethnopoetics

IV

VI

i

11

Ε A Β C D Ε A Β C A Β C D {E}

F A Β C D Ε F G Η I

J 111

A Β C Ε F G Η I

J

abc abc abc abode abc abcde abc abc

381-82 383, 384-85,386-87 388-89, 390-92, 393, 394-95, 396 397, 398, 399 400-01,402,403, 404-08,409 410,411, 412 413, 414, 415,416,417 418, 419, 420 421,422, 423

abc abc abc a ab: "A" "B" "C" abcde abc abcde abcde abc abcde abc abc a be d< abc abcde abcde abc abc abc ab abc ab ab abc a

424, 425, 426 427-29,430-32,433 434, 435, 436 437-40 441, 442-43 444-46,447-49,450-52 453-55,456-58; 459-63, 465-65; 466, 467-68 469-71,472-76,477-79 480-84 485, 486-90,491 492, 493, 494,495, 496 497, 498,499, 500, 501 502-03, 504-05, 506-07 508, 509, 510, 511, 512 513, 514, 515 516, 517, 518 519, 520, 521; 522, 523 524,525, 526 527, 528, 529, 530, 531 532, 533, 534, 535, 536 537, 538, 539-540 541, 542, 543 544, 545, 546 547, 548 549, 550, 551 552-54, 555 556, 557 558-62,563, 564 565

333

334

Dell Hymes

Notes 1. In what now seems an early book (Hymes 1981), chapter 3, "Breakthrough into performance", written in 1971, is revised as chapter 6 to show the difference. Cf. also an early and later chapter on the same story, "Seal and her younger brother lived there" (8, 9), and a later version of chapter 4, "The deserted boy", itself (1994a). 2. Whether or not this text was so called by its teller is uncertain. The superscription on the manuscript is ambiguous, and what is published has no discussion of titles. 3. Lävi-Strauss (1981:202,589) distinguishes the role of the loon in one area of western North America from another (north : south) in terms of relation to exogamy rather than incest. In Jo Bender's version the opposition is posed within the story itself, exogamy winning. 4. The "prose poem" is increasingly popular, but it avoids the appearance of intentional lines. It does seem that intentions differ and have different effects, at least with good writers. I have tested classes with a short prose poem by the Native American poet, Diane Glancy, and a short lined poem by William Stafford. The students were given each poem in two forms, that devised by the author and another devised by me (spaced lines for Glancy, a paragraph for Stafford). So far almost all of them prefer what the author intended. 5. Unfortunately, the preceding stanza was printed incorrectly (p. 169). One line is misassigned and one line is missing. The stanza should read (Crab is speaking): "How small these salmon are." Her fingers were broken off. She hung those "small" fish on other fingers. Her fingers were broken off. All her fingers were broken off. 6. "Monotelic" is a suitably obscure word for this with a nice ring. See Hymes 1968. 7. Demetracopoulou's article of 1933 does not have the complete speech, presumably through some mischance. A crucial part of instructions as to how the great bird comes is omitted, with the important line, "she who made us kinless". The English text in the earlier collection from her fieldwork with Cora DuBois (DuBois and Demetracopoulou 1931) does, as of course does the unpublished Wintu ms. This is another example of the need to check all sources. 8. The sequence I recorded in 1954 from Hiram Smith has distinct interpretations of individual stories, but Mr. Smith was unsure of the whole sequence and followed that available in Wishram Texts. 9. Curtis (1911:12, note) says that "This and the following story were related as part of the transformer myth, but they doubtless should be considered as separate stories of a later period." He published them after a myth of "The Transformers", told by a downriver (Chinook) woman, one in which Coyote has no part. The two stories evidently belong within the Coyote cycle, and where I have placed them, for Coyote is fooled by fleas (in a different way) in Louis Simpson's cycle at that point, just after the two incidents about white salmon, just before the incident of not recognizing something from a woman's body as desirable food (Chinook salmon). There is nothing about chronology to separate these two incidents from the narrator's cycle.

Sang epic and Native American ethnopoetics

335

10. In Louis Simpson's cycle "The news about Coyote", in which Coyote sucks himself, follows the one involving fleas, and immediately precedes the old woman who offers salmon. Coyote creates a wall to keep the news of what he has done from getting out, but the news rises over the wall, and people he comes to, having heard it, refuse him food. (In Victoria Howard's Clackamas version, Coyote builds a pile of rocks, which the news breaks open.) The Curtis version elides the news, retaining the hole and a place name, in the middle of the story of the old woman. Having left her, Coyote smells something sweet, as if cooking: He ran about, calling, "Wait for me; Do not eat all the food, I am hungry!" ]He ran about so distractedly ]that he bumped his head on a pile of stones, ]and made a hole in them, and from this accident the place received a name. 11. For a close comparison of five versions of "The news about Coyote", see Hymes 1996a. For the order of cycles in Shoalwater and Kathlamet Chinook, Coos (Oregon coast), Quinault (Washington coast) and Winnebago (originally Wisconsin), see Hymes ms. a. 12. On thought couplets structuring as poetry ancient Egyptian texts assumed to be prose, see Foster 1980. 13. I analyzed Mamie Offield's story "Coyote trades songs and goes to the sky" (Bright 1957:193-94) while preparing this paper. 14. Melville Jacobs took down some narrative dictation in Tillamook from Mrs. Pearson, but it has not been analyzed for verse patterns. I hope it will be and that the possible difference of a male audience will be indicated. 15. Let me mention an interesting article and book, just brought to my attention, the article (Booth 1978) by Brian Swann, the book (Rubin 1995) by a visit of its author to my university. The two contrast. Booth is concerned with the fact that performance is indeed sequential in time, but that in it emerge patterns that are "time-independent" - annular (ring) and chiastic patterns in ballads, influenced in origin, he suggests, by music, such that "lines must participate in the whole rather than lead step by step to the conclusion." He cites brain research that the right brain is the primary locus for song, Gestalt perception, what is "appositional" (Booth 1978: 383). Rubin's main concern is that "oral traditions, like all oral language, are sequential", and his goal is "a theory of serial recall in oral traditions". Technically, his theory "is based on the idea of cue-item discriminability developed from the interference-theory literature and later studies of cuing" (Rubin 1995: ch. 8, "A theory of remembering for oral traditions", p. 175). He nonetheless takes generic constraints seriously, and has a sense of humor. His interest in constraints is primarily how they effect stability over time, recall and generation of new instances. In "Epic and formulaic theory", Rubin equates

336

Dell Hymes

Lord's "themes" with psychologists "scripts", and reports a search of the Iliad and Odyssey to see if the same actions repeat and whether they repeat in the same order. In a note (18) he remarks: This attempt to add to the great accumulation of knowledge on Homeric epic by psychologists who cannot translate the word Odysseus without a dictionary is fair revenge for the numerous scholars who have invented theories of what memory can and cannot do in order to support their theories of the composition of epic (Rubin 1995: 225). In the end, the fact that Homeric tradition is dead and South Slavic dying limits usefulness to the ideas put forth here. Although epic exists in other places as a living tradition (Foley 1990), to pursue these ideas further, we turn to two very different genres. (Rubin 1995: 224.) One is counting-out rhymes (ch. 10), the other ballads, specifically "North Carolina Ballads" (ch. 11). Here ring-composition and other constraints are tested by recall of old ballads, and also by having people compose new ballads. The basic view taken here [of ballad constraints and their effect on recall, stability and generation of new instances], is shared, often with a few distinctions, by most ballad scholars. What has been added from psychology is a set of concepts, mechanisms, and methods to make these ideas more specific, more testable, and often, through the magic of quantification, considerably duller. (Rubin 1995: 293.) Rubin thinks his findings help reconcile the debate as to ballads being products of memorization or re-creation. He acknowledges difficulty in adapting experimental methods to ring composition and the like except in a general way, and that care must be taken not so lose what cannot be put in such a synthesis. (Rubin 1995: 294.) 16. The second sentence in paragraph 3 seems to invert the terms "folklore" and "literature", either in the translation or the original. Here is the sentence in its immediate context: "All verse folklore is always sung. The form of oral verse is alien to folklore; it is possible only in literature. Therefore, when musical epic folklore becomes written literature, it first loses its musical form and sometimes its verse form as well." Surely the second sentence should read: "The form of oral verse is alien to literature; it is possible only in folklore." 17.1 am grateful to my colleague, Natalie Konenko, for directing me to these sources. 18. See Murko (1951), Zirmunskij (1962), Romanska (1963) and Angelova et al. (1971). I am grateful to Liliana Daskalova-Perkowski for the information in this note and the main text. 19. The sister is not named in the story and the word "loon" does not occur. The older brother's name consists of öal-im- 'beautiful' (as object with -m) and -lelu-heres 'be made (to be), be transformed into'. Cf. laq-um leluheres 'transformed into a wild goose (laq-)'. 'To be transformed into a loon, to be made to be loon', would begin with kiri:w- 'loon'. The full word would be kiri:wleluheres. If the title at the head of the ms. reflects Wintu throughout, in Wintu it would be Calimliliiheres 'ut Laqumleluheres.

Sung epic and Native American ethnopoetics

337

References Afanasyev, Aleksandr 1945 [1826-71] Russian Fairy Tales. Translated by Norbert Guterman. New York: Pantheon Books. Angelova, R., L. Bogdanova, T. Romanska, E. Stoin and S. Stoykova 1971 Bülgarski junaski epos. [Bulgarian heroic epos.] Sbornik za narodni nmotvorenija i narodopis, Vol. 53. Sofia. Anttonen, Pertti 1994 Ethnopoetic Analysis and Finnish Oral. Verse. In: Anna-Leena Siikala and Sinikka Vakimo (eds.), Songs Beyond the Kalevala: Transformations of Oral Poetry, 113-37. (Studia Fennica Folkloristica 2.) Helsinki: Finnish Literature Society. Babcock, Barbara 1985 "A Tolerated Margin of Mess": The Trickster and His Tales Reconsidered. In: Andrew Wiget (ed.), Critical essays on Native American literature, 153-85. Boston: G. K. Hall. Reprinted from Journal of the Folklore Institute 9:147-86 (1975). Booth, Mark W. 1978 The Ballad and the Brain. The Georgia Review 32(2): 371-86. Bright, William 1957 The Karok Language. (University of California Publications in Linguistics 13.) Berkeley and Los Angeles: University of California Press. Brightman, Robert 1989 Tricksters and Ethnopoetics. International Journal of American Linguistics 55: 179-203. Curtis, Edward S. 1911 The North American Indian, Vol. 7. Plimpton, Mass: Norwood. Reprinted, New York: Johnson Reprint Corporation, 1970. Dauenhauer, Nora Marks and Richard Dauenhauer 1987 Haa Shuka, Our Ancestors. Tlingit Oral Narratives. Seattle: University of Washington Press. Daskalova-Perkowski, Liliana 1994 The Bulgarian Folktale: Is It Prose Or Poetry? Bulgarski Folklor 5: 45-55. Demetracopoulou, Dorothy 1933 The Loon Woman Myth: A Study in Synthesis. Journal of American Folklore 46:101-28. DuBois, Cora and Dorothy Demetracopoulou 1931 Wintu Myths. University of California Publications in American Archaeology and Ethnology 28(5): 279-403. Berkeley. DuBois, Thomas A. 1993 From Maria to Marjatta: The Transformation of an Oral Poem in Elias Lönnrot's Kalevala. Oral Tradition 8(2): 247-88. 1994 An Ethnopoetic Approach to Finnish Folk Poetry: Arhippa Perttunen's Nativity. In: Anna-Leena Siikala and Sinikka Vakimo (eds.),

338

Dell Hymes

Songs Beyond the Kalevala: Transformations of Oral Poetry, 138- 79. (Studia Fennica Folkloristica 2.) Helsinki: Finnish Literature Society. 1995 Finnish Folk poetry and the Kalevala. New York: Garland. Evers, Larry and Felipe S. Molina 1987 Yaqui Deer Songs /Maso Bwikam. A Native American Poetry. Tucson: Sun Tracks and the University of Arizona Press. Foley, John Miles The Singer of Tales in Performance. Bloomington: Indiana University 1995 Press. Foster, John L. 1980 Sinuhe: The Ancient Egyptian Genre of Narrative Verse. Journal of Near Eastern Studies 39(2): 89-117. Foster, Michael K. 1974 From the Earth to beyond the Sky: An Ethnographic Approach to Four Longhonse Iroquois Speech Events. Canadian Ethnology Service Paper no. 20. Ottawa: National Museums of Canada. Fowler, Catherine S. and Robert C. Euler 1992 Kaibab Paiute and Northern Ute Ethnographic Field Notes. In: William Bright (ed.), The Collected Works of Edward Sapir, X, 779902. Berlin and New York: Mouton de Gruyter. Guss, David M. (ed. and trans.) 1980 Watunna. An Orinoco Creation Cycle (by Marc Civrieux). San Francisco: North Point Press. 1989 To Weave and Sing. Art, Symbol and Narrative in the South American Rain Forest. Berkeley and Los Angeles: University of California Press. Hinton, Leanne 1977 Havasupai Songs: A Linguistic Perspective. Dissertation. University of California, San Diego. (Published 1984, Tübingen: Guntec Narr.) Hymes, Dell Linguistic Features Peculiar to Chinook Myths. International Jour1958 nal of American Linguistics 24: 253-57. 1965 Some North Pacific Coast Poems: A Problem in Anthropological Philology. American Anthropologist 67: 316-41. Reprinted as chapter 1 of Hymes 1981. 1968 Linguistics - the Field. In: David Sills (ed.), International Encyclopedia of the Social Sciences, Vol. 9,351-71. New York: The Macmillan Co. and the Free Press. 1974 Foundations in Sociolinguistics. Philadelphia: University of Pennsylvania Press. 1976 Louis Simpson's "The Deserted Boy". Poetics 5:119-15. Reprinted as chapter 4 of Hymes 1981. 1981 "In Vain I tried to Tell You": Essays in Native American Ethnopoetics. Philadelphia: University of Pennsylvania Press.

Sung epic and Native American ethnopoetics 1982

1984 1985 1986 1989 1992a 1992b 1993a

1993b

1994a 1994b 1995

1996a 1996b

1996c 1996d

339

Narrative Form as "Grammar" of Experience: Native American and a Glimpse of English. Journal of Education 164(2): 121-42. Reprinted in Hymes 1996c, ch. 6. Bungling Host, Benevolent Host: Louis Simpson's "Deer and Coyote". American Indian Quarterly 8(3): 171-98. Language, Memory, and Selective Performance: Cultee's "Salmon's Myth" as Twice Told to Boas. Journal of American Folklore 98: 391434. Some Subtleties of Measured Verse. In: June Iris Hesch (ed.), Proceedings of the 15th Spring Comference, Niagara Linguistics Society, 1985,13-57. Buffalo, NY: The Niagara Linguistics Society. Tlingit Poetics: A Review Essay. Journal of Folklore Research 26(3): 236-48. Helen Sekaquaptewa's "Coyote and the Birds": Rhetorical Analysis of a Hopi Coyote Story. Anthropological Linguistics, 34: 45-72. Use All There Is to Use. In: Brian Swann (ed.), On the Translation of Native American Literatures, 83-124. Washington, DC: Smithsonian Institution Press. Inequality in Language: Taking for Granted. In: James E. Alatis (ed.), Language, Communication, and Social Meaning, 23-40. Washington, DC: Georgetown University Press. Reprinted and revised as chapter 10 of Hymes 1996c. In Need of a Wife: Clara Pearson's "Split-His-Own-Head". In: Anthony Mattina and Timothy Montier (eds.), American Indian Linguistics and Ethnography in Honor of Laurence C. Thompson, 127-62. Missoula: University of Montana. Ethnopoetics, Oral Formulaic Theory, and Editing Texts. Oral Tradition 9(2): 330-70. Sun's Myth. In: Brian Swann (ed.), Coming to Light. Contemporary Translations of Native American Literatures of North America, 273- 85. New York: Random House and Vintage Press. Reading Takelma Texts: Francis Johnson's "Coyote and Frog". In: Roger D. Abrahams (ed.), Fields of Folklore.. Essays in Honor of Kenneth S. Goldstein, 90-159. Bloomington: Trickster Press. (Printed separately by Trickster Press, 1998.) Coyote, the Thinking (Wo)man's Trickster. In: A. James Arnold (ed.), Monsters, Tricksters and Sacred Cows. Animal Tales and American Identities, 108-37. Charlottesville, VA: University of Virginia Press. Consonant Symbolism in Kathlamet and Shoalwater Chinook. In: Don E. Dumond (ed.), Chin Hills to Chiloquin. Papers honoring the versatile career of Theodore Stern. University of Oregon Anthropological Papers 52. Eugene: University of Oregon. Ethnography, Linguistics, Narrative Inequality. Towards an Understanding of Voice. London: Taylor and Francis. Ankara Rhetoric. Ethnopoetic Suggestions. In: John D. Nichols and Arden Ogg (eds.), nikotwäsik iskwähten, päskihtepauih! Studies in

340

Dell Hymes

honour of Η. C. Wolfart, 263-80. Winnipeg, Manitoba: University of Manitoba. 1999 Boas on the Threshold of Ethnopoetics. In: Lisa Philips Valentine and Regna Darnell (eds.), Theorizing the Americanist Tradition, 84107. Toronto: University of Toronto Press. ms. a Coyote, the Thinking Person's Trickster. Manuscript in preparation. ms. b Loon Woman. (He-who-is-made-beautiful and She-who-becomesLoon.) To appear in a volume edited by Herbert Luthin. Hymes, Dell and Henry Zenk 1987 Narrative Structure in Chinook Jargon. In: Glenn Gilbert (ed.), Pidgin and Creole Languages: Essays in memory of John E. Reinecke, 445-65. Honolulu: University of Hawaii Press. Jacobs, Elizabeth D. 1959 Nehalem Tillamook Tales. Ed. M. Jacobs. Eugene: University of Oregon Books. Republished at Corvallis: Oregon State University Press, 1990. Jacobs, Melville 1958 Clackamas Chinook Texts. Part I. International Journal of American Linguistics 24(2). Indiana University Research Center in Anthropology, Folklore, and Linguistics Publication 8. Bloomington, Indiana. Jakobson, Roman 1945 On Russian Fairy Tales. In: Aleksandr Afanasyev, Russian Fairy Tales, 631-51. Trans, by Norbert Guterman. New York: Pantheon Books. Keeling, Richard 1992 Cry for Luck. Sacred Song and Speech among the Yurok, Hupa and Karok Indians of Northwestern California. Berkeley and Los Angeles: University of California Press. Kroeber, Alfred L. 1946 A Karok Orpheus Myth. Journal of American Folklore 59:13-19. 1948 Seven Mohave Myths. Anthropological Records 11(1). Berkeley and Los Angeles: University of California Press. 1951 A Mohave Historical Epic. Anthropological Records 11(2). Berkeley and Los Angeles: University of California Press. 1972 More Mohave Myths. Anthropological Records 27. Berkeley and Los Angeles: University of California Press. 1976 Yurok Myths. Berkeley and Los Angeles: University of California Press. ms. Hupa Myth Analysis. Bancroft Library. See pp. 7ff. "Informant A. Emma Lewis". I am indebted to Victor Golla for this information. Kroeber, Theodora 1959 The Inland Whale. Bloomington: Indiana University Press. Kuusi, Matti, Keith Bosley and Michael Branch 1977 Finnish Folk Poetry: Epic. Helsinki: Finnish Literature Society. Lävi-Strauss, Claude 1981 The Naked Man. Introduction to a Science of Mythology, 4. New

Sung epic and Native American ethnopoetics

341

York: Harper and Row. (Translated from L'homme mi. Paris: Plön, 1971.)

McLendon, Sally 1977 Cultural Presuppositions and Discourse Analysis: Patterns of Presupposition and Assertion of Information in Eastern Porno and Russian Narrative. In: Muriel Saville-Troike (ed), Linguistics and Anthropology, 153-89. Georgetown University Round Table on Languages and Linguistics 1977, Washington, DC: Georgetown University Press. 1982 Meaning, Rhetorical Structure, and Discourse Organization in Myth. In: Deborah Tannen (ed.), Analyzing Discourse: Text and Talk, 284-305. (Georgetown University Round Table on Languages and Linguistics 1981.) Washington DC: Georgetown University Press. Murko, M. 1951 Tragom srpsko-hrvatske narodne epike, Putovanja u godinama 19301932, knj. I, II. [In Search of the Serbocroatian Folk Epic, Travels in the years 1930-1932, Books I, II.] Zagreb. Nichols, John 1991 "Chant to the Fire-Fly": A Philological Problem in Ojibwe. In: H. C. Wolfart (ed.), Linguistic Studies Presented to John L. Finlay, 11326. Algonquian and Iroquoian Linguistics, Memoir 8. Winnipeg, Manitoba. Pitkin, Harvey 1985 Wintu Dictionary. University of California Publications in Linguistics 95. Berkeley and Los Angeles: University of California Press. Propp, Vladimir 1984 Theory and History of Folklore. Translated by Ariadna Y. Martin and Richard P. Martin. Edited by Anatoly Liberman. Theory and History of Literature 5. Minneapolis: University of Minnesota Press. Rubin, David C. 1995 Memory in Oral Traditions: The Cognitive Psychology of Epic, Ballads, and Counting-out Rhymes. New York: Oxford University Press. Romanska, Tsvetana 1963 Predanijata za Krali Marko vuv folklora u juznite slavjani - obsta xarakteristika [The legends of Krali Marko in the folklore of the South Slavs - Ageneral characterisation]. Slavisticni studii, 379-91. Sofia. Rothenberg, Jerome (ed.) 1986 Shaking of the Pumpkin. Traditional Poetry of Indian North Americas. Revised edition. New York: Alfred van der Marek Editions. Sapir, Edward 1909 Wishram Texts. Together with Wasco Tales and Myths collected by Jeremiah Curtin and edited by Edward Sapir. Publications of the American Ethnological Society 2. Leyden: E. J. Brill. 1910 Song Recitative in Paiute Mythology. Journal of American Folk-Lore 23:455-72. Reprinted in part in: David Mandelbaum (ed.), Selected

342

Dell Hymes

1912

1930 1992 Shaul, David ms. Sokolov, Υ. M. 1950 Swartn, Brian 1993

writings of Edward Sapir, 463-67. Berkeley and Los Angeles: University of California Press, 1985. Reprinted in full in: Regna Darnell and Judith Irvine (eds.), The Collected Works of Edward Sapir, X, 541-58. Berlin and New York: Mouton de Gruyter, 1994. The Takelma Language of Southwestern Oregon. In: Franz Boas (ed.), Handbook of American Indian Languages, Part 2, 1-296. (Bureau of American Ethnology Bulletin 40). Washington: Government Printing Office. Issued separately; the complete Part 2 appeared in 1922. Reprinted in: Viktor Golla (ed.), Takelma Texts and Grammar. The Collected Works of Edward Sapir, Vol. 8. Berlin and New York: Mouton de Gruyter. Texts of the Kaibab Paiutes and Uintah Paiutes. Proceedings, American Academy of Arts and Sciences 65(3): 297-535. Reprinted in Sapir 1992:315-555. Southern Paiute and Ute linguistics and Ethnography. In: William Bright (ed.), The Collected Works of Edward Sapir, X. Berlin and New York: Mouton de Gruyter. Hopi traditional narratives. Russian Folklore. Trans, by Catherine Ruth Smith. New York: Macmillan. Song of the Sky. Versions of Native American Song-Poems. 2nd edition. Amherst: University of Massachusetts Press.

Tedlock, Dennis 1972 Finding the Center. Narrative Poetry of the Zani Indians. New York: Dial. 1983 The Spoken Word and the Work of Interpretation. Philadelphia: University of Pennsylvania Press. 2irmunskij, V. M. 1962 Narodnyj geroiceskij frpos, sravnitel'no-istoriceskie ocerki [Folk Heroic Epos, Comparative-Historical Studies]. Moscow - Leningrad.

Generic models, entextualization and creativity: epic tradition on the Southern Cook Islands Anna-Leena Siikala Textualization has played and even today still plays a central role in building cultural identity on the basis of oral traditions. As Michael Silverstein and Greg Urban have stated in the introduction to their joint publication Natural Histories of Discourse (1996), "the natives" (including us) "engage in processes of entextualization to create a seemingly shareable, transmittable culture" (Silverstein and Urban 1996:2). Different questions of textualization have gained increasing attention since the pioneering works of Dell Hymes (1981), Dennis Tedlock (1983), Elizabeth C. Fine (1984), etc. who concentrated on the textualization of oral discourse, specially folklore texts. We could also ask, as Greg Urban (1996: 21-44) and John B. Haviland (1996: 45-78) do, how "the natives" themselves tackle these problems. Both have created an experimental context in which people with little previous experience and trained by the researcher produce text-artifacts on the basis of recorded instances of discourse. In these experiments, a phase of discourse is transformed into a new form, a text-artifact, which is, according to Silverstein and Urban, "opposed to discourse as decontextualized is opposed to contextualized" (Silverstein and Urban 1996: 4). The idea of a text as a decontextualized inscription of speech is of course justified if we define a text - as Paul Ricoeur does - as any discourse fixed by writing (Ricoeur 1991:106). But, in our conventional textual practices w e do not textualize any parts of discourse as such. Ricoeur continues: What is fixed by writing is thus a discourse that could be said, of course, but that is written precisely because it is not said. Fixation by writing takes the very place of speech, occurring at the site where speech could have emerged. This suggests that a text is really a text only when it is not restricted to transcribing an anterior speech, when instead it inscribes directly in written letters what the discourse means. (Ricoeur 1991: 106; italicization by the present writer.) When Urban's and Haviland's "natives" were given the task of transcribing a recorded phase of discussion, they were actually given an order to give it meaning. No wonder that Greg Urban notices that for a master of mythic tradition it is easier to entextualize, that is to de- and recontextualize, to give a meaning and structure to, a mythic text than an everyday discussion (on the concept of entextualization, see the critique by Lauri Honko [ 1 9 9 8 : 1 4 1 -

344 Anna-Leena Siikala 51]). Wanpo, belonging to the "elders", who transformed a myth through deletion and substitution, was not only copying it but using the original as a stimulus for creating his own original story (Urban 1996: 36). Urban further claims that discourse setting itself apart from the local is more replicable than discourse with several deictics and devices connected to the local context. He refers to m/ths, which are easily replicable: a kind of distilled discourse which in many societies belongs to the essential part of culture. And he asks: Does some discourse make for better culture? If Urban had continued his line of thought, he would have moved from the conceptual sphere of discourse defined by linguistic inquiry into another conceptual domain of discourse constituting the power relationships of society in the spirit of the works of Michel Foucault and Pierre Bourdieu. Some discourse does, indeed, make for better culture. Myths and historical traditions carry the memory of the past, they represent an authoritative discourse bigger than society and form a useful tool for legitimating power relations and thus giving substance to the ethnic and national processes. The European Romantic era of the 18th and 19th centuries testified several nation-building processes in which the textualization of symbolically relevant oral traditions had an important, sometimes even central role. The post-colonial interest in reviving ethnic traditions includes, in different parts of the world, entextualizing projects which have been criticized because they seem to follow the Romantic models of the old Europe. On the other hand, the desire to fix the socially and politically valuable parts of oral discourse, myths and history, in inscription - in clay, papyrus, palm leaves, paper, etc. - whenever it is possible, can be followed for thousands of years in human history. The invention of tradition in the Europe of Romanticism was not a new invention, if an invention at all. Different Central Polynesian island groups have in recent years published collections of their oral heritage in addition to the collections published in the 19th century. In the Cook Islands, the tumu koreros, masters of oral history, have joined forces in the pursuit of creating a picture of the history of the islands. A collection of the korero, 'history' was, organized in 1974 and as a result discussions with tumu koreros, local experts on oral traditions, were recorded and preserved in the archive of Rarotonga. Besides that there has been a number of publications concerning the oral traditions of the outer islands. The publications compiled by local collectors are built on a European model of narrative collection: individual narratives centering around a hero or a special event are randomly grouped together. For a person who knows the uses, cultural position and value of epic narratives, the collections do not seem to make sense. The meaning conveyed in oral performances is lost in the published narratives. What has gone wrong? The European model of legend collection does not provide room for representations of inner relations

Generic models, entextaalization

and creativity

345

of narrative corpus or alternative ways of creating coherence for it. When the mythic historical discourse is entextualized as bits and pieces representing random phases of its (re)contextualization, the meaning of discourse grasped and recreated in different ways by every qualified performer and his or her audience is lost. The collections do not observe discursive practices, the generic models of korero. It seems that the discussion on entextualization has not paid enough attention to the generic differences, or to the intertextual relationships of discourses to be transformed into texts. We could, for example, ask which is the best way to build a publication which mirrors the character and cultural meaning of a local epic tradition. Instead of asking how "the natives", who have no knowledge of Western ways of entextualization, would do it in a random test situation, we can ask how they have done it when dealing with a discourse they understand and regard as vital in their lives. There is an abundance of non-Western manuscript cultures using different traditionbased entextualizing strategies in order to transmit culturally valuable knowledge to subsequent generations. What is the relationship between these manuscripts and oral discourse or other forms of performance? For there often is a relationship which, furthermore, is often also crucial for the production of manuscripts, and "texts" created using more advanced techniques when possible. Even though I cannot answer all these rhetorical questions, I will try to illuminate some points by observations based on fieldwork with Jukka Siikala on Mauke and Atiu, small Central Polynesian Islands belonging to the Southern Cook Islands, in 1983-84,1988 and 1995.

Myth, history and genealogy The oral tradition known as korero 'history' represents the speech of ancestors, the knowledge inherited from the authoritative past. The korero can be characterized as mythic historical knowledge and it contains both cosmological and historical information. Remembering the past is crucial for our sense of identity (Lowenthal 1990: 197-200), be it an individual or a group identity. A sense of a shared "history" seems to be necessary to the formation of an ethnic group. Following the "invention of tradition" discussion of the eighties Kirsten Hastrup says that in a case where the shared history is absent, traditions may be invented for the purpose of distinction (Hastrup 1987:258; see Hobsbawm and Ranger 1992 [1983], Anderson 1983). A concept of the past of one's own group and of the fundamental essence of the universe is one of the most basic forms of human knowledge. Before the art of writing was invented, oral narratives and poems were the primary tools for preserving this knowledge. The controversy over the explanation

346 Anna-Leena Siikala of the mythical knowledge codified in oral tradition - evident as a preference for either the mythical or the historical interpretation - ultimately illustrates the difference between the concepts of history held by Western researchers and the members of the culture studied. In traditional cultures myth and reality are not placed in different compartments as they are in Western thinking. The fact that comparative religion has underlined the religious dimension of mythology has often prevented us from seeing its historical function, its position as the "sacred" history of one's own group. Mythical world history is just as much the history of humans and deities as of forefathers and superhuman heroes. Both myth and history are symbolic forms of connecting the past and present. Christine Seydou (1990) points out that epics with a mythical bias are a characteristic feature of communities which lack a centralized power structure and in which the devices for communal identity are founded on a system of kinship. Epics with a historical bias are in turn to be found among communities in which a framework for identity is provided by state-like systems and hierarchical social relations. According to Kirsten Hastrup "the myth represents the history of oral culture, while history is the myth of literal culture" (Hastrup 1987: 260). She states that history and myth connect the past and present in different ways. Myths are discontinuous in a way that the intermediate time between the sacred time of origin and the present situation disappears, while history tries to make the process from past to present visible and comprehensible (Hastrup 1987: 261). She denies the thought that history is more true than myth. Hastrup compares oral myth and written history, which leads her to difficulties. She does not pay attention to the fact that in oral communities there is also an awareness of a history involved in secular events. Neither does she remember the religious dimension of the myths. On the other hand, myth and history may be included in the same narratives. As parts of the same concept of reality, myths and concrete historical events may have merged to form a single entity. Kirsten Hastrup's idea that myth lacks a chronological order (which she sees as a sign of history) does not match the Polynesian cosmological narratives. Even though the "historical time", happenings in the world of human beings, is not part of the remote past of the time of creation, it does have a connection - expressed as family ties or by genealogies - to these happenings. Claude Levi-Strauss noted this special feature of Polynesian mythology in his work The Savage Mind: "Some Polynesian mythologies are at the critical point where diachrony irrevocably prevails over synchrony, making it impossible to interpret the human order as a fixed projection of the natural order by which it is engendered; it is a prolongation, rather than a reflection, of a natural order" (Levi-Strauss 1973: 233; see Sahlins 1985). The sense of history owned by different cultures is expressed by means of different genres and in different contexts. Stephen Hugh-Jones has empha-

Generic models, entextnalization and creativity

347

sized the difference of contextual relevance between the domains of myth and history (Hugh-Jones 1989:55). On the other hand, matters classified under mythical consciousness are manifest in many different ways, not merely as oral narration (cf. Dumezil 1970: 3). Alongside the cyclical, ritual repetition of primeval events described, for example, by Mircea Eliade (1954) we find genealogical lists stretching back for dozens of generations, as on the Cook Islands, to what are reputed to be the divine forefathers. Valerio Valeri claims that the relationship between present and past is articulated by specific cultural modes, which he describes as paradigmatic and syntagmatic or metonymic in Hawaiian culture. The genealogies represent the latter type of relationship, in which past and present in turn emphasise the analogic relations between certain events of the past and certain events of the present. An analogical relationship between special events is possible by implicit mediations of a scheme or rule common to them both which makes them comparable. (Valeri 1990:157.) Valeri refers to the custom of passing on the concept of social rules to new generations in the form of historical or mythical narratives. The very fact that events belong to the past, that they are unalterable, gives them a position of authority vis-a-vis the present (cf. also Hastrup 1987: 264). The two modes of connecting the present and past mentioned by Valeri are actually closely related in Central Polynesian narrative art. In the Cook Islands the authority of the past is mediated to new generations with genealogies to which myths and legends are connected. One example is the cosmology of Mauke, an outer island of the South Cook Islands, collected in 1882 by William Wyatt Gill and published in 1911 under the title "Growth of the Land of Ma*>uke". It differs considerably from the other, larger cosmogonic genealogy of Mauke, "A legend of the Ancestors of the Growth of 9 Akatoka-Manava-te-rae-o-te-pa^u", given by Stephen Savage (Tivini Haueti) and Samuela (S-TeRei-More-Taunga o-te-tini) in 1916 (Savage, Ε Buka Tuatua Enua, 1916, The Cook Islands Archives) even the main scheme is about the same. Whereas Savage's cosmology concentrates on the growth of different layers of earth from each other and emergence of the seasons, Gill's version contains more clearly genealogical information connected to heroes: Ko Area te katiri ο te au mea kaatoatoa. Ko Atea ka no^o i te valine, i a Paparoai-te-^itinga; anau ta aaua ko Te Tumu: 32. Atea = Paparoa-i-te-^itinga 31. Te Tumu = Papa-i-te -^opunga 9 Eke Te Tumu e! neke mai, Kare au e neke atu, Ko Te Tumu au no te ''enua. Ko Te Tumu ka no ? o i te valine i a Papai-te-^opunga, anau ta raua ko Tumu-te

348

Anna-Leena Siikala

nekeneke. 9 E Papa e! neke mai, Kare au e neke atu, Papa au no te ^enua. 30. Tumu-te-nekeneke = Tumu-te-oioi 29. Tumu-tikei =Tumu-aro 28. Toka-rukuruku = Toka eaea 27. 9 Uke = Te Pua^i-angauta 9 Uke umu ο te va^arua kino. Kua tae mai ki te ao nei, kua tangata. Kua no^o ^a 9 Uke i te valine, i a Te Puai">angauta, 9 anay raua ko 26. Tara-matie-toro = Tura Na raua i katiri te tangata i Ma^uke e Atiu. Ε tangata Atiu 7 a Tura, e kua ^aere mai 7 a ia ki Ma^uke nei e rave'' i tetaPi tama^ine a 9 Uke, i a Tara-matie-toro. 9Anau ta raua ko: 25. Purea = Tarangi-^enua-manu 24. Kura =Vai-ko'>uko'>u i te Kauariki 23. Tiutiu =Tuakau 22. 9 Ona-ariki = Rangi-marie 21. Patu-kura = Totere 20. Tu-ariki = Te Tau 9 E toko-ono te po mai; e 25 papa-uki tangata mei i a 9 Uke mai e tae ^ua mai ki a Pare-pora. Atea was the origin of all things. Atea stayed with a woman named Paparoa-i-te-itinga and Te Tumu was born to them: 32. Atea = Paparoa-i-te-^itinga 31. Te Tumu = Papa-ite-te-^opunga Ο te Tumu, come closer, I will not go away from you, I am the source of the land. Te Tumu stayed with the woman Papa-ite-^opunga and to them was born Tumu-te-nekeneke. Ο Papa! come closer I will not go away from you I am the foundation of the land. 30. Tumu-te-nekeneke = Tumu-teoioi 29. Tumu-tikei = Tumu-aro 28. Toka-rukuruku = Toka-eaea 27. 9 Uke = Te Puai^angauta 9 Uke-umu of the evil deep. He came to

Generic models, entextnalization and creativity

349

this land and day and became a man. 9 Uke stayed with a woman named Te Puai-angauta and to them were born 26. Tara-matie-toro = Tura It was these who were the origin of the people of Ma^uke and Atiu. Tura was from Atiu and came to Ma^uke to marry this child of 9Uke, Tara-matie-toro. To them were born: 25. Purea = Tarangi-^enua-manu 24. Kura = Vai ko^uko^u i te Kauariki 23. Tiutiu = Tuakau 22. 9Ona-Ariki = Rangi-marie 21. Patu-kura = Totere 20. Tu-ariki = Te Tau There are six from the po here; and 25 generations of men from 9 Uke down to Pare-pora. (Gill 1911:134-35.) Te Tumu literally means the trunk or a main part of anything from which something grows. Stephen Savage translates the concept as 'a foundation, root, cause, origin, source' (Savage 1962: 413). Te Tumu, the source of the cosmic growth, is a child of Atea, the origin of all things, and Paproa-i-teitinga, The Earth of Sunrise. Jukka Siikala has in his work Akatokamanava, Myth, History and Society in the Southern Cook Islands (1991) seen the cosmic genealogy as an emerging process in which Te Tumu, the source, generates the human beings by gradually elevating them from darkness to light. The version given by Savage shows how the genealogical generations represented by named gods are born one after another from the ancestral divine parents. The main pattern is reminiscent of the Hawaiian cosmogony Kumulipo in its main structure (see Beckwith 1981). The number of species is, however, limited in shorter Maukean cosmological genealogies. In Gill's cosmology, which describes the origin and the growth of the land and the appearance of gods and people, there are 26 generations before the ancestors of the Maukeans, Uke and his wife Te Puai^angauta, who form the turning point in the cosmic development. Life buds out from the dark night, po, and the genealogical line of layers of earth and gods is transformed into a line of human beings: Uke "came to this land and day and became a man". In giving the names of the descendants of Uke's daughter Tara-matie-toro and her husband Tura, the Atiuan ancestor chief, the text specifies six generations from the po; altogether there are "25 generations of men from 9 Uke down to Pare-pora". A mythical genealogy turns into a historical one reaching from the divine ancestors to the living people. It is interesting to see that the exact number of generations is mentioned. When discussing the length of historical time, Maukeans counted that the intermediate time between the

350

Anna-Leena Siikala

po and the present day consists of 25-30 generations. That is a common length of the family genealogies which do not contain the cosmological events but begin from the first ancestor after po.

Generic models of mythic-historical discourse The masters of mythic-historical discourse are the tumu koreros. Tumu korero can literally be translated as the source of tradition. Tumu koreros are guardians of tribal traditions, they are historians and genealogists, the main pillars and final authorities of the orally transmitted tribal lore, history and genealogical structure. Their aura of authority is needed in organizing tribal rituals. The position is inherited from one or several former specialists of tradition. Often the candidate already has a reputation of a person who knows. For example, Pirioi Tairea from Mauke, a good singer and an extremely talented narrator, who in 1983-84 had no position as a tumu korero but who enjoyed performing tua taPtos as well as humorous tales, was selected in the late 1980s. The only acknowledged tumu korero of Mauke in the early eighties, Tengaru Tobia, told of his selection in a modest way: I started in 1974 when I was with the Mauke Tere Party in Rarotonga, and there was a meeting and Tautara Purea asked me to join the meeting for the Korero. I listened to all the korero from the Outer Islands and I was thinking whether this is true or not. I could not really believe it and when I came back to Mauke, he asked, Tautara asked me to join the other three Tumu Koreros, that was Tararo Ariki, Aiturau and Tangi Purea... All these three are dead now and there is no one else to take over. So, I am only an Assisting Tumu Korero but just as well I learned some from them and now I am carrying on with it. (Tengaru Tobia 1983.) Over the years Tengaru, or Papa Tete, as people prefer to call him after his long career as a tumu korero, has institutionalized his position, created his own view of the truth in korero, gained self-assurance and elegance in his performance, which could be observed e.g. in 1995 in Rarotonga, where he told a long narrative of Taratoa, a hero whose "ownership" Atiu and Mauke, two neighbouring outer islands, are at the moment disputing, in order to testify that the hero originally represented Mauke. Tumu koreros teach tua tai^to narratives in maraes, meeting houses and also in schools. Every child learns the old narratives of his family. Nowadays the most important form of performing the tua taPtos are action songs and drum dances. The main parts of the legends are danced and sung. The form of the Cook Island tua tai^to, old narratives, is similar to the Scandinavian sagas. Short poems (songs), pe?es, are situated at crucial points in the prose text. The pe9e expressing emotion, describing a memorable moment or sometimes giving knowledge of the genealogy or names of belongings of a

Generic models, entextualization

and creativity

351

hero, is a sign of authencity showing that the performer masters the tradition. The best tumu koreros master the historical continuum of the narrated events and can produce a shorter or a longer version of a narrative. The key to the narrative tradition is the genealogy and the key words are the names of heroes. There is a narrative of Uke, a narrative of his daughter Kaitini, a narrative of his daughter's son Moenau, etc. The main events in the lives of ancestors can also be told in a summarizing, saga-type narrative. So, we could say that the korero, the oral history of the kin group, is transmitted to new generations as (a) strictly replicable genealogies, papa9anga and (b) myths and legends, tua taPto, 'old narratives', telling of the deeds of ancestors who form the genealogy. Researchers of Polynesian folklore have noted the close connection of genealogies and narratives dealing with mythic history (Chadwick and Chadwick 1986 [1940]: 2 4 2 ^ 3 ; Chariot 1987: 11). Gregory Schrempp formulates the quality of the connection in the New Zealand Maori tradition in the following way: "...the genealogical and prose cosmogonies bear the following complementary relation: they recount the same process, the former from the standpoint of demonstrating continuity, the latter from the perspective of recognizing and emphasizing discontinuity" (Schrempp 1985: 23). Schrempp stresses the notion that in the genealogical ordering, discontinuity is subordinated to continuity. The prose form specially marks the phases of separation (wehe) in the genealogy. (Schrempp 1985:23.) Because the same main heroes of mythical history are the dramatis personae of traditions on many different islands, they are also connected to different sets of genealogies. As a consequence of the connectedness of the narratives through genealogy and geography, the corpus of narratives cannot be defined in the usual folkloristic way: seeing it as a collection of separate narratives. Even though the individual narratives are named by referring to the main hero, and even though individual narratives can be performed separately, the script of the performance of a qualified narrator is always reproduced on the basis of larger structures of knowledge concerning the internal connections and mutual hierarchies of narrated and not-narrated. For what is not said is an important part of korero, too. A proper knowledge of the genealogies and connected narratives can be used as a valuable weapon in disputes over land rights and titles. For that reason silence is as important a part of genealogical knowledge as its performance in the right context and moment. The role of silence in the korero is manifested in several ways. Firstly, the genealogical knowledge is, due to its value in land disputes, considered to be extremely confidential and potentially secret information. The traditional knowledge of genealogy, history and heroes belongs to a lineage, to the line of descent from the common ancestors. Secondly, the Southern Cook Island cultures, seemingly outgoing and open, are cultures of silence where differentiated knowledge mastered by specially selected and trained people is

352 Anna-Leena Siikala strictly kept from outsiders. This concerns not only the common tradition of lineages but also the intangible heritage inside the family. Only one of the children will be selected to inherit, e.g., the ethnomedical knowledge transmitted in a family and the role of healer connected to it. The other members of the family have to rely on his or her expertise and help. Thirdly, there is a division of tasks between tumu korero and va9ara, the main speaker of the island, who is a specialist in oratory and narrative performance. Sometimes these two roles meet in one person: Papa Aiturau, the most authoritative specialist in Mauke korero during the seventies, was at the same time the va^ara of the island. He passed the information of korero and tasks to Mapu Taia, the principal of the school, who has actively increased his knowledge from several other sources. In 1984 Mapu Taia explained the role of oral tradition in his practice of va^ara, saying: "When you speak, when you are welcoming a tere party, you must have these things ready. So, you incorporate these (narratives) into your speech." Mapu Taia points to the generic intertextuality of korero, the role of entextualization and recontextualization in the discourse production of va^ara (see Bauman and Briggs 1990: 73-74; Briggs and Bauman 1992: 148). The speaker must have an ability to entextualize korero discourse, to create a "coherent, effective and memorable text" which can be performed in different contexts. When invoking the korero by the use of performative strategies, allusions and narrative episodes, he invests himself with the authority and power of the genre. Va9ara 'has things ready', decontextualized, from previous instances of discourse, and he "incorporates", recontextualizes relevant narrative parts into his performance according to the demands of the situation. The tua tai?to corpus of the Southern Cook Islands could be described as a network of potentially performative episodes, narrative events which can be decontextualized and recontextualized according to conventional and individual performance strategies. These strategies involve not only narrating but all the culturally preferred means of expression including, today, in addition to telling, singing, dancing and acting also pictorial presentations. All the forms of performance are realized using, besides, different sets of performing strategies. Complicated performances sometimes need negotiations by several experts of tradition; quite common is the co-operation of a tumu korero with a tumu imene, master of songs, in creating a performance in which singing and acting are crucial. The horizontal dimension of the korero network is made up of episodes connected to different ngatis (lineages) and islands; the vertical dimension in turn creates genealogies. Martha Beckwith has suggested that genealogies in Kumulipo could be read not only vertically but also horizontally; on genealogy and related information (Beckwith 1981; Shore 1982:83.) In performing a narrative, the tumu korero can follow the vertical line connecting the lives and deeds of different heroes to form a sequence in time, or he may follow

Generic models, entextualization and creativity

353

the deeds of a hero on different islands, weaving the episodes into a large narrative entity. Genealogies act as redactional outlines - to use the term of John Chariot - which are needed in producing large tradition complexes orally. Chariot studied the Hawaiian Kamapua9a literature and stated that "the outline provided the framework of a complex, referring to its individual stories, complexes, or sections in their proper places. This outline could be memorized, and the complex could then be told by expanding some or all the sections as one reached them." (Chariot 1987:67.) He found several forms of "redactional outlines", a typical form being a list. Korero is a kind of metadiscourse in which the content is mediated in using different strategies, genealogies marking the vertical line of descent, tua tai*>to narratives, the horizontal chain of events, and pe?es, recited or sung poems, the memorable moments of action and emotion. In ura performances, the events of korero are reanimated by dance and song created by tumu imene, the song master, for different occasions. Plays and scenes of the post-colonial era use the narrative world and knowledge of korero as well. Korero could also be described as metagenre consisting of the same contentual matter and the same functional field but represented and performed in different forms. Because the forms of korero are not random but related in a patterned way which also offers scripts for performances, the korero is based on generic models mastered by tumu koreros. When speaking of the Southern Cook Island oral tradition, it is easy to agree with William F. Hanks (1987) in his description of genre when he conceives - in contrast to formalist conceptions of genre which stress the immanent, thinglike character of oral genres - of genres as orienting frameworks, sets of expectations and interpretive procedures for the production and reception of discourse. For Hanks, genres are not text or discourse structures but models for the ways people relate and use language. (Hanks 1987:670; see also Briggs and Bauman 1 9 9 2 : 1 4 2 ^ 3 . ) The life of korero can be described by using Hanks's view of oral genres: "The idea of objectivist rules is replaced by schemes and strategies, leading one to view genre as a set of focal or prototypical elements, which actors use variously and which never become fixed in a unitary structure" (Hanks 1987: 681). A tumu korero has an idea of vertical and horizontal connections of the tua tap to narratives and their episodes, of the role of pe?es as legitimations of narrative truth and markers of the culmination points of narrated events. The inner coherence of korero can be viewed in terms of intertextuality in Bakhtin's (1985,1986) sense. Every performance of tumu koreros is an individual statement or interpretation of korero discourse dependent on the former statements and interpretations of the same discourse. Without the connection to korero, tua tai^to narratives have only denotational structural meaning which can be open to any world preferred by the interpreter, but which at the same time lack all the uniting power and

354 Anna-Leena Siikala

meaning from the island point of view. Narrative collections following the European model of publication offering a selection of random individual "legends" have no sense, because they have lost the inner meaning of korero discourse, even though reading them may be entertaining. The failure to catch the inner coherence of the "native" discourses in published "legend" collections obscures not only the intrinsic power of narratives but also the grand poetry emerging from their culturally defined values and world view. First attempts at textualization: puka papa?anga The collections made by missionaries or ethnographers working in the Cook Islands during the last century or the European-type narrative collections are not the only attempts to textualize the korero. The islanders themselves have been textualizing their oral traditions for over a hundred years. When writing was introduced into the island culture, it provided a means to immortalize the most valuable parts of the intangible culture. The puka papa?angas of the Cook Islands were created to preserve the genealogical information and epic tradition in manuscript form. Jukka Siikala has handled the use and meaning of these manuscripts, which are considered to be important family property: These books are most carefully written by hand and contain several kinds of information... The main content is genealogies proving the necessary testimonies for the land court in establishing land rights... However, the books also contain other kinds of material, most of which has direct relevance to the genealogies... historical events are carefully recorded in the family books in the form of the historical narratives generally known on Ma^uke. Although anonymous and undated, these narratives are firmly linked with everyday life making repeated references to topographic features on Ma^uke, while the meaning of these landmarks is regularly expressed in short allusive references to the narratives.... Besides these historical chronicles, an essential part of every book on Ma^uke seems to be some kind of history of the chiefs of the island... (Jukka Siikala 1991: 16-17.)

The family books of the Cook Islanders seem to follow the general Austronesian pattern of the manuscript cultures analysed in Bali by James Boon. The contents of the books are clearly divided into two types: lists of genealogical and topographical names, and connected narratives, calendrical and seasonal information etc. The length of narratives, the amount of detail and important information, such as pe?es, vary, as does the skills of the writers. The purpose of entextualization is revealed in a manuscript of Stephen Savage (The Cook Islands Archives). In 1916 Samuela-TeRei-More-Taunga-

Generic models, entextualization and creativity

355

o-te-tini showed his puka papa?anga to Stephen Savage, who wrote down the long narrative performed by Samuela and based on the genealogical notes and narratives in the book. At the end of his long narrative Samuela says to Savage: My friend, that is all I have to say about this book. The book my father wrote is as big as the Bible. The work of Tangaroa when he created the islands is all in that book. When he created us and all the other islands. Not this book, they are not all in here. In my father's book is the Ariki genealogy. It is very clear because he came from it. If only we had that book w e will not have any doubts. This story about the creation came from the beginning. Our narrators [are korero = the house, institution, of korero] have not reached it. Look my friend, at our korero. It says that it was Atea and Papa. But when it comes down to Papa's genealogy one's different from the other. They are not the same. There are many narrators but only one ancestor and that is Atea and Papa. When it comes down to Papa's genealogy they all have different legends. (Savage 1916: 32-33.)

Samuela refers to different family and island traditions which form conflicting interpretations of the same genealogies and narrative events. A genealogy forms an unbreakable chain which links everybody to the origin, to the divine world of gods and ancestors. Both a person and the whole ethnic group have a history which defines all the important factors constituting the identity. Because the genealogies reveal the ownership of the land, the right to chiefly titles, etc., they are constantly discussed, reinterpreted and recreated in everyday life. Richard Tapper has found the same kind of reinterpretation to be typical of the genealogical practice in Iran and Afghanistan. When discussing the popular discourses of identity, he writes: "Most groups do have genealogies or stories of common origins, but these represent not a pure line of descent from the origin or founders of the group but a disputed and changing ideological charter for present-day social and cultural relations among members of the group and a basis for claims as against other groups... Essential to these processes is the ambiguity of the names and labels concerned". (Tapper 1989: 237-38.) The ambiguity of the concept of the past, the discussion and reinterpretation of memorized knowledge are not typical features of genealogies only but can be applied to myth and history, too. The concept of history is a question of negotiation which both reflects and guides the definition of identity. Local politics affect the interpretations of mythic and historical traditions in the southern Cook Islands (see Siikala 1991:140-42). Jonathan Friedman says that the politics of identity consists in different ways of anchoring the present in a viable past. The past is thus "constructed according to the conditions and desires of those who produce historical texts in the present." (Friedman 1992: 207.)

356 Anna-Leena Siikala Because conflicting family traditions need to be negotiated again and again in disputes over land and titles, the authority of tumu koreros is constantly in dispute. The following statements are common in Mauke: "Tumu koreros are ashamed to tell the truth" (Teariki Maeva 1983) or "tumu koreros are liars, making up narratives" (Ngatuakana Aerenga 1984). The negotiations and disagreements of tumu koreros are also commonly known. Tera Moekaa, for example, from Mauke said: "There are some old people who know; but they have different stories these tumu koreros and the old people. So sometimes the tumu korero says O h , your story is wrong; where did you get your reference?' That's what they argue." (Tera Moekaa 1983.) Her husband Tara described the relationship of tumu koreros in the following way: "Because these people the korero they have different stories. You go to the other korero - different story. Those korero you tell yours and I tell mine and where it doesn't agree we make it agree." (Tara Moekaa 1983.) The creation of a manuscript form of korero was undoubtedly connected to the disputes over genealogies and the striving to testify the suitable descendency and legitimate land rights of chiefs and ordinary people. Puka papa9angas did, however, become a valuable instrument for those who performed tua tai?tos and other forms of korero. Most people interested in tradition have their family books. The tumu korero of Atiu, George Mateariki, for example, performed the whole series of narratives connected to the cosmic genealogy of Atiu following the order of happenings from the beginning of the world. As a script for his performance he had his puka papa">anga, a hand written book which contained lines of names. Picking out name after name, he entered into the narrative world connected to each of the heroes, creating during the narration a long saga of adventures. When we talked about the relationships of Atiu and Mauke, two competing islands, he used the genealogical line as his mnemonic aid, selecting the characters relevant to the occasion. His eyes closed, he entered into complicated life histories of selected heroes, explaining the events and occasionally checking a detail from the narratives written in his puka. George Mateariki was reacting to our need for knowledge; he created an Atiu history from the beginning of the world until the present day from the perspective of interisland political relationships. The tumu korero of Mauke in 1983-84, Tengaru Tobia, used, during the first discussions, his puka as a guide to his memory A talented narrator, he did not read the texts but checked during the discussions that all the relevant information was given in the right order and that the formulations of pe^es were right. Vaine No^omakere, a Maukean woman with several mataiapo titles and considerable knowledge of island history, always had her puka with her when she performed those tua tatfto she selected to introduce to us. It must be mentioned that it is not possible to interview the Cook Island tumu korero in the conventional folkloristic way. When tumu korero knows

Generic models, entextualization

and creativity

357

that somebody needs information, he or she prepares him/herself beforehand and sends word fixing a suitable date. The encounter is totally ruled by the performing tumu korero, the master of knowledge, who does not necessarily prefer answer to open questions even though (s)he anticipates the needs of the person seeking information. The situations where tua tai^tos were told were performances highlighted with mastery of vivid narrating techniques. The session ended with the last phrases of the narrative; the discussion of details came later. A similar preparatory phase including examination of puka papa^anga and possibly consultation with other masters belongs to the performances organized in connection with different social gatherings. People used to performing - va?ara of Mauke, Mapu Taia, and Tengaru Tobia in his later years as well as Pirioi Tairea - did not need the puka or preferred to perform without it. The use of a puka during performance would diminish the aesthetic effect and credibility of the performance. Talented and experienced narrators use it beforehand when the larger frames of knowledge or detailed information must be internalized. The narratives of puka are decontextualized as needed and recontextualized in different performances (see Briggs and Bauman 1992). Puka offers a script for organizing information both vertically and horizontally in giving the genealogical and schematic knowledge of narratives (on the cognitive processes of narration, see A.-L. Siikala 1990). It helps in memorizing the pe^es and in coding mythic information. Debate on the memorizing of great epics has been concentrating on oral performance and on the relationship between written and oral sources. Researchers have been far less interested in the role of manuscripts in the life of epic narration. Yet the manuscript form as a mnemonic instrument has an impact on the performance of the epic. Manuscripts crystallize the internal relationships of individual texts and give alternative models for the composition of larger entities. The narrative of Samuela represents one possible variant of the Central Polynesian epic.

An attempt to build an epos Samuela Terei was born on 9 December 1854 in Manua, Samoa, as the first son of a pastor, Tauna-o-te-tini, and his Rarotongan wife. Samuela Terei's father had a wide knowledge of Island groups. He had stayed in Samoa, in New Caledonia for five years, in Lifu, Mare and Rarotonga before going back to Samoa in 1852. Samuela Terei himself stayed in Samoa for 24 years before coming to Rarotonga on 1 October 1878. In 1895 he visited Mangaia with his wife and found signs of some primeval events included in his narrative. In 1916, when Samuela's narrative was written down by Stephen Savage, he was 62 years old and spoke both Samoan and Rarotongan Maori.

358

Anna-Leena Siikala

The 36-pages manuscript is interesting in many ways. It represents the special form of preserving a tradition: it is a manuscript based on an oral performance which is in turn based on a hand-written puka papa9anga created to replace the lost puka of his father, memorized by the narrator. Another source of tradition is mentioned at the end of narrative: a book in the ngati Tauanuu's are korero of Tui-manua-tere-ma-Samoa-atoa in Manua. Samuela seems to know the book because he states that some parts of the book resemble those of Jews in the Bible, some are different. The third book introduced at the end of the manuscript obviously refers to the Manuan manuscript: The Tauanuu man who wrote this book was Tauanuu-Sorai-^autere. He and a Mataiapo called Turifua. They were men from Oakari. This Mataiapo was fetched as witness. Sometime later Tauanuu-Sorai-^autere died and he was replaced by his son Varu. It was he, Tauanuu-Varu, who finished this book with Turifanua Metua Mataiapo, the father of Turifua Tangipo. (Savage 1916: 35.)

The passage reveals the importance of origin and authorship in korero: all participants are mentioned according to their contribution to the writing process. The authority of a performer is based on the authority of his predecessors. Father's book is described as "big as the Bible"; it contained the ariki genealogy according to his father's lineage. The comparison to the Bible is not made by chance; his book as well as his father's contains "the story of the creation". At the beginning of his narrative, Samuela borrows the voice of his father and ultimately - the voice of the Bible - at the end of the performance reveals that the echoes of his narration even have a wider sounding board. Even though we do not have access to the situation of Samuela's narration, it is possible to reconstruct the event on the basis of observation on the performances of later owners of puka papa^angas. The narrator does not only tell the story (s)he has in his puka, which is not necessarily written by him/herself but which (s)he obviously knows well enough to perform it, but discusses different aspects of the narrative and puts forward his/her own feelings and experiences. Savage's manuscript is written in Rarotongan Maori, which was the mutual language of Savage and Samuela and which seems to have been the language of Samuela's puka. He mentions that he wrote names in Samoan because "they are shorter than ours" (Savage 1916: 7). Samuela had been living in the Cook Islands for 32 years before his encounter with Savage. The last chapters of the manuscript reveal that he became interested in his genealogy at a mature age and possibly compiled his own puka according to his father's (and possibly his own by Tauanuu lineage) example using the new language. Samuela's narrative is a monologue containing not only the narrated events or lists of important names but also Samuela's own comments and inter-

Generic models, entextualization and creativity

359

pretations. The voice of Savage is missing, but the comments of Samuela are directed to him creating a dialogue in the text. Even though the content is full of cosmological knowledge, the narrative is not an orthodox piece of korero. It is more a personal view of the entire world history which puts the narrator in the middle of the scene of events. The cosmology is woven to form a world history and at the same time a life story of the narrator. The manuscript begins with a detailed description of the creation of the cosmos and gods and the origin of the island populations: In the beginning there was no sky, no land. Tangaroa was the God. He had nowhere to stay because there was no sky and no land. This Tangaroa had no father and mother - he stayed in the air. Tangaroa said: "Let there be a Papa." There appeared one. This became his threshold. Instead of just walking around in the air he now had something to stand on. And he saw this was good. He stretched out his right hand and another god appeared just like him. He called him Tangaroa also. He stretched out his right hand again and another god appeared again. He also called him Tangaroa. He turned left and another god appeared. He did this until there were many Tangaroas in the air. They became an army of three gods. Tangaroa named his first son Tangaroa Refuri (Tangaroa Ariuriurikore in our language), the other he named TangaroaSavari (Tangaroa-Teretere in our language). Tangaroa spoke to the Papa and another Papa appeared. This was called: Papa takoto followed by Papa toro Papa Raua^au (Papa Rauakau) Papa One Papa Karakaraea (clay sand) Papa Tu Papa Karetu There were seven altogether, Papa's first generation... (Savage 1916: 1-3.)

The first part of the text contains a narrative of the creation of the cosmos, gods, human beings, islands and natural phenomena. It continues by telling how the gods had a meeting where all but a couple of gods were allocated a portion of space. Everybody got a place to live in and that also established the order in the world of humans, their descendants. Comparison of Samuela's Maori text with "The Samoan Story of Creation" published by John Frazer in Journal of the Polynesian Society in 1892 in both Samoan and English shows that the first part of Samuela's narrative follows the myth Frazer got at Malae-la (where Samuela's father worked before coming to Manua in 1852) not only in structure but also in many details, sometimes even in coding (despite the language). Both myths begin by placing Tangaroa, the god, in the air, in the expanse, where there was no sky,

360

Anna-Leena Siikala

no land: Ko Tangaroa te atua ('Tangaroa was the God') / Ο Tagaloa le atua ('Tangaroa the god'); Kare e rangi, kare e enua ('no sky, no land') / uana ο ia e leai se Lagi, e leia se Nu?u ('not any sky, not any country'). The phrase is repeated in both myths: in Samoan using the different coding e leaifo?i le Sami, ma le Lau-?ele?ele ('there was also no sea, and no earth') and in Maori word by word kare e rangi, kare e enua. In both myths the first threshold of Tangaroa is Papa, the Rock, and the creation is performed either by using the word or "reaching out a hand". The beginning of both variants resembles the creation myth of the Bible, which may explain the similarity of phrases. The sentence Kua kite aia, i te meitaki ('He saw this was good') links the description of the creation act to the Biblical discourse. On the other hand, the creation of the layers of the universe, heavens and gods follow the same non-Biblical pattern in Samoan and in Maori variants. A good example of a striking similarity is the creation of man, in which the same phrases are used (Savage 1916: 4; Frazer 1892:176). The first part of Samuela's narrative does not follow the traditional pattern only in the main content consisting of a chain of myths of origin, but also in sequencing the events, using list formulas to show the genealogical relationships between narrated events, metaphors, naming and in phrasing the important parts of the narratives, e.g. creation acts. Samuela seems to be well aware of the Samoan oral tradition telling about the beginning of the world. Richard Bauman and Charles L. Briggs have stated in talking about generic intertextuality: "The invocation of genre... provides a textual model for creating cohesion and coherence, for producing and interpreting particular sorts of features and their formal and functional relations all the way from particular poetic lines to the global structure of the narrative" (Briggs and Bauman 1992: 147). By using generic models Samuela creates the special meaning for his text, and formulates its status: it is a recontextualization of prior, authoritative discourses mediated to him by his father and other experts. Furthermore, Samuela does not only invoke generic models but minimizes the intertextual gap between his own performance, based, as some comments show, on entextualized narrative of his puka, and its generic precedents. This kind of strategy is a powerful instrument for creating textual authority (Briggs and Bauman 1992:148). Then, in the middle of the narration, on page fifteen of the manuscript, something happens. Samuela has turned his attention from mythical events to a period of wars between island groups. Turi became Tangaroa's shadow. During the time of a battle he was the war god. He would lead the army into battle. He would fly in the front. He was like a bird (Kuri). Rongo, a god also, is the image of Tangaroa from Manua. Anything that would happen, Rongo would tell about it before it actually happens (like a prophet).

Generic models, entextaalization and creativity

361

Everything that happened, he had told about it before. He also followed the battles. When he flies with the Kuri, in the front, the army would be happy. When Rongo turns away the army would turn back. Rongo is like a bird (Ruru). These were the images of two Tangaroas, the other is Kuri and the other is Ruru. Turi in Samoan and Kuri in Rarotongan. When the Ruru is seen, the people would say "Here is Tangaroa, he is Rongo". Then they would look at what the bird would tell them. They would see where he came from and where he would go to. These were all signs. This was the same for Kuri. When he is seen, the people would say, "Here is Tangaroa, here is Turi." Then they would look at what he would tell them. These were their roles. From the times of the heathens until this day. Their land never slaved to another island. They carried any weapons. The islanders were all killed. No one would return to their home, Tonga fought against them three times and they never won.... This is my own story of what I have actually seen. I am Samuela Terei, a son of Tauna-o-te-tini... (continues with father's short life history). I was born in 1854, on 9 September. When I was 13, there was a battle between these two islands, this was the year that afire started from under to ocean between Taua and Orosenga. It was during that battle that I actually saw the powers of these two birds, I was only a child but I was old enough, 13 years. Like mentioned before, it is true, just as the battle started the Ruru arrived. These two birds would not land anywhere else, they would go straight to Maraes, Marae-Vauvau, Marae-Tere and sometimes Marae-a-totoa... (Savage 1916:16.)

The historical war between Tonga and Samoa, in the year of volcanic eruption, merges with battles testified by Samuela, as a boy of 13. Before describing the events, he introduces himself in a proper way. The details given of his father's life are important, because his father connects Samuela to the Manuan ariki lineage and at the same time to Manuan history. The thorough comments on the bird shapes of Rongo and Tangaroa show that these battle omens have special meaning for Samuela which later becomes obvious in the words "this is my own story of what I have actually seen". The emotional charge of an unusual experience - Samuela has seen signs of Gods, the birds, during the pressing time of battle - transforms the text from general world history into his own life history. Samuela returns after the intervention of personal feelings to his mythical considerations. The phrase "This happened from that time until this day" (Savage 1916:17, cf. "From times of the heathens until this day", p. 13) describes the everlasting character of the bird signs and, at the same time, builds a bridge from Samuela's own experiences to the events of primeval time. With the words "I will go back to Rongo and Turi" he returns to the world of the Beginning. The events of cosmology and later history create the island world, islands which form the scene of Samuela's life. The connecting links between the Beginning and Today are the names given by the God Tangaroa during his creation acts.

362

Anna-Leena Siikala

World history is rooted in the landscape by names and signs which testify to the primeval happenings. Samuela's narrative continuously follows the korero model in its form. The basic metaphor of cosmology is the genealogy, articulated by using list formulas and the phrase anau mai = was born. Gods are born from the body of Tangaroa, the layers of the world are born from each other. After birth everything is named. Samuela refers to the importance of names by saying "As for the names he saw, they remained as signs for him." (Savage 1916: 23.) In the last part of his narrative, he brings his own experiences and evaluations into his narration more and more. The structure of the narrative becomes looser and the coherence no longer lies in the replication or the reproduction of the precedent discourses. Instead of minimizing the intertextual gap, Samuela enlarges it by references to his own life framed with explicit markers of authorship: "This is my own story" (Savage 1916: 15,20). Widening the intertextual gap gives room for a creative thinker who debates the heard and seen. Like Tengaru Tobia later, Samuela wonders what the truth is in korero: "I think, these are my own thoughts", "I think, this is what happened" (Savage 1916: 23, 24). Samuela's pursuit of the truth is connected to the process of the invention of self. The reconstruction of the inherited knowledge of the korero serves rhetorics of self-making as well as the historical awareness of lineage members. The crucial point of Samuela's narrative lies in the lines: "TangaroaRefuri was the Ariki of the ninth heaven. He gave his name to the Ariki of Manua. This was what he said, 'My Ariki title I give to you. My power I give to you, my name I give to you. This is the name I give to you and that is Tui. Here is your full name: Tui-o-Manuatere-ma-Samoa-aloa'." (Savage 1916: 11.) Samuela was the first son, first born, in a line of descent which through his father was connected to the ariki genealogy of Manua: "In my father's book is the ariki genealogy. It is very clear because he came from it." (Savage 1916: 32.) The quest for identity, and in this case, the prestige and feeling of importance connected to a hierarchically high lineage, is a motive for Samuela's intellectual speculations. He has to fill in the gaps in the genealogies, to find features connecting distant events and signs which testify to the truth. The description of vivid dialogue with a man called Taivairanga in Mangaia, where Samuela searched for signs of Rongo, does not follow the conventions of puka manuscripts: In 1895 I went to Mangaia, I saw everything that I had heard about that island. There were octopuses there. A man called Taivairanga took us to the mountain that looks like a man who was lying face down from the head to his hands and legs. The water was flowing from it as if this man was urinating, lying face down. I asked about the name of the mountain and was told that it was called Rangi Motia, torn heaven. It was strange, the front was placed at the back and the back in front but still it looked the same. There was no change.

Generic models, entextnalization and creativity

363

I asked who is this man lying face down? He replied, that is Rongo. I asked, where did he come from? He answered, that is our ancestor. I asked again, where did he come from? He replied, he is an ariki, he is from here. I asked, if he is an ariki, why did not the people of this island bury him in the soil? He replied, because he will not fit. Look at the island, it is as long and wide as he is. We have a good ariki, he is feeding us. Look, the head is on one side and on the other side are some food storage places. He is giving food for his children to eat. Look at his armpits, those are the storage places for the food, and the water, he is feeding his children with his milk from breasts. Look at his feet, there are also some storage places and also water. His children are eating and we are drinking his urine. Then my wife laughed as she also came with me. My wife was from that island called Manua. She also knew the narrative (tuatua) of Rongo. This was my last question, "What is this island's first name?" He replied, "Auau." I asked, "What does it mean?" He replied, "I am not sure." I looked at the lay-out of the island, there were holes for the octopus. I think, these are my own thoughts, it was the first man who arrived on this island. He knew about the legend of Rongo, he had looked at the island like I did when I arrived on Mangaia and he had lied that Mangaia was his own island and the man lying down on the mountain was his father. As for the names he saw, they remained as signs for him. The stories we hear from other people, these people may be the children of this first man who had lied. So the lies of this first man were believed by everyone from that time until now. We knew about the god Rongo, he was by himself on his part of the rock. He died on this rock and he became a mountain as he was floating on this rock, which became an island. This is just as I thought. I think this was what happened. Some of the people of Mangaia say Ruanuku was the first man and some say it was Tangaengae and his four brothers, Teakaiaroe, Temanavaroa, Timatekore and Vari. This is what I think. I think Ruanuku was the first man. He came from Manua and he knew about the story about Rongo. He can then say that Rongo was his father so the story then would not be wrong as he knew about it. We will now leave the story about Rongo and go back to the time when Tangaroa-Teretere and Tangaroa-Faatutupunuu created the islands... (Savage 1916: 21-24.) Because Samuela knows the Samoan narrative of Rongo punished by Tangaroa to float away lying on a stone, he cannot believe the historical explanation given by the Mangaian man. The exactness with which the dialogue is described tells that the matter is important to Samuela. According to his knowledge, Rongo was a god, not the ancestor of the Mangaians. W h o w a s lying? Samuela comes to the conclusion that the first inhabitant of Mangaia was Ruanuku (one of the generally known heroes in Central Polynesia and N e w Zealand) from Manua, the original homestead of all islanders. He was the liar. Disbelief, doubts, inversion, borrowing and integration of disparate ideas and contexts are tools of creative imagination necessary in producing a coherent, personally convincing discourse on mythic history. Invention has as great a role as the convention in the reconstruction of tradition. Roy Wagner

364 Anna-Leena Siikala sees that invention and convention stand in a dialectical relationship to another: it is an invention that constantly recreates the convention, collective orientation in culture, and convention that continually facilitates its own reinvention (Wagner 1981:52). The creative reconstruction of mythic-historical knowledge both in individual performances and at the level of "mental text", as Lauri Honko formulates the totality of internalized tradition and associated performance strategies (Honko 1995:131-32), is a condition for its perpetuation. The fact that Samuela gives the gods and heroes names in both Samoan and Maori shows how he consciously compares different island traditions in creating his own view of the world history. He also compares the three cosmologies he knows: the Rarotongan, the Manuan and Old Testament: Their name is Tangaroa-Faatutupunuu, the Jews' name is Jehova, theirs is Tangaora-Refuri, the Jews' name is Iesu, theirs is Tangaroa-Tetetere, the Jews' name is Vaerua Tapu. When we talk about this, there are lots of the Ariki families and also the ordinary families. These are not clear. This is the same with the Jews. Theirs started with Po Akiriva. Ours starts with: Po Akiriva - Enua uri Po Akiriva - po tango Po Akiriva - aa atu, aa atu. (Savage 1916: 36.)

His final sentence is: "I do not believe in ours, I believe in Manua's" (Savage 1916:36). He has found - after comparison - the basis for his world view and self definition. The published version of the Manuan cosmology testifies that Samuela's narrative begins by following traditional pattern. He has extended the narrative so that his life scenery, all the islands important to him, are included in the narrative as well as those of his own experiences which have a connection to the sacred genealogy of the earth. As a good tumu korero does in the considerations creating the base for oral performances, Samuela compiled an epic, an epic revealing the ariki genealogy of Manua, connecting himself to the beginning of the world through the line of the ruling chiefs.

To conclude: from superstories to superdiscourses The knowledge Samuela conveyed in his narrative is part of a socially significant discourse structuring an understanding of the world. His attempt to entextualize his knowledge of world history can be seen as an attempt to build an epos. It could not be Samoa's or the Cook Islands' Kalevala because of its personal bias; it could not serve the purposes of a nation-building project. Nevertheless, it sheds light on the problem of the textualization of Central Polynesian oral narration.

Generic models, entextualization and creativity

365

The view held by the 19 th century Romantics of an ancient pro to-epic that survives in fragments in folk poetry and that can be put together again to form one great entity was a source of inspiration to Elias Lönnrot and other compilers of the great world epics. The functionalists and performancecentered folklorists have in turn stressed the necessity of seeing folk poetry as individual folklore items connected to the performers and the performance situation. In the light of Cook Island epic narration, it seems that the individual texts make sense only when their generic intertextuality is taken into consideration. They should not be seen as individual performances in random situations, but as addresses, either statements or interpretations, to larger socially meaningful, sometimes even crucial discourses. We can better understand the nature and meaning of epic narration through Samuela's failed epos than through modern narrative collections. The problem of textualization is elsewhere: whose path of ancestors should we follow? Whose interpretation of authoritative tradition should we accept? Processes of the construction and invention of Tradition have been especially visible in the creation of the world's epics (see Honko 1990a). The European literary eposes differ both from one another and from the epics of Asia and Africa not only in content but also in the way and the period they were noted down. The relationship between the finished epos and the collection of oral poems telling of the deeds of the same hero is also a problem. Nevertheless, there are common features in these traditions. Lauri Honko calls the epic, as Brenda Beck did in 1982, a superstory, a folk genre carrying an image of first-rate narration (Honko 1990b: 13; Beck 1982: 196). Folk epics are according to Beck and Honko long poems often running to thousands of lines sung or recited by "professional" tradition bearers. Their special value, sacredness or cultural significance place them above all other stories. Myth and history are woven into them to form true, trusted events, so that they become mirrors of the world, the nation and its early heroes. The common features of the world's superstories can be outlined in the following way: 1) A long epic narrative or a cluster of independent poems. The processes of combining and dismantling. 2) Poetic form, prose form or a combination of these. 3) Performance: professionals or specialists in oral art; literature. 4) Interaction of oral and literary traditions. 5) Mythic historical content. Mythical or historical bias according to the society in question. 6) Openness to multiple interpretations. 7) Special cultural value; sacredness.

366

Anna-Leena Siikala

8) Male tradition, even though women might act as performers. 9) Connection to the history of kinship/ethnic/national groups: political and ideological values. 10) Power, hegemony, political dreams. Most of these features considered to be typical of epic also characterize larger mythical-historical discourses which are non-coherent but marked by generic intertextuality, such as the korero of the Southern Cook Islands, and which form fertile soil for building the eposes. It seems that the poetic majesty of the epic genre has hidden the cultural domain it represents - to which it is a channel or a path - and at the same time the multidimensional discourse which is the source of the political power of the epic. In order to understand the creative reconstruction of epic tradition, and the problems of entextualizing that tradition, we should examine besides the epics the generic models of "superdiscourses", hot or marked forms of culturally significant discourses behind the finished epics. Superdiscourse should then be understood as a historically established understanding of the social hierarchy and power relationships and their execution. This discourse is not a reflection of ruling power, or hegemony but it is a basis and at the same time the generic force in the social processes establishing and legitimating the order of a society. Nor is it a discourse in the sense of Michel Foucault, finalized by hegemony, but open to new ideas and interpretations of changing hegemonies which make it durable in time.

References Anderson, Benedict 1983 Imagined Communities. Reflections on the Origin and Spread of Nationalism. London: Verso. Bakhtin, Mikhail M. 1985 The Dialogic Imagination: Four Essays. Michael Holquist (ed.); Caryl Emerson and Michael Holquist (trans.). Austin: University of Texas Press. 1986 Speech Genres and Other Later Essays. Caryl Emerson and Michael Holquist (ed.); Vern W. McGee (trans.). Austin: University of Texas Press. Bauman, Richard and Charles L. Briggs 1990 Poetics and Performance as Critical Perspectives on Language and Social Life. Annual Review of Anthropology 19: 59-88. Beck, Brenda E. F. 1982 The Three Twins. The Telling of a South Indian Folk Epic. Bloomington: Indiana University Press.

Generic models, entextnalization and creativity

367

Beckwith, Martha 1981 The Kumidipo. A Hawaiian Creation Chant. Translated and Edited with Commentary by Martha Warren Beckwith. Honolulu: The University Press of Hawaii. Briggs, Charles L. and Richard Bauman 1992 Genre, Intertextuality, and Social Power. Journal of Linguistic Anthropology 2 (2): 131-72. American Anthropological Association. Chadwick, Η. Munro and N. Kershaw Chadwick 1986 (1940) The Growth of Literature. Vol. Ill, Part. II. The Oral Literature of Polynesia. Cambridge: Cambridge University Press. Chariot, John 1987 The Kamapua?a Literature. The Classical Traditions of the Hawaiian Pig God as a Body of Literature. (Monograph Series, No. 6, published by The Institute for Polynesian Studies.) La^ie, Hawaii: The Institute for Polynesian Studies, Brigham Young University, H a w a i i Campus. Dumezil, Georges 1970 The Destiny of the Warnor. Chicago: Chicago University Press. Eliad, Mircea 1954 The Myth of the Eternal Return. (Bollington Series 46.) New York: Pantheon Books. Fine, Elizabeth C. 1984 The Folklore Text: from Performance to Print. Bloomington: Indiana University Press. Frazer, John 1892 The Samoan Story of Creation. Journal of the Polynesian Society, vol. 1:164-88. Wellington. Friedman, Jonathan 1992 Myth, History and Political Identity. Cultural Anthropology vol. 7, nr. 2:194-210. Gill, W. Wyatt 1911 Extracts from Papers of the Late Rev. W. Wyatt Gill, L.L.D. Journal of the Polynesian Society, vol. 20:116-51. Wellington. Hanks, William F. 1987 Discourse Genres in a Theory of Practice. American Ethnologist 14 (4): 668-98. Hastrup, Kirsten 1987 Presenting the Past, Reflections on Myth and History. Folk, Journal of the Danish Ethnographic Society, vol. 29:257-67. Copenhagen: The National Museum of Denmark. Haviland, John B. 1996 Text from Talk in Tzotzil. In: Silverstein and Urban 1996: 45-78. Hobsbawm, Eric and Terence Ranger (eds.) 1992 [1983] The Invention of Tradition. Canto: Cambridge University Press. Honko, Lauri (ed.) 1990a Religion, Myth, and Folklore in the World's Epics. Berlin and New York: Mouton de Gruyter.

368

Anna-Leena Siikala

Honko, Lauri 1990b

The Kalevala and the World's Epics: An Introduction. In: Honko 1990a: 1-26. 1995 Multiformit ja pitkän eepoksen arvoitus. Sananjalka 37: 117-45. Turku. 1998 Textualising the Siri Epic. (FF Communications 264.) Helsinki: Academia Scientiarum Fennica. Hugh-Jones, Stephen 1989 Waribi and the White Men: History and Myth in Northwest Amazonia. In: Tonkin et al. 1989: 53-70. Hymes, Dell 1981 "In Vain I Tried to tell You": Essays in Native American Ethnopoetics. Philadelphia: University of Pennsylvania Press. L£vi-Strauss, Claude 1973 The Savage Mind. Chicago and London: The University of Chicago Press. Lowenthal, David 1990 The Past is a Foreign Country. Cambridge, New York, Port Chester, Melbourne and Sydney: Cambridge University Press. Luomala, Katharine 1949 Maiti-of-a-Thousand-Tricks: His Oceanic and European Biographers. (Bernice P. Bishop Museum Bulletin 198.) Honolulu: Bernice P. Bishop Museum. Mauke korero s.a. Cook Islands Archives. Rarotonga. Ricoeur, Paul 1991 From Text to Action. Essays in Hermeneutics, II. Trans, by Kathleen Blarney and John B. Thompson. Evanston, Illinois: Northwestern University Press. Sahlins, Marshall 1985 Islands of History. Chicago and London: The University of Chicago Press. Savage, Stephen 1916 Ε Buka Tuatua Enua. Papaanga Tupunua, Papaanga Ariki, Mataiapo, Kopu Tangata. A manuscript. Cook Islands Archives. Rarotonga. 1962 A Dictionary of the Maori Language of Rarotonga. Wellington: Department of Island Territories. Schrempp, Gregory 1985 Tu Alone Was Brave: Notes on Maori Cosmogony. In: Antony Hooper and Judith Huntsman (eds.), Transformations of Polynesian Culture, 17-37. Auckland: The Polynesian Society. Seydou, Christine 1990 Identity and Epics: African Examples. In: Honko 1990a: 403-23. Shore, Bradd 1982 Sala'ilua: A Samoan Mystery. New York: Columbia University Press.

Generic models, entextnalization and creativity

369

Siikala, Anna-Leena 1990 Interpreting Oral Narrative. (FF Communications 245.) Helsinki: Academia Scientiarum Fennica. Siikala, Jukka 1991 ?Akatokamanava. Myth, History and Society in the Southern Cook Islands. The Polynesian Society in Association with the Finnish Anthropological Society. Auckland. Silverstein, Michael and Greg Urban (eds.) 1996 Natural Histories of Discourse. Chicago and London: The University of Chicago Press. Tapper, Richard 1989 Ethnic Identities and Social Categrories in Iran and Afganistan. In Tonkin 1989: 232-17. Tedlock, Dennis 1983 The Spoken Word and the Work of Interpretation. Philadelphia: University of Pennsylvania Press. Tonkin, Elizabeth, Maryon McDonald and Malcolm Chapman (eds.) 1989 History and Ethnicity. (ASA Monographs 27.) London: Routledge. Urban, Greg 1996 Entextualization, Replication, and Power. In: Silverstein and Urban 1996: 21-44. Valeri, Valerio 1990 Constitutive History: Genealogy and Narrative in the Legitimation of Hawaiian Kingship. In: Emiko Ohnuki-Tierney (ed.), Culture Through Time: Anthropological Approaches, 154-92. Stanford, California: Stanford University Press. Wagner, Roy 1981 The Invention of Culture. Revised and Expanded Edition. Chicago and London: The University of Chicago Press.

Index Aarne, Antti 109 Abasi, Bakari, a folk poet 249 c Abd al-Rahmän al-Abnüdi, a folk poet 271 Abu Bakari Bin Mwengo, a folk poet 249 Academic Transcription/Orthography 133,136 141, 142, 143 accuracy in documentation, publication 3, 8-9, 11, 29, 31, 36, 40, 50, 75, 240-43 Adam/Adamu 248, 261 adaptation of tradition 13, 15, 17, 20, 22-24, 26, 34, 44, 142, 210, 217-18, 222, 316 Agboghidi epic 49, 282-84 Aeneid epic 257 Afanasyev, Aleksandr 318 African epics 7, 37, 46-48, 237-87, 365 Ahanon, Robinson 284 Aiguobasimwin, a prince, later Oba Eweka the second 282 Ainu epic 130-31 Argonauts 317 Ariki 348-49, 355, 361-64 Ajjeru subepic 227, 228, 230 Akamatsu, Tsutomu 123 Alamji-mergen epic 135-37 Ali, Yahya, a singer 250 Allen, John W. T., and Mrs. 248, 254 allusive, implied knowledge 13,16,22, 37, 40, 50, 76, 83-84, 189, 196, 231, 352,354 Aloni, Antonio 67 Alpämis epic 109 Altai-Tuvinian, Altai-Tuvinians 133 Altai-Tuvinian hero-tales 130,133-34 Altaian (heroic) epic 131-33 Alur epic/language/people 247 Amanzolov Erezep-zi'raw, a singer 120 American folk heroes 237 American Indians 280, 281 Andersson, Theodore M. 89

Andrzejewski, B. W. 241 Anglo-Saxon epic, see Beowulf Ankole epic/language 247 Annamese 44 Anochin, A. 131 Antoine, Robert 206, 209 Anusäsanaparvan 194, 199 apprenticeship, see specialization in epic traditions Arabic 272, 322 - pan-Arab oral tradition 264 Aranyakända 196,198 archives, creation of 4,11, 30,66, 74, 78, 82, 84, 104,131, 137, 143, 273, 344 Archives of the Center for the Folk Arts at Harvard University 273 Archive of the Serbian Academy of Sciences 74 Arjuna (as Pärtha) 195 art of writing, the introduction of 4, 39, 57, 345 Asian epics 73 Asiatic Society of Japan 131 Asvamedhikaparvan 199, 201 Atea 347, 349, 355 Athena, the patron goddess 66 Atlakviöa, an older Atli poem 91-92 Atlamäl, a younger Atli poem 89, 9 2 93 audience reaction, interaction with real or ideal audiences 7, 12, 17, 19, 2 3 28,35,39-41,45-46,48,51,59-60,62, 64, 66, 83-84, 90, 95, 106, 112, 129, 132, 153, 156, 167, 193, 202-03, 209, 217-18, 220, 224, 226, 231, 254, 257, 263, 265, 269-74, 284-85, 298, 314, 335) 345 - see also projected audiences (in editing epics), reception audio recording 13, 28-30, 7 7 - 7 9 , 1 0 4 06,114,130-33,135,138-39,143,146, 148, 219, 228, 231, 239, 251-53, 263, 265, 271, 273, 275, 284

372

Index

audiovisual recording 9, 13, 28-30, 35, 45, 106, 129, 143, 163, 219, 228, 231, 263 authenticity 3, 8, 45-46, 132, 148, 202, 237-45 authorship of oral singer 4-5, 14, 24, 36-38, 41,43, 58, 89-90, 95,206, 358, 362 Avrorin, V. A. 144-45 Avvakum (Archpriest A.) 164 Ayodhyäkända 194, 198, 203, 206, 212 Ayzanstadt, M. 139

Babylonians 61 Bajgoric, Halil, a singer 83 Bakhtin, Mikhail 281-82, 285, 353 Bälakända 194, 197, 198, 200, 203-04, 209, 212 Baldaev, S. P. 135 Balinese 44 Ban! Hiläl ("sons of Hiläl"), a Bedouin tribe 263-64, 271 Bantu 58, 247 Barnakov, B., a singer 135 Bartok, Bela 80, 122 Batchelor, J., Revd. 130-31 "Battle of Kosovo" (in 1389) 75, 84 "The Battle with the Monster Sire Ninata" 134 Bauman, Richard 15, 105, 360 Beck, Brenda E. F. 365 Becker, Jörg 144 Beckwith, Martha 352 Belcher, Stephen 280 Beldy, Ivan Torokovic, a shaman 16887 Beldy, Lindzja, a shaman 168-87 Beldy, Marija Vasil'evna, a shaman 168, 170-71,173-85 Beljaev, V. 104 Ben-Amos, Dan 48-49, 61, 293 Bender, Jo 294, 306, 316, 317, 325, 327, 334 Bengali 44 Beowulf epic 7, 39, 40, 41, 71, 82, 8384, 95, 255, 258, 273

Beramma, a god 228 Bergen rune 95 Bergman, Ingmar 93 Bhalo, Mohammed, a singer 249 Bharata 207 Bhärata epic 210 Bible 60, 242, 281, 355, 358, 360 Biebuyck , Daniel 244, 279 Billavas 222, 226 Bishop Njego5 84 Bjarkamäl 91 Boas, Franz 11, 293, 298, 312 Bogatyrev, Petr 5 Böhtlingk, Otto 133, 139-40, 143-44 Bol'sezemlja dialect 147 Bonjour, Adrien 83 book as a medium 17-18, 30, 34-35, 39, 46, 59-60, 71-72, 76, 85, 91, 94, 139, 221, 243-45, 275, 354-56, 358, 362 book culture 60, 66, 354-56, 358 Boon, James 354 Bosley, Keith 323 Bourdieu, Pierre 344 Bowra, C. M. 203, 279 Bradbury, R. E. 49, 283 Branch, Michael 323 Briggs, Charles 15, 360 Bright, William 313, 321 British Institute of Recorded Sound 251 Brockington, John 43-44,195,197,198, 200, 201, 212 Brockington, Mary 212 Buettner, C. G. 247 Buffalo, John Rush 314-15 Bulatova, Ε. V. 137 Bulgakova, Tatjana 161, 164, 172, 173, 181 Bulgarian tradition 292, 297, 317, 31920, 322 Burak 256-57 Burjat epic 42,129,134-37 Burne, Charlotte S. 237 Bynum, David 80

Calumniated Wife 109 Cambodian 44

Index Cantilena, Mario 67 Carpenter, Rhys 65 Casträn, Μ. Α. 145-46,151, 161 Catenacci, Carmine 67 CD-ROM 35, 46,114, 123, 243-44, 272 Center for the Folk Arts, Cairo, Egypt 272, 273, 274 Central-Asian epics 7, 38 Chariot, John 353 Chinese 4 4 , 1 4 4 , 1 6 4 , 280 Chinookans 293, 297, 302 - Kathlamet Chinook 298, 313, 335 - Wishram Chinook 302, 303, 310 Christianity, Christians 7,165, 248 chronographic (vs. episodic) epic 6465, 67 Church Missionary Society 247 Church Slavonic 147 El Cid epic 7, 39, 258 Cincius, V. 1.137 Civdavsl epic 137 Clover, Carol J. 89 Codex Regius 91 cohesion in narrative 16, 22, 24, 33, 44, 48, 360 collateral action (in the performance of oral epics) 12,22,24-26,45,217-219, 231 collateral genres (as part of epic performance) 12, 34-36, 273 collation 8, 29-30, 253 collective creation 4-6, 37 collector's impact 3, 8, 17, 30, 40, 42, 46,57,61,65, 71-74, 77-81,83-85,91, 104-05, 131, 133, 140-42, 148, 15556, 224, 226, 238-45, 251, 254, 258, 265, 272, 275, 292, 344-45, 351, 354 communicative event (as replacement of text) 16,114,122 competence - cultural 12, 20-21, 84, 165-68 - narrative 8, 22, 220, 317 - performative 12, 17, 20-21, 218-19, 318 - thematic 17 compilation, composite epic 3, 8-9,14, 37, 45-47, 202, 206, 239-40, 253, 344,

373

358, 364-65 composition, see epic discourse as language, mental text, sequencing composition-in-performance 12,24,77, 267-68 compromises in documenting, editing and publishing oral epics 9, 29-31, 34,42,47,50,130,218,240, 264,27374 context 7, 12-13, 15, 17-19, 20-29, 3 3 39, 42, 45-47, 58, 60, 73, 79-80, 90, 93, 95-96, 113-14, 122, 163, 217-18, 220, 223-33, 238-43, 245, 258, 26566, 268-69, 271-74, 301, 313, 319, 343-47, 351-52, 357, 360, 363 - see also context of documentation, "strong" context context of documentation - natural 17, 29, 46, 80, 218, 227-28, 230, 233, 238-41, 243, 258, 266, 271 - induced 17, 23, 29, 34, 45-46, 218, 231, 233, 238, 240-41, 243 - interview 13,17, 29, 34, 42,163, 218, 231, 266-67, 356 contextualization 15, 36-37, 76, 274, 343, 345, 352, 357, 360 Cook Islands Archives 347, 354 Cook Islands epic tradition 343-66 Cooke, B. G. 123 copying 7,11,39, 51,60, 62,66,152, 344 copyright (of oral work) 37 co-text 3, 33 "Coyote and Eagle's Daughter" 314-15 "Coyote and Frog" 310, 314 "Coyote, Jack Rabbit and Eagle's Daughter" 315 Crystal, David 111, 123 Cultee, Charles, a narrator 313 cultural identity 27-28,51,75,226,299, 343, 345-46, 355 Curtin, Jeremiah 302 Curtis, Edward S. 302, 334, 335 cycle of epic lays 40,65,74,76,104,109, 129, 134, 206, 223-24, 248, 265, 29192, 302, 334-35

374

Index

Danish Academy 59 Danish ballads 93 Dammann, Ernst 248, 249 Dasaratha 195, 198, 207 Daskalova-Perkowski, Liliana 317,322, 336 Davids, T. W. Rhys 212 decontextualization 15,36-37,343,352, 357 Dede Korkut, Book of 108 Delgani epic 137 Demetracopoulou, Dorothy 294, 31516, 334 Densmore, Frances 293 Derridean grammatology 14-15 description (as narrative unit) 17-19, 33, 73,197, 204, 207, 220 Dhrtarästra 202 dictation (of epics) 7, 8, 28-29, 39-40, 60, 63, 66-67, 73, 75, 78-79, 131-33, 136,138-39,141,147-48,153-54,156, 202, 220-21, 232-33, 241, 249 difficulties in documenting oral epics 9, 42, 47, 111, 115,135,141,151, 224, 226, 255, 268, 270-71 Djerdjelez Alija 84 Dmitriev, P., a singer 134 Doane, A. N. 14 Dobrovol'skij, Β. M. 129-30, 138, 14648 documentation strategy 3, 9 , 1 3 , 1 7 , 2 3 , 28-31, 35-36, 40-41, 45, 47-48, 73, 103,105,107, 111, 115, 122,163, 217, 231-33, 240, 252, 263, 272-74 Donner, Kai 161 Dorson, Richard 239 Double Scene 95 Draupadi 207 Dravidian languages 219, 222 Dronke, Ursula 92 DuBois, Cora 315, 334 DuBois, Thomas A. 311, 317, 322 Dugarov, A. S. 135 Dunn, C. J. 130 Durväsas 207 Dzabbarov, A. 121

The Eddas 7, 39 - Snorri's Prose Edda 41, 90-91 Eddie elegy 92-96 Eddica minora 94 Edige epic 108-09, 123 editing, editor's impact 3, 7 - 1 0 , 1 7 , 3 0 31, 33-34, 36-37, 39-41, 45, 48, 51, 57-58,60-61,67,74-78,80,82,84,85, 103-07,109-15,122-23,130-31,132, 135,136-37,139-43,145-46,148-53, 196, 223, 253-55, 259, 263, 266, 26974, 279, 285, 293 - see also mental editing Edo 48, 49, 284-86 Edo epic 281-86 Egharevba, Jacob 282 Egill Skalla-Grimsson 94-95 Egils saga 95, 96 elegy 41, 89, 93-96 Eliade, Mircea 168, 347 Emel'janov, Ν. V. 140,143 epic, categories of - literary epic 5-9, 14, 37-39, 43, 73, 76, 81-82, 209-11, 254, 258, 275, 279, 282, 365 - oral epic 3, 6 - 9 , 1 7 - 2 2 , 2 4 - 2 5 , 2 9 - 3 1 , 33-47,51,57,59-62,71-72,75-77,8182, 84-85, 103-06, 109, 115, 122-23, 135, 193-94, 201, 203, 205-07, 21727,238-39,245,254,263-64,267,270, 273-75, 291 - tradition-oriented, traditional, oralderived epic 6-9, 14, 29, 37, 71, 76, 82-83, 85 - see also long oral epic epic cycle, see cycle of epic lays epic discourse as language 5, 8 , 1 7 , 1 9 , 21, 26, 46-47, 66, 76, 79, 82, 85, 193, 202, 209,218,232,241,243,249,25455, 267, 272, 281, 284, 297-98, 315, 317, 353 - see also epic idiolect, epic register epic heroes, heroines 22, 28, 42, 65, 75, 81, 84, 92,109, 111, 113,117,121,134, 140,149,150,153,155,195,207,22022, 225-26, 232, 237, 239, 248, 261, 267, 271, 273, 282, 307, 318, 318-19,

Index 344, 346, 347, 350-53, 356, 363-65 epic idiolect 8,17, 21, 31, 79, 218, 220 Epic Occasion 156 epic register 7 , 1 7 , 1 9 , 2 1 , 3 5 , 7 9 , 8 5 , 2 1 8 , 220, 320 episode (as narrative unit) 16, 19, 2 4 25, 28, 30, 37, 47-48, 64-66, 83-84, 170, 172, 193, 203, 208, 224-25, 228, 233, 245, 252, 265-66, 269, 271-74, 302, 304, 352-53 episodic epic, see chronographic (vs. episodic) epic fir Soyotox ("The Solitary Hero") epic 140 Erdely, Stephen 80,122 Erdälyi, Istvän 150 Ergas-säzanda Zakirov, a singer 122 fergis, G. U. 139-43 fcrzakovic, B. G. 120 Esemurat-fc'iraw Nurabullaev, a singer 108 ethnomusicology 32, 4 7 , 1 6 1 , 1 7 2 , 244, 267, 274 ethnopoetics 12,25,49, 50,85,105,107, 111, 146, 291, 295, 311, 323 European epics 38-41, 57-99, 258, 280, 365 Even (Lamut) epic tradition 137-38,139 Evenk (Tungus) epic tradition 4 2 , 1 3 8 39,143,164 Exirit-Bulagat (West Burjat) 129, 13437 Exirit-Bulagat epic, see Alamji-mergen extended text 12,16, 36, 292, 302

Fäfnismäl 92 fair copy 8, 42, 74,152 Fäzil Yoldä§-ogli, a singer 109 field-assistants to collectors 40, 42, 48, 71, 77,133, 241, 271, 284-85 field reports on oral textualization 4, 30, 37, 47,132 fieldwork 4 , 9 , 1 4 - 1 5 , 1 8 , 2 9 , 3 9 - 4 0 , 4 5 46,48,50-51,57-61, 67,73, 77-79,85, 108, 129-130, 132, 135-36, 140-41, 150, 156, 161-67, 171-72, 238, 240,

375

242-43,263-65,270,275,284-85,314, 318, 334, 345 "Fight at Finnsburg" 83 filler words 32,148, 312 Fine, Elizabeth C. 11, 12, 85, 105, 343 Finnegan, Ruth 279-80 "Finnish School" in folkloristics 10-11 Finnish-American Literary Heritage Foundation 311, 323 Finnish-Tulu team of scholars (project team) 28, 45,156, 219 Finno-Ugrian Society 145, 161 Fitzwilliam Museum in Cambridge 107 fixity, fixed form, plot, text, wording 3, 5,11,14-16, 23, 42-43, 47, 63-64, 67, 106,116,137,201, 203, 205, 209, 21112, 220-21, 313, 343-44, 353 Fokos, D. R. 149-50 Foley, John Miles 19, 40, 67, 320 folkloristics 4, 10-12, 91, 104, 237-39, 244, 351, 356 formula, formulaic 8,17-19, 22, 28, 33, 44, 81-83,194-96,198-207, 212, 220, 228, 245, 267, 273, 318, 360, 362 - see also oral-formulaic theory Foucault, Michel 344, 366 Frazer, John 359 Friberg, Eino 323 Friedman, Jonathan 355 Fula epic 155

Ganesa 202 Gangä 207 Ganguli, Κ. Μ. 202 Ganin, Α. Μ., a narrator 318 Geakcaval epic 137 Geiker, Gara, a shaman 171 Geiker, Nesulta, a shaman 168,170,171, 174-86 Genette, Gerard 64 genre 7 , 1 0 , 1 2 , 1 7 - 2 2 , 3 2 , 3 4 - 3 6 , 3 8 , 4 1 43, 50, 72, 73, 77, 84, 90, 93-94, 129, 145,148,155,163,170,197, 217, 219, 221-22, 227, 232, 245, 263, 267, 270, 273-74, 281, 292, 317, 319-20, 336, 346, 352-53, 360, 365-66

376

Index

George, Larry 296 Gesar/Geser epic 7, 89,134,136 Gilgamesh epic 7 Gill, William Wyatt 347, 349 Glancy, Diane 334 Goethe, Johan Wolfgang von 285 Golden Horde 108 Goody, Jack 6 Govorov, I. I. 141 Gowda, Deevappa 231 Gowda, Κ. Chinnappa 219, 231, 232 Greek (archaic) 56, 59, 60, 66, 83, 252, 255, 280, 281, 317, 322 Grierson, G. A. 212 Grimm Brothers 40, 74 Grincer, P. A. 204, 205-06, 212 Gromkovskaja, L. L. 131 Gromov, A. M., a singer 137-38 GuÖrun 91-92 Gujarati 44 Guliga narrative cycle 224 Guss, David M. 301, 302 Hajdü, Päter 148 Halpert, Herbert 162 Hamadi Bin Hasani Sayidi, a folk poet 249 Hanks, William F. 353 Hara, Minoru 199, 212 Haring, Lee 280 Harivamsa 209 Harris, Joseph 41, 89, 91, 96 Harris, William 59 Hastrup, Kirsten 345, 346 Hatto, Arthur Τ. 4 2 - 4 3 , 1 4 2 , 1 4 3 Havasupai narrative 291 Haviland, John Β. 343 Hawaiian cosmogony 349 Hawwäs, c Abd al-Hamid 272, 274 Heathobard feud 83 Hebrew traditions 280-81, 309-10 Hefter, J. 144 Heimskringla 89 Hein, Norvin 196 Helgakviöa Hundingsbana II 93 Henry, John 237

Herekali/Tambuka epic 254, 262 Heraklios epic 252 Herskovits, Melville 241 Herzen Institute of Northern Peoples in St. Petersburg 161 Heubeck, Alfred 67 HEURO project on South Slavic epics 85 Heusler, Andreas 93-94 Hezhen dialect 144,164 Hickman, B. 130 Hinde, R. A. 123 Hinduism 209 Hinton, Leanne 291 Hipahipa, a Supreme Hero 155 Hipparchus 64, 65-66 historic-geographic method 10-11, 18 Hittites 61 Hoijer, Harry 315 Homer 57, 58, 61, 63, 64, 65-66, 67, 77, 78-80, 89, 255, 273, 281 Homeric epic(s) 7 , 3 8 - 3 9 , 5 1 , 5 7 - 6 7 , 81, 82, 83, 204, 256, 281, 336 "Homeric" performance 220 Homeric poems 59, 61, 65, 71, 77, 78, 81,82 Homeric question(s) 58, 66 Homerists, Homeric studies 14, 38-39, 57-58, 60, 61, 65, 67, 203 Honko, Anneli 62, 219, 231 Honko, Lauri 58, 62, 94, 106, 154, 156, 219, 231, 245, 291, 343, 364, 365 Hopi (Indians) 296, 299, 309, 310, 321 Hopkins, E. Washburn 193 Howard, Victoria 297, 302, 313, 335 Hultkrantz, Ake 165 Hugh-Jones, Stephen 346 Huns 91 Hymes, Dell 19,49-50,105,321-23, 343 Hyper-Kalevala project 85

Ibonia epic 280 identification with epic personages (in ritual, epic performance) 27,35,222, 229, 231 identity representations and epics 2 7 -

Index 28, 35, 51, 75, 226-27, 230, 257, 343, 345-46, 355, 362 Iditua, a singer 48, 282-85 Iliad 38-39, 57-67, 77, 79, 80,155, 206, 210, 219, 336 Ilya Muromets 318-19 Imekana 131 incantation 43,168, 233 Indian (Indie) epics 3 , 7 , 3 8 , 4 3 - 4 5 , 1 9 3 234, 256, 319 individuality of oral epics 6 , 1 6 , 2 2 - 2 3 , 85, 220-21 Innes, Gordon 253 Institute of Pedagogy, Gorno-Altaisk 132 interest in documentation (inside/outside community) 3,10,17,28-29,51, 69, 221, 344 internalization, learning of epic tradition 17, 19-22, 27, 218, 226, 228-30, 266, 268 interpretation (by audience, reader, singer, scholar) 3, 17, 20, 22, 27-28, 30-31,34-35,42,48,50-51, 71,76,84, 92, 167, 171, 207, 270, 272, 275, 281, 284, 286, 298-99, 315, 346, 353, 355, 358, 365-66 intersemiotic translation 13, 30, 35-36, 71, 76, 218 - see also translation intertext, intertextuality 3, 16-17, 22, 27-28,35,40,50,93,345,352-53,360, 362, 365-66 invocation 13, 29, 34, 43, 217, 232, 292 Inyo-kutavere, a singer 42,154-55 Ionov, V. M. 144 Iranian epic 7 Irkismond' epic 138-39 Ishan people 283 Islam 165 Islamic tradition 245, 247, 258, 264 Jacobi, Hermann 193, 212 Jacobs, Elizabeth (Beth/Bess) 314 Jacobs, Melville 302, 335 Jakobson, Roman 5, 110, 318-19

377

Jakut, Jakuts 133, 138-44,161 Jakut epic 104,121,139-44 Jakut heroic epic Qi'is Debeliye 104 Jakut Institute of Folklore 143 Jamal dialect 147 Jamcarano, C. t . 135-37 Jangar epic 7 Jaräsamdha 206 Jätaka 209 Javanese 44 Jaya 210 Jensen, Minna Skafte 39, 51 Jesus (Isa) 248, 259, 261 Jews 248, 280, 281, 358, 364 Johnson, Frances 310, 314 Johnson, John William 45-46, 219, 280 Jones, Jack 154-55 Joseph, a prophet 259, 262 Judaxin, Κ. K. 152-53 Jumaadi narrative cycle 224 Kacic-Mio§ic, Andrija 84 Kahombo, C. Mateene 279 Kaikeyi 207 Kakridis, Johannes 64 Kalevala epic 3, 7-8, 37,46,61,224,258, 273, 311-12, 323 Kalkin, A. G. 131-32 Kannada 44, 28, 219 Karadzic, Vuk 40, 71-76, 77-81, 83-84 Karakalpaks 41, 107,115,121 Karakalpak epic tradition and language 106-23,152 Karakalpak singer 106-23 Karjalainen, K. F. 151 Kashmiri 44 Kauravas 207 Kazakh, Kazakhs 1 0 7 , 1 0 8 , 1 0 9 , 1 5 1 - 5 3 Kazakh Academy of Sciences in AlmaAta 104 Kazakh Department of the Sinkiang Academy of Social Sciences in Ürümchi 104 Kazakh epic 109,120, 206 Keats, John 107 Khanty (Ostjak) oral tradition 149-51,

378

Index

161,163,166 Khotanese 44 Kieleväinen, Vaassila, a singer 46 Kijuma, Muhammad 251 Kindaichi, K. 131 Kirghiz heroic epic 42, 104, 108, 109, 121, 129, 151-54 Kirk, G. S. 67, 79 Knappert, Jan 8, 46-47, 261-62 Ködäkc^n epic 138 Kökötöydün asi (The Memorial Feast for Kökötöy-khan) epic 42,151-53 Komi 161 Konenko, Natalie 336 Kooti Cennaya epic 22, 220, 222, 223, 225-26, 229 Koran 60, 250, 259 Koranic prophets 248 Korguzalov, V. V. 135 Köroglu epic 103, 104, 123 Kotan Utunnai, the epic of 131 Kraljevic Marko / Krali Marko (Royal Marko), see Prince Marko Krapf, Ludwig 247 Krnic, KreSimir 204 Kroeber, Alfred L. 42, 154-55, 291-92, 302, 315-16 Kroeber, Theodora 294 Krohn, Julius 10 Krohn, Kaarle 10 Krsna 195 Kubodera, I. 131 Kulun Kullustuur epic 140-42 Kumara, Siri's son, a priest in the Siri possession rituals, 20, 24, 27, 220, 227, 230-31 Kunti 195, 205, 207 Kuprijanova, Ζ. N. 129-30, 145-48 Kusa and Lava 203, 210-11 Kuusi, Matti 323 Lach, Robert 120 Lachmann, Karl 57 Lachmannian textual criticism 107 Laksmana 195 lament 18,20,41,84,94-95,172-73,187,

330-31 Baltic-Finnic lament 18, 20, 94 funeral lament 94 lament in Nanaj 172, 173,187 lament in Native American languages 331 - lament in Old Norse 95 - "Old Man's Lament" in Beowulf 95 - Yugoslavian lament 84 Lamut epic tradition, see Even Lane, Edward 247 Laotian 44 Lapsui, Anastasia 148 Latacz, Joachim 58 Laughead, William B. 237 Lebedev, V. D. 137-38 Lebedeva, E. P. 144-45 Lehtisalo, Toivo 129,145-46,148 length as a criterion of epic 34, 44, 63, 73, 77,80,144,210,218,223-24,23132, 243, 258, 265, 273-74, 292, 350 - see also long oral epic Leningrad Oriental Institute of the Academy of Sciences of the U.S.S.R., Oriental Archive 151 Lävi-Strauss, Claude 294, 334, 346 Lianja epic 7, 257, 279 Library of Alexandria 51, 66, 67 Library of the University of Helsinki 145 Liedertheorie 57 Ling Chun-sheng 144 Liongo cycle 248, 256, 261 literacy, see orality (vs. literacy) literary epic, see epic, categories of long oral epic 3-5, 8-9, 18, 22, 24, 29, 33, 36, 41, 44-45, 49, 61, 63,139,171, 217-26, 229, 244-45, 259, 262, 265, 279, 283, 291, 301, 350, 355-56, 365 "Loon Woman" 294-95, 297, 300-01, 304, 306, 310, 313, 316, 325-30 Lopatin, I. A. 168,170, 171 Lord, Albert B. 4, 12, 40, 57, 59, 61, 63, 67, 71-72, 76, 77-85. 3, 204,205, 207, 241, 280, 336 Lönnrot, Elias 8, 46, 57, 61, 224, 31112, 365

-

Index Lönnroth, Lars 95 Lusiadas epic 257 Maadaj kara epic 132-33 Magan Sisökö, a singer 239 Mahäbhärata epic 7, 43-44, 193-206, 208-12, 256 Majimaji epic 248, 261 Makhtumquli, a Turkmenian poet 103, 108 Malayalam 219 Malaysian 44 Malozemlja dialect 147 Malygin, P. N. 141 Manas epic 7, 89,121,153, 206 - the Kirghiz Semetej (Manas cycle) 129 Manasin, the Kirghiz film 104 Mande/Mandekan/Mandika epic 45, 239, 243-45, 253 Mandekan 241 Mandelbaum (Edel), May 314 Maninka and Bamana praise-poems 239 Mansi (Vogul) 149,161 Manäut's variant of Abai Geser Xubun 136 Manuan cosmology 364 Maori - Rarotongan Maori tradition, cosmology 50, 357, 358-61, 364 - New Zealand Maori tradition 351 Mapu Taia, a singer 350, 357 Marathi 44 Margulan, Alkey 151 Marjanovic, Luka 84 Marsh, Sadie, a narrator 295 Masaibu epic 257, 261 master text 5,15, 37, 46, 253 Masudi Al Buhriy, a folk poet 249 Mateariki, George 356 Mateene, Kahombo C. 279 Ma9kwem-kwapäive, Gentle Hero 155 Mauke/Maukean/Maukeans 50, 345, 349, 350, 352, 356, 357 - Maukean cosmological genealogies

379

50-51, 343-366 McLendon, Sally 302, 322 meaning - construction of 12-13,16-17,19,2628, 35, 41, 49-50, 72, 79, 82, 92, 165, 238, 241-42, 250, 255-56, 275, 28485, 294-95, 297, 299-300, 302, 304, 310,313-15, 343-45, 353-54, 360-61, 365 - transformal 14, 30, 48 measured (vs. metrical) speech 49,284, 296 Medjedovic, Avdo, a singer 80 Meletinskij, Ε. M. 145 melody 22, 25, 32, 35, 80, 83, 108, 115, 118-21,135, 221, 291-92, 312 memorization 12, 19, 21, 51, 78, 169, 249, 268, 336, 353, 355, 357-58 mental editing 22-24, 28, 218-20, 222, 233 mental text 17,22-24, 34,37, 41,45-47, 58,66-67,106,132,135,153-56,21821, 223-24, 232-33, 245, 364 Michailo Potyk 318 Middendorf, A. T. von 140 Middle High German 89 Milton, John 7, 89 Mixajlov, V. A. 135-36 Mladenovic, Zivomir 72, 74, 75 Mnyampala, Matiasi 259 Moculin epic 144 Moculin, a prince 144 mode of performance 17,23,25,41,44, 111-12, 221, 223-24, 228 Mohave epic of Inyo-kutavere 42,15455, 210, 291-92, 302 Möhr, Wolfgang 93-94 Mokrouga, N. G., a singer 137 Muhamadi epic 251, 261, 262 Muhammad, Prophet 247,248,256,257, 259, 261, 262, 272 multiforms 17-19, 23, 33, 220, 232 multiple documentation 13, 29, 40, 47, 78, 80, 84,163, 269 Murko, M. 319, 336 Museum in Alexandria 57 music 3, 30, 32, 35, 41-42, 47, 103-05,

380

Index

107-08,113-16,120-23,130-31,135, 138-39, 145-48, 163, 172-73, 206, 210-11, 226-27, 240, 243, 252, 257, 263-64,267-68,272,274,291-93,319, 335-36 musical accompaniment 25, 46-47, 73, 77-78, 80,103-04,108,122,172,175, 239, 243, 257, 264, 280, 283 musical notation 3, 119-20, 122, 138, 148,163, 272, 274 Mustajbeg of the Lika (licki Mustajbeze)81 Mwengo Bi Asumani, a folk poet 249 Mwindo epic 7, 244, 279-80 Myers, W. E. 302 Myreeva, A. N. 42,138-39,143 myth 10, 23-24, 45, 50, 64-65, 94, 149, 163,166-69,171-72,205-06,217,220, 222-23,226,229-30,256-57,291-92, 295, 297-99, 301, 304, 306, 309, 313, 315-17, 334,343-47,349-51,355,357, 359-61, 363-66 Nabii Yusufu, the tale of 259 Nagy, Gregory 67 Naika, Gopala, a singer 22, 24, 28, 30, 34, 35, 45, 219-21, 227-33 Namsaraev, X. 137 Nanaj (Gold) 43,144-45,161-89 Nanaj shamanic epic 161-89 Nannaya 202 Närada 203 Nasser, Ahmad, a folk poet 249 national epic and personal elements 37, 48, 51, 245, 284 Native Americans 49, 281, 295 Native American - ethnopoetics 291-342 - languages 297, 321-22 - tales 84 - narrative 29 - mythology and spoken narrative 295 Neckel, Gustav 94 Nedic, Vladan 72, 75 Nenec (Samojed) oral tradition 129-30, 144, 145-48, 161

- Tundra and Forest Nenec 146,148 Nenola-Kallio, Aili 94 Nettl, Bruno 118 Nevskij, N. A. 131 NHK (Japanese Broadcasting Corporation) 130,131 Nibelungenlied 7, 39, 57, 258 Niemi, Jarkko 148 Nkundo people 257, 279 Noghay heroes, a cycle 109 Norna-Gests dättr 41, 91, 95 North-American epics 291-336 Norwegian and Northumbrian Christian artists 91 Nosov, G. 141 notation of performance 8, 12-13, 36, 111-12,122-23, 315 Notopoulos, James 63 Novgorodov, S. A. 144 Novi Pazar songs 80, 81 N'urgun epic 142-43 Ob-Ugrian epics 149-51 Oceanian epics 50-51, 343-66 Odologie epic 49, 282-83 Odyssey epic 38-39, 57-67, 77, 80, 81, 206, 257, 336 Offield, Mamie, a narrator 313, 335 Oistrakh, David 114 Okpewho, Isidore 280 Old Believers 161,164-65 Old English 14, 58, 89 Old English elegy 94 Old Germanic literature 94 Old Man Milija, a singer 73 Old Norse 89-96 Old Norse-Icelandic poetry 41, 89-96 Old Saxon 89 Old Testament cosmology 364 Ol'sk dialect 138 Onenka, Bogden, a shaman 170 Ong, Walter J. 4, 6, 90 oral epic, see epic, categories of oral performance 3-10, 12-31, 33-41, 43-51, 58-61, 63-67, 71-83, 85, 8991, 94-95, 103-08, 111-12, 114-16,

Index 118,120-22,131-38,141-43,146-48, 168-74, 176, 178, 180-88, 203, 205, 208, 210, 217-26, 228-30, 232-33, 237-41, 243-45, 252, 256, 258-59, 263-75,280-81,284-86,292-93,29798, 300, 304, 311-12, 314-20, 334-35, 344-45, 350-53, 355-58, 360, 364-66 oral poetry 4, 8, 13, 16, 38, 40, 45, 57, 63, 74, 91, 94, 105, 122-23, 145, 193, 202, 254 oral text 3,5, 9 - 1 0 , 1 2 - 1 3 , 1 5 - 1 6 , 1 8 , 2 3 , 26, 29, 31, 33-37, 41, 45, 4 7 ^ 8 , 105, 146, 209, 218, 221, 223-26, 233, 241, 251 oral textualization 5, 6, 13-15, 21-22, 24, 29, 42-43, 47, 50, 218, 227 oral work (vs. literary) 5, 32, 34, 36-37, 73,85 orality (vs. literacy) 3-4, 6, 8, 14, 3134,36,44,59-61,89,93,197,201,204, 208, 211, 233 oral-derived epic, see tradition-oriented epic oral-formulaic theory 12,40,57-58, 63, 67, 77, 81, 335 oral/literary interface 37, 39, 50 O'Reilly, Edward 237 Oriya 44 Orosin, a singer 142-43 Ormr, skald of Barrey 90 Orthodox Church 147,165 Ottoman Empire 75 Oxotsk dialect 137-38

Päbüji epic 155 Pagis, Dan 280 Paiute (Southern) 291, 322 Palnaadu epic 64 Pälsi, Sakari 161 Panathenaic festivals 64-66 Panathenaic Rule 64 Pändavas 206, 207 Panel on Oral and Semiliterary Epics in Mysore (1995) 193 Panjurli narrative cycle 224 Papa Aiturau, a singer 350, 352

381

Päpay, Jozsef 149-51 Paproa-i-te-itinga 349 paralinguistic features 12, 36, 41, 109, 111, 114-15,122-23 Paräsara 207 Parry, Milman 12,40, 57, 61, 63, 67, 7172, 76, 77-85,122, 204-05, 207, 241 Parry Collection at Harvard University 81, 273 Parry-Lord acoustic records 83 Parry-Lord songs 81, 82, 84 pauses in the singing of long oral epics 13, 29, 39, 60, 111-14, 116, 132, 143, 228, 231, 233, 250, 253, 296 path of composition 19, 25, 28 Peabody, Berkley 210 Pearson, Clara, a narrator 313-14, 316, 335 Pekarskij, Ε. K. 139-44 Pentikäinen, Juha 43,182, 184 performance, see oral performance performance arena 19, 27, 169, 172 performance career 20-21, 24, 46, 218, 220, 350 performance context, see context, context of documentation, strong context performance in writing 7 - 8 , 1 0 , 37, 44, 46, 50, 63, 76, 356, 358 - see also written (secondary) textualization performance record (Fine) 12,18,30,36, 41 performance report (Fine) 12,30, 36,41 performance situation 12-14,17,24,26, 29, 37, 217-19, 229, 358, 365 performance strategy 17, 24-25, 28-29 performative competence, see competence performative style 17, 25-26, 44, 46, 228, 230, 232 Perttunen, Arhippa, a singer 311-12 Petri, Β. E. 135 Petrovna, Marija, a shaman 162 Phaeacians 64 Philippi, Donald L. 131 phrases 17-19, 23, 28, 33, 48, 79, 81-83,

382

Index

120,138,156, 220, 243, 250,268, 272, 275, 283, 285,298, 300, 320,331, 357, 360-62 Pirioi Tairea, a singer 350, 357 Pisistratus 63, 64, 65 Plato 92,107 Platonic text 122 plot, see storyline Podrugovic, TeSan, a singer 73 poet, see traditional poet poetic line 31, 49, 320, 360 politics and oral poetry 12,40,50,63-65, 73, 80-81,153,265,273,344,356,366 Polynesian mythology, narrative art 34647, 351 - Central Polynesian epic 357 - Central Polynesian oral narration 364 pool of tradition 17-19, 22-23, 59, 64, 66, 199, 220 Popov, A. A. 142 Porthan, Henrik Gabriel 10 Potanin, G, N, 134 praise song 109, 239, 247, 250, 272, 279 Prakrit 44,193, 212 prayer 13,34,43,173,217,223,230,250, 292 Prince Marko (or Kraljevic Marko) 7576, 84, 319 projected audiences (in editing epics) 40, 72, 76, 84,136, 237-38, 242-45 Propp, Vladimir 3, 205, 318, 319 prose epic (including prosimetric epic) 25, 41, 63, 89-90, 108, 111-12, 115, 122, 131-32, 139, 142, 145-46, 151, 155,166,170, 212, 239, 248, 253, 271, 279-80, 284, 292, 295, 296, 319-20, 334-35, 350-51, 365 prosodic features 12,31-32,42,46,109, 111-12,114-15,122-23,130,133,146, 243, 255, 312 publication strategy 8,17,29-31,34,37, 40,46-47,50,71-81,85,130-32,13842,146-52,220,238-39,243-45,26364, 271-75, 298, 344, 354, 364 Pugurcin, Ν. I. 147 punctuation in oral text 32-33, 114-16 Puränas 200, 202, 208

Puränic style of narration 201, 222 Purusha dance 226 Pushkin, A. S. 108,123 Puxov, I. V. 132,139,142-43 Pyrerka, A. P. 145 Quinn, Judy 89, 96 Qoblan epic 106-10,113, 116-20, 123 Qurtqa 113,117,118 Radlov, V. V. (Radioff, W.) 133, 152-54 Räghava 195,198 Rai, Β. A. Viveka 219,222,224,225,232, 234 Räma 44, 195-96, 198, 203, 207-08 Ramanujan, Α. Κ. 44, 207, 211 Rämäyana epic 7, 43-44, 193-212 Rämopäkhyäna 200, 209 Rarotongan Maori, see Maori Rasi'l-Ghuli epic 251 Ravana 44 Razzaq-bax§i Qazaqbaev, a singer 122 readability of oral text 32-33, 35 reception 16-17, 20-22, 26-28, 36, 38, 40, 71, 76, 81, 224, 353 recitation (of epics) 7-8, 23, 25-26, 32, 40, 42-43, 45-47, 64, 73, 78, 90, 95, 103-04,120-23,130-31,145,173,177, 180-82, 193, 195, 202-03, 209-11, 225-26, 239, 241, 247, 249-52, 25859, 264, 291, 353, 365 reader's role 5,30-33,35-38,40,59-60, 62, 66, 72, 74-76, 80, 83-84, 92, 115, 123, 143, 238, 241, 253, 270, 273-74, 283 refrain 22, 25, 194, 201, 207, 221-22, 227-30 Reguly, Antal 42, 149-51,153,161 Reichl, Karl 41-42,105, 106,121,123 repeatable expressions, repetition 8, 18-19, 21-22, 26, 31-32, 35, 44, 64, 116, 119-20, 154, 166, 184, 201-02, 207, 211, 222-23, 228-29, 253, 293, 297, 300, 305-06, 310, 312, 315, 32324, 331, 336, 347, 354, 360

Index repertoire 6, 8,19, 23, 34, 48, 61, 67, 79, 106, 108-09, 131, 137, 166, 208, 220, 223, 245, 249-51, 264, 268-69, 319 Return Song 80-81 Reuben, Nettie, a narrator 313 revival of tradition 43,45,167-69,17173, 220, 255, 344 Reynolds, Dwight F. 47-48, 322 Regetnikova, A. P. 104 Rexroth, Kenneth 293 Ricoeur, Paul 13, 343 ritual (in the performance of oral epics) 12-13,20,22-24,26-28,34-35,43,45, 114-15,129,156,162-64,166-73,188, 206, 217, 219-24, 225-32, 283, 292, 318, 347, 350 Roghair, Gene 32, 64 Roland, The Song of 7, 39, 258, 273 Romanticism 4-6,10, 37, 103, 344, 365 Rongo, a god 360-63 Rouget, Gilbert 122 Rubin, David C. 335-36 Rudnev, A. D. 136 Russification 134,163 Rwanda epic/language 247

Said Karama, a folk poet 259 Salbykov, Elbon, a singer 136 Salko Moric, a singer 79 Samoan 50, 357, 358, 360, 361, 363, 364 Samojed, see Nenec Samuela Terei (Samuela-TeRei-MoreTaunga-o-te-tini), a singer 50-51, 347, 354-55, 357-65 Sanskrit epics 43-44, 193-212 Sanskrit literature 194,197 Santali 44 Santanu 207 Säntiparvan 194,199 Sapir, Edward 11, 291, 293, 298-300, 302, 316 Saryar epic 108,123 SatyavatT 207 Saussure, Ferdinand de 5 Savage, Stephen 50, 347, 349, 354-55, 357, 358-59

383

Saville-Troike, Muriel 105 Sayaqbay Qaralayev, a singer 104 Scerbakova, A. N. 145 Schiller, Friedrich 285 School of Oriental and African Studies, University of London 241, 251-52 Schrempp, Gregory 351 SCHS (Serbo-Croatian Heroic Songs) 40, 80-84 Schücking, Levin 94 scribe 6-9,14, 36-37, 43-44, 58, 61, 63, 66-67,152-53, 210-11, 254 "Seal and Crab", a Kathlamet myth 298 Seesamma, a singer 229 Sejkin, Ju. 1.135,139 Semen Savin, a singer 138 Semetej, see Manas Semyonov, Elijah, a narrator 318 Sen, Nabaneeta 204 sequencing (epic events, poem texts) 16-18, 23, 25, 32, 37, 47, 75, 119-20, 219, 223-24, 233, 266, 269, 271, 29194,296-97,301-02,304,306,310, 314, 323-24, 334, 352, 36 Serbo-Croatian performance 320 Seydou, Christine 346 Shähnäme epic 7 shaman 108, 112, 161-73, 183, 185-86, 188-89 shared elements of tradition 9, 17-19, 21, 153, 171, 197-99, 208, 218, 220, 318, 345 - see also epic register Sharma, R. K. 203 Shaykh cAbd al-Wahhab, a singer 268 Shaykh Biyali Abu Fahrn!, a singer 265 Shaykh Tähä Abu Zayd, a singer 48, 268, 269-274 Sidahome, Joseph E. 282-83 Sieper, Ernst 94 SigurÖarkviÖa in meiri 89 Siguröarkviöa in skamma 93 Siberian epic 104, 129-51, 153-54, 156, 161-89 Silämaka, a Super-hero 155 Silverstein, Michael 343 Simkevic, P. P. 168, 170

384

Index

Simpson, Louis, a narrator 300, 302, 310, 316, 334, 335 singer's role (poet/performer, ritual leader, narrator/editor) 9 , 1 2 - 3 4 , 3 6 42,44-49,51,58-67,71,73,78-81,83, 103-04,106-12,114,116,118,120-21, 123,129-32,134-35,137-38,142,150, 203-04, 217-23, 225-33,247,249-51, 253-54, 263-64, 268-70,274,319-20, 350 Sinhalese 44 Siikala, Anna-Leena 20,50-51,168,170 Siikala, Jukka 345, 349, 354 Simson, Georg von 202 Siri Archive of RRC (Regional Resources Centre for Folk Performing Arts, M.G.M. College, Udupi, Karnataka, India) 30 Sirat Barn Hiläl ("The Epic of the Bani Hiläl Bedouin Tribe") 7, 47, 263-76 Siri, the protagonist of the Siri epic 220, 222, 227, 228, 229-31 Siri epic 4, 7, 19, 20, 22, 24-25, 27, 2 8 35, 38, 45, 51, 210, 219-21, 222, 223, 226-33 Siris, the Siri women in the Siri possession ritual 20, 24, 27, 227, 229-31 Sita 44, 207 Slavic epic 318-19 Slyomovics, Susan 60, 264 Smith, Hiram 315, 334 Smith, John D. 196 Snorri's Heimskringla 89 Snorri's Prose Edda, see The Eddas Sodani epic 42,139 Sodasaräjaklya 201 Söhnen, Renate 206, 208 Sokolov, Υ. Μ. 318, 319 Solon 64 Solovej the Robber 318 solidarity to tradition 5,14, 37 Somali epic 45, 241, 243 Sonatorrek by Egill Skalla-Grimsson 94-96 Songolo-Sartul' 137 Son-Jara epic 239, 244 - see also Sunjata

Sonne 227, 231 Sons of Homer 65 Sontoxonov, S. S., a singer 135 Sotho epic/language 248 Sound and Video Analysis Instructional Laboratory (SAVAIL) at the Folklore Institute, Indiana University, Bloomington, Indiana 244 South Slavs 77, 319 South Slavic epic / heroic song 40, 7 1 85,122, 265, 280, 319-320, 336 specialization in epic traditions 1 8 , 2 0 21,150, 251, 317, 350, 352, 365 speech event 14,19, 27, 30, 36, 41, 220 Sprenger, Ulrike 93, 95 Stafford, William 334 Stalin, Joseph 43, 165, 168, 171 standardizing, harmonizing oral text 6, 19,23,26-27,33,39,60,109,131,195 Starritt, Julie, a narrator 313 Steinitz, Wolfgang 150-51 stereotypes in research 4-6, 16, 38-39, 45, 48, 217-18, 292, 313 Sternberg, Lev Ja. 168,170 Stolac songs 78-80, 83 "strong" context 45, 217, 226 "story of Tulunaadu" 220, 221 - see also Siri epic storyline, plot 6 - 8 , 1 7 - 1 9 , 2 2 - 2 5 , 2 8 , 4 4 , 47-48, 64, 67, 108, 123, 205-06, 208, 218, 222-23, 226, 228, 230-33, 242, 244,270,274,284,292,294,312,31819 St. Paul 242 Sul'gin, Β. M. 131 Sumerian epic, see Gilgamesh Sunjata epic 7, 89, 219, 253 - see also Son-Jara Surazakov, S. S. 132-33 Suxbaatar, C. 133 Svargärohanaparvan 202 Swahili epic 46, 247-59, 261-62, 279 Swann, Brian 335 symbols (as key for reception) 16, 22, 28,222-23 synchronic (vs. diachronic) studies 6, 15-18

Index synchronic (vs. diachronic) variation 17, 23-24, 26 system of tradition, of poetic forms, signs 6, 8, 12, 21-23, 33-36, 47, 92, 205, 218, 312, 319 Tacitus 95 Takelma language/tradition 299, 309, 310, 314, 315, 322 Tamil 44 Tangaroa, a god 355, 359-64 Taplin, Oliver 63 Tapper, Richard 355 Tara Moekaa, a singer 356 Tararo Ariki, a singer 350 Taratoa, a long narrative of 350 Tarvainen, Lyuba 172 Taube, Erika 129-30,133 Taylor, Archer 280-81 Tedlock, Dennis 105, 293, 343 Te Puai^angauta 348-49 Te Tumu 347-49 Telugu 44, 202 Tengaru Tobia (Papa Tete), a singer 350, 356, 357, 362 Tera Moekaa, a singer 356 TereScenko, Ν. M. 145-48 text, literary (printed) vs. oral 6-8, 11, 35, 37-38, 66, 106,115, 270 text-criticism 10,11, 44 text paradigms 3-18, 30, 36, 38, 41, 76 - see also co-text, oral text, intertext, master text, mental text textuality 13, 16, 37, 89, 90, 93, 233 textualization 3-7, 9 , 1 3 - 1 6 , 18, 21-22, 24, 28-30, 32, 36-43, 45-48, 50, 7172, 76, 82, 84, 89-93, 95-96, 104-05, 115, 122, 129, 133, 136, 140-41, 1475 0 , 1 5 6 , 1 6 3 , 1 6 9 , 1 7 2 , 1 9 3 , 211, 21718, 221, 226-27, 229,231-32, 238-39, 242-44, 247, 263-64, 270, 275, 27981, 283, 317, 320, 343-45, 352, 354, 364-65 - see also oral (primary) textualization, written (secondary) textualization texture 10,19, 31, 35, 49, 312

385

Thai 44 Thayer, Ernest 237 Thomas, Rosalind 59 Thompson, Stith 109 Thorns, W. J. 237 Thonga (Tsonga) people 247 Tibetan 44,134 Timofeev, a singer 140-42 Tillamook tradition 313-14, 322, 335 Tkacik, N. P. 137, 139 TKU Archive, University of Turku, Finland 30 Todorov, Tsvetan 94 Tonkawa language/tradition 299,31415, 322 traditional poet 64, 78, 83, 92, 94, 103, 108, 195, 249, 251, 257-58, 264-65, 269, 271, 274-75 traditional referentiality 35, 76, 82, 84 traditional rules 14, 17-19, 21, 23, 37, 46-47, 233, 255 tradition-oriented epic, see epic, categories of - see also intersemiotic translation Trager, George L. I l l , 123 transcribing, transcription 3 , 1 1 , 1 3 , 1 7 , 30-32, 34, 36, 41, 43, 46, 48-49, 7173,75, 77, 80, 8 2 - 8 3 , 1 0 4 - 0 7 , 1 0 9 - 1 4 , 118, 120-23, 130-33, 135-37, 139, 141-44, 147, 149-53, 163, 169, 173, 205, 238, 240-41, 243, 253-54, 263, 265, 271-73, 284, 293, 299-300, 343 translation 3, 17, 24, 30-36, 38, 43, 48, 80,82,104-05,130,134,138,142,144, 146, 149-52, 155, 163, 173, 202, 207, 240-44, 253, 255-56, 263, 267, 27175, 284, 293, 295, 300, 307-08, 31517, 321, 323, 336 Trofimov, N. G., a singer 42,138-39,143 Trojan War 317 truthfulness of performance 17, 25-27, 63, 350, 353, 356, 362 Tswana epic/language 248 Tucker, A. N. 252 Tudenov, G. O. 135 Tuloxonov, M. 135-37 Tulu 18, 31, 32, 35, 45, 219-20

386

Index

Tulu(naadu) people, Tulu speakers, the Tuluva 4, 32, 34,35,219-21,222,226 Tulu epics 7, 20, 21, 25, 33, 45, 217-34 Tungus epic tradition, see Evenk Turkic oral epics 41,103-23,132 Turkmen epic 104,110,123 Turkmen(ian) singer 103,108,123 Turks 75 Turku Conference on the Textualization of Oral Epics (1996) 4, 61, 71, 155, 211, 234 Tuvinian 133 type scene, theme (Lord) 18-19,44, 81, 111, 153, 204, 208, 211, 336 Uke/Uke 348-49, 351 Ulanov, A. 1.135 Ungin 134 untold epic as a mental backdrop 24, 45, 50 Upadhyaya, Susheela 224, 234 Upadhyaya, U. P. 224, 234 Uralic languages 161 Urban, Greg 343-44 Uspenskij, V. S. 104 Utenzi wa Enjili Takatifu epic 259,261 Uttarakända 194, 197, 198, 200, 203, 209, 210, 212 Uvarovskij, A. Ju. 140 Uzbek 108-10,112 Uzbek singer 112,121,122 7

Vaidya, P. L. 195 Vaine No^omakere, a singer 356 Vaisampäyana 202 Valeri, Valerio 347 Valixanov, Sultan t o k a n tingisovic 151-53 Välmlki 43-44, 203, 206,209-10 Vämb^ry, Hermann 103 variation, variant 10-11, 16-19, 23, 26, 30, 34, 39, 43-44, 47-48, 75, 78, 92, 104, 110-11, 118, 132, 136, 140, 144, 146, 193, 195-200, 204-05, 218, 224, 233, 239, 251, 263, 268-69, 271-72,

274,292-94, 311-12, 357, 360 - see also synchronic (vs. diachronic) variation Vasil'ev, V. N. 141-42 VasileviS, G. M. 137-38 Vasilkov, Ja. V. 204, 206, 208 Vedic literature 209, 222 Veenker, W. 146 Venetus Marcianus 51 Veselovskij, Ν. I. 151 de Vet, Thirfcse 67 Vico, Giambattista 281 Vinogradov, V. S. 121 Virgil 89, 273 Visnu/Vishnu 195, 228 Vladimir (Prince V.) 318 voice, the singer's 3, 14-15, 25-26, 3435, 38, 41-42, 50, 78, 95, 103, 108, 110-12,114,120-23,131-32,174,224, 228, 231-32, 247, 250, 267, 270, 358 Vujnovic, Nikola 71, 78-79, 83 Vyäsa / Kpsna Dvaipäyana Vyäsa 43, 202, 207, 210

Wagner, Roy 363 War of Tabuk epic 256 wedding song 81, 221 Werner, Alice 251 West, Martin 65 West Germanic poetry, early medieval 93 whole epic, the difficulty in obtaining 23-24, 28-29, 33, 38, 45-47,131,149, 151, 218, 220-21, 225, 231-33, 245, 253, 265, 271, 292, 294, 298, 301, 356 Wilson, Duncan 74-75 Wintu language/tradition 294-95,297, 300, 301, 304, 307, 315-17, 322, 334 Wolf, F. A. 57 work song 22, 219, 221, 225-27, 229 written (secondary) textualization, codification 3, 6-8, 13-15, 30, 35-39, 4243, 47, 50-51, 57-59, 65, 67, 73, 90, 94, 105,109,141,147,193,203,205-06,209, 211, 218,254-55,270,343,354 Wyndham, John 282

Index Xangalov, Μ. Ν. 134 Xaratuur xaan epic 137 Xomonov, Μ. P. 135,136 Xori 134,136-37 Yaqui song 293 Yesufo, Mussa 284 Yuddhakända 194, 196, 198, 200, 203, 212 Yudhisthira, son of Kuntx 195, 205 Yugas 208 Yugoslavian oral tradition 58,61,63,77, 78, 84, 280

387

Yurok myths 315-16 Yusufu Ulenge, a folk poet 259 Zajatuev, Mensxen, a singer 137 Zenk, Henry 311, 321, 323 Zerahska-Kominek, Slawomira 123 Zirmunskij (Zhirmunsky), Viktor M. 123, 205, 336 Zsirai, Miklös 149 2umabay-zi'raw Bazarov, a singer 10608, 110,112, 120-21 Zumthor, Paul 14 Zuni (Indians) 296, 322

List of contributors Dan Ben-Amos

Professor, Department of Asian and Middle Eastern Studies, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

John Brockington

Professor, Sanskrit, School of Asian Studies, University of Edinburgh, Edinburgh, U.K.

John Miles Foley

Curators' and Byler Professor of Classical Studies and English, Director, Center for Studies in Oral Tradition, University of Missouri, Columbia, Missouri, USA.

Joseph Harris

Professor of English and Folklore, Department of English and American Literature and Language, Harvard University, Cambridge, Massachusetts, USA.

Arthur T. Hatto

Professor Emeritus of German Language and Literature, University of London, London, U.K.

Lauri Honko

Professor Emeritus of Folkloristics and Comparative Religion, Director, Kalevala Institute, University of Turku, Turku, Finland.

Dell Hymes

Professor Emeritus, Department of Anthropology, University of Virginia, Charlottesville, Virginia, USA.

Minna Skafte Jensen

Professor, Center for graesk-romerske Studier, Syddansk Universitet / Odense Universitet, Odense, Denmark.

John William Johnson

Professor, Department of Folklore and Ethnomusicology, Indiana University, Bloomington, Indiana, USA.

Jan Knappert

Dr., School of African and Oriental Studies, University of London, London, U.K.

390 List of contributors Juha Pentikäinen Professor of Comparative Religion, Department of Comparative Religion, University of Helsinki, Helsinki, Finland. Karl Reichl Professor, Englisches Seminar der Universität Bonn, Bonn, Germany. Dwight F. Reynolds Chair, Islamic and Near Eastern Studies, University of California, Santa Barbara, California, USA. Anna-Leena Siikala Academy Professor, Academy of Finland / Department of Folklore Studies, University of Helsinki, Helsinki, Finland.

Trends in Linguistics. Studies and Monographs Edited by Werner Winter Mouton de Gruyter · Berlin · New York 87 Peter Harder, Functional Semantics. A Theory of Meaning, Structure and Tense in English. 1996. 88 Language Contact in the Arctic. Northern Pidgins and Contact Languages. Edited by Ernst Häkon Jahr and Ingvild Broch. 1996. 89 A Bibliography on Writing and Written Language. Edited by Konrad Ehlich, Florian Coulmas and Gabriele Graefen. Compiled by Gabriele Graefen and Carl W. Wendland, in collaboration with Georg F. Meier and Reinhard Wenk. 1996. 90 Historical, Indo-European, and Lexicographical Studies. A Festschrift for Ladislav Zgusta on the Occasion of his 70th Birthday. Edited by Hans H. Hock. 1997. 91 Henning Andersen, Reconstructing Prehistorical Dialects. Initial Vowels in Slavic and Baltic. 1996. 92 Natural Phonology. The State of the Art. Edited by Bernhard Hurch and Richard A. Rhodes. 1996. 93 Hans H. Hock and Brian D. Joseph, Language History, Language Change, and Language Relationship. An Introduction to Historical and Comparative Linguistics. 1996. 94 Insights in Germanic Linguistics II: Classic and Contemporary. Edited by Irmengard Rauch and Gerald F. Carr. 1997. 95 Seiichi Suzuki, The Metrical Organization of Beowulf. Prototype and Isomorphism. 1996. 96 Linguistic Reconstruction and Typology. Edited by Jacek Fisiak. 1997. 97 Advances in Morphology. Edited by Wolfgang U. Dressier, Martin Prinzhorn and John R. Rennison. 1997. 98 Language Change and Functional Explanations. Edited by Jadranka Gyozdanovic. 1997. 99 Modality in Germanic Languages. Historical and Comparative Perspectives. Edited by Toril Swan and Olaf J. Westvik. 1997. 100 Language and its Ecology. Essays in Memory ofEinar Haugen. Edited by Stig Eliasson and Ernst Häkon Jahr. 1997. 101 Language History and Linguistic Modelling. A Festschrift for Jacek Fisiak on his 60th Birthday. Edited by Raymond Hickey and Stanislaw Puppel. 1997. 102 Robert S. Bauer and Paul K. Benedict, Modern Cantonese Phonology. 1997. 103 Studies in Middle English Linguistics. Edited by Jacek Fisiak. 1997. 104 Culture and Styles of Academic Discourse. Edited by Anna Duszak. 1997.

105 New Approaches to Chinese Word Formation. Morphology, Phonology and the Lexicon in Modern and Ancient Chinese. Edited by Jerome L. Packard. 1997. 106 Codeswitching Worldwide. Edited by Rodolfo Jacobson. 1997. 107 Salish Languages and Linguistics. Edited by Ewa Czaykowska-Higgins and M. Dale Kinkade. 1997. 108 The Life of Language. Papers in Linguistics in Honor of William Bright. Edited by Jane H. Hill, P. J. Mistry and Lyle Campbell. 1997. 109 English Historical Linguistics and Philology in Japan. Edited by Jacek Fisiak and Akio Oizumi. 1998. 110 Marta Harasowska, Morphophonemic Variability, Productivity, and Change. The Case ofRusyn. 1998. 111 James Dickins, Extended Axiomatic Linguistics. 1998. 112 Advances in English Historical Linguistics. Edited by Jacek Fisiak and Marcin Krygier. 1998. 113 Fragments of the Tocharian Α Maitreyasamiti-Nätaka of the Xinjiang Museum, China. Transliterated, translated and annotated by Ji Xianlin, in collaboration with Werner Winter and Georges-Jean Pinault. 1998. 114 Language Change. Advances in Historical Sociolinguistics. Edited by Ernst Häkon Jahr. 1998. 115 Jacob L. Mey, When Voices Clash. A Study in Literary Pragmatics. 1998. 116 Productivity and Creativity. Studies in General and Descriptive Linguistics in Honor of Ε. M. Uhlenbeck. Edited by Mark Janse with the assistance of An Verlinden. 117 Philip Baldi, The Foundations of Latin. 1998. 118 Numeral Types and Changes Worldwide. Edited by Jadranka Gvozdanovic. 1999. 119 Birgit Anette Olsen, The Noun in Biblical Armenian. Origin and Word Formation - with special emphasis on the Indo-European heritage. 1999. 120 Bob Morris Jones, The Welsh Answering System. 1999. 121 Eva Koktova, Word-Order Based Grammar. 1999. 122 Nancy A. Niedzielski and Dennis R. Preston, Folk Linguistics. 2000. 123 David Prager Branner, Problems in Comparative Chinese Dialectology. The Classification ofMiin and Hakka. 2000. 124 Gender in Grammar and Cognition. I: Approaches to Gender. II: Manifestations of Gender. Edited by Barbara Unterbeck, Matti Rissanen, Terttu Nevalainen, and Mirja Saari. 2000. 125 Brigitte Bauer, Archaic Syntax in Indo-European. The Spread of Transitivity in Latin and French. 2000. 126 Codeswitching Worldwide II. Edited by Rodolfo Jacobson. 2000. 127 Analogy, Levelling, Markedness. Principles of Change in Phonology and Morphology. Edited by Aditi Lahiri. 2000. 128 Textualization of Oral Epics. Edited by Lauri Honko. 2000.