The oral Traditional Background of Ancient creek Literature 0815336829, 9780815336822

Edited with an introduction by an internationally recognized scholar, this nine-volume set represents the most exhaustiv

151 104 24MB

English Pages 408 [409] Year 2001

Report DMCA / Copyright


Polecaj historie

The oral Traditional Background of Ancient creek Literature
 0815336829, 9780815336822

Table of contents :
Title Page
Copyright Page
Series Introduction
Volume Introduction
Section A. Definitions of Ancient Greek Oral Traditions
1. Homer, Parry, and Huso
2. Studies in the Epic Technique of Oral Verse-Making II: The Homeric Language as the Language of an Oral Poetry
3. Hesiod, Odysseus, and the Instruction of Princes
4. The Simile of the Cranes and Pygmies: A Study of Homeric Metaphor
5. Genre and Generation in the Odyssey
Section B. Prehistory and Oral Traditions
6. "Reading the Texts": Archaeology and the Homeric Question
7. A Mycenaean" Akhilleid"?
8. A Tale of Two Cities: The Miniature Frescoes from Thera and the Origins of Greek Poetry
9. The Rise of the Greek Epic
10. The Descent of the Greek Epic: A Reply
Section C. Explanatory Models
11. Towards a Generative View of the Oral Formula
12. Enjambment and Binding in Homeric Hexameter
13. Speech Introductions and the Character Development of Telemachus
14. Eurykleia and Odysseus's Scar: Odyssey 19.393-466
15. Discourse and Performance Involvement, Visualization, and "Presence" in Homeric Poetry
Section D. Oral and Textual Traditions
16. The Textual Criticism of an Oral Homer
17. The Gardens of Alcinous and the Oral Dictated Text Theory
18. The Homeric Poems as Oral Dictated Texts
19. Irreversible Mistakes and Homeric Poetry
Copyright Acknowledgments

Citation preview

Greek Literature Volume 1

The oral Traditional Background of Ancient creek Literature

Edited with introductions by

Gregory Nagy Harvard University

~~ ~~o~;~;n~~;up LONDON AND NEW YORK

Series Content Volume I


Volume 2



Volume 4


Volume 5


Volume 6


Volume 7





The editor wishes to thank the following scholars for their help and encouragement: Benjamin Acosta-Hughes, Victor Bers, Emmanuel Bourbouhakis, Casey Due, Mary Ebbott, David Elmer, Corinne Pache, Jennifer Reilly, Panagiotis Roilos, David Schur, Roger Travis, T. Temple Wright, Dimitrios Y atromanolakis.

Published i n 2001 by Routledge 711 Third Avenue, New York, NY 10017 Published in Great Britain b y Routledge 2 Park Square, Milton Park, Abingdon, Oxon, OX1 4 4RN Routledge i s an Imprint of Taylor & Francis Books , Inc. Copyright * 2001 by Routledge All rights reserved. No part of this book may be reprinted or reproduced or utilized in any form or by any electronic, mechanical , o r other means, now known o r hereafter invented , includin g an y photocopying an d recording, or in any information storage or retrieval system, without permission i n writing fro m the publishers . 10987654321 Library of Congress Cataloging-in-Publication Data Greek literature / edited with introductions by Gregory Nagy. p. cm. Includes bibliographical references . Contents: v. 1. The oral traditional background of ancient Greek literature — v. 2. Homer and Hesiod as prototypes of Greek literature — v. 3. Greek literature in the archaic period : the emergence of authorship — v. 4. Greek literature in the classical period : the poetics of drama in Athens — v. 5. Greek literature in the classical period : the prose of historiography and oratory - v . 6. Greek literature and philosophy — v. 7. Greek literature hi the Hellenistic period — v. 8. Greek literature in the Roman period and in late antiquity — v. 9. Greek literature in the Byzantine period. ISBN 0-8153-3681-0 (set ) - ISB N 0-8153-3682-9 (v . 1) - ISB N 0-8153-3683-7 (v. 2) - ISB N 0-8153-3684-5 (v . 3) - ISB N 0-8153-3685-3 (v . 4) .- ISBN 0-8153-3686-1 (v. 5) -- ISBN 0-8153-3687-X (v. 6) - ISB N 0-8153-3688-8 (v. 7) - ISB N 0-415-93770-1 (v. 8) - ISB N 0-415-93771-X (v. 9) - ISB N 0-8153-21. Greek literature-History and criticism. I . Nagy, Gregory. PA3054 .G74 2001 880.9-dc21


ISBN 0-8153-3681-0 (set) ISBN 0-8153-3682-9 (v.l) ISBN 0-8153-3683-7 (v.2) ISBN 0-8153-3684-5 (v.3) ISBN 0-8153-3685-6 (v.4) ISBN 0-8153-3686-1 (v.5) ISBN 0-8153-3687-X (v.6) ISBN 0-8153-3688-8 (v.7) ISBN 0-4159-3770-1 (v.8) ISBN 0-4159-3771-X (v.9) Publisher's Note The publisher has gone to great lengths to ensure the quality of this reprint but point s out tha t some imperfections in the original may be apparent.


v11 1x 1 15 65 85 129

139 157 165 191 213

217 261 281

Series Introduction Volume Introduction

Section A. Definitions of Ancient Greek Oral Traditions l.Homer, Parry, and Huso Albert B. Lord 2.Studies in the Epic Technique of Oral Verse-Making II: The Homeric Language as the Language of an Oral Poetry Milman Parry 3.Hesiod, Odysseus, and the Instruction of Princes Richard P. Martin 4.The Simile of the Cranes and Pygmies: A Study of Homeric Metaphor Leonard Muellner 5.Genre and Generation in the Odyssey Laura M. Slatkin Section B. Prehistory and Oral Traditions 6. "Reading the Texts": Archaeology and the Homeric Question E. S. Sherratt 7.A Mycenaean" Akhilleid"? L. R. Palmer 8.A Tale of Two Cities: The Miniature Frescoes from Thera and the Origins of Greek Poetry Sarah P. Morris 9.The Rise of the Greek Epic Martin West 10.The Descent of the Greek Epic: A Reply Martin West Section C. Explanatory Models 1l.Towards a Generative View of the Oral Formula Michael N Nagler 12.Enjambment and Binding in Homeric Hexameter Matthew Clark 13.Speech Introductions and the Character Development ofTelemachus Deborah Beck


v1 301 303

Contents 14.Eurykleia and Odysseus's Scar: Odyssey 19.393-466 Irene J F. de fang 15.Discourse and Performance Involvement, Visualization, and "Presence" in Homeric Poetry

Egbert. J Bakker


Section D. Oral and Textual Traditions 16.The Textual Criticism of an Oral Homer Graeme D. Bird 17.The Gardens of Alcinous and the Oral Dictated Text Theory


18.The Homeric Poems as Oral Dictated Texts


19.Irreversible Mistakes and Homeric Poetry


Copyright Acknowledgments


Martin West

Richard Janko Gregory Nagy

Series Introduction

This nine-volume set is a collection of writings by experts in ancient Greek literature. On display here is their thinking, that is, their readings of ancient writings. Most, though not all, of these experts would call themselves philologists. For that reason, it is relevant to cite the definition of "philology" offered by Friedrich Nietzsche. In the preface to Daybreak, he says that philology is the art of reading slowly: Philology is that venerable art which demands of its votaries one thing above all: to go aside, to take time, to become still, to become slow- it is a goldsmith's art and connoisseurship of the word which has nothing but delicate cautious work to do and achieves nothing if it does not achieve it Iento. But for precisely this reason it is more necessary than ever today; by precisely this means does it entice and enchant us the most, in the midst of an age of "work," that is to say, of hurry, of indecent and perspiring haste, which wants to "get everything done" at once, including every old or new book:- this art does not easily get anything done, it teaches to read well, that is to say, to read slowly, deeply, looking cautiously before and aft, with reservations, with doors left open, with delicate fingers and eyes. (This translation is adapted, with only slight changes, from R. J. Hollingdale, Friedrich Wilhelm Nietzsche, Daybreak: Thoughts on the Prejudices ofMorality [Cambridge, 1982].) Nietzsche's original wording deserves to be quoted in full, since its power cannot be matched even by the best of translations: Philologie namlich ist jene ehrwiirdige Kunst, welche von ihrem Verehrer vor Allem Eins heischt, bei Seite gehn, sich Zeit lassen, still werden, Iangsam werden- , als eine Goldschmiedekunst und -kennerschaft des Wortes, die Iauter feine vorsichtige Arbeit abzuthun hat und Nichts erreicht, wenn sie es nicht Iento erreicht. Gerade damit aber ist sie heute nothiger als je, gerade dadurch zieht sie und bezaubert sie uns am starksten, mitten in einem Zeitalter der "Arbeit," will sagen: der Hast, der unanstandigen und schwitzenden Eilfertigkeit, das mit Allem gleich "fertig werden" will, auch mit jedem alten und neuen Buche:- sie selbst wird nicht so Ieicht irgend womit fertig, sie lehrt gut lesen, das heisst Iangsam, Vll


Series Introduction tief, riick- und vorsichtig, mit Hintergedanken, mit offen gelassenen Thiien, mit zarten Fingern und Augen lesen ... (Friedrich Nietzsche, Morgenrothe. Nachgelassene Fragmente, Anfong 1880 bis Friihjahr 1881. Nietzsche Werke V.1, ed. G. Colli andM. Montinari[Berlin, 1971], 9.)

This is not to say that the selections in these nine volumes must be ideal exemplifications of philology as Nietzsche defined it. Faced with the challenge of describing their own approaches to Greek literature, most authors of these studies would surely prefer a definition of "philology" that is less demanding. Perhaps most congenial to most would be the formulation of Rudolf Pfeiffer (History of Classical Scholarship I [Oxford, 1968]): "Philology is the art of understanding, explaining and reconstructing literary tradition." This collection may be viewed as an attempt to demonstrate such an art, in all its complexity and multiplicity. Such a demonstration, of course, cannot be completely successful, because perfection is far beyond reach: the subject is vast,· the space is limited, and the learning required is ever incomplete. Finally, it is important to keep in mind that disagreements persist in the ongoing study of ancient Greek literature, and thus the articles in these nine volumes necessarily reflect a diversity of opinions. There is ample room for disagreement even about the merits of representative articles, let alone the choices of the articles themselves. It is therefore reasonable for each reader to ask, after reading an article, whether it has indeed been true to the art of philology. The editor, a philologist by training, has his own opinions about the relative success or failure of each of the studies here selected. These opinions, however, must be subordinated to the single most practical purpose of the collection, which is to offer a representative set of modern studies that seek the best possible readings of the ancient writings.

Volume Introduction

The focal point in this first of nine volumes about Greek literature, centering on oral traditions, is Homer, a prehistoric figure conventionally viewed by classical civilization as a prototypical poet of the Greeks. Some may even assume that research on Greek oral traditions should apply to no one but Homer. And yet, as the readings in section A of this volume suggest, the entire history of early Greek literature is based on oral traditions. The evidence for the oral traditional basis of ancient Greek literature is both internal and comparative. The decisive impetus for research has been the comparative evidence of living oral traditions. The two most prominent names in the history of this research are Milman Parry (collected papers published posthumously in Parry 1971) and Albert Lord (definitive books published in 1960, 1991, 1995). The first article in this volume and in the whole series, "Homer, Parry, and Huso" (Lord 1948), provides a vivid account of Parry's discovery procedures. Parry had started by studying systematically the internal evidence of Homeric poetry, as reflected by the texts of the Iliad and the Odyssey, even before he set out to observe firsthand living oral poetic traditions in the former Yugoslavia (first in the summer of 1933, and then from June 1934 to September 1935). The article by Parry (1932) in this volume, "Studies in the Epic Technique of Oral Verse-Making II: The Homeric Language as the Language of an Oral Poetry," where he elaborates on his concept of the "formula" (p. 6), is an example of his systematic approach to analyzing patterns of regularity in Homeric form and content- at a time when he had not yet observed comparable patterns in living oral poetic traditions. Parry's sudden and premature death in December 1935left his student Lord with the task of undertaking the systematic comparisons that Parry had only begun. These comparisons culminated in what remains the most definitive book on the subject of oral poetry, Lord (1960 [2000]; see Mitchell and Nagy 2000). How Lord's book extends from Parry's unfinished work is recounted in Lord (1948, article 1). In the history of Greek literature, the term "oral" applies not only to Homer. Nor does it apply only to epic, which seems, at first, the prototypical poetic genre in the history of Greek literature. The cumulative finding of ongoing anthropological research is that oral poetry and prose span a wide range of genres in large-scale as well as small-scale societies throughout the world and that epic is not a universal type of poetry, let alone a privileged prototype (Nagy 1990:17-51). There is no justification for assuming that epic poetry was the first genre of Greek civilization. lX


Volume Introduction

Although the epic poetry of Homer is the earliest attested genre, at least in its transcribed form, in the history of Greek literature, the contents of this poetry refer to or even "quote" from a plethora of other genres typical of oral traditions, such as love songs, laments, invectives, spells, boasts, and praise songs (Martin 1984 [article 3]:30-31; on laments, see especially Alexiou 1974; on prayers, see Muellner 1976). Among a variety of examples is the poetry of divination, as reflected in Homeric similes (Muellner 1990, article 4). Thus epic was not the only extant form of ancient Greek poetry that derived directly from oral traditions (Nagy 1990:414-437). Still, in the history of research on ancient Greek literature, the. single most important body of internal evidence showing traces of oral traditions has been the text of Homeric poetry, in the form ofthe//iad and the Odyssey. The article by Parry (1932) in this volume shows that his interest in epic led him to look for oral traditions underlying other ancient Greek genres, as represented, for example, by the "lyric" poetry of Sappho and of Alcaeus. Such directions in Parry's lines of interest were cut short, however, by his death. The posthumous publication of his papers, collected by his son Adam Parry (1971), situated Milman Parry's work in a scholarly context that confined the question of oral traditions to Homer, virtually excluding the rest of Greek literature. In Adam Parry's introduction to his father's book (Parry 1971), genres other than epic are not actively considered; moreover, there is a pronounced aversion to engaging with the comparative evidence of oral poetics (see also Parry 1966, included in volume 2; for further discussion, see Nagy 1999 [article 19]:266267). By contrast with the discontinuities inherent in the publication of Parry (1971), the work of Lord continued systematically the comparative methodology of Milman Parry, with applications to "lyric" (Lord 1995:22-68) as well as epic (Lord 1991). In terms of this methodology, to draw a line between Homer and the rest of ancient Greek literature is to risk creating a false dichotomy. There is a similar risk in making rigid distinctions between oral and written aspects in studying the earliest attested forms of Greek literature in general (Lord 1965:105-106). Besides examining Homer, volume 1 addresses the complementary importance of Hesiod as a foundational figure in the history of Greek literature. Homer and Hesiod are symmetrical "culture heroes" of Greek civilization, as we hear directly from the so-called father of history himself, Herodotus (2.53.1-3; Nagy 1990:215-217). The comparative evidence shows that Hesiodic poetry, like Homeric, derives from oral poetic traditions, as analyzed in this volume by Martin (1984), "Hesiod, Odysseus, and the Instruction of Princes" (see also Edwards 1971). Moreover, the Hesiodic 1beogony and Works and Days are complementary, as poetic compositions, to the Homeric Iliad and Odyssey; such structural complementarity provides valuable clues for defining the genres and subgenres of Homeric as well as Hesiodic poetry. This subject is addressed in this volume by Slatkin (1987), "Genre and Generation in the Odyssey" (see also Muellner 1996:52, 119-120). Homeric and Hesiodic poetry can even cross over into each other's poetic forms, as when the Odyssey reenacts the poetic genre known as "the mirror of princes," which otherwise typifies Hesiodic rather than Homeric poetry (Martin 1984, article 3).

Volume Introduction


Reconstructing ancient Greek literature backward in time from Homer and Hesiod, scholars are faced with a vast variety of problems and controversies. Although it may seem obvious that oral traditions must be the basis for the development of Greek literature as ultimately defined by Homer and Hesiod, a major question is: How are we to define these two figures themselves? The answer to this question is not at all clear. To say simply that Homer and Hesiod are the earliest authors of Greek literature is hardly adequate. The question remains: How are we to define the authorship of Homer and Hesiod in terms of the oral traditional heritage that shaped their poetry? In the ongoing search for answers, scholarly interest has consistently gravitated toward Homer and toward the genre that defines him, epic, at the expense of Hesiod. In the history of scholarship, it is in fact customary to speak exclusively in terms of the "Homeric Question." A similar question- or set of questions- is just as timely in the case of Hesiod as well as other early figures in the history of Greek literature. Still, most of the research in the oral traditional background of Greek literature gravitates toward Homer and the Homeric Question. The readings in section B of this volume reflect that fact. The Homeric Question cannot realistically be reduced to a single unified "question," as if all experts could agree on a definition of that should come as no surprise, then, that the answers, as offered by a variety of experts, are multiple and even contradictory. It would be misleading to attempt a synthesis of the conflicting views. For the reader to make an informed judgment, it is preferable to concentrate on the methods applied and on the results achieved. A powerful means for reconstructing the oral traditional prehistory of Homeric poetry is provided by the discipline of archaeology (see in general Snodgrass 1987). As we see from the overview of Sherratt (1990, article 6), the external dating criteria provided by the existing archaeological evidence point to many centuries of evolution for the oral poetic tradition that culminated in the Homeric Iliad and Odyssey. A major point of convergence for archaeology and the study of Homeric poetry is the issue of the Trojan War- or, more accurately, Trojan Wars- and the degree to which the Iliad and the Odyssey reflect the realities of the late second millennium B.C.E. (see Sherratt 1990, article 6; also Morris 1989, article 8). Homeric poetry, in the process of evolving as an oral tradition, reflects the realia of Greek civilization all the way from the middle of the second millennium B.C.E. to the seventh century B.C.E. and perhaps even later. Such an assessment, taking into account the testimony of (1) Homeric poetry as an ongoing system of communication and (2) the successive layers of archaeological evidence, represents an evolutionary model (see Sherratt 1990, article 6). The archaeological evidence is supplemented with the important testimony of the so-called Mycenaean Linear B tablets, the earliest attestation of the Greek language in writing; Palmer (1979, article 7) argues that we see here across section, dating back to the Mycenaean civilization of the second millennium B.C.E., of a phase of overall Greek civilization that decisively shaped the evolution of the Homeric tradition (on the name of Achilles as a reflex of "Mycenaean epic," see Nagy 1994).


Volume Introduction

Another powerful means for reconstruction is art history. The evolving traditions of visual arts, going as far back as the middle of the second millennium B.C.E. and even beyond, can be compared as parallel to the evolving traditions of the verbal arts as represented by Homeric poetry. A most dramatic illustration is the cross section provided by the miniature frescoes of Thera, discussed in "ATale of Two Cities: The Miniature Frescoes from Thera and the Origins of Greek Poetry" (Morris 1989, article 8). On these frescoes, which are dated well before the middle of the second millennium B.C.E., we can find representations of various themes that match corresponding themes in Homeric poetry, and the resulting visual/verbal correspondences can lead to the conclusion that at least some of these Homeric themes, such as the "tale of two cities" as represented on the Shield of Achilles in Iliad XVIII, were well over a thousand years old before they were recorded in written versions of the Homeric Iliad (for more on the Shield, seeNagy 1997). Yet another means, perhaps the most powerful of all, is linguistics (Nagy 1974, Muellner 1976, Frame 1978; see in general Watkins 1995). The application of historical linguistics to the diction of oral poetry yields new techniques of reconstruction, where the terminus of a given reconstruction backward in time can stop short of a "proto-language" phase (see, for example, Nagy 1994 on the name of Achilles, where the terminus of the reconstruction stops short of "proto-IndoEuropean"). Two papers of West (1988 and 1992, articles 9 and 10) survey the evidence provided by linguistics for the derivation of Homeric poetry from IndoEuropean poetic antecedents (for similar conclusions but different perspectives, see Nagy 1974, supplemented in Nagy 1990). Such reconstructions of Homeric poetry from Indo-European models need to take into account the lateral influence of Near Eastern languages and civilizations, especially in the eighth and seventh centuries B.C.E. (West 1988 [article 9]:169-172 and West 2000b; see also especially Hendel 1987). It is commonly assumed about oral poetry that it must be disorganized and incoherent in comparison to written poetry. Empirical observation ofliving oral traditions refutes such an assumption: degrees of poetic competence and skill may vary greatly, but the capacity of oral poetry for mechanical and aesthetic virtuosity has been confirmed in studies spanning a variety of cultures (Martin 1989, Foley 1998, Mitchell and Nagy 2000). To the extent that Homeric poetry is derived from oral traditions, its mechanics and aesthetics may differ from what is found in verbal arts that depend on the technology of writing (Muellner 1996). Accordingly, special models are needed for analyzing and explaining the potential cohesiveness and artistry of oral poetics. Section C offers a sampling of such models. Nagler's 1967 paper, "Towards a Generative Viewofthe Oral Formula" (article 11), views oral poetics as a cognitive system, supplementing Parry's model of the Homeric formula by way of linguistic models derived ultimately from generative grammar (for other models of the formula, see Nagy 1990:17-51). Clark (1994, article 12) expands Parry's metrical frame for defining the Homeric formula, demonstrating the existence of functional formulas that stretch far beyond the confines of a single Homeric verse. Beck (1998/9, article 13) shows that Homeric

Volume Introduction


variations of the formula can be driven by Homeric narrative, which has a built-in capacity for long-term thematic development. The brief paper of J ong (1985, article 14), which amounts to a preview of her book on Homeric "focalization" Gong 1989), is clearly not intended by the author as any kind of illustration of oral poetics; still, the devices of "narratology" as she describes them can be reinterpreted in terms of oral poetics (see, for example, Martin 1989). In contrast withJong, Bakker (1993, article 15), in an analysis of Homeric diction, shows how the self-presentation of Homeric poetry requires "live" performance, so that the very language of this poetry presupposes an oral tradition. A final but potentially vital question has been reserved for section D: If it is true that Homeric poetry derives from an oral traditional background, how did it become a textual tradition in the first place? Bird's paper (1994, article 16) offers a general assessment of the problem (see also Nagy 2000). According to one solution, as proposed by West (2000a, article 17), the "authors" of the Iliad and the Odyssey (who, according to this solution, were two distinct poets) had a hand in the recording of these poems, perhaps even intervening in the actual process of writing them down. This solution was evidently designed as an explicit alternative to the one proposed by Janko (1998, article 18), for whom the Iliad and the Odyssey as we have them derive from texts dictated by "Homer" himself, supposedly sometime in the second half of the eighth century B.C.E. As an alternative to both these solutions, Nagy (1999, article 19) proposes an "evolutionary model." At stake in choosing among all these alternative solutions is the definition of Homeric reception itself: How were the Iliad and the Odyssey understood by the classical world in the days of a figure like, say, Aeschines in the fourth century B.C.E.? This question, as carefully analyzed by Due (2001, article 20), leads to other vital questions. How was Homer received by classical Greek civilization writ large? Was he still the embodiment of the living word in performance? Or had he become a mere corpus of writings, the remains of the word that once was alive but since had died, hundreds of years before the emergence of the author we know as" our" Homer? Further Readings Alexiou, M. 1974. 7he Ritual Lament in Greek Tradition. Cambridge. 2nd ed., 2001, with new Introduction by P. Roilos and D. Yatromanolakis. Lanham, Md. Carlisle, M., and Levaniouk, 0., eds. 1999. Nine Essays on Homer. Lanham, Md. Day, J. W. 1989. "Rituals in Stone: Early Greek Grave Epigrams and Monuments." Journal ofHellenic Studies 109:16-28. Edwards, G. P. 1971. 7he Language ofHesiod in Its Traditional Context. Oxford. Foley, J. M. 1998. "Individual Poet and Epic Singer: The Legendary Singer." Arethusa 31:149-178. Frame, D. 1978. 1he Myth ofReturn in Early Greek Epic. New Haven. Hendel, R. S. 1987. "Of Demigods and the Deluge: Toward an Interpretation of Genesis 6:1-4." Journal of Biblical Literature 106:13-26.


Volume Introduction

Jensen, M. Skafte. 1980. 1he Homeric Question and the Oral-Formulaic 1heory. Copenhagen. J ong, I. de. 1989. Narrators and Focalizers: 1he Presentation ofthe Story in the "Iliad. .., 2nd ed. Amsterdam. Lamberterie, C. de. 2001. "Milman Parry and Antoine Meillet." In Loraux, Nagy, and Slatkin 2001:409-421. Latacz, J ., ed. 2000. Homers !lias. Gesamtkommetar. L2. Commentary on Ilzad I by J. Latacz, R. Niinlist, and M. Stoevesant. Munich and Leipzig. Loraux, N., Nagy, G., and Slatkin, L., eds. 2001. Antiquities: Postwar French Thought IlL Paris. Lord, A. B. 1953. "Homer's Originality: Oral Dictated Texts." Transactions ofthe American Philological Association 94:124-134. Rewritten, with minimal changes, in Lord 1991:38-48 (with an" Addendum 1990" at pp. 47-48). - - - . 1960. 1he Singer of Tales. 2nd ed., 2000, with new Introduction by S. Mitchell and G. Nagy. Harvard Studies in Comparative Literature 24. Cambridge, Mass. - - - . 1991. Epic Singers and Oral Tradition. Ithaca. - - - . 1995. 1he Singer Resumes the Tale, ed. M. L. Lord. Ithaca. Martin, R. P. 1989. The Language ofHeroes: Speech and Performance in the "Iliad. ..,Ithaca. Mitchell, S., and Nagy, G. 2000. "Introduction to the Second Edition." In Lord 2000:vii-xxix. Muellner, L. 1976. 1he Meaning of Homeric EYXOMAI through Its Formulas. lnnsbrucker Beitrage zur Sprachwissenschaft 13. Innsbruck. - - - . 1996. 1he Anger ofAchilles: Menis in Early Greek Epic. Ithaca. Nagy, G. 1974. Comparative Studies in Greek and Indic Meter. Harvard Studies in Comparative Literature 33. Cambridge, Mass. - - - . 1990. Pindar's Homer: 1he Lyric Possession ofan Epic Past. Baltimore. - - - . 1994. "The Name of Achilles: Questions of Etymology and 'Folk Etymology."' Illinois Classical Studies 19 (Studies in Honor ofMiroslav Marcovich, vol. 2):3-9. - - - . 1997. "The Shield of Achilles: Ends of the Iliad and Beginnings of the Polis." In Susan Langdon, ed., New Light on a Dark Age: Exploring the Culture of Geometric Greece, 194-207. Columbia, Mo. - - - . 2000. Review of M. L. West, ed., Homeri !lias I (Stuttgart and Leipzig, Bryn Mawr Classical Review 00.09.12. 1998). http:/I Palmer, L. R. 1980. 1he Greek Language. Atlantic Highlands, N.J. Parry, A., ed. 1971. TheMakingofHomeric Verse: 1heCollectedPapersofMilmanParry. Oxford. Snodgrass, A. M. 1987. An Archaeology ofGreece: The Present State and Future Scope ofa Discipline. Berkeley and Los Angeles. Watkins, C. 1995. How to Kill a Dragon: Aspects ofIndo-European Poetics. Oxford and New York. West, M. L. 2000b. 1heEastFaceofHelicon: WestAsiaticElementsin Greek Poetry and Myth. Oxford. For West 2000a, see the main list of readings in this volume.

Volume Introduction


Wolf, F. A. 1795. ProlegomenaadHomernm. Halle. Translated 1985,Prolegomenato Homer, with Introduction and notes by A. Grafton, G. W. Most, andJ. E. G. Zetzel. Princeton.

Homer, Parry, and Huso*



began his first study of Homeric style 1 with a quotation from Ernest Renan which epitomized the method which he was later to follow so scrupulously. Renan's words have become familiar to all of Parry's students. 'Comment saisir Ia physionomie et l'originalite des litteratures primitives, si on ne penetre Ia vie morale et intime de Ia nation, si on ne se place au point meme de l'humanite qu'elle occupa, afin de voir et de sentir comme elle, si on ne Ia regarde vivre, ou plutot si on ne vit un instant avec elle ?' 2 'La litterature de chaque pays et de chaque epoque', Parry continued, 'n'est comprise comme elle doit l'etre de fac;on naturelle que par !'auteur et son public contemporain. II existe entre eux un fonds commun d'experience qui permet a !'auteur de mentionner tel objet, ou d'exprimer telle idee, tout en etant sur que son public se represente bien le meme objet et saisit les nuances de !'idee. L'auteur, et c'est Ia une partie de son genie, tient compte a tout instant des idees et du savoir de ceux auxquels il soumet son oeuvre; done la tache de celui qui, vivant a une autre epoque, veut apprecier cette reuvre avec justesse, consiste precisement a retrouver le savoir varie et les groupes d'idees que l'auteur supposait appartenir naturellement a son public ... si le principe n'est que trop apparent, son application rigoureuse est des plus rares, etant complexe au point d'etre impossible a realiser de maniere tout a fait satisfaisante : la critique se propose la un but qui est la perfection meme.' 3 This was the goal which Parry set for himself from the very beginning and it was this which led him by logical steps from the painstaking analysis ofHomeric style to thr~ investigation of the oral nature of the South Slavic epic, from Homer to Huso. In L' Epithete traditionnelle [TE, above], he showed that the nounepithet combinations in the Homeric poems were part of a vastly intricate pattern offormulas which the poet (or poets) had available to enable him (or them) to express a given idea within the limits of the verse. ILMAN PARRY

• First published in American Journal_ of Archaeology 52 (1941!), 34-44. Reprinted by kind 1 TE, above. permission of the editor and Dr. Lord. • L'Avenir de la science. p. 292; TE, p. 2 above. 3 TE, p. 2 above. 8141816




Homer, Parry, and Huso


Because of the lack of extant Greek material it was impossible to prove that all of the noun-epithet combinations were formulas, but the proportion which were demonstrably so was great enough that it was obvious that such a style could not be the work of a single poet. It must have been years in the making and it must have required the efforts of many poets. Hence this style was traditional and was thus set apart from the style of other epic poets such as Apollonius or Virgil. They were imitating Homer, but they were composing in a different way. With them the epithet was a literary device used to impart an 'epic' flavor to their verses, but in the Homeric poems the traditional, or ornamental, epithet was forced on the poet by the exigencies of the verse and was an integral and necessary part of the style. The requirements of the verse-making created the formulaic, traditional, style. When approached from this angle many of the difficultiems in the poes were solved, or at least were readily understood, without any departure from the critical method which Parry ever had before him. In his supplementary thesis 1 he considered two types of metrical irregularities from this point of view : the hiatus and the short vowels which had to be given the value oflongs because of their position in the verse. In putting the formulas together to form verses the poet sometimes found that a metrical irregularity occurred at the point of I juncture ; a formula ending with a vowel, for example, had to be joined to a formula beginning with a vowel and to elide would leave the verse short a syllable; or a formula ending with a short vowel followed by a single consonant was to be joined to a formula which began with a vowel. But the poet would rather ignore this irregularity than change the formula, which was for him the proper way of expressing the idea. He was not willing to depart from the traditional phrase. Another cause of such metrical irregularities, Parry pointed out, was the construction of formulas by analogy with others. This is important, because it illustrates the way in which formulas are created. The most common examples, and the most obvious, arise from a change of case in a noun and its adjective. MepcnrE!; av8pW7ToL occurs at the end of E 288 and the last syllable of the adjective must be considered long. To understand how this happ~ned the verse should be compared with the several instances where a line ends with the formula fLEprYrrWV av8pomwv (1: 342, E 490, y 2I7).z The modification of the formula caused a metrical irregularity, which the singer was willing to overlook in favor of the formulaic pattern. One could continue in this way to show how Parry applied the touchstone of the traditional style to the apparent inconsistencies which critics had found in Homer, 3 but it is not the intention of this brief article to 1


FM, above. See HG, and DE, above.


• FM, pp. 1978 above.

Homer, Parry, and Huso review in detail the writings of Milman Parry. What I would like to point out is that in his early works he was thinking of a traditional style which was created by the poet's need for an easy way of making verses. The oral nature of that style had not yet occurred to him. He did not yet realize that this need for an easy versification arose from the fact that the poems were composed orally. The first expression of this is found in his article on enjambement in Homeric verse. '. • . Homer was ever pushed on to use unperiodic enjambement. Oral versemaking by its speed must be chiefly carried on in an adding style. The Singer has not time for the nice balances and contrasts of unhurried thought: he must order his words in such a way that they leave him much freedom to end the sentence or draw it out as the story and the needs of the verse demand.' 1 The idea had been sown in fertile soil and it grew rapidly. Parry's two articles in the Harvard Studies2 show an amazing development. He had in the meantime been delving into other heroic poetries, as even a cursory glance at the footnotes in the second of the two articles shows. These two articles, together with his French theses, present fully his theories and convictions about the Homeric poems up to the time of his research on the Yugoslav epic. The first of these two studies deals with style and the second with language. Since style is concerned with the form of thought, the basis for a discussion of Homeric style is the formula. Parry showed that the formula and the systems offormulas are peculiar to Homeric style. 'It is of course th,e pattern of the diction which, as in the matter of the authorship of the style, proves by its very extent that the Homeric style is oral. It must have bt.en for some good reason that the poet, or poets, of the Iliad and the Odyssey kept to the formulas even when he, or they, had to use some of them very frequently. What was this constraint that thus set Homer apart from the poets of a later time, and of our own time, whom we see in every phrase choosing those words which alone will match the color of their own thought? I The answer is not only the desire for an easy way of making verses, but the complete need of it. Whatever manner of composition we could suppose for Homer, it could be only one which barred him in every verse and in every phrase from the search for words that would be of his own finding. Whatever reason we may find for his following the scheme of the diction, it can be only one which quits the poet at no instant. There is only one need of this sort which can even be suggested-the necessity of making verses by the spoken word. This is a need which can be lifted from the poet only by writing, which alone allows the poet to leave his unfinished idea in the safe keeping of the paper which lies before him, while with whole unhurried mind he seeks along the ranges ofhis thought for the new group of words which his idea • DE, p.




HS, above; and HL, above. Hh2


Homer, Parry, and Huso


calls for. Without writing, the poet can make his verses only if he has a formulaic diction which will give him his phrases all made, and made in such a way that, at the slightest bidding of the poet, they will link tpemselves in an unbroken pattern that will fill his verses and make his sentences.'' This is an enlargement of the brief statement I have already quoted from Parry's article on enjambement. When he came to study the Homeric language as the language of an oral poetry he gave a much fuller description of this style, and then proceeded to show by the same type of reasoning that the language of an oral poetry is made up of archaic elements, foreign elements, and artificial elements, and he applied this to the Homeric language. In this paper, too, he was thinking for the first time not only of formulas but of whole traditional passages.2 This is a point which was later to assume even greater importance in his thinking, although he did not live long enough to set his ideas down on paper. I know, however, that he had formulated them pretty exactly. But at this point in • 932 he had reached a crisis. He wrote: 'To prove that there were one or many poets, and to show what passages were taken whole from the tradition and which were made anew out of single formulas or verses, we must turn to the study of other oral poetries where the processes of composition can be studied in actual practice and in a greater body of poetry than we have for the Greek epic. When, by the exact analysis of oral poems in reference to their tradition, we have grasped in detail just how the oral poet works and what it is that makes a poem good or bad in the judgment of himself and his hearers, we shall then, but only then, be able to undertake to study the authorship of the Iliad and the Odyssey, and to try to apportion that which is due to the tradition and that which is due to the author.' 3 Up to this point he had been talking of oral poetry from a theoretical basis and from accounts of others who had done field work. With rare exceptions (for example, the research ofMurko and Gesemann in South Slavic poetry and Radloff in Turkish) Parry had no great faith in the reports of these investigators. He was too thorough a scholar (as anyone will attest who has taken the trouble to follow the close reasoning and careful analyses in his writings) and too devoted to method to rely upon the observations of others when it was possible for him to observe the phenomena of oral poetry himself. There was nothing else to do then but to learn Serbo-Croatian (he chose the Yugoslav field because in those days it was the most accessible of the still living oral epics), to have a recording apparatus built which would satisfy the needs of continuous recording, and to go to Yugoslavia. With the financial assistance of the American Council of Learned Societies and 1


HS, p. 317 above. HL, p. 334 above.



HL, p. 361 above.


Homer, Parry, and Huso


of the Milton Fund and Clarke Bequest of Harvard, and with the cooperation of the Yugoslav government, he surmounted all the difficulties of such an undertaking and returned in the fall of I 935 with one of the most remarkable collections of oral poetry ever made. We all know that he did not have the opportunity to commit the results of his field studies to writing. However, in the fall of I935 he did begin a book to which he gave the title 'The Singer of Tales'. Seven typewritten pages of the first chapter of that book, typically entitled 'Aim and Method', are still in my possession and, obviously unfinished though they be, they are published here for the first time. They are the last words which he wrote on the subject of oral poetry. 'This book is the report of a study which I made in the years I 933935 of the heroic songs of the Slavic speaking peoples of L;.e Balkan peninsula. It happened at the time that I gathered much lore and music of many sorts, and made beginnings of what I hope may be in time a full and much needed work on Southslavic heroic and lyric song, but the study reported here was planned and carried out with no such large end in view: its business was to find out how the singers of the heroic tales learn and practice their art. But if I thus narrowed the scope of the search it was because I believed that we needed a very particular kind of knowledge before we could go much farther in our understanding of a whole vast and often very famous body of poetry, namely of what is properly to be called the song of unlettered peoples, but has been variously named as folk, or popular, or primitive, or traditional, or merely early poetry. Briefly, the aim of the study was to fix with exactness the form of oral story poetry to see wherein it differs from theform ofwritten story poetry. Its method was to observe singers working in a thriving tradition of unlettered song and see how the form of their songs hangs upon their having to learn and practice their art without reading and writing. The principles of oral form thus gott..:n would be useful in two ways. They would be a starting point for a comparative study of oral poetry which sought to see how the way of life of a people gives rise to a poetry of a given kind and a given degree of excellence. Secondly they would be useful in the study of the great poems which have come down to us as lonely relics of a dim past: we would know how to work backwards from their form so as to learn how they must have been made. Thus this book is meant not only for the Slavist but as well for the folklorist or anthropologist who has to do with the songs of any unlettered people, and even more for the students of such "early" poems as the Iliad and Odyssey, or the chansons de geste, or Beowulf. 'I am not of course the first to try to find out how an oral poem comes into being and passes from one singer to another, and what changes it undergoes in the course. No more am I the first to try to use living I


Homer, Parry, and Huso


unlettered song for better understanding of "early" poetry. Indeed, no small part of my faith in the method I have followed has come from my belief that the work of other scholars has tended ever more and more towards this method, until the time had come for someone to attempt a rigid formulation and use of it. It is even more than likely that someone else would have done this before had it not been for the lack of the mechanical means: it has only been in the last few years that the science of electrical sound recording has given us an apparatus of such a sort that it can record songs of any length and in the large numbers needed before one can draw conclusions, and finally which can make records which are so good that the words on them can be accurately written down for the p:.~rpose of close study.

The Oral Form of Unlettered Song 'The critics, groping for the rules by which they should group the varied works of the world's literature, have come to see more or less clearly that literature falls into two great parts, but they have not yet agreed upon the real nature of these two parts, nor upon the terms which should be applied to them. Such names as folk-literature, or popular, or primitive I literature have much truth in them but they are not finally good because, not to speak of the strange use of the word "literature", they are purely negative terms and mean at the best nothing more than the talk and song of men who have not the education of a self-styled civilized people, while at the worst they either betray a scorn which it would be hard to justify for certain ranks and forms of society, or else a wistful belief in questionable theories which make of the "common" man and the "simpler" stages of society the springhead of art. With our great anthropological knowledge we can now see that such terms get us little further than the point men reached in the seventeen hundreds when they believed that savages had a poetry which was more "natural" th~n their own. We are readier, now that we know more, to set lore against literature. These two words entangle us in no doubtful theories, but they do suppose that the use of writing brought about the one greatest change in man's artful use of words. It would seem true, however, that learning the use of writing is the one greatest cultural happening in the life of a people. 'If we put lore against literature it follows that we should put oral poetry against written poetry, but the critics so far have rarely done this, chiefly because it happened that the same man rarely knew both kinds of poetry, and if he did he was rather looking for that in which they were alike. That is, the men who were likely to meet with the songs of an unlettered people were not ordinarily of the sort who could judge Sundly how good or bad they were, while the men with a literary back-



Homer, Parry, and Huso

47 1

ground who published oral poems wanted above all to show that they were good as literature. It was only the students of the "early" poems who were brought into touch at the same time with both lore and literature. Early poems come from the time debatable between the lack and the usc of writing, and if the pride of the nations in the genius of their past has led above all to the vaunting of these poems for the same sort of merits in them as one finds in great literary poetry, the little that is known about how they were made keeps pointing the other way to poets who made little or no use of writing. So from the start the songs which men gathered in modern times from unlettered peoples were likened to the early poems. The theories which resulted, for good or for worse, have been of any number of sorts, but broadly they may be put under the headings of those which stressed origins, content, or practice. 'We can see now that the critics began working from the wrong end when they leaped upon the oral poems to wrest from them the answer to the question which had so long vexed them: who was their author? Their question, unhappily, was the wrong one, because in it they failed to see that an oral poem undergoes two kinds of creation, that of the man who first makes it and that of the man who sings it each time. We are able to get pretty fully at the creation of the singer, but we are slapped at every turn when we try to get our hands upon the creation of the maker (and very properly, for it is the only way we will learn that we are looking for the wrong thing), but it was nevertheless the creation of the maker which the first critics of the oral poems were sure they had found. It was simply that they had never heard or dreamed of a poem being otherwise thanfixed, so they left the factor of the .fluidity of the oral poem altogether out of their equation and got the answer that whereas the author of a written poem is so-and-so, the author of the oral poem is the people. The solution to the age-long problem they then set forth in a pleasant setting of ideas on the genius of the folk, which were then prompted largely by the democratic faith in man, but which have since, not without a certain irony, become the heart of the various doctrines of nationalism. The view which came belatedly, because it is only by careful study that we can work away from habitual ideas, that uneducated and uncivilized man is not really very unlike the man of the modern world, and that the genius of the oral poet is not really any different from the I genius of the written poet, is still so rare that it is limited almost to the experts. Nevertheless it was this first great eagerness to know the origins of the oral poems which turned men to them and brought into being the work of such great collectors as Karajitch and Radloff whose work with true singers led them to see for the first time the true nature of creation in unlettered song. 'It was soon seen that there was much against the theory of folkorigin in such a simple form and further work was carried on along two


Homer, Parry, and Huso


lines. First, the method of comparative linguistics was carried over into the field of early and oral poetry, and a great deal ofsound study showed how parts of stories, or even the whole story, keep turning up in poems of different regions of a country, and in different centuries, and even in different countries. It was a method which could not be used as rigorously as in the field of linguistics for fixing the "common" form from which all the other forms must have come in the course of time and travel, because the material was too varied, and there was no way of telling whether two poems with the same story came from a common source, or whether one of them came from the other. Nevertheless the search itselfwas of a sort which kept to the songs themselves, and even though it failed to give us a body of Aryan ballads it showed much about the life of the theme in oral poetry, though the reasons for that life can be well understood only when one has seen the use of it by the singer in making up his song. The second line of study which sought to get at the origins of oral poetry took up more carefully the practice of the poetry itself as a way to this end, but it will be better, before looking at what was done by scholars along this right road, to glance at what was done along another. 'The students of the early European poetries, as these became better known, were struck by the great likeness in their thought, in such a way that passing over their oral nature, which they were more or less ready to grant, they came more and more to treat them as common examples of heroic poetry. This was very good as far as it went. Certainly no reader of these ancient poetries can help being struck by the fact that they have chiefly to do with the prowesses of men of strength and courage, whom the poets believed to have lived in a more or less distant past when human powers were greater, and whom they called by a special term which we translate as "hero". These critics were wrong, however, when they went on and supposed that heroic poetry was one of the necessary stages in the growth of literature. They failed to keep apart history and poetry. Poetry is heroic only because it is created by a people who are living in a certain way, and so have a certain outlook on life, and our understanding of the heroic will come only as we learn what that way of living is and grasp that outlook. We find, for example, that cattle-lifting is a common theme in the ancient European poetries, but it is found there because of no law of poetry, but because these people happened to live in a way which led them to the stealing of cattle on the one hand and to the practice of poetry on the other. The heroic element in early poetry is not a problem of lore, but one of anthropology and history, and the students of heroic poetry have done a very great deal in showing how the social background is mirrored in the poetry, but far less in showing us the nature of the poetry itsel£ The same is true for a number of other works which follow what is usually called a sociological method: inasmuch as


Homer, Parry, and Huso


they explain the content of the poems by the social life of the people they show us nothing about oral poetry which could not also be true of literary poetry. When, however, such works point out thefonction of song in a given society, that is another matter, for they are then beginning to get at social factors which bear upon the form of song. But to this I shall shortly return. 'To get back now to the critics who worked further along the line of origin: these were I chiefly the students of the English and Danish ballads, that is, of poems which were written down fairly lately, and which even now are sung in such out of the way regions as the Faroes Islands and the mountains of Kentucky.' When we were still in the field in Dubrovnik during the winter of I934-5 Parry began to dictate a running narrative account of his two Yugoslav trips, inserting each text into the narrative in its chronological order and providing each text with a commentary. He covered only eight texts (the last unfinished) and a period from the fifth to the twentysecond of July, I933· These are only field notes, but they are far from rough. I believe he intended some time to complete them. The present plan is to include them in the introductory pages to volume one of the published collection. I mention them here because Professor Parry gave to them the title Cor Huso. Cor Huso Husein was a blind singer whom we never met because he had already gone to his reward in I935, but who had become a legendary figure among the singers in the Sandzak of Novi Pazar, where we first heard ofhim, and in Montenegro. A full account of what we know about him will be given in its proper place, but to Parry, I believe, he symbolized the Yugoslav traditional singer in much the same way in which Homer was the Greek singer of tales par excellence. Some of the best poems collected were from singers who had heard Cor Huso and had learned from him. The Parry Collection of South Slavic Texts, now the property of the Harvard University Library, contains over 3,580 twelve-inch phonograph records and more than I2,500 texts, of which approximately 750 are recorded. Generally speaking, there are four types of text: (I) Instrumental; there are a few records illustrating the music of various instruments used in the Balkans; (2) Women's Songs; lyric and short narratives sung, usually unaccompanied, by the women and young people for their own entertainment or at social gatherings; about I I ,ooo of the total number of texts in the collection are of this type and some 250 are on records ;1 r The music of many of the recorded women's songs has been transcribed by the late Prof. Bela Bartok, who, with the permission of Harvard University, was commissioned by Columbia University to do this work. Seventy-five of these, together with a detailed analysis by Prof.



Homer, Parry, and Huso


(3) Narrative poems of adventure sung to the accompaniment of the onestringed gusle; these are the core of the collection and, although the actual number of texts is smaller than that of the women's songs, they are much longer; (4) Conversations held before the microphone with the singers, who tell of their lives and of how they learned to sing. The interest of the Homeric scholar naturally centers around the third group of texts. It would seem that the best way to determine what oral form is and to analyze the oral technique of composition and transmission is to take the songs of a given singer and relate every formula and passage in them to the songs of other singers first in the same district and then in other districts. If the material were abundant enough one would thus be able to show exactly what elements in that singer's songs are the common traditional property of all the singers and what elements are 'original'. It was with this in mind that a plan of publication for the Parry Collection was laid down several years ago. The material is to be published by districts and by singers within a district, and correspondences between the poems within the district and those outside are to be noted in an appendix. The first volume is nearing completion. It will contain, beside a lengthy introduction which will give an account I of the collecting, a representative selection of poems gathered from the region of Novi Pazar. It is obvious that before publication could be undertaken the words had to be transcribed from the phonograph records and a series of indices had to be devised so that the twelve thousand odd texts could be handled readily. Some of the transcription was done in Yugoslavia during I934 and I935 by Mr. Nikola Vujnovic, himself a singer from Hercegovina, who had been Professor Parry's chief assistant in the field. In the spring of I937 approximately 500 records were copied at Harvard and I took them to Dubrovnik that summer where Mr. Vujnovic transcribed them. But there were still many records left untranscribed, and in I 938, with the combined assistance of the American Council of Learned Societies, the Society of Fellows, and Harvard University, Vujnovic was brought to Harvard for a period of eighteen months. By the spring of I 940 all the records of the collection, with the exception of a handful in Albanian and Turkish and one or two which Vujnovic could not understand either because of poor recording or of dialect differences, had been transcribed. During the period in which the transcription was going forward a set of Bart6k and texts edited and translated by myself are soon to be published by the Columbia University Press. The reader is referred to this forthcoming volume for a fuller account of these songs. Several of the women's songs were also transcribed musically by Prof. Samuel P. Bayard of State College, Pa. and it is planned to publish six of these in the not too distant future.


Homer, Parry, and Huso


practicable indices was devised and completed. Up to that time there had been only a rough log of each text as it was collected : a number was assigned to each in the field and an entry made of the name of the singer, the place and date of recording, and the numbers of the records. There are now four separate indices, on three-by-five cards. One is a master index of texts listed numerically ; another is a master index of recorded texts arranged numerically, with guide cards for the centers in which the texts were collected ; a third is an index of singers with the songs gathered from each entered on the card; and the fourth is an index of the first lines of the songs giving the text number of each version, the name of the singer, and the place of collection. With this apparatus one can readily review the songs contributed by any singer, or survey the songs collected in a single district, or study all the variants of any given song. There was also a third task of a clerical nature which had to be done. This was typing the texts so that one could work from a typed copy, thus leaving the original manuscripts to be handled as little as possible. This work was started in Dubrovnik and continued later in Cambridge, Mass. There are still a large number to be typed, but this can be done as their turn for publication comes up. It can be done properly only by somebody who has at least an elementary knowledge of Serbo-Croatian. Within the framework of the larger study of oral form are two related problems the investigation of which will add immeasurably to our understanding of those poems which have come down to us from the past. In the Parry Collection there are many instances of the same song from a single singer both in a recorded and in a dictated version. The process of dictating was unnatural for most of the singers. Accustomed to compose their verses rapidly to the rhythm of the instrumental accompaniment the majority of the singers found it difficult to dictate good lines. In fact, some of them found it impossible and it was necessary to give them the gusle and ask them to sing a line and then stop and wait for the scribe to record it before passing on to the next. Since the collector was seeking a normal ten-syllable line, the dictatsd version tends to be more perfect metrically than the sung version, because in the heat of normal oral composition metrical irregularities are frequently glossed over in the singing by adding an extra grace note or drawing out another for two beats. With the material at our disposal in the Parry Collection we shall be able to determine very exactly what the differences are between the sung and the dictated versions of a song. The second problem is of peculiar interest to Homeric scholars. Most of the songs taken I down from unlettered peoples are short, usually only a few hundred lines, in rare instances reaching two thousand lines. This is far from the sixteen thousand lines of the Iliad or the thirteen thousand


Homer, Parry, and Huso


of the Otfyssry. During the summer of I935, while collecting at Bijelo Polje, Parry came across a singer named Avdo Mededovic, one of those who had heard Cor Huso in their youth,.whose powers of invention and story-telling were far above the ordinary. He was encouraged to take all the time which he wished, to rest whenever necessary, and to sing as long a song as he could. He sang for a week and our turntables rolled for about two hours in the morning and two hours in the afternoon, with short breaks every twenty minutes or half hour for a cup of Turkish coffee or some stronger refreshment. At the end of a week the song was still unfinished, but the singer's voice had gone, so medication was ordered and after a week's rest Avdo continued. Another week sufficed to complete the song, which ran to I 3,33 I lines. 1 Another song from the same singer, this time in a dictated version, runs to about the same length. 1 One should not seek the same type of excellence in these long songs as one finds in the Iliad and the Otfyssey, yet in spite of repetition of incident and lengthy catalogues (it will be noted that these are characteristic also of the Homeric poems) Avdo's songs are well above the average. We must posit for the Homeric poems, I believe, a more flourishing tradition than the one which prevailed in Yugoslavia in I 935, and a much richer one. But it is the length of Avdo's poems and the way in which that length was obtained which are of particular value. To illustrate the leisurely style of these songs I quote from a translation of the dictated 'Song of Meho, Son of Smail' which Professor Parry began. After a brief introduction the story begins: Now to you, sirs, who are gathered here I wish to sing the measures of a song, that you may be merry. It is a song of the olden time, of the deeds of the great men of old and the heroes over the earth in the time when Suleyman the glorious held empire. Then was the empire of the Turks at its highest. Sixty provinces it had and Bosnia was its lock, its lock it was and its golden keys, and a place of all good trust against the foe. Now they gathered together in Kanija in the gay tavern as the custom long had been. At that gathering were thirty barons, the chief men of all the city of Kanija, and four and twenty of the Emperor's lords. At the head of the gathering was the Duke Hasan Tiro with his fifty men of war and beside the Duke at his left side, Count Orner of Kanija, the old man. Beside the old man were two of the Emperor's marshals, and beside the marshals was Sifrich lord Hasan. Next to Hasan was his nephew lord Mehmed, the dear son of Smail the Pilgrim, and brother's son to Sifrich Hasan .... Two squires served the wine, one the squire of Kanija's count, the other of their marshal, even the warriors Hasan and Huseyin. Beneath their anns each held a goatskin of wine, and in his right hand a great measuring cup. Ever in order did they serve their chiefs, Duke Hasan and the great men of the realm. When they had served every man then did they thrust their hands behind their 1

Parry Collection, Texts nos. 12389 and 12441.


• Ibid. Text no. 68o2.

Homer, Parry, and Huso



~hes and stood at homage to their lords, that their lords might find their drink the sweeter. Now when the lords had drunk of their wine they put the wine glasses aside, for the wine had flushed their faces, and took up the brandy bowls. Brandy is ever a talker, and of those barons and those lords of the Emperor not one was a man who had to borrow .... 1

Among the projects which Professor Parry had set for himself was a series ofarticles for the American Philological Association under the general title 'Homer and Huso'. He published an abstract of the first ofthese,a which he called 'Homer and Huso I : The Singer's Rests in Greek and South Slavic Heroic Song' ; and I later followed the pattern of the abstract and I wrote an article on the subject. 3 This and two later articles4 were modelled along the general lines of Professor Parry's earlier paper entitled 'Whole Formulaic Verses in Greek and South Slavic Heroic Song.'s They were intended to illustrate the ways in which the two poetries could be most profitably compared. While in Novi Pazar Parry had recorded several Albanian songs from one of the singers who sang in both languages. The musical instrument used to accompany these songs is the gusle (Albanian lahuta) but the line is shorter than the Serbian decasyllabic and a primitive type of rhyming is regular. It was apparent that a study of the exchange offormulas and traditional passages between these two poetries would be rewarding because it would show what happens when an oral poetry passes from one language group to another which is adjacent to it. However, there was not sufficient time in 1935 to collect much material or to learn the Albanian language. While in Dubrovnik in the summer of 1937 I had an opportunity to study Albanian and in September and October of that year I travelled through the mountains of northern Albania from Shkodre to Kuksi by way ofBoge, Thethi, Abate and Tropoje, returning by a more southerly route. I collected about one hundred narrative songs, many of them short, but a few between five hundred and a thou...La, aLaAEKTWL" TO a· aUTO Kat ~LKaLapxo~. 1 This hypothesis as developed by A. Fick won both more favor and more scorn than it deserved.2 The favor it should not have had because Fick in putting it into practice used a method far too arbitrary, and those who scorn it now do not see that it first brought into prominence the two facts on which the whole problem of Homeric language hinges, namely that Homer's poetry can with no very great change be turned from Ionic into Aeolic, and that the non-Ionic forms are kept as a rule only when Ionic itseli has no forms which could take their place. K. Witte, when he wrote that the Homeric language is the work of the Homeric verse, gave the better reason for this, but it was Fick nevertheless who made the needed if false step, and we shall see what a large amount of truth there was after all in his views.

The Homeric Language as a Poetic Language Witte was able to show long lists of words from ~he .Iliad and Odyssey in which Aeolic or older forms stood beside Ionic forms, al-

ways with a difference of metrical value, and he wa.s further able to show that these different forms were suited for use in different places in the verse. 3 As we shall see,4 he failed to consider Arcado-Cyprian, and had no notion of an Aeolic poetic language, and so was wrong in thinking that some of these Ionic forms were only Ionic and so could not have been used by Aeolic poets. But this misunderstanding in no way affects the soundness of the principle which he drew from his evidence: F. Osann, Anecdolltm Romanum (Gissae, 1851), p. s; cf. p. 280. August Fick, Die homerische Odyssee in der ursprunglichen Sprachform wiedrrhergestellt (GOttingen, 1883); Die homerische Ilias (Gottingen, 1886); Die Enlstchung der Odyssee (Gottingen, 1910). 3 K. Witte, Home~os B) Sprache in Pauly-Wissowa, VIII (1913), coll. 2213-2247. The subject has been further developed by K. Meister, Die homerische Kunstsprache (Leipzig, 1921). 4 Pp. 26-27, 32. I



The Epic Technique of Oral Verse-Making


the Homeric poems were composed in a poetic language wherein old and foreign forms had been kept and new forms brought in by reason of the help they gave the epic poets in making their hexameters. These poets ever sought a language which was easier to handle, and for that reason ever made use of the fact that the older or foreign form of a word was to a Greek, as Aristotle tells us, more poetic than the form used in everyday speech. 1

The Homeric Language as an Oral Poetic Language In one way, however, the theory of Witte, even with the further work done on it by Meister, is unfinished: they have logically proved that the language of Homer is the work of the Homeric verse, but they have not at all shown how the verse in this case could have such power. It did not have it in the later Greek epic, nor in Roman hexameter verse, nor in short do we find elsewhere in ancient or modern literature (with the very notable exception, however, of the early poetry of the nations) any but the slightest traces of the verse-form acting on the language of the poetry. Clearly a special language for the hexameter could come into being only when poetry was of a very different sort from that which we ourselves write, and which we know to have been written throughout the history of European literature. To say that the Homeric language was the work of the Homeric verse thus implies a poetry which is, at least to our way of thinking, of a very special kind, so that while the theory may be proved it cannot really be understood until we know just what this poetry was. It is my own view, as those who have read my studies on Homeric style know, that the nature of Homeric poetry can be grasped only when one has seen that it is composed in a diction which is oral, and so formulaic, and so traditional.2 So it is for the language of the Iliad and Odyssey: if we know what an oral diction is we shall have the larger Rhetoric 1404b 10. L'epithete traditionnelle dans Hom ere (Paris, 1928); Les jormules et Ia metrique d'Homere (Paris, 1928); The Ho.neric Gloss in Transactions of the American Philological Association, LIX (1928), pp. 233-247; Enjambement in Ho.neric Verse, ib. LX (1929), pp. 2oo-22o; Ho.ner and Homeric Style in Harvard Studies in Classical Philology, XLI (1930), pp. 73-147; The Traditional Metaphor in Homer in Classical Philology, XXVIII (1933), pp. 3tr-43· 1 2



Milman Parry

background which the theory of a language made to fit the hexameter calls for. At the same time the Homeric language when thus explained by the diction will in turn give us the history of that diction. 2. THE TRADITIONAL POETIC LANGUAGE OF



The Formula

In a society where there is no reading and writing, the poet, as we know from the study of such peoples in our own time, always makes his verse out of formulas. He can do it in no other way. Not having the device of pen and paper which, as he composed, would hold his partly formed thought in safe-keeping while his· unhampered mind ranged where it would after other ideas and other words, he makes his verses by choosing from a vast number of fixed phrases which he has heard in the poems of other poets. 1 Each one of these phrases does this: it expresses a given idea in words which fit into a given length of the verse. Each one of these fixed phrases, or formulas, is an extraordinary creation in itsel£.2 It gives the words which are best suited for the expression of the idea, and is made up of just those parts of speech which, in the place which it is to fill in the verse, will accord with the formulas which go before and after to make the sentence and the verse. Each formula is thus made in view of the other formulas with which it is to be joined; and the formulas taken all together make up a diction which is the material for a completely unified technique of verse-making. 3 Finally, the formulas of an oral poetry are not each one of them without 1 Cf. A. van Gennep on the Serbian epic (La question d'Homere, Paris, 1909,p. 52): "Les pocsies des guslars sont une juxtaposition de cliches, relativement peu nombreux et qu 'il suflit de posseder. Le developpement de chacun de ces cliches se fait automatiquement, suivant des regles fixes. Seul leur ordre peut varier. Un bon guslar est celui qui jouc de ses cliches comme nous avec des cartes, qui les ordonne divcrscmcnt suivant le parti qu 'il en veut tirer." Cf. also F. S. Krauss, Slavisclle Volksforsclrungcn (Leipzig, 1908), pp. 183-184, and John Meier, Werden und Leben des Volkscpos (Halle, 1909), pp. 17-19. 2 For a fulJcr definition of the formula see L'epithete traditionnelle, pp. 15-17; II omer and II om eric Style, pp 8o-84. 1 For the technique of composition by formulas see L 'epithete traditionnelle, pp. 8-19, 45-145; Lcs formules et Ia metrique d'Homere, pp. 1o-13, 17-23, 48-52; llomer and Homeric Style, pp. 84-89, 14o-147.


The Epic Technique of Oral Verse-Making


likeness to any other; in that case the technique would be far too unwieldy. They fall into smaller groups of phrases which have between them a likeness of idea and words, and these in turn fall into groups which have a larger pattern in common, until the whole diction is schematized in such a way that the poet, habituated to the scheme, hits without effort, as he composes, upon the type of formula and the particular formula which, at any point in his poem, he needs to carry on his verse and his sentence. 1 A single man or even a whole group of men who set out in the most careful way could not make even a beginning at such an oral diction. It must be the work of many poets over many generations. When one singer (for such is the name these oral poets most often give themselves) 2 has hit upon a phrase which is pleasing and easily used, other singers will hear it, and then, when faced at the same point in the line with the need of expressing the same idea, they will recall it and use it. If the phrase is so good poetically and so useful metrically that it becomes in time the one best way to express a certain idea in a given length of the verse, and as such is passed on from one generation of poets to another, it has won a place for itself in the oral diction as a formula. But if it does not suit in every way, or if a better way of fitting the idea to the verse and the sentence is found, it is straightway forgotten, or lives only for a short time, since with each new poet and 1 For the schernatization of the formulaic diction see L 'epithete traditionnelle, pp. 19-24, 85-94; Homer and Homeric Style, pp. 84-89, 14o--147· W. Radloff (Proben der Volkslitteratm der niirdlichen Tiirkischen Stamme. V. Der Dialect der Kara-Kirgisen, p. xvii) gives the words in which an oral poet tells of the ease with which he composes: "lch kann tiberhaupt jedes Lied singen, denn Gott hat rnir diese Gesangesgabe ins Herz gepflanzt. Er giebt rnir das Wort auf die Zunge, ohne dass ich zu suchen habe, ich habe keines rneine.r Lieder erlemt, alles entquellt meinern Innem, aus rnir heraus." This is a commentary on two passages in Horner:

-rw• -yap /Ja. Oeos 1rlip• aw«~s, Arcadian hLep~v, Homeric i aijv, •Ap71v, Me'Y77v); the suffix -repos in the sense of one of a pair of things (Arcadian rwppevrepov 'YEIIOS, Homeric fJ1)AUTEpaL, fJewrepaL, a'YpOTEpos, etc.); Arcadian OWJ..ta, aeALOs, EI1XEfJov, il.J..taTaj Arcadian and Cyprian -rrroALs (Homeric -rrroALs, -rrroAEJ..tos, cf. Eustathius 6 o6 -rrroAEJ..t.OS Kv-rr plwv). 2 The following Homeric words are found in neither Ionic nor Aeolic inscriptions, but in Arcado-Cyprian; they occur in the Greek literature we know only as poetic words: In Arcadian and Cyprian alcra, ~p-rrw, elrx.w"Aa, olros. In Cyprian fliva~, aJIW'YW, aprvw, avrap, EAos, loe, 1r011LS, 117rEOS, xpavw. In Arcadian a-rrvw, aCTK1)fJ~s, OEaJ..taL, KEAEvfJos, AEVI111w. 3 If we consider how small a part of Arcado-Cyprian vocabulary it is which we know from the inscriptions, this number of poetic Homeric 1 On Arcado-Cyprian in Homer see H. W. Smyth, The Arcado-Cyprian Dialect in Transactions of the American Philological Association, XVIII (1887), pp. SQ-I33i C. M. Bowra, Homeric Words in Arcadian Inscriptions in Classical Quarterly, XX (r.926), pp. 168-176; Hoffmann, op. cit., I, pp. 276-283. 2 It is hard to see what sure conclusions can be drawn from the Arcadian name of a festival 'EKo-r6v{Jota.. Arcado-Cyprian {J6Xo!J.a.t is also found at Oropus and Eretria. Of the forms given above the following are found in West Greek: liEXwf; i1~J.a.-ra. (in Aetolian); the pairing --rEpof (in Elean). 3 "Ep ...w and O.p-rbw occur also in West Greek. Nv is found in Arcadian, in Cyprian, and in Boeotian, but its use in no one of these three places, is that found in Homer. Hesycltius glosses ovvov ••• KuTpw• opo,.av, which is some reason to take tpwvvwf, the epithet of Hermes, as Arcade-Cyprian. If E. Forrer's translation of the Boghaz-Keui tablets (Orientalische Literaturzeitung, XXVII, 1924, pp. II4-n8) is correct they show Ko£pa.110f to be an "Acltaean" word.


The Eplc Technique of Oral Verse-Making words which we find in current usage in Arcadia and Cyprus is highly significant. It can be understood only by assuming that the Homeric diction comes, for a large part at least, from Arcado-Cyprian poetry. The most stable part of an oral diction is its vocabulary, since it is usually easier for a singer to change a form on the model of his own language than it is for him to give up one word and find another, and his art of verse-making is chiefly the art of using the traditional poetic words. The Aeolic element in the Homeric language seems indeed to have been more one of morphology than of vocabulary.

The Aeolic Element The belief was held at one time that the Aeolisms in Homer were really only older forms of Ionic, but this was due to a misunderstanding of the nature of linguistic change, since most of the forms in question are due to two separate treatments of one original form. Thus the dative in -tuut was formed on the analogy of the tu- stems (-yl:vtu-ut, ~eXtu-ut); the first aorist in -uu- on the analogy of u-stems (UJXacr-cra, heXtu-ua); the perfect active participle in -wv, -ovTa, was formed after the present participle; 1rluvpt~ for Ionic Tl:uuapt~, and cf>~p for Ionic O~p, show different treatments of *q'lf and *g,lz~t; ta for Ionic 1-1£a. shows a complete absence of the initial *sm; ap-yEvvbs and il!-1!-IES show different treatments of *-uv- and *-uw from those which gave Ionic cf>attvbs and ~JLEL~. "01r1rws, 07r7rot, etc. is an innovation of Lesbian, seemingly made after oTn (original *oon). In view of the number of these certain Aeolisms it is clearly better to take also as Aeolisms those forms which and might be earlier forms of Ionic, e.g. the genitives in -iio and in -ow. Indeed the number of Homeric forms which are not Ionic but are found in oth_er dialects is such that it seems to outweigh that of the archaic and artificial forms. The following traits of the Homeric language are Aeolic: (r) In Lesbian (Aeolic of Asia Minor), Thessalian, and Boeotian, the dative in -Euut; 8tpu- instead of 8apu- (Homeric 8tpuLT7}s, 8tpcrlXoxos, cf. Thessalian 8tpulTas, 8EputMXEtos, etc.); ia instead of Ionic pla.; the patronymic adjective instead of the genitive of the father's name (Homeric TtXa.M(.:."'ws, N71X~ws, etc.); the treatment of labiovelars as labials even before front vowels (Lesbian 1rl:crcrvpEs according to Hesychius, cf. Balbilla 1rl:uvpa, Sappho 7r~Xvt, Boeotian 7rETTapEs,



Milman Parry Thessalian 7rEp:rre, 7recpetpaKovres, Homeric 7rLCJupes, 7rEJJ.7rw{3oXa, 7rEpt7rMJJ.evos, ~1r"Xe, E7rAEro,, etc., 1re"Xwp, 1rehwpws, cf>~p, and the variant reading cpXLif;erat for fJXLif;erat in p 221); declension of 'IIF-stems in -7jos, -1it (Lesbian {3aCJLX.,os, Thessalian {3aCJtheZos, Boeotian {3arrt"XeZt); liJJ.JJ.Es, VJJ.JJ.Es, etc. (Lesbian and Thessalian, no evidence for Boeotian); the apocopated forms of the prepositions; the dual (Thessalian oe£JJ.eve, ao[?]arotv, Boeotian E7r0tECJ'raTav, avefJhav, Lesbian [liv]ope, TW E7rtCJrara). 1

(2) In Lesbian and Thessalian the development of u followed by a liquid or nasal into double liquids or double nasals respectively (Lesbian EJJ.JJ.Evat, ZovvvCJw, •AXX.,Kros, Thessalian EJJ.JLEv, !::uovvvuot, Homeric

E/.1./.I.EVaL, EJLJJ.EV, c/JLhOJJ.JLELo.,s, ~JJ.JJ.Ope, i:pe{3evvos, ap"{Evvos, a"{avvLcpos, 'Evveov, ahh7IKTOS 1 lAAa{3e, Eppeov, f:{;ppoos, KaTappew); the change of *Tk and*(}~ into uu (Lesbian and Thessalian ouuos, etc., Lesbian JJ.Euuos, Homeric ouuos, JLf:uuos, etc.); KaXXos instead of Ionic KaMs in compounds, and in the comparative and superlative (Lesbian KaXXLKA71L, etc., Thessalian KaXXtcppovvretos, etc., cf. Boeotian Kahfos, Homeric KaXXtavauua, KahXwv, KaAAL"{vvatKa, etc.); KE instead of liv. (3) In Lesbian 07r7rws, 07r7rOL 1 orn, etc.; the infinitive of non-thematic forms in -JJ.evat (Lesbian EJJ.JLEvat, EOJJ.Evat, fJf.JLevat, OOJJ.EVaL, Homeric EJJ.p.evat, EOJl.EVaL, (Jf.p.evat, 00/.I.EVaL, etc.); aJJ.{3pOT71V (Homeric TJJJ.f3porov, {3poros, aJJ.{3poutos); ta- from &a- (cf. Lesbian Zovvvuw, Sappho ra.

01 I:Xe~aJJ.aV, Homeric ra.oeos, tarpecp~s, etc.). (4) In Lesbian and Boeotian the aorist in -uCJ-. (5) In Thes5alian and Boeotian the thematic and non-thematic infinitive in -JJ.Ev (Thessalian EJJ.JJ.Ev, fJeJJ.ev, OOJLEv, Boeotian OOJJ.ev, etc., Thessalian Kpevvi:JJ.ev, 1rpauui:JJ.ev, Boeotian cpepeJLev, etc., Homeric EJJ.JJ.EV, fJf.JLev, ObJJ.ev, etc., a"{eJJ.ev, cpepeJJ.ev, etc., and the variant readings 7roAeJJ.LtEJJ.Ev II 834, aKOUEJJ.EV T 79. etc.); the genitive in -OtO (Thessalian of Pelasgiotis and Perrhaebia Ilauuouveloto, 7rOAEJLOLo, etc.). (6) In Boeotian the genitive in -iio (Boeotian 'Apturl:ao, etc.); the genitive in -'&wv {Boeotian Opaxp.awv, etc., cf. Thessalian KOLVaouv, etc.); rol, ral instead of ol, al.2 1 Cf. A. Cuny, Le nombre duel en grec (Paris, 19o6), pp. 454-466, 487-505. 2 The vocative in -a. is attested only in Lesbian verse, which also has -a, so that there is nothing to show us which was the common form. The evidence for -• as Aeolic is too slight to be given much weight (d. Bechtel, op. cit., I, p. 269).


The Epic Technique of Oral Verse-Making Of all these Aeolic traits the only ones which are found in other dialects are the following: in Cyprian KEi in Arcado-Cyprian the declension in -ijros, -ij[L, etc.; in West Greek the non-thematic infinitive in -JJ.Ev, JJ.~uuos, etc., ToL, mi; and in Arcadian and \Vest Greek apocope in the prepositions. 'A-rr, br, and inr, however, are Thessalian only. 1 The only cases of apocope in Ionic inscriptions are two occurrences of np. The Homeric words which have Aeolic (and original) ii where one would look for Ionic 71 are dealt with below (p. 36). The traces of the digamma in Homer are likewise Aeolic, and allow us to say from which of the three Aeolic groups the poetry passed to the Ionians, but before dealing with this sound we must understand the nature of the Lesbian poetic language. The Traditional Language of Lesbian Lyric Poetry

The same forces which created the poetic epic language of Homer created the poetic lyric language of Sappho and Alcaeus. The scant remains of these two poets do not allow us to show, as we can do for Homer, that their diction is formulaic, and so oral and traditional. We do know, however, that Solon and Theognis were still following an oral tradition of iambic poetry,2 and that they lived at that time, always so precious for our own knowledge of oral poe tries of the past and present, when verse-making was oral but writing known and used as a means of recording and keeping. 3 All that we know of the use of writing in Greece at the beginning of the sixth century points to the same thing for Sappho and Alcaeus. Yet while we may still feel some doubt as to the way in which they made their verses, there is not the least doubt that their poetic language was drawn from an oral tradition: One might add -O.fu in Cyprian as equivalent to Boeotian -iio. Cf. Homer and Homeric St)•le, pp. QI--., (Oxford,

1925), pp. xxviii-lxxvi; 'A"Nca!ov MEA7J (Oxford, 1927), pp. xxviii-xciv. I On F- in Homer d. Les formules et la mitrique d. H omere, PP· 43-s6.


The Epic Technique of Oral Verse-Making


sound. Thus we find in Sappho -yXwuua "l:a-yE, ¢alvEraL oi,l and in Alcaeus 1riJEVj.J.OJJa oivw,, V1r0 Ep"(OJI (beside aj.J.VO"n8os "l:p-yov ), Xuu' liTEp E(hv ( _ v _ v v) ; likewise the trace of the initial digamma is seen in Sappho's tE,7rE, and in Alcaeus' a]1rvEl1r1J[' and EI:LvauuE, where the meter kept the unshortened form beside El7rov and~~'!!"[ in Sappho and EtJrE and ~X1rEro in Alcaeus. There are in the two poets some 33 places where an initial digamma would spoil the meter. The keeping of rp- in some form which is noted in our manuscripts as {Jp- (Sappho {3p08ov, {3p&.8,vos, {3paKEa)- how it was sounded is doubtful -was a poetic device to keep for these words the power of lengthening the foregoing syllable. Had such a treatment ever been a part of the spoken language the poets would have used it when the second element of a word began originally with *rp-, and Herodian would have quoted from Lesbian poetry *l:{3p&.-y1J and *a{3p1JKros (*f.-rpa-yl), *aFPlJKro~), and not Eupa-yl) and avplJKro~, as he did. These two forms are beyond doubt, because they show the treatment which the spoken language gave to the group vowel-digamma-consonant-vowel, as in oEi•w (*8f:ruw), vavo~ (*viiruo~), and EupvuLXaos (*E-rpvul-Xaros) of the inscriptions. Nevertheless such forms as Eupa-yl) and avpl)Kros could never have been a part of the spoken language, since -rp- would then have been treated as an initial and not an intervocalic sound-group, and they would have become *f:pa-ylJ and *apl)Kros. Eupa-y71 and a.vplJKTo~ can only be understood as poetic forms made to keep the metrical value of *f:rpa-yl) and *afplJKros. The following forms in Homer show this Lesbian treatment of the digamma: 8Evw beside Ionic Ui; XEVW, tXEVO.II, etc., beside Ionic XEH, EXEO.JJ, etc., a7rOVpas (*a7rOfpas), 0.11"1'/UPO.S (*ali"Efpa~) 1 ro.Aaup,vos (*ro.Aafp,pos) 1 Ka"Xavpoif; (*Ka"Xafpoif; ).

A like treatment was given in Aeolic to the group vowel-consonantdigamma-vowel. Er•o.8E (from *€uro.8E) is cited by Choeroboscus as 1 The reading of a papyrus fragment of Sappho (Lobel a 3, 6) is without value, since antiquity, no less than our own times, had its grammarians who, failing to understand the hiatus, wished to restore the digamma, e.g. Apo!lonius Dyscolus, who quotes XVua• iirtp ftOtv to show that the third personal pronoun began with a digamma. Likewise John the Grammarian (Hoffman, op. cit., II, p. 2 r 7) states that the Lesbians wrote ro'ivov, but Aulus Gellius wrote the quotation from Alcaeus rvtiJp.ova olvw.>.' li-yEfJ' l>.r,Kot p.f.v 'A7ro>.>.wv 'ApTtp.tat ~lw, xa.LpETE a· VJJ.Etf 'll"clCTO.t, l:p.E'io af. Ka.l JJ.ETinrtufJE p.,.qua,q(}' o7r7rOTE KEJI TU E'll"txfJovLwv]{)W'II"WP tpfjlia' allfLP"'Tat ~EtiiOf Ta.>.a.'II"ELptOS eMwv· W KOVpa.t, TiS a• Vp.p.tll all~p jfatCTTOS aotawv f.pfJME 7rW>.E'iTa.t, Ka.l TEWt TEp7rEufJE p.O.>.tuTa.; VJJ.EtS a· EU p.O.>.a. 'll"clCTO.L V'II"OKplva.ufJE ac/J' ~~· TVc/JMf av1,p, olKEt ai XLwt till 71"0.t7ra.>.oECTCT"'t Tov 7riiua.t JJ.ETO'II"tq(}EJI O.ptCTTEVoVCTtll O.o,aa.L. ~p.Etf a· ~P.ETEpOII K>.EOf OLCTOJI.EII /)(1(1011 E71" ala.v Q.pfjpW'II"WII CTTpEc/JOp.Eq(}a, 'II"O>.EtS EU IIO.LETO.WCTO.f" ol a. E7rl a~ 'II"ELCTOIITa.t E'II"El Ka.l h1,rvp.6v ECTTtJI •1 0

1 Vv. 146-152, 165-176. The great majority of the manuscripts read inroKpl•a.trlle in v. 171, which is correct, the metrical fault being a guarantee of the text (cf. Les jormules et la mitrique d'/lomere, pp. 13-16). The reading 6.4>' Jt~o~i• of some nine manuscripts, instead of 6.4>' u~"• which is given by the other manuscripts and modern editors, is exactly suited to the pride which oral poets everywhere have in their own skill; likewise the variant JtJ.&Erepo" in v. 174 is to be preferred to iJ~o~irepo~. The variant reading was due to the feeling of the scribes- which has also been that of modern editors'- that Homer could not have been so immodest.


The Epic Technique of Oral Verse-Making


Here we clearly have to do with a singer and no reciter; he is singing to Ionians; and he says that there are many other singers and that each of them has his own songs. One is forced to grant that this Hymn, in which we find the very same poetic language as in the Iliad and Odyssey, was the work of an Ionian, in the diction common to all other heroic poets of h,~ time, at a moment when the tradition of epic poetry was still that of creation and not of recitation.

The Theory of an Aeolic Diction Accepted Moreover the theory of an Aeolic Iliad and Odyssey rests altogether on one supposition, which is false, namely that the ease with which the poems can be turned into Aeolic proves they must have been more or less entirely as they stand the work of an Aeolic poet: really it proves only that the formulaic diction was Aeolic. As was said above (p. 6), oral poetry is altogether made up of traditional formulas and series of formulas, each of which is an artifice for making the verse and the sentence. The singer has learned these formulas by hearing them in the mouths of older singers, and he makes his own poetry out of them from beginning to end, since the only way he can compose is by thinking in terms of the formulas. Thus while the poems of an oral poetry are ever each one of them in 8. never-ceasing state of change, the diction itself is fixed, and is passed on with little or no change from one generation of singers to another. This is why we find that even those Ionic words which in themselves are metrically different from the Aeolic words are used in the Iliad and Odyssey in such a way that the change to Aeolic is still possible: in each case the word is used in fixed formulas and types of formulas which are traditional artifices of verse-making. 'Il~Ets, for example, can be changed to il~~Es in 73 out of the 81 times it occurs, which could by no means be due to hazard, and could not happen in the verse of any poet who was making each verse out of his own new words. Homer, however, was using the word in fixed phrases; so we find .q~Ets ~(E) 41 times as the device for beginning a sentence, and in 34 of these 41 it is a device for beginning the verse as well. 'II~Ets ~E" accounts for 9 other occurrences of the word, falling 8 times at the beginning of the verse. •A;\;\' 11-yEfJ' r,p.Ets np begins the verse twice. This, when we have set aside the 13 cases


Milman Parry


of ~J.I.E~s at the verse-end, leaves ~J.I.E~s before a consonant in only 8 places where the phrase is not clearly fixed in the diction. It is of course largely hazard and the length of the Iliad and Odyssey which bring it about that we can observe the use of a fixed phrase in a number of places and so analyze the technique of its use. 'HJ.I.E~s o(€) is itself often used to make up longer formulas. One of the most needed artifices of the singer is that of ready phrases of different metrical values to join his sentences on to one another. Homer's technique of the conjunctional formula is vast and complex, and, as in the case of the epithet, easily permits analysis into systems of great length and great simplicity. One such series of formulas is that for expressing the essential idea but we: ~J.I.E~s o' (2o times), ~J.I.E~s o€ (9 times), ~J.I.E~S o' aV (3 times), ~J.I.E~S o' a.vr' (once), ~j.l.e'ts o' a.VTE (3 times). Each of these phrases expresses the same essential idea but has a different metrical value. These formulas are then in turn used in other formulas. We find at the beginning of the verse ~J.I.E~s o' EJ.I.J.I.EJ.I.a.wres (N 785, if; 127), ~J.I.E~S o' EUTa.OTES Oa.uJJ.a!:oJ.I.EV (B 320, n 394), ~J.I.E~SOE ¢pa.!;wJJ.e0' 81rws (~ 14, Z 6r, if; rq). Twice we have the pair of verses ~J.I.E~S

o' o~r·

1rpLv 'Y'

E7rL "Ep-ya. 7rfLPOS 'Y' tJ.I.Ell OVTE 'lr'T/t /i.)l.)\1/t


'YfJJJ.a.uOa.t 'Axa.twv ti,t K' i;(J(:)I.71tut,

(ft 127-128, u 288-289). We have the system ~J.I.E~s of beLua.vres (, 236, 396), ~J.I.E~s of K)l.a.Lovres (t 294), ~J.I.E~s oE laxovres (o 454). 1 Among the nine uses of ~J.I.E~s J.l.fll we find TJJ.I.E'Ls J.l.fll ra ~aura. otEt7rOJ.I.EV (A 706, J.l. 16), ~IJ.E~s J.I.Ell ')'ap ('Y 262, 276). 2 Nor are the 13 cases of ~J.I.E'Ls at the verse-end due wholly to chance: the word is regularly used there as a means of filling in the last foot of a verse in which the fifth foot has ended with -oJ.I.Ev, -OJ.I.Ell ~J.I.E~s, making, as it were, merely a longer personal ending. Thus we find t7r€¢voJ.I.Ell ~J.I.E~s (K 478), Et7rOJ.I.Ell TJJ.I.E~s 1 'HJ.&E'iS OE OEl.traVTES and TJj.IE'is OE taxovTEs when o' au and o' auT'' which in the heroic style are equal in meaning to the simple ot, might have been used, show how keenly the singers felt the accepted irregularities as positive features of the epic versification. They are among the many Homeric phrases which bear witness to the oral nature of the diction at the same time that they prove the soundness of the traditional text. 2 The repetition of a more or less unco=on formula at a short interval, as in the case of iJJ.&E'is ot o•l.travTEs and Ti.u•, 25

>}J.}J.Ill'"· 38, 39 _,,, 26, 38

o}v, 24, 37

l!v, 24, 25, 35 o}vu&, 38 -ijvu&, 26 ijv8uve, 33 llvt'YI)v•oxijes, I9 n. 3 -ijOf, 28 '".for, 37 n. I _,,, 26 !}s, 24 and n. 2, 37 Bupu-, 27 (Jit.ptros, 38 Beci, 35 Biu&vu&, I9 n. 3 BiJ.tpE1 ftfll yaia 1roAbu fjlou, olJpEUI 3( 3pvs &"P"' fttll Tf 4>tpfl fjaAauovs, fttUU7J 3( ftEAluuas· Elfo7ro«OI 3' &,n ftaAAois «aTafjEfjpl8au&' !m' ahov. (Od. 19.106-ll4)

This passage, apparently a digression in the interview between husband and wife on Ithaca, will turn out to be a key element in Odysseus' revelation. We have seen, then, that Old Irish and Greek instructional poetry-the Audacht Morainn and the Works and Days-both contain one inherited theme concernin~ the abundance resultin~ from the power of a good king, and that the same theme appears in the Odyssey. The point which. I wish to make involves the same spread of attestation, with the substitution of the Theogony proem for the Works and Days passage. In Audacht Morainn, it is implicit that the fir flathemon is verbally expressed by the king. There is evidence, in fact, from the twelfth century that a king actually recited or assented to the proverbial statements and injunctions of a tecosc during his inau~uration ceremony. It may help us to appreciate the social parallels in the Greek and Irish situations for instruction-poetry if we note that such Irish inaugurations involved the passing of a rod, symbol of sovereignty, from a poet to the new king: the rod thus resembles the skeptron, not only in its close connection with kingship, but also in its relation to authoritative poetic speech. The Irish king is certified by the poet; reciprocally, the poet is maintained by the king and tribe. A similar relationship lies behind the proem to the Theogony, a poem resembling an inauguration ode for Zeus and detailing (as did Irish odes) the king's genealogy and battle prowess. Hesiod's gift of a skeptron from the Muses (Th. 30) is a gift from their father, Zeus, and gets as its recompense the gift of the Theogony itself, which authorizes the divine king.I5 Explicitly as well, the Audacht presents instructions on the role of speaking in kingship. After discoursing of the prosperity that the Ruler's Truth accomplishes, Morann gives Feradach a string of admonitions: Arethusa 13 (1980) 203-16; on kingship as a relic institution at the composition of Homeric epic, see R. Drews, Baslleus: The Evidence for Kingship In Geometric Greece (New Haven and London 198.'3) 105f., 144; on heritage of Greek kingship, see Detienne (above, note 10) 42-50. Ill On the Irish inauguration ceremonies and other king-rituals, see Kelly (above, note II) 14; M. Dillon, "The Consecration of Irish Kings," Celtlca lO (1973) 1-8; and F. Byrne, Irish Kings and High Kings (London 1973) 7-27, esp. 15ff. On th!' meaning of Hesiod's skeptron, see Duban (above, note 2) 7 and Puelma (above, note 7) 94, both of whom also notice the parallel action of po!'ts and kings in the description of the Muse's gifts (Th. 79-103).



Richard P. Martin Apair fris, (a) ha trocar, (h) bad firion, (c) bad chosmuil, (d) bad chuihsech, (e) bad fosath, (f) bad eslabar, (g) bad garte, (h) bad fialainech, (I) bad 8essach, (j) bad lessach, (k) bad eitir, (l) bad inric, (m) bad suthnge, (n) bad /oruste, (o) bad /irbrethack Tell him, let him he (a) merciful, (b) just, (c) impartial, (d) conscientious, (e) firm, (f) generous, (g) hospitable, (h) honourable, (I) stable, (j) beneficent, (k) capable, (l) honest, (m) well-spoken, (n) steady, (o) true-judgtng. (AM sect. 55, ed. Kelly)

The final three are of particular interest here: "let him be well-spoken, steady, true-judging." For, semantically, we have here the exact pairing of the Greek phrase asphaleos agoreuet. 16 The notion of true judgment, central to the Hesiodic description of ideal kingship in the Theogony passage (cf. especially Th. 85-90), in the AM immediately follows the phrases concerning speech-as is proper for the ideology of a culture in which tribal king and judge are closely allied. 17 Examples can be multiplied to show that speaking well and truthfully is traditionally important advice in Irish Instruction of Princes. The Senbriathra Ftthail ("proverbs" of the pagan judge Fithal), another tecosc text, has among its formulas the following: adcota mtltenga brithemnacht ("A sweet tongue begets judgment"). We can note in this formula the association, as in the Theogony proem, of the king's speech with judgment. Furthermore, the word mtltenga-literally "honey-tongue" -appears to

16 The adverb asphalPos is t>specially important because, in the diction of the Theogony, only Gaia and Ouranos l~t>ar tht> adjective form as epithet: as Duban notes (above, note 2, p. 19), this implies that "the king's 'unerring' pronouncements are the moral safeguard to the universe's physical stability"; he parallels the king's effectiveness in Od. 19.109££., a passage which we have already set>n is derived from archaic prince-instruction material. Old Irish foruste, the word translated here as "steady," might be further glossed "unmoved because rooted in tradition," as the noun forus, from which the adjective is derived, has particular reference to abstruse knowledge, especially of legal precedents and lore: cf. Dictionary of the Irish Language (Dublin 1913-76) "F," col. 373 s.v., sections b and c. On the importance of such "steadiness" in the context of kingship, cf. the evidence of a third cognate tradition (ancient India) as seen in, e.g., RV 10.173.4, noted by B. Schlerath, Das Konigtum im Rig- und Atharvaveda (Wiesbaden 1960) 118. The lack of cognate linguistic exprenions for these identical themes in the three traditions cited (Irish, Greek, Indic) should not be taken as proof that the complex of ideas regarding kings developed late and independently in each dialect of Indo-European, as semantic structures are often retained despite surface lexical renewal: cf. C. Watkins, • Aspects of Indo-European Poetics," in Indo-European Studies IV, ed. C. Watkins (Cambridge, Mass. 1981) 764-99, esp. 780, 794. 17 The Audacht presents the ruler not only as possessing the mystical fir .flathemon, upon which justice is based, but also as setting legal precedents in the matter of fees, obligations, honor-price, and so forth (cf. sections 47-52 AM). At the same time, the existence of independent judges appears to be acknowledged (cf. section 23, "Let him not exalt any judge unless he knows the true legal precedents," Kelly [above, note 11]8-9).


Hesiod, Odysseus, and the Instruction of Princes


contain the exact notion found in meilikhios, a key adjective in the Greek passages in question. IS The "Battle of Airtech" (Cath Airtig) contains the tecosc given by Conall Cernach to Cuscraid Mend Macha as the latter assumes the kingship. After the instructions concerning proper gift-bestowal and assemblyholding, the new king is enjoined as follows: 19 Bat eolai in gec:·h herlae ar narhat ainfis i nnach dan conruidfe a frilacra f riut. Bat firen firhrethach cen forhrisiu n-indsciu etir tethrai tren (ocus) trug.

Be skilled in every tongue, so thou be not ignorant in any art that one will speak in argument with thee. Be just and righteous in judgment, not supressing speech between the tethra of the strong and the weak.

As in Audacht Morainn, instruction on speech is immediately followed by advice concerning judgment (with the same word used: firbrethach = "true-judging"). It should be pointed out that in both the Audacht and the Battle of Airtech, the instructions to the king are embedded in a narrative, as I believe is also the case in the Odyssey. In my final Irish example, the embedded tecosc is less well integrated with its narrative, a sign that the story-teller is indeed borrowing from another genre to fill out his tale. In a section of the Wasting Sickness of Cu Chulainn (Serglige Con Culainn), the hero instructs his foster-son Lugaid, who has been chosen as next king of Tara by an unusual bull-feast divination ceremony: 20 18 On this formula see R. Smith, "The Senbrlathra Hthad and Related Texts," Revue Celtlque 45 (1928) 14. Although melltkhlos is probably not related to the Greek word for "honey" (melt), as Chantraine showed ("Grec melltkhlos," in Mel. Rolsacq I [Brussels 1937] 169-74), there is no doubt that an association between the two word-groups has developed by the time of the composition of the Theogony proem (cf. lines 83-84: the Muses pour dew on the king's tongue and his words pour forth 'melllkha'). Compare also the Homeric description of Nestor's voice at II. 1.249, a poetic detail which may be of Indo-European date, on which see R. Schmitt, Dichtung und Dichtersprache ln lndo-germanlscher Zelt (Wiesbaden 1967) 256. Finally, on the association of honey and true speech, seeS. Scheinberg, "The Bee Maidens of the Homeric Hymn to Hermes," HSCP 83 (1979) 16-23; and on the symbolic value of honey in the Theogony in relation to speech, see P. Pucci, Hesiod and the Language of Poetry (Baltimore and London 1977) 19-21. 19 "The Battle of Airtech," ed. and transl. R. Best, Erlu 8 (1916) 173. The obscure word tethral may mean "paragons": cf. Dictionary of the Irish Language (above, note 16) s.v. "2 tethra." Compare the similar injunction, also in context of judging, contained in Tecosca Cormaic (Instructions of Cormac): eolas cech berla, "knowledge of each language," is one of the essentials for the tribe's welfare. (Text in Rook of Lelnster, vol. 6, ed. A. Sullivan [Dublin 1983]343c13.) 20 Sergltge Con Culalnn, ed. M. Dillon (Dublin 1975) 9; the translation is by M. Dillon, "Wasting Sickness of Cu Chulainn," Scottish Gaelic Studies 7 (1951) 57, who notes uncertainties ad Joe. Once again, the addres.'iee is a foster-son: cf. above, note II.


Richard P. Martin

38 26. Ni Ni Ni Ni Ni

fresnesea co labur. aisneisea co gl6rach. fuirse. chuitbe. faithchither sen6ri.

Be not haughty in contradiction. Be not loud in telling. Do not play the buffoon. Do not mock. Do not threaten old men .... (SCC 278ff.)

This portion of Cu Chulainn's instructions centers, once more, on the proper verbal behavior for kings. Although the contexts differ in the Wasting Sickness and the Odyssey passages, both Cu Chulainn and Odysseus instruct their young prince interlocutors to speak in a respectful, restrained manner. In Greek terms, this means having aidos; this brings us to the second theme shared by the Odyssey and Theogony passages with which we began. Aidos is an important topic of Prince Instruction, as can be seen more clearly when we turn from the shame culture of ancient Ireland to that of archaic Greece. Hesiod's Works and Days, as we have seen, is an explicitly instructional composition. It is significant, given this context, that the poem dramatizes the negative exemplum of the Fifth Age in terms of aidos: the era is to be marked by a complete failure of reciprocal relationships (WD 182-91) usually observed between family members, xeinoi, and parties to oaths: ~I ' I ·~· ul«"l u~' fV ~fpcr&' «a&I a&uf.tiS' OVIC la-Ta&, fJA6.1/fn ~, 0 ICaiCOS' Tlw apflova f/>wTa I'VBo&cr& noA&oir ~Vf7rf.tiV, ~7r\ ~· llp«ov 31'ftTa&. (WD 192-94)

Aidos is clearly a matter of proper verbal behavior here; the man who lacks this quality speaks "crooked words" in harming his betters. The description of the Fifth Age closes with another reference to aidos, this time in the symbolic departure from earth of the goddess, accompanied by Nemesis (WD 199-200). Finally, before proceeding to the specific instructions of the poem's latter part, Hesiod directs to his brother Perses a short exhortation concerning aidos: al~l.Jr ~' o/Jic /lya8~ «fXP"ll'fVOV llv~pa ICOI'l(n, '• I >I • I I ·~ t 1 I a&uwr, 71llt'f' familiar with archaic Greek compositions in the gf'nrt>-such as tlw Kheiri'nws hupothekai-would delight in Homer's creation of dramatic ironif'~ throughout this scene. For Odysseus the strangf'r yet speaks like a king. to those who know. Wf' have sf'en that Odysseus' words to Euryalos contain two themes promint>nt in princP-instmetions elsewhere, which account for the presence of similar phrasPology. Rut the dramatic context itself in Odyssey R marks Odysseus' discourse for what it is. Stanford, noting the apparent digrf'ssion in the spf'f'ch to Euryalos, compares it to a "short sermon on the divf'rsity of gifts" (adducing I Corinth. 12.4ff.), and his emphasis on the authoritative tonf' of the speech is correct: Odysseus is delivering the archaic Greek equivalent of quoting St. Paul, using the topic and tone of traditional instructional poetry J12 What is explicit in Th. 86ff.-that it is the king who speaks faultlessly, winning aidl)s and solving a neikos in the agora thereby-is acted out in thf' Odyssey. Odysseus makes a general statt;!llent to his opponent (I 76f.) whom the poet has already characterized as being the best looking of the nobles (8 116). Although the contrasting character in Odysseus' short speech, the anonymous man to whom the gods have given the gift of speech, appears to be simply a foil, we must see that it is none other than Odysseus himself. This is disguised self-reference. 33 Odysseus does not have to say that he (unlike Euryalos) can speak well and with aidos: he demonstratPs this in his reply. If Odysseus conceals the reference to the inenlity of the good speaker, we recall that at this point in Scheria he still masks his own identity. The two hidden facts are significantly intertwined. Odysseus the faultless speaker is Odysseus son of Laertes, king of Ithaca. ~2 W B Stanford. The Ody.~sey of Homer (Edinburgh 19592) vol. I, ad loc. In fad,

in the Iliad is a r"gular sermonizer about kingship in particular: of the four f'xamples of gnomic v!'rsf's which mention the role of basileus in the Iliad, Odysseus SP"aks three (II. 2. 19fi; 2 204fL 19.IR2f.; the fourth is by Calchas, II. 1.80); for the collection, see F. Ahrens, (;nomen in griechischrr Dtchtung (Halle 1937) l2ff. Only in the Odys.YPY are Odysseus' pronouncements disguised. 33 Cf. T. Rf>rres, "Das zeitlich" Verhiiltnis von Theogonie und OdysSPe," Hermes 103 (1975) 134. who vif'ws O y£ ~~:vv£~ 7Toaa~ lzpyol E'7rovro. 8£fT7T£CTlrw a, rii> Y£ xapw KarlxEV£V , AB~V'I/. rov ~' '/TctVTH Aaol f7TEpxop.£vov 8y£vvro . ... (Od. 2.10ff.) {Jfj

~~:ap7TaAlftw~ ~' tft7TATJVTO {3porwv lzyopal TE ~~:at E'~pa1 lzypoftlvwv· 7TOAAot ~' 8y~cravro iMvrtr vlov Aalprao ~a'lppova. rii> lJ' llp' 'MJ~v71 0£CT7T£CTl'I/V KaTfX£VE x&piV KEcf>aAfi H Kat ~f'OI~ .... (Od 8.J6ff.)

34 See R. Martin, Healing, Sarrifice and Battle: Amechania and Related Concepts in Early Greek Poetry (Innsbruck 1983) 65-7n on the Greek notion of contest as solution to

difficulties. 35 The primary and most extended expression of this "looking-on" motif comes at II. 12.312 (the people look on Glaukos and Sarpedon as gods). It is part of a larger theme, that of the "honor of gods": one is honored (II. 9 155), repaid (II. 9 ..102f.), welcomed (II. 22.435) like "a god." Other instances of the theme: at 11. 22 ..194 Hector is prayed to "like a god"; at Od. 7.11-a scene preparatory to the passage under discussion-Aikinoos, clearly a king, is said to be heard "like a god" by the demos. Note that, in the two passages cited in the text, the pouring of kharis by Athena is closely associated with the audience's favorable reaction to the figures of Telemachus and Odysseus. This accords precisely with the association of kharis and aidos in H. lJem. 214-15, quoted above: the possessor of kharis transmits aidos (and therefore receives it also) Finally, let me dte Pindar 0/. 1.30f., xap1s ~' Hwfp llwavra TfV)(fl rlz /A:f{>.,xa 8varols, in order to indicate the traditional nature of the dictional equivalences among terms which describe reciprocal behavior: atdos + kharts (H. Dem. 214f) kharts + ta meiltkha (01 I 30f) to meiltkhon + atdos (Theognis R5f.; 36.5) atdot metltkhtei (Th. 92, Od. R 172).



Richard P. Martin

In the first passage, the context is concerned with the uncertainty as to whether the young prince will be king of Ithaca and gain a permanent place on the throne of his father, which he now temporarily occupies (2.14). Interestingly, his speaking ability is singled out for mention as the characteristic that most marks him for the position of king. Antinoos says: T'', ~ p.aA.a ~~ fTf ~IMtTICOVfTIV 8t'Ot aVTOt btayOp'JV T' lp.n•a& ~Cat 8apuaA.Iwr h.yopfVfW p.~ ul y' Jv h.p.!f!u1.A.f!J 'I8a"n ~au&A.ija Kpovlwv wo1~umv, II ro1 yfvffi 1rarpw1·ov Junv. (Od. 1.384-87)

So too in Scheria thPre is the possibility that the new arrival will knock the islanders out of contention and become Nausicaa's husband. The poet plays with the resonances of a folklore motif, the royal marriage test. 36 Royal marriage seems far from conn~ted to the topic of prince-instruction, but as it turns out, in the Odyssey marriage and kingship have everything to do with one another. Penelope and the rule of Ithaca are indissociable. I pointed out earlier that Odyssey 19 contains an extended fragment of the Ruler's Truth ideology, in form resembling something directly taken from prince-instruction poetry. Who gives this apparent digression within the poem? None other than the disguised king again, Odysseus. Both in his talk with Penelope and in his reply to Euryalos, Odysseus allows his interlocutor to know that he is the king-provided the interlocutor recognizes the genre in which Odysseus is speaking. 37 In Book 8, he seals his coded signal with a clear sign consisting of a spectacular discus-throw. As if to contrast with the final revelation of his status, Odysseus begins the throw while still wrapped in his pharos (8.186); the throw literally lowers the status of the Phaeacians as it reveals Odysseus' own: they cower to the ground beneath the flung discus (190), which 36 Such rf"SSnances were certainly possible for an ancient audience: witness the multiform, attested in cult, in which Odysseus competes in a foot-race to win Penelope and makes dedications to Athena for his victory: Pausanias 3.12.4, discussed in M. Detienne, "The Sea-Crow," in Myth, Religion and Society, ed. R. Gordon (Cambridge and Paris 1981) 24f. (=a translation of the chapter in Les Ruses de l'lntelllgence: Ia meUs des grecs [Paris 1974) 202-43.) Further on such folktale elements in the competition for Penelope: U. Hoelscher, "The Transformation from Folktale to Epic," in Homer: Tradmon and lnVf'ntlon, ed. B. Fenik (Leiden 1978). 37 Odys.'!t'us' parallel role as poet, manipulating genres of discourse, once more makes it easy to lose sight of the artful composer Homer, who puts the genre of instruction-poetry into Odysseus' mouth for purposes of the wider composition. The layering evident in Odysseus' speech at Od. R.l67ff. is thus another piece of evidence (like Odysseus' lies and his adventure-telling) for the theory that the Odyssey's overriding concern is with the practice of narrative per se: see T. Todorov, "Primitive Narrative," in PoeUcs of Prose, trans). R. Howard (Ithaca, N.Y. 1977) 53-65 (a translation of Poetique de Ia prose [Paris 1971)).


Hesiod, Odysseus, and the Instruction of Princes


flies "past all the semata." While the discus surpasses these "distinguishing marks," the result of the throw is itself a sema, one which Athena validates in person with a proclamation containing a double meaning: Kai K, a.\ao~ TOl, ~ilV(' lJLaKpium TO CT~iJ-a a!A-f#>af/>owu· f1T(t oi! n fJ.ffJ.l"YfJ.,UOU ECTTtU op.i>.cp. a.\.\a 1TOAV 1TpWTOU' CTV liE 8apuu TOVlif y, l1.(8.\ou· oi! n~ lf>al~KWV rolif y' t(fral oVlJ' t!1Tfp~uu. (Od. 8.195-98)

The key to the riddle is contained in an unusual catachresis, for only here (196) in Homeric poetry is homilos used to refer to inanimate objects rather than to a throng of persons, usually warriors. The significance of the fact becomes more clear when we note that homilos can be used to describe the spectators at a contest (as at Il. 23.804, 813) and actually is used of the Phaeacian onlookers in this very contest at Od. 8.109. Athena's brief epinikion commemorating Odysseus' throw can be decoded to mean that Odysseus stands out from the "crowd" just as his discus is "not mixed in with the homilos." The message of Athena is exactly that which Odysseu~ transmits concerning himself (the "speaker") in his words to Euryalos: cf. 8.172, pHa ot 1rpf.1m a:ypopf.vounv. "Even a blind man" can distinguish Odysseus' sign, says Athena; the statement is ironically half-fulfilled soon after the contest when Demodocus, the blind bard, tells the story of Odysseus in the Trojan Horse but does not recognize the person who requested the tale, Odysseus (8.487ff.). This "sign" that has "surpassed all others," then, is iconic for Odysseus' status as surpassing speaker, athlete, and king. In Rook 19, a sema again follows Odysseus' coded use of princeinstruction in the conversation with Penelope. This time the sign that alerts the audience to the presence of other coded messages is the brooch that Penelope recognizPs from Odysseus' dt>scription. 38 Following Nagy's recent demonstration that poetry itself can be a sema in epic, I now suggest that in both the Odyssey 8 and 19 Ppisodes, Odysseus is delivering two semata: thosp designated as such (the discus-throw, the brooch) and the undesignated use of a gPnrP of discourse as a code to his listener. 39 In Scheria, only ken-minded Alcinoos succeeds in interpreting the sign, though the audit>nce of the poem appreciates its meaning: it requires a king to know one. In Rook 19. suspense arises from the use of the hidden sema: will PenelopP accomplish the task of rt>ading Odysseus' 38 On the ironies of this entire scene see R. Fenik, Studies in the Odyssey (Wiesbaden 1974) 20-2.5; on Odysseus' clothes as a coded message, see G. Nagy, "Serna and Noesis: Some Illustrations," Arethma 16 (198.1) 36f. :!9 On poetry as sema: Nagy (above, note 38) .'51 The general symmetry between the Phaeacian and Jthacan recognition scenes is outlined by Fenik (above, note 38) .53-.5.5.



Richard P. Martin

genre? But the queen, clever woman, king's proper wife (herself like a good king: Od. 19.108£.), must no doubt have already understood the speaker for what he is-not beggar but king, instructor of princes. 40 40 A shorter version of this paper was delivered at the APA convention in Philadelphia in December 1982. For useful suggestions, I wish to thank the anonymous TAPA referees and Gregory Nagy.



autap E1t£t 1COOJlTJ8Ev aJl' Tl'YEJlOVEOuyov Kai a9iaq>atov o~j3pov, KA.ayyfi ta{ YE 7tetovtat E1t • 'flKEavot:o poac.ov like birds, just as the shriek of cranes arises in the sky, the ones who, fleeing storm and endless downpour, fly with a shriek over the streams of Okeanos

(3.2-5) Some aspects of this developing image are by now familiar: the shriek of these birds, who in fact have a windpipe up to five feet in length with which they produce piercing, sonorous, trombone-like sounds; the flight from violence that their shriek signals, even if here the violence is not that of predators but of the weather; and also, their presence near a river, in this case the Okeanos. But another aspect of the image is not familiar at all: these cranes are high in the sky, a position reserved for predators in other bird similes. The reference here is to a piece of lore also mentioned in the Works and Days of Hesiod, that the annual southerly migration of the cranes marks the onset of winter: q>pa~Ea9at o' EUt' av Y£pavou q>C.OVTJV E7taiCOU EUpEt"[l ~aioaA.o~ TioKTloEv KaUmA.oKaJ.up 'Aptaovn. ev9a J.1EV l)(SEOl Kat 7tap9£vot QMpEml3tat ropxEuvt', &Ut1A.rov E7tt Kap7tcJ> XEtpa~ exovtE~. trov 0' ai J.1Ev AE1tta~ o9ova~ exov, oi o£ Xt trova~ Etat' £uvvt1tou~. ~Ka onA.~ovta~ £A.a{cp· Kat p. ai J.1EV KaAa~ O''tE> 31 • Perhaps the same can be said of the function of genre, if for poet we substitute socicty32 • (Columbia University New York)


31. M. Douglas, Implicit Meanings, London, 1975, p. 273. 32. An early version of this paper was delivered at a symposium in honor of George E. Dimock, jr. at Smith College in November 1985. My ongoing thinking about the subject owes much to stimulating discussions with M.D. Carroll, A.E. Johnson, N. Loraux, and L. Muellner, and to valuable suggestions from P.E. Easterling, P.-Y. Jacopin, and S.L. Schein.


'Reading the texts': archaeology and the Homeric question E.S. SHERRATT* Specialists in Greek literature have long argued about how Homer's epics were formed and just what they represent. The question is a pressing example of the larger general case -the relationships between archaeology. history and oral literature in many periods and places. E.S. Sherratt, in offering an archaeological perspective on Homer, sets out to establish the stratigraphy of its text and its history. My basic conclusion is that certain 'event-related' structures do not restructure easily. It is my hypothesis that a certain clustering of anomalous features will show the trace of the survival of a structure from 'life'. Pure text were it to exist, would. in contrast. present almost no barriers to quite arbitrary restructuring. ARDEMR 1988: 31

Introduction The Homeric epics- the Iliad and the Odysseyare among the oldest European literary documents. Traditional sources and linguistic evidence suggest that they were substantially fixed in the form in which they have come down to us sometime around 7'00 BC. at the dawn of literate Greek history (Janko 1982); but it is clear that they contain echoes of an even older, orally remembered past. As such. they are bound up with the problem of the relationship of 'history' to oral tradition, which is not by any means confined to ancient Greece (Vansina 1975: Sahlins 1985: ch. 2). The perceived status of the events. society and material culture presented in the epics has fluctuated between history. legend. myth and fantasy. reflecting both the changing attitudes of literary scholars to the nature of their composition (the 'Homeric Question'). and the varying desires of historians and archaeologists to make use of the historical. social or ideological information they potentially contain. They are an area where literature, archaeology and his•

tory meet, in texts which have often seemed to provide some of the most tantalizing glimpses of protohistoric societies whose material remains are known from the archaeological record. But the precise relationships between these glimpses and the formation of the texts which crucially affects the way we use the latter - are still a matter of continuing debate. Students of literature or linguistics. historians and archaeologists have each had their own wa) of approaching the problems. and. as a result, have tended to arrive at what often appear to be incompatibly different answers. Literary approaches Central to this are the implications of the 'Homeric Question' which, in its wider sense (the circumstances of their formation and final composition), is not just a problem for those with a literary interest in the Homeric epics but one which also has some bearing on the historical status of their content. Since Milman Parrv in the 1920s demonstrated the presence of th~ characteristics of orally composed and transmitted poetry in the epics, it has been widely (though not universally) accepted that they are ultimately the product of an oral bardic tradition (Parry 1971: for a discussion of subsequent work on this aspect of Homer see Householder & Nagy 1972: 19ff.). Within this constraint, however. opinions have varied among literary critics and others as to the precise implications

3 Woodside, Wood Green. Witnev. Oxfordshire ox8 6oQ.

ANTIQUITY 64 (1990): 807-24




of what Parry & Lord [1960) have termed the epics' oral-formulaic mode of composition. in particular over the relative roles of a longstanding tradition of transmission and of the creative genius of one poet- Homer- in shaping the Iliad and the Odyssey as we know them. This is essentially a q_uestion of emphasis. For some Homerists the latter is of prime importance, to the extent that the epics may be seen as essentially the work of one poet who. sometime in the years around 700 BC, travelled about Greece collecting and combining a wealth of topographical detail with various tales and traditions [some inherited in verse form, others not) and composed them into poetic works which transcend the inheritance of both subject matter and technique which lay behind them [Taplin 1986: 70f.: Rubens & Taplin 1989). Others, over the years, have taken a less cataclysmic view. Most, however. insist - quite justifiably- on the integral unity of each of these two long epics. and maintain that a single individual was responsible for shaping the final structure of each and ensuring its integrity through various internal linking and unifying elements, such as the prophetic cross- references contained in many of the speeches [Rutherford 1985).

Archaeological approaches For those concerned primarily with the historical and cultural background to the epics as revealed through the archaeological record. greatest interest has lain in identifying the chronological period in which the greater part of their material or cultural content- if not their actual composition - can best be set. Here too there have been distinct shifts of emphasis. shifts which have ranged from the Late Bronze Age [Mycenaean) era in Greece to various points in the Early Iron Age (FIGURE 1). From the time when Schliemann first claimed to have found the graves of Agamemnon and his companions within the citadel at Mycenae in 1876 (Schliemann 1880: 336-45). there arose a growing conviction among archaeologists [if not among Classical scholars) that the material setting and historical background of the epics was essentially that of the Late Bronze Age Mycenaean world. Although Schliemann's identification of Homer's Mycenae with that of the early Mycenaean Shaft Graves was soon rejected as too early by about 300 y~ars - lar"e!y on the

basis of the historical Greeks' own traditional dates for such events as the Trojan War- the belief in a basically Late Bronze Age setting persisted through such writers as Allen [1921), Nilsson [1933}. Page [1959), Wace & Stubbings [1962), Blegen (1962). !vlylonas [1966: 213ff.). and on down to Luce [1975) and Wood [1985) in the last two decades. Meanwhile. Lorimer (1950) and Gray [1947; 1954; 1955; 1958; 1968), while maintaining the Late Bronze Age basis of the events and much of the world portrayed in the poems, stressed the archaeological grounds for regarding the epics as heavily interwoven with the material culture of later periods. By the 1950s. on the other hand, Finley [1954 [1956]; 1957) was arguing strongly for a primary setting in the 10th and 9th centuries - though less on grounds of the material record than on the type of society he saw portrayed in the epics. Later, Snodgrass [1971: 389; 1974) was to argue for a selective mixture of [palatial) Mycenaean and 8th-century elements with little in between: while the emphasis above all on the 8th century was also taken up by Kurtz & Boardman (1971). and- again concentrating mainly on the poems' social institutions- by :\!orris [1986). Finally, Dickinson (1986) returned to the chronological setting proposed by Finley (though in this case more explicitly in terms of the material record) and placed the background and creation of the epics in the period between c. 1200 and the end of the 9th century. These apparently quite different conclusions represent the answers to different questions. each of which arises from a different view of the social function of the epics and stresses the importance of different aspects of their content. At one extreme. many archaeologists have been content to regard them as little more than entertaining accounts of historical or semihistorical events, people or societies which can be assigned to chronological periods with the help of the archaeological record. What is important here is the assumption of an accretion of continuous but gradually fading traditions (in poetic or other form) which underlie and inform the finished epics and which can be traced through the social or material reflections thev contain. At the other extreme, other archae~logists and historians have stressed the primary role of epic in the establishment and enhancement of social and political structures. and conc:uJed :hat the -lrt







Early Historic







Iron 900

} } } }

} } } } }

Age Protogeometric


"D a r



(Submycenaean) r------Mycenaean (Late Helladic) IIIC



Late 1300

- --





A g e"

} } ) } }

Mycenaean (Late Helladic) IIIB


Mycenaean (LH) IliA


Age 1400

Mycenaean (Late Helladic) II


Mycenaean (Late Helladic) I


(Early Palatial)


Chart showing archaeological di,•isions of the Greek Late Bronze :\ge and Early Iron Age with approximate absolute dates.


of the poets lav in weaving a complex web of social ideolog\' and material symbolism in which even the smallest detail played a part. On this view (which presupposes a sustainedly attentive and reflective audience) the final version as an integrated whole is all that counts (cf. Morris 1986); and the task of the archaeological or historical interpreter is to identify the messages encoded in the poems and match them to the social and political circumstances most likely to have produced them.

The archaeology of the texts It seems to me that there must be some systematic method of combining aspects of these

contrasting approaches. while keeping sight of the implications both of oral delivery and of an oral-:·ormulaic technique of composition. which appear to impose a restraint on the extremes of both views. To begin with. we can consider three sample passages from the Iliad. The first. from Iliad xxiii. concerns a prize offered in the games which accompanied Patroklos· funeral (Iliad xxiii.826-35, translated Lattimore (1951)): Now the son of Peleus set in place a lump of pig-iron. which had once been the throwing-weight of Eetion in his great strength: but now swift.footed brilliant Achilleus had slain him and taken




the weight away in the ships along with the other possessions. He stood upright and spoke his word out among the Argives: 'Rise up. you who would endeavour to win this prize also. For although the rich demesnes of him who wins it lie far off indeed. yet for the succession of five years he will have it to use: for his shepherd for want of iron will not have to go in to the city for it. nor his ploughman either. This will supply them.

Aineias' shield (where it now seems to be stuck] and brings it back to deposit it at Achilles' feet (xx.321-4). Apparently having learnt no caution, Achilles again throws his spear at Polydoros and impales him so that the spear ends up right through his body (xx.413-18). One might have thought that this too would have caused difficulties for our hero. but just a few lines further on (x.x.446) there he is with spear in hand again, this time with no explanation as to how it was retrieved. In the same passage Hector appears to have similar difficulties with spears. Having earlier armed himself with an eleven-cubit-long spear (something We have here what seems to be a rather odd like 4 m in length] (Iliad viii.493f: repeated in situation. :\ lump of unworked iron (the Greek vi.318f.; cf. xiii.830), Hector too indulges in solos outokhoonos is unclear. but conveys the some spear-throwing in the thick of battle. In sense of something which is rough and this case ,\thene, deflecting the spear from its unshaped and produced- literally 'self-cast'- target, brings it back like a boomerang to Hecwithout the intervention of metalworking tech- tor's feet (xx.438f.]. niques) has been regarded as a prized possesThe third passage concerns Aias' shield. an sion for a long time. first as the favourite extraordinarv affair which. as Aias enters the throwing-weight of a king and hero, then as battle in Iliad vii.219 is described as tower-like something worth taking as a spoil of war. then as (elite purgon). It is made of seven layers of worth having as a prestigious prize. Yet it is oxhide to which an eighth layer of bronze has suddenlv - almost as an afterthought - recog- been added. apparently as an afterthought nized a; having its prime desirability in a (vii.223). As if this were not enough. this shield. potential for utilitarian use. as a source of a few lines further on in the thick of the fight. agricultural and pastOJ"a! tools. suddenly acquires a boss (messon epomThe second passage comes in Iliad xix-xx. a phalion: vii.267) which has no part in the long account of the battle between ,\chilies and original description. Hector too has a very add the Trojans. In Iliad xix.369-91 Achilles arms shield. at one point described as extending from himself before the battle. He puts on greaves and his neck to his ankles (Iliad vi.11/) and at a cuirass. takes up a sword. a huge and massive another as completely circular (vii.250]. That is shield (described in detail in a lengthy digress- a shield worth trying to imagine! ion). and finally arms himself with a spear (Iliad While these oddities have often been disxix.38/-91. translated Lattimore): missed as examples of the inconsistencies and discrepancies one might expect in poetry simul~ext he pulled out from its standing place the spear of taneously composed and recited orally. the his father. archaeological record suggests that what we huge. heavy. thick. wh1ch no one else of all the have in each case is the juxtaposition or superAchaians could handle. but ,\chilleus alone knew how to wield imposition of more than one chronological reflection. Taking iron first. the de\·elopment of it. the Pel ian ash spear which Cheiron had bmught to his iron use and technology- and resulting cultural attitudes to iron- show a relatively clear pattern father from high on Pelion. to be death for fighters in battle. in the Aegean. In the period between c. 1600 and c. 1200 we have several small objects of iron A little further on. in the thick of the battle. in Aegean contexts. of a size and nature which Achilles throws this spear at Aineias (Iliad required a minimum in the way of working xx.273f.). He misses and it sticks firmly in the (Pleiner 1969: 8-9). Most of these are personal ground (xx.279-80), and Achilles is left without ornaments or other trinkets: iron rings with gold a spear. Divine intervention comes to his aid in or gold-plated bezels. bronze rings plated with the form of Poseidon who pulls the spear out of gold and iron. and iron studs set in gold



(Buchholz & Karageorghis 1973: 26-7). The frequent combination of iron with gold suggests that iron was regarded at this time as an exotic luxury with intrinsic value as a precious metal. no doubt enhanced by its obvious magnetic properties (which may also be obliquely reflected in a recurrent line in the Odyssey: xvi.294, cf. xix.13). From c. 1200 BC onwards. the first small iron blades appear in the Aegean in the form of knives with bronze rivets which are almost certainly imports from the East Mediterranean. The breakthrough in blade technology in Greece itself comes sometime around the middle of the 11th century when the first all iron dagger (or curtailed sword) appears, closely followed by full-sized iron swords [Snodgrass 1971: 217 ff.; 1980). Not long after. by the beginning of the 1Oth century the technology required for more difficult objects like spearheads (which. unlike their bronze counterparts. could not be cast) had been mastered. and. once this was in place, there is little doubt that objects like axes. ploughshares etc .. which are more rarely found in archaeological contexts. were also made. To sum up: the first part of the passage in Iliad xxiii concerning the prize would seem to accord best with an attitude to iron which prevailed between the 16th and 12th centuries. while the second part belongs to a time from c. 1000 on when iron tools were regularly produced in Greece. The spears of the second passage present a less clear picture, but one which nevertheless also corresponds to a distinct chronological pattern. The spearheads of the early Mycenaean period are vast affairs - fully comparable with Hector's eleven-cubit spear (cf. e.g. Dickinson 1977: 70; Karo 1930-33: plates LXXII.215. XCVJ.902, 910. 933. XCVIU49). Representational evidence of this period - which is likely to correspond closely with the self-image which a contemporary warrior class wished to project -shows them in use in close combat (Crouwel 1981: 121; cf. e.g. Karo 1930-33: plate 24; Sakellariou 197 4: Lorimer 1950: figure 8 and possibly figure 7 (but cf. Lorimer 1950: 144-6); and. indeed, it is hard to imagine how else they could be used. The introduction and use of a throwing spear is more problematic. Although much smaller javelins occur in graves from early in the Late Bronze Age, there is no representational evidence that the early Mycenaeans thought of themselves as using


these in militarv contexts: and indeed the ideal of close comb~t indicated both by representations and the nature of fighting equipment tends to exclude this. During the palatial period of the later 14th-13th centuries representational evidence for paired spears - a good sign that their bearer might be thinking of throwing one of them - is associated exclusivelv with hunting scenes. Towards the end of the 13th century new and smaller types of spearheads enter the Mycenaean repertoire (Desborough 1964: 66-7; cf. Sandars 1978: 91 ff.), and for the first time in 12th-century (post-palatial) representations we see paired spears carried in what are apparently military settings (Crouwel 1981: plate 59; Popham & Sackett 1968: figure 43), and at least one instance of spears being thrown in battle (Dakoronia n.d.). Paired spears of equal size begin to appear in graves from the 12th century onwards (Snodgrass 1964: 136). and are the norm in 8th-century Attic representations of warfare where spears are also sometimes seen hurtling through the air (Ahlberg 1971 ). Thus in Iliad xix-xx we again seem to have some chronological superimposition: with Achilles starting with a spear which would seem most at home in the 13th century or considerably earlier, and finishing with one which acts as though it belongs in the 12th century at the earliest. and possibly several centuries later. Finally. shields. The unwieldy. tower-like (rectangular or figure-of-eight) hide body shields of the early Mycenaean period, extending from neck to ankle. and known only from representational evidence, are all of a package with the close combat weaponry associated with them: and they show every sign of gradually becoming redundant once bronze body armour- of the kind found in a late 15th- or eariy 14th-century tomb at Dendra (Buchholz & Karageorghis 1973: no. 712). and depicted on the Knossos Linear B tablets of similar date and on the later tablets from Pvlos (Ventris & Chadwick 1973: 375. 380. i.deograms 162-3) becomes the warrior norm towards the end of the pre-palatial or early in the palatial period. Although figure-of-eight shields continue to appear as decorative - or possibly symbolic motifs in ivorv and faience work. there is no sign in the and 13th centuries that they were actually part of regular warrior equipment used in battle. In fact, the tablets from Knossos




and Pylos which record military equipment are strikingly silent on the subject of shields, as are representations of military scenes in this period. This is not really surprising. Anyone wearing armour like that found at Dendra would have quite enough to manage without manoeuvring a shield as well. At the very end of the 13th century - as part of the same new package of equipment which included smaller throwable spears and the slashing sword - we see the appearance of smaller hand-held targes (e.g. Lorimer 1950: figure 9). These, and somewhat larger shields in a variety of circular, sub-circular or other shapes. intended to protect the trunk, appear quite frequently on 12thcentury and later representations (e.g. Crouwel 1981: plates 53, 59: Lorimer 1950: plates Il.Z. lll.3; Catling 1977: figure 34) and continue to appear on those of the 8th century [cf. e.g. Lorimer 1950: figure 14). By the mid 11th century we have the first appearance in graves of metal shield bosses with long projections, which are most easily interpreted as a response - more or less delayed - to the new fighting tactics which the new types of offensive equipment entailed. Again. then. in the case of Aias' shield it looks as though we have some chronological layering, with one element of the shield lying most easily in a 16th-15th century context. and another perhaps some four or more centuries later. So far. so good. Even those archaeological commentators on Homer who argue most keenly for a date after c. 1200 for the material setting of the poems and the beginning of their formation have conceded that there are a few elements in the epics which can be projected back to an earlier date (Dickinson 1986: 28-30 ). What thev do not tell us. however, is how this actually ~omes about. Dickinson. who rejects the idea of a barciic hexameter tradition going back bevond 1200, resorts to the idea of traditional t~les and memories. and raises the possibility of heirlooms and material relics [Dickinson 1986: 22. 27, 28: cf. Kirk 1960: 190f.); :\!orris (1986: 89f.) suggests conscious archaizing to differentiate the epic from the real world: while Knox (1973: 21). in an article which argues for the essentially 9th-8th centurv nature of the architecture in the Odyss~y. suggests that such features, some of them genuinely recalling ~lycenaean things. were addP.d as de,;criptive J,,tail to provide

authenticity (cf. also Kirk 1960: 191). I do not wish to argue here at length against Dickinson's (and others') rejection of the possibility of a pre-12th-century bardic hexameter tradition. Such a view is far from universally accepted (Kirk 1962: 120; Chadwick 1976: 182f.: West 1988), and Dickinson's argument seems to rest partly on a misunderstanding of the nature and effects of the language changes which take place in a living tradition of continuous oral transmission. For the other, I find it hard to believe that epic audiences, who could stomach such strange contradictions as the neck-to-ankle circular shield of Hector. or the transformation of Agamemnon- in the space of a single bookfrom being lord of all Argos and many islands to ruler of a small kingdom stretching in quite the opposite direction (Iliad ii.108. contrast 569-76). would care very much about a few details added as occasional. and apparently quite arbitrary. 'distancing effects' of a very minor and often inconsistent nature (:Vlorris 1986: 89f.). or put in for the sake of a Disneyland authenticity. In order to see how it does work. we have to look briet1y at the internal structure of the poems. FIGL'RE 2 summarizes the main structural elements of which the epics are made up. :\o chronological expectations or assumptions apart from the most self-evident- are built into this diagram. It is merely a breakdown of the various structural levels recognized by Homeric scholars from :Vlilman Parry onwards and their generalized relationships to each other. At one level are the epics in the form in which they have come down to us: and here I assume. for the sake of argument (but see further below). that once they entered the hands of the rhapsodes in the 7th century and later they remained more or less unchanged in form and content (cf. Kirk 1962: 1976). except for odd tinkering fur the purposes of parochial political propaganda of the kind hinted at by the ancient critics· account of the Salamis entrv in the Greek Catalogue (/Jiad ii.558) for which Strabo records a different ;-..!egarian version [cf. e.g. ~lonro 1963: 271-2). At the opposite end of the scale are the lowest level building blocks identified by Parry: the formulaic lines. line-endings and epithets particularly characteristic of a tradition of oral composition. ln between are the various episodes or main and sub-plots. genre scenes tsuc:J. as acnim;. fightin;s. assembly, council.




Epics in finished form Links

11 .

Formu a1c scenes


Episodes (main plots and sub-plots)


Necessary description

Non-formulaic seines



Incidental descliption

Minor characters



Formulaic lines


Major charjcters

Formulaic endings

Diagram showing the hierarchy of the main structural components of the Homeric texts. from lowest level elements (at the bottom of the diagram) to the epics in their extant _form.


sacrifice and feast scenes - some regularly formulaic in layout and vocabulary. others less so), descriptions (both incidental and essential to the plot with which they are associated). speeches. characters (major and minor). retrospectives (passages which emphasize the genealogy or pedigree of a particular character or object), and the famous Homeric similes. FiGL'RE 3 is an arrangement of these various structural elements on the basis of their relative stability or instability. Given the two main characteristics of an oral-formulaic technique of composition-in-performance - the freedom to create a unique work on every occasion combined with a tendency to phraseological or formal conservatism at certain levels of construction- it seems probable that certain types of elements were more prone to regular alteration than others. AnJ here too the needs and STRUCTURAL MORE


expectations of epic audiences must have played a part. Some elements, such as the similes. one of whose main functions is an illustrative one (to illustrate something which is antique, exotic or generally outside an audience's own experience), are more likely to change regularly as cultural surroundings and experiences alter (cf. Shipp 1972). Other elements seem. on the face of it. less likelv to be subiect to constant alteration. These in~lude det~il or description essential to the plot of any particular story: highly formulaic genre scenes whose convenient repetitive pattern is conducive to a certain amount of inertia; and prefabricated formulaic lines. line endings and epithets which can be used again and again in widely different contexts without the bard having to stop to thin;. (cf. Gray 1947). In addition. there are a number of structural elements which do



Episodes (main plots and sub-plots) Formulaic scenes Necessary description or detail Formulaic lines Formulaic endings Retrospectives Major characters

Episodic links Speeches Incidental description or detail Similes Non-formulaic scenes Catalogues Minor characters Fil:URE 3.


Structural components


the Homeric epics (cf.


Fld ~aritirue er.1phasis in th~

settlement patterns of Mainland Greece from the 12th century on. During the same period we see the spread of the costly and time-consuming practice of cremation- most spectacularly and 'heroically' in the early 10th-century heroburial at Lefkandi and subsequent horse and weapon burials (Popham et al. 1982; Pophamet al. 1989: 118---23). but also rather earlier in the recentlv discovered 12th-centurv cremation tumul~s at Chania near Mycenae (Catling 1985: 21). The third period is the later 8th century, the period associated with the establishment of the historical Greek city-states. This is characterized in some regions by some no less wealthy and impressive burials (Snodgrass 1971: 268, 271; cf. Schefold 1966), and by a renewed emphasis on military and funerary ideals as portrayed in representational art. It is also, however, a time which witnesses other new phenomena. These include, on the one hand. the growth of wider regional and supraregional religious centres (Snodgrass 1971: 421): and, on the other. a burgeoning of interest in offerings. not at contemporary or near- contemporary graves. but at chamber tombs of the :Vlycenaean age (cf. Snodgrass 1987: 159ff.; :VI orris 1988; Whitley 1988). While at one level a proliferation of rich dedications at inter- regional religious centres can still be seen as supplements to- if not substitutes for- the individual and familv statements made bv lavish burials, the very g;owth of these centre; indicates something new: a sense of regional and possibly wider identity which marks a transition beyond the family- or group-based interests of a classic heroic society. and beyond even the concern for internal definition of the newly emergent polities themselves. f\t the same time. the growth of local ·hero· [or ancestor) cult. focussed on tombs of the evidently distant past. marks a transition away from legitimation grounded in the present. Heroes in general were settling into the past (cf. Nagy 1979: 115-16). From now on their value to the interests of family, community and wider groupings lay in possession and conservation of the heritage of tradition they already provided. The scene was set for the spontaneous transubstantiation of kleos (epic glory) into a ready-formed body of polydynamic myth (cf. Vemant 1982a; 1982b: 41-2). In terms both of material culture and of the wider pictur~ we can construct from the




archaeological record. there are some distinct differences between these three periods. Yet in terms of the social ethos and mores of their specifically heroic ideals- based as they are on comparable political and dynastic aspirations and similar methods of laying claim to thesethe differences are likely to be less marked, and not easily distinguishable [cf. Rowlands 1980: 22). In any case, the chronological layering evident in aspects of the epics' material culture offers a warning of the a priori dangers involved in constructing a composite picture [including the social institutions) of any synchronous society from the texts as a whole (cf. Snodgrass 1974; Coldstream 1977: 18). particularly where this depends on a combination of essentially








circumstantial arguments derived from reading between the lines, or from reading a large number of different lines together. While elements of the poems are infinitely separable, they are not - at least for our purposes infinitely combinable. i Early Mycenaean beginnings I suggest that all three of these periods, to a greater or lesser extent. contributed some formative input to the Homeric epics as we know them (FIGURE 4). The first period - that of the pre-palatial and formative palatial Mycenaean world - saw the initial creation of the bardic tradition which formed the basis for later developments. Its traces are visible in a limited but

ARCHAEOLOGICAL ~------.--------.-------,--------~------, FIGHTING HOUSING BURIAL METAL OTHER

----1--Last stage

Later 8th




of active generation and consolidation

---- 1-------------1----i----Active generation

PostPalatial (l2t.h to early 8th centuries


recreation of inherited tradition in several regions,

leaving original remnants









throwing spears; horned helmets; slashing sword;

(longer range fighting)

~aintenance 1-of

(later 14th- inherited 13th tradition cent~ries with lit~le modi fieat ion





Pitched roof;

earthen floor; relative



epic in Peloponnese

Maritime enterprise Utilitarian & "Phoenician iron activities"



Bronze corselet?; boar's tusk Pre-Palatial helmet; Complexity; & Early Creation of large staircases; Palatial

(16thearly 14th


(single) upper spear; chambers; body shield flat roof (close

Intrinsicvalue iron




Stages in the history of epic formation and their material and cultural imports.


"Trojan• War?



significant number of material or cultural elements, above all in an array of close-combat military equipment (cf. FIGURE Sa): the large thrusting spears, the tower or man-covering shields, the thrusting swords, and probably the boar's tusk helmet which, though still occasionally found in graves as late as the 12th and 11th centuries, disappears from representational art of a military nature after 1200 BC. The typical 'epic' use of chariotry in warfare, as transport for warriors to the front (often obliquely referred to in lines with a recurrent stock ending: e.g. Iliad iii.29; iv.419), may first have entered the tradition in this early period (Crouwel 1981), as - towards the end of the period - may the first indications of bronze body armour (FIGURE 5b). Traces of this period may also be visible in the Odyssey's confused hints of palatial complexity and of flat-roofed. dressed-stone palaces with upper storeys. since already by the end of Late Helladic II and the beginning of Late Helladic IliA (late 15th to early 14th century) buildings of this nature can be found in Greece (Kilian 1987). At this stage, it seems probable that the creation of epic was concentrated in the early Mycenaean core area, above all in the Peloponnese whose early kingdoms in the Argolid, Messenia and Laconia are reflected in the main royal personae of the Iliad and in the grand tour of palaces undertaken by Telemachos in the Odyssey (cf. Masse 1980: 9: Gray 1958). The complicated, intensely agnatic structure of royal inheritance displayed by the Atreid dynasty in the epics (Loptson 1986) seems particularly characteristic of an expansive 'heroic' society (c.f. Rowlands 1980: 18ff.), and indeed the 15th century appears to have seen a considerable expansion of political (particularly perhaps Argive) power outside the original Mvcenaean heartland, which culminated sh~rtly after 1400 BC in a series of geographically widespread destructions (Doxey 1987: cf. also Gatling 1989). Among these is a destruction at Troy, and it seems not impossible that a story of the - or at least a - Trojan war (more than one is mentioned in the Iliad) entered the epic tradition at this time (Vermeule 1986). At the very least the presence of what has been interpreted as a recurring 'siege motif in the art of the early Aegean Late Bronze Age (Vermeule 1964: 100f.; 1986: 88-9; cf. Negbi 1978; Laffineur 1983) suggests that the siege of a walled city

was an important theme in the acta (or agenda) of those who counted (or wished to count) during this period. ii Palatial maintenance It is likely that the results ofthis early period of

epic formation were subsequently preserved in a less actively creative bardic tradition during the palace period of the later 14th-13th centuries (Late Helladic IllA2-B). The almost complete absence (even as residual traces) of reflections in the epics which accord with our general picture of life during the floruit of the Mycenaean palaces with their complex bureaucratic administrations, knowledge of writing (albeit of limited application) and centrally controlled industrial production (Finley 1956: 165££.; 1957; Morris 1986: Rubens & Taplin 1989), together with the lack of features of material culture described in the poems which can be tied down specifically to this time, suggest that this was not a period which contributed much in the way of significant input. It is a time which, in several respects, lacks some of the most typical 'heroic' markers. Burials, though still in communal (family) tombs. are notably poorer in grave goods, particularly precious metal and bronze; and while representational art (often now in the form of architectural frescoes) continues to play a prominent part, much of the main emphasis (particularly on pictorial pottery) is on nonmilitarv or svmbolic scenes of chariots and bulls. it was time when (as the Pylas tablets suggest) the complex internal social structure of the kingdoms no longer needed creative definition; and when there were other more systemic means bv which the dominant sector of society could quietly maintain the validity of its position (Kilian 1988). In such a context the main social function of epic is likely to have been the preservation ofthe status quo, perhaps best achieved by the maintenance of a relatively stable tradition in the hands of court poets whose main task (like that of Demodocus in Odyssey viii.489) was to ·tell the tale correctly (kala kosmon)', and to preserve the general themes and forms (and already antique setting) of an existing tradition in much the same way as, in arguably similar circumstances, the official rhapsodes of the 7th and 6th centuries were bound more or less to follow the Homeric canon. This is not to say that the poet's art




during the palatial period was in any WJY confined to straightforward memorization with no scope for improvisation and elaboration, or that its social function was any less importantmerely that it was not engaged in actively reshaping a tradition which best met the needs of its patrons in its existing form. An important part of its social efficacy may have lain in the idea of direct lineal connection between the palace rulers and the 'heroic' forefathers. the founders of the dynasties, and an emphasis on retrospective passages which stress the notion of long continuity could well have been a particular feature of epic transmission during this time. One thinks here of the trouble which the 13th-century rulers of Mycenae took to refurbish the long-disused Circle A of the Shaft Graves and include it within their citadel wall (:'vlylonas 1966: 94-5). iii Post-Palatial re-creation The second period of active generation - the post-palatial era of the 12th to 9th and early 8th centuries - was probably responsible for the


greate~t contribution to th8 epics as we know them (cf. Kirk 1962). This period was one both of new creation and of active transformation, resulting in an updating of the general material culture of the inherited tradition with remnants of the old tradition left primarily in the most stable elements of the structure. It was responsible for the emphasis on cremation as a burial rite; for the equipment and tactics (slashing swords, horned helmets, double throwing spears, bossed shields) characteristic of a slightly longer-range, more mobile style of fighting (FIGURE Sc); and for iron in the form of everyday objects. perhaps accompanied by bronze as a persistent, if often unattainable. ideal for the ultimate in aristocratic weaponry (cf. Snodgrass (1989). and note that it is tools. not spears and swords. that the iron prize is destined to furnish). It was also responsible for the glimpses we get of Odysseus' and others' palaces as relatively simple buildings: for the maritime 'Phoenician' activities (opportunistic seaborne trading, raiding, slave-taking) of the Phoenicians themselves and certain other char-





FIGURE 5. Three generations of Homeric heroes. a-b Pre-palatial and early palatial (16th to early Hth century). c Post-palatial (12th to early 8th century). d late Bth century. Composites drawn from representational and archaeological evidence.




acters in the epic (including Odysseus); and almost certainly for the introduction of completely new heroes. This period is likely, after various vicissitudes, to have seen the gradual crystallization of the general outlines of many of the episodes found in the existing poems; and of numerous other episodes and cycles of episodes which lacked the direct relevance tci later historical conditions (the settlement of the Troad and the movement up into the Black Sea area in the late 8th and ith centuries, and Greek hostilitv to Persia in the later 6th and early 5th centu;ies) which were to prove particularly advantageous to the survival of the Iliad and Odyssey. The outlines of the Greek Catalogue, with its emphasis on Central Greece and its relatively high proportion of place names relatable to centres which became prominent only in the post· palatial period, probably belongs - along with its Trojan counterpart- to this period (cf. Kirk 1985), and particularly perhaps to its later part when a growing sense of local or regional definition and identity foreshadow the territorially based competition of the early historical period. These developments almost certainly took place in several regions of Greece, probably at different times during the course of this long period [cf. West 1988; Janko 1982: 92), which may account both for the dialectal variety and for the appearance of regional comprehensiveness (in terms of the Greek Mainland at least) which can be seen in the epics as we have them. iv From statement to possession: the end of the line This process of re-creation culminated in the third and last period - the later 8th century which is usually accredited with the final formation of the Iliad and Odyssey as we know them. It coincided with rapid colonial activity (including that in the Troad area) and the beginning of renewed changes in fighting equipment and tactics; and the occasional glimpses of planned 'ideal' colonial layout, of characteristic colonial foundation stories (Odyssey vi.3-10), and possibly of incipient hoplite tactics and equipment (e.g. Iliad xvii.354f.; xix.374-9; vii.223; xi.36-7; but cf. Latacz 197i) probably entered the epics at this point (cf. FIGURE 5d). However. the very scarcity of indubitably late 8th-century ret1ections fKirk 1962: 282) sug-

gests that the active generation of heroic epic along previous lines was now drawing to a close, and that some other process - above all perhaps one of consolidation- was beginning to take its place. Part of this process may have been an increasing emphasis on existing epic traditions, not just as expressions of definition for individual families or groups of elites, but as possessions of the wider communities whose sense of self-identity was progressively enhanced in various ways during this period. The transparent attempts to reconcile inconsistencies between the Greek Catalogue (which is above all concerned with localities and their people) and other episodes of the Iliad, while leaving other types of inconsistency untouched, may reflect this process (cf. Iliad ii.686-94, 699-710. 721-8). The circumstances under which a selective range of episodes coalesced. and the final composition of the integrated Iliad and Odyssey as we know them took place, remain very unclear. We do not know where and when this happened, nor whether it was a gradual, cumulative process (0iagy 1979) or done in a single stage by one man. Traditional sources suggest that Homer was a native of either Chios or Smyrna. a location which has long seemed consistent with the predominantly Ionic dialect of the extant epics. Recently, however. West (1988: 165ff.), stressing the contribution of West (rather than East) Ionic to the completed poems. has suggested that Euboea may have been the most important region during this last phase of their development- a proposal which might accord with the archaeological and historical picture we have at this time of Euboean wealth and colonial enterprise, and of its distinction as the setting for the first war in which much of Greece was involved (cf. Jeffery 1976: 63ff.). Yet again, the fact that some of the earliest explicit and unambiguous references to incidents as narrated in the Iliad and Odrssey occur on 7thcentury Attic. Corinthian and .-\rgive pottery might suggest that this corner of the Eastern Mainland was quite closely associated with the emergence of the epics' final form in the years around 700 BC. If. on the other hand, we follow others in believing that the poems we know were composed by a professional travelling poet for performance at a festival at some supra-regional religious centre such as Delos or Oly~npia [Rubens & Taplin 1989: :C9: d. :'-iagy



1979:8 n.l). then the origin and home of Homer becomes immaterial. We begin to glimpse the mechanism by which the epics have already been transformed into what may be regarded as pan-Hellenic possessions. thus completing the process which brought an end to epic generation as an active instrument of definition for a self-consciouslv heroic element in society. The emphasis had s-hifted from statement to p-ossession. From now on the creative function of the bard (aoidos) gave way to the relaying role of the rhapsode. To these final stages of composition we can ascribe many of the most strikingly integrative features of the finished poems: the links between the episodes and many of the speeches. The role of literacy in facilitating or influencing this last operation has been much debated (e.g. Kirk 1962: 98-101: 1976: 122ff.: Goody 1987: ch. 3; l\'lorris 1986: 121ff.: cf. Finnegan 1988). The introduction of the alphabet in the mid 8th century is surely connected with the potential of writing as yet another instrument for elite self-definition (Stoddart & Whitley 1988): and the hexameter graffiti found on late 8th- or early 7th-centurv sherds at Athens and elsewhere bear witn~ss to an earlv association between literacy and poetry. Yet it remains unlikely that poems quite so long as the Iliad and Odyssey were committed to writing in their entirety at this time. and unnecessarY to invoke literacv as a practical aid to their co~ position. While s~me form of written transmission from the time of Homer himself down through the ith and 6th centuries might be necessary to preserve a verbatim text. it is doubtful that this is of very great importance. It seems likely that. even without the aid of writing. the main forms and themes and integrity of the poems would have been transmitted with little noticeable change (Kirk 1962: 319--20). This is what is implied in the verv existence of the tradition of Homeridae (Sons ·of Homer). rhapsodes who claimed descent from Homer and guardianship of his heritage. The absence of anything referable to the material culture of 7th- or 6th-century Greece makes it clear that the epic tradition ;s last role as an active instrument for heroic self-definition was over. As in the palatial


period of the 14th and 13th centuries, it had taken on a conservative function in which conservation of the tradition itself was allimportant. Now, however, the added dimension of pan-Hellenic possession gave it a new permanent stability which was no longer capable of further transformation. Conclusion This reconstruction of the history of Homeric epic suggests that it parallels the patterns of material and cultural change manifested in the archaeological record, and that the two texts literary and archaeological- can indeed be read together. The phases of active generation correspond to periods during which competing groups of rising elites seek to define their image and lifestyle through such devices as ostentatious burial and both visual and verbal representations of military and other prowess complete with the latest. most prestigious equipment. With the establishment of institutionalized power structures, the long-term visual effect of architectural elaboration above ground takes the place of the short-term display involved in richly equipped burials: and the role of epic changes to one which stresses the heritage and stability of the ·history' it represents. Political disintegration and recurring sectional conflict bring a renewed impulse to redefine those elements of society which have most to gain from the distinctive projection of their own distinguished self-image and selfjustification. This is accomplished not only by new creation. but bv the transformation of an existing oral epic tradition in order to dress it in more recognizably contemporary garb. Only those elements of the tradition which, for technical or contextual reasons. are most resistant to restructuring· preserve remnants of previous creation. These act as fossilized traces of the successive contexts which formed the epics as we know them. and which can themselves be read in the archaeological record. :\cknowledgements. ~y thanks to Christiane SourvinouInwood ..-\nthony Snodgrass and oihers unnamed for much helpful comment and ad\•ice on successive drafts of this paper: and to ... listair Sherratt for FIGL'RE 5.




References AHLBERG. G. 1971. Fighting on land and sea in Greek Geometric art. Stockholm: Svenska lnst. i Athen. Skrifter udg. av Svenska lnst. i Athen 16 .. ALLEN. T. 1921. The Homeric catalogue of ships. Oxford: Clarendon Press. ARDENER. E. 1988. The construction of historv: 'vestiges of creation', in E. Tonkin, M. McDon~ld & M. Chapman (ed.), History and ethnicity: 22-33. London: Routledge. ASA Monograph 27. BLEGEN. C. 1962. The Mycenaean age: the Trojan war. the Dorian invasion, and other problems. Cincinnati [OH): University of Cincinnati. Bt.:CHHOLZ, H.-G. & V. K.\RAGEORGHIS. 1973. Prehistoric Greece and Cyprus. London: Phaidon. GATLING, H. 19i7. The Knossos area. 1974-76. Archaeological Reports 1976-77: 3-23. 1985. Archaeology in Greece. 1984-85. Archaeological Reports 1984-85: 3-69. 1989. Some problems in Aegean prehistory 14501380 BC. Oxford: Leopard's Head. CHADWICK. H.M. 1912. The heroic age. Cambridge: Cambridge University Press. CHADWICK. j. 1976. The Mycenaean world. Cambridge: Cambridge University Press. GOLDSTREAM. j.N. 1977. Geometric Greece. London: Benn. CROIJWEI.. ). 1981. Chariots and other means of land transport in Bronze Age Greece. Amsterdam: Allard Pierson Series 3. DAKORONIA. F. n.d. Late Mycenaean ships from Kinos in Locris. Mvcenaean Seminar. l.Jniversitv of London Instiiute of Classical Studies. 25 ..\pril 1990.

DAVIES.). 1984. The reliabilitv of the oral tradition. in L. Foxhall & j. Davies (ed.), The Trojan War: its historicitv and context: 87-110. Bristol: Bristol Classical.Press. DESBOROI:GH. V. 196-1. The lost M1·cenaeons and their successors. Oxford: Clarend~n Press. DICKINSON. 0. 1977. The origins of :.fycenuean ch·ilisation. Goteborg: Paul Astrorns Forlag. Studies in Mediterranean :\rchaeology XLIX. 1986. Homer. the poet of the Dark ,-\ge. Greece and Rome 33: 20-37. DOXEY. D. 1987. Causes and effects of the fall of Knossos in 1375 BC. Oxford Journal of Archaeology 6(3): 301-24. ELIAS. :-.1. 1978. The civilising process: the history of manners. Oxford: Basil Blackwell. FINLEY. M. 1956. The world of Odvsseus. Revised edition. London: Ghatta & Wind~s. 1957. Homer and Mycenae: property and tenure. Historic 6: 133-59. FINNEGAN. R. 1977. Oral poetry: its nature. significance and social context. Cambridge: Cambridge Uni vers ity Press. 1988. Literacy and orality: studies in the tech-

nology of communication. Oxford: Basil Blackwell. GoODY, J. 1987. The interface between the written and the oral. Cambridge: Cambridge University Press. GRAY, D. 1947. Homeric epithets for things, Classical Quarterly 61: 109-21. 1954. Metal-working in Homer, Journal of Hellenic Studies 74: 1-15. 1955. Houses in the Odyssey, Classical Quarterly (n.s.] 5: 1-12. 1958. Mycenaean names in Homer. Journal of Hellenic Studies 78: 43-8. 1968. Horner and the archaeologists. in M. Platnauer (ed.), Fifty years (and twelve) of classical scholarship: 24-31. Oxford: Basil Blackwell. 1974. Seewesen: Archaeologia Homerica: die Denkmiiler und das friihgriechische Epos I. Kapitel G. Gottingen: Vandenhoeck. Hot:SEHOLDER. F. & G. NAGY. 1972. Greek: a survey of recent work. The Hague: Mouton. JANKO, R. 1982. Homer, Hesiod and the Hvmns: diachronic development in epic diction.· Cambridge: Cambridge University Press. jEFFERY. L. 1976. Archaic Greece: the city-states c. 700-500 ac. London: Methuen. K.\RO. G. 1930-33. Die Schachtgriiber von Mykenai. Munich: Verlag F. Bruckmann. KILIAN. K. 1987. Altere mykenische Residenzen. Schriften des Deutschen Archaalogenverbandes 9: 120-4. Mannheirn. 1988. The emergence of wanax ideology in the Mycenaean palaces. Oxford Journal o.f Archaeology 7[3): 291-302. KIRK. G. 1960. Objective dating critcr:a in Homer, Museum Helveticurn 17: 189-205. 1962. The songs of Horner. Cambridge: Cambridge l:niversitv Press. 1976. Homer and the oral tradition. Cambridge: Cambridge University Press. 1985. The Iliad: a commentary I. Cambridge: Cambridge University Press. K:...ox. :1.1. 1973. :l.legarons and megara: Homer and archaeology. Classical Quarterly {n.s.) 23: 1-21. Kt.;RTZ. D. & J. BOARDMAN. 1971. Greek burial customs. London: Thames & Hudson. L.\FFINEt.;R. R. 1983. Earlv Mvcenaean art: some evidence from the West. Ho~se in Thera. Bulletin o.f the Institute :J( Classical Studies 30: 111-22. L.\TACZ. ). 1977. Kompfpariinese. Kamp.fdarstellung und Karnp.fwirklichkeit in der !lias. bei Kallinos und Tyrtaios. Munich: Verlag C.H. Beck. Zetemata: Monographien zur Klassischen Altertumswissenschaft. Heft 66. LATTIMORE. R. (trans.) 1951. The Iliad of Horner. Chicago [IL]: The University of Chicago Press. LoP:so:-.. P. 1986. Argos Achaikon. L'Antiquite clossique 55: 42-65.



LoRD. A. 1960. The singer of tales. Cambridge (.\1.-\): Harvard University Press. LORIMER, H. 1950. Homer and the monuments. London: Macmillan. LUCE, J. 1975. Homer and the heroic age. London: Thames & Hudson. MOI'RO, D.B. (ed.). 1963. Homer: Iliad Books I-XII. 5th revised edition. (First published 1884.) Oxford: Clarendon Press. MORRIS, I. 1986. The use and abuse of Homer. Classical Antiquity 5: 81-138. 1988. Tomb cult and the 'Greek Renaissance·: the past in the present in the 8th century BC, Antiquity 62: 750--61. MORRISON, J. & R. WILLIAMS. 1968. Greek oared ships 900-322 BC. Cambridge: Cambridge l"niversity Press. MasSE, C. 1980. Ithaque au la naissance de la cite. Archeologia e Storie antica 3: 7-19. MYLOI':AS, G. 1966. Mycenae and the .\lvcenaean age. Princeton (!'IJ): Princeton University Press. NAGY. G. 1979. The best of the Achaeans: concepts of the hero in Archaic Greek poetry. Baltimore (MD): Johns Hopkins l'niversity Press. NEGBI. 0. 1978. The 'Miniature Fresco' from Thera and the emergence of Mycenaean art. in C. Doumas (ed.). Thera and the Aegean world I: 645-56. London: Thera and the Aegean World. NILSSON, M. 1933. Homer and Mvcenae. London: Methuen. . P.\GE. D. 1959. Historv and the Homeric Iliad. Berkeley (CA): Univ"ersity of California Press. PARRY. A. (ed.) 1971. The making of Homeric 1·erse: the collected papers of ;\'/ilman Parry. Oxford: Clarendon Press. PLEII'ER, R. 1969. Iron working in ancient Greece. Prague: National Technical Museum ..-\eta l-ilusei Nationalis Technici Pragae 7. POPHAM, M. 1987 .•-\n early Euboean ship. Oxiord Journal of .-\rchaeology 6: 353-9. POPHAM. M., P. CALLIGAS & H. SACHTT. 1989. Further excavation of the Toumba cemetery at Lefkandi. :\rchaeological Reports 1988-89: 117-29. POPH.\).1, M. & H. SACKETT (ed.) 1968. Excavations at Lefkandi. Euboea 1964-66. London: Thames & Hudson. POPH.\M, M., E. TOI.:LOL'P.\ & H. SACKITT. 1982. The hero of Lefkandi. Antiquity 56: 169-74. ROWLANDS. M. 1980. Kinship. alliance and exchange in the European Bronze .-\ge. in ). Barrett & R. Bradley (ed.), Settlement and society in the British Later Bronze Age: 15-55. Oxford: British Archaeological Reports. British series 83. RUBENS, B. & 0. TAPLIN. 1969. An Odvssev round Odysseus. London: BBC Books. . . RUIJGH. C. 1957. L"element acheen dans Ia langue epique. Assen: Van Gorcum. RUTHERFORD, R. 1985. At home and abroad: aspects of


the structure of the Odyssey, Proceedings of the Cambridge Philological Society (n.s.) 31: 133-50. SAHLINS, M. 1985. Islands of history. Chicago (IL): University of Chicago Press. SAKELLARIOU. A. 1974. Un cratere d'argent avec scene de bataille provenant de la /Ve tombe de !'Aeropole de Mycenes. Antike Kunst 17: 3ff. SANDARS. N. 1978. The Sea Peoples: warriors of the ancient Mediterranean 1250--1150 Be. London: Thames & Hudson. SCHEFOLD, K. 1966. Die Grabungen in Eretria im Herbst 1964 und 1965. Antike Kunst 9: 106-24. SCHLIE).!ANN. H. 1880. ,'vfvcenae: a narrative of researches and discoveries at Mycenae anct" Tiryns. (Reprint. 1976. New York: Arno Press.] SHIPP. G. 1972. Studies in the language of Homer. 2nd edition. Cambridge: Cambridge University Press. SNODGRASS. A. 1964. Early Greek armour and weapons .from the end of the Bronze Age to 600 BC. Edinburgh: Edinburgh University Press. 1971. The Dark Age of Greece. Edinburgh: Edinburgh University Press. 197-l .•'\n historical Homeric society?. Journal of Hellenic Studies 94: 1 H-25. 1980. Iron and early metallurgy in the Mediterranean. in T. Wertime & ). Muhly (ed.), The coming of the age of Iron: 335-7-l. ;\lew Haven (CT): Yale University Press. 1987. i\n archaeology of Greece. Berkeley (CA): Universitv of California Press. 1989. The . coming of the Iron Age in Greece: Europe's earliest bronze/iron transition. in M.L. S!Jrensen & R. Thomas fed.). The Bronze AgeIron Age transition in Europe: 22-35. Oxford: British Archaeological Reports. International series 483(i). STODDART, S. & ). WHITLEY. 1988. The social context of literacy in Archaic Greece and Etruria. Antiquity 62: 761-72. T.'\PLIN. 0. 1986. Homer. in). Boardman.). Griffin & 0 . .\lurray (ed.), The Oxford history of the Classical world: 50-77. Oxford: Oxford l'niversitv Press. VAl\!SIN.\, 1. 1975. Oral tradition as history. :V!adison (WI): University of Wisconsin Press. VENTRIS. M. & ). CH.\DWICK. 1973. Documents in .'v/ycenaean Greek. 2nd edition. Cambridge: Cambridge University Press. VER~!El:LE. E. 1964. Greece in the Bronze Age. Chicago (ILl: University of Chicago Press. 1986. Priam's castle biazing. in .\1. Mellink (ed.). Trov and the Troian war: 77-92. Brvn Mawr: Bry;, :VIawr College. VERNANT. ).-P. 1982a. The brigins o.f Greek thought. Ithaca [NY): Cornell University Press. 1982b . .\'lyth and society in ancient Greece. London: Methuen. VESTERG.\ARD. E. 1987. The perpetual reconstruction of the past, in I. Hodder (ed.). Archaeology as




long term history: 63-7. Cambridge: Cambridge University Press. WACE, A. & F. STUBBINGS (ed.). 1962. A companion to Homer. London: Macmillan. WATKINS, C. 1987. Linguistic and archaeological light on some Homeric formulas in S.N. Skomal & E.C. Polome (ed.). Proto-Indo-European: the archaeology of a linguistic problem: studies in honor of Marija Gimbutas: 280--98. Washington (DC):

Institute for the Study of Man. WEST. M. 1988. The rise of the Greek epic, Journal of Hellenic Studies 108: 151-72. WHITLEY, ]. 1988. Early states and hero cults: a re-appraisal, Journal of Hellenic Studies 108: 173-82.

Wooo. M. 1985. In search of the Trojan war. London: BBC Books. .



A Mycenaean 'Akhilleid'? The Mycena.ean origins. of the 'Heroic' world that forms the background of the Homeric epics have long been argued by scholars from different angles - myth, religion, archaeology, language, and literature. In particular 1\Iartin Nilsson noted 1 that much of Greek mythology centred round places like Mycenae and Tiryns which later were of small importance and he drew the conclusion that the stories were largely of Mycenaean origin. The decipherment of the Linear B script by ~Iichael Ventris in 1952 gave new stimulus and brought important new material, which was at once both encouraging and discouraging. For while it was at once clear that Homer's 'Heroic' society bore little resemblance to the centralized bureaucratic world that gradually emerged from study of the tablets, Nilsson's main argument was strongly supported by the occurrence in the new texts of many 'Heroic' names which later passed out of common use. Somehow the story-tellers had preserved a fund of genuine Mycenaean onomastic material. So it was understandable that in the early heady days after the decipherment under the impact of names like 'Alektru6n son of Eteu·oklett·es' there should have been some O\'ereagerness to find Mycenaean echoes and evidence of early ::\Iycenaean poetry in the tablets. Among the heroes who supposedly went back to the early :Mycenaean period is Ajax (who is 'dated' Ly his body shield), and apparently his name occurs in the form of aawa on a Knossos tablet (C 973), at the time believed to be of 15th century date (see below). However, sober textual philology will first compare aau·a with aawaja, the name of a woman at Pylas (En 70.22, Eo 160.2), and such a feminine derivative tells against positing a stem Aiwant- for a3 wa. Nor can it in any case provide evidence for 15th century mythology, since the word in the Knossos text is probably the name of an ox. In this context it is quite possibly a short form of aaworo = Aiwolos, which also occurs as an ox-name several times at Knossos (Ch 896 etc.). Yet when criticism has done its worst, the fact remains that the bulk of the 'Heroic' names on the tablets is firmly established. Still, the key question in assessing their contribution to the history of Epic is when such names ceased to be used in ordinary life. It may have been in the Sub-::\Iycenaean period. In a sober survey of the question of the ::\lycenaean origins of the Epic G. S. Kirk comes to the modest conclusion 2 ' ••• The available evidence about the existence of Mycenaean narrative poetry is indecisi>e, though there are certain indications in its favour'. Only a handful of objects or practices that appear in the Epics can be identified as 'Achaean' with any certainty. Albin Lesky 3 , too, while believing in the existence of epic minstrelsy at the Mycenaean courts, adds ,Es bleibt uns versagt, von mykenischer Heldendichtung eine greifbare Vorstellung zu ge\\inuen. Das betrifft ihre Form ebenso ''ie ihren Gehalt ... Hierin ist auch durch die Entzifferung von Linear B keinerlei \Vande) eingetreten'. Kirk observes that until the archaeolo1

2 3

The Mycenaean Origin of Greek Mythology (Berkeley 1932). The Songs of Homer (Cambridge 1962), 120. RE Suppl. XI ( 196 i ), 694.




gical exploration of Greece is more nearly complete pride of place must be given to the study of the epic language. In this necessarily brief contribution we single out the key names, for it can hardly be denied that the central figures of the two great Epics are men called Akhilwus and Odysseus. On this topic Kirk~ writes: 'Per;;;onal names in Homer are a fascinating but deceptive subject. Those ending in -eus (Linear B e-u) are very old, and the stems to which they are attached are sometimes difficult to explain as Greek. Among the great Homeric heroes Achilleus and Odysseus have this old kind of name - and Idomeneus; usually the -eu.s names belong to heroes of the past, like Tydeus, Oineus, Atreus, Neleus ... '. A similar opinion is expressed in the new edition of the standard work on Homeric word-formation 5 : ,Die meisten dieser und anderer Heroennamen sind offenbar vorgriech., vgl. zu 'AzL(f.))..~u; und 'Oilu(cr)cr~u~ Frisk s. vv. Zur iilteren Heroengeneration gehoren 'A-rp~u~, Tuil~u~, ll1JA~u~." It is our contention that most of these names can be shown to be Greek by the morphological procedures explored by Risch, and further that there is nothing which points to a non-Greek and preGreek origin. We take the latter point first. The assignment of the -eus names to a pre-Greek language is not based on any positive evidence. This may be brought out by considering the parallel case of the pre-Greek place-names like Parnasaoa. With these ample testimony has been brought for the occurrence of this particular name and others of a similar morphology in Asia Minor. The localities can be mapped and the distribution shown to be mainly in the southern half of Asia Minor 6 • Eminent Anatolian specialists have examined the languages of this region and E. Laroche has summarized his findings 7 : 'the Anatolian languages ... provide a total explanation of these three suffixes ... the territory over which the names in -anda, -aasa, and -wanda occur includes the South, Southwest, and Central Anatolia, but not the North, Northwest, and East beyond the Euphrates.' The suffixes in question can be added I. to Hittite 2. to Luwian 3. to Common Anatolian words and 4. to indigenous loanwords adopted by the Anatolian languages, such as parna-. Nothing of the sort has been adduced in support of pre-Greek origin of the personal names under discussion. Where have we the evidence for a -eua suffix 1 It would be strange if the early Greek invaders picked up so many place-names, to say nothing of ecological and cultural words, from the -aasa- language but owed these personal names to a different quite unknown source. Nor can we detect the source of the stems to which this suffix was added: 'Arp-, N"l]A-, 'AXLA(A)-, 'Oilucrcr-, etc. In brief, the determination of the -eua names as of pre-Greek origin is void and ex nihilo. As we wrote some fourteen years ago 8 , this conclusion is simply a confession of failure. Scholars are merely saying that with the given stock of anthroponym components and the known procedures of morphological analysis they have been unable to fit these names into established Greek and Indo-European patterns. They despaired too easily. In a note on the alleged 'difficulty' attaching to these names Kirk comments 9 : 'The difficulties are perhaps diminishing' with a reference to my Minoans and c Op. cit., 118f. 6 E. Risch, Wortbildung der homerischen Sprache (Berlin, 2. Aufl. 1972), 158. 1 See for instance F. Schachenneyr RE 22, 1954, 1511-12, Karte 7. 7 Acta Mycenaea (ed. M.S. Ruiperez) (Salamanca, 1972) I, 127. 1 The Interpretation of Mycenaean Greek Texts (Oxford 1963). 79. 1 Op. cit., 396.


A Mycenaean "Akhilleid"!


Mycenaeans' (London 1961), 147f. In fact a solution had already been proposed in 1956 10 , which sought to demonstrate that these names unbuttDned easily if the morphological procedures set forth in Risch's first edition of his Wortbildung were applied to them. There is no fundamental change in the second edition 11 , so a brief summary is called for. Nu;tor is an agent noun like Mentor. On the model of Meneliiwos we can construct *Nu;e-liiwos, which will, of course,) ~Veheliiwos, and this satisfactorily accounts for Linear B neerawo 12 • Full names .can be shortened. Stheneliiwos) Sthenelos, and such short forms often have suffixes attached, notably -eus and -on, e. g. Eurustheus son. of Sthenelos. This pair of names serves to illustrate a number of relevant points. One is what may be called the Leitmotiv principle so often operative in dynastic name-giving in 'Heroic' society. Both names contain the element sthenos, and their full form would be Euru-slhenes and Sthene-liiwos. The short form of the first analyses as Euru-sth-eus and that of the second as Sthene-l-os. About the latter another point of importance must be made: neuters-stems commonly appear as a first element with a compositional suffix in -i 13 ; Oidi-podes (oiooJ.w, xti.Of.LwA.oy'"" uroiXEia ur~v Y urtpo«v«Aalluc~ 'Apxtru:.rovt«1] (Diss. Athens

Polytechneion 1988); R.L.N. Barber, The Cyclades in the Bronze Age (Iowa City 1987) 46-52,201-21. • Barber (supra n. 3) 50-51, 216; cf. Doumas, Thera 117.

"There is no direct or convincing evidence that specifi< rooms were used as shrines or sanctuaries.,

'Thera IV, 20-25, 49-51, color pis. A-C; Thera VI. 11-13 on the remodeling that reduced the size of Delta 2

A. Michaelidou, "The Settlement of Akrotiri (Thera): A Theoretical Approach to the Function ofthe Upper Storey," forthcoming in P. Darcque and R. Treuil eds., L'habitat egeen prehistorique (BCH Suppl. XVI). I am very grateful to Ms. Michaelidou for allowing me to read and cite her manuscript, and for criticizing my own text. ' Thera VI, 38-41; S. Petersen, "A Costuming Scene from the Room ofthe Ladies on Thera," A}A 85 (1981) 211. Less certain is whether this painting


and its access.

' N. Marinatos, Art and Religion 93-94 for the religious interpretation; M.B. Hollinshead, "Room D2 and the Swal· lows of Akrotiri, Thera," A}A 92 (1988) 253-54 and "Th< Swallows and Artists of Room Delta 2 at Akrotiri, Thera." A}A 93 (1989) 339-54, for a new explanation. Hollinshead identifies the season as summer, not spring, and attributes

the room's decoration and its contents (Thera IV, 42-43: storage jars, cooking pots, grills, sickles) to a late phase of

the room, the

adjacent alcove with a papyrus fresco, and the entire archi-

occupation, not an early shrine.

tectural unit a "shrine" where such a ceremony was actually

performed, as argued by N. Marinatos in Art and Religion





aesthetic taste for interior decoration inspired by the natural world, appropriate to a culture which first applied engineering to the natural environment for the purposes of human comfort-indoor plumbing, lightand air wells, rooms that can be adjusted for climate. As an elegant and prosperous town, Akrotiri included attractive interiors much as in Greek and Roman houses at Delos or Pompeii, where natural landscapes decorated the walls of private homes. Social pretensions may have inspired the Cycladic population of Akrotiri to emulate Minoan culture in private interiors, including scenes of Minoan religious practices, much as Romans and Campanians affected Hellenic tastes in cult as in art. 1° Current perspectives on Roman painted walls in domestic settings may be more relevant to context than Egyptian or Minoan iconography commonly invoked in religious interpretations of the Thera frescoes.'' It seems timely to heed Renfrew's recent reservations on identifying ancient cult evidence, and to acknowledge the same ambiguities in Aegean cult contexts recently acknowledged by archaeologists in the Near East. 12 These preliminary observations on context serve to introduce the wall paintings from the West House as a decorative complex in a private home, without the assumption that it represents a cult center." Compared to Xeste 3, the West House is modest in the number of its rooms and the absence of such Minoan features as

a lustral basin or generous ashlar masonry. But it contained fine, imported Minoan pottery, much of it found in fragments with the miniature fresco, plus Minoan luxuries such as a lavatory, in addition to its fresco decoration." The ground floor rooms face the Triangular Square, and a doorway which leads directly to a stairwell for access to the upper story (or stories). The street-level quarters are typically devoted to 'service' functions, while the upper floor has more spacious rooms, two of which (4 and 5, on the west side) were decorated with frescoes during the town's latest phase of repair. Room 4 is a bathroom complete with a latrine connected to the town's sewer system, as well as a clay bathtub and bronze vessels for heating and washing water." A partition wall separated the bathroom area, the lower walls of which were protected with a "splash" coating of plaster painted yellow ochre, from the rest of Room 4 and its frescoes. Bowls of red pigment suggest the room was still being repainted in the period after the earthquake and before the final eruption.•• Its window was attractively painted to imitate inlaid marble on the sill and marble jars with lilies against the jambs, resembling illusions for decorative effect in Roman painted interiors." Around the walls of Room 4, a set of eight painted panels were found: above a dado of imitation marble is an oxhide stretched across a frame of three vertical posts topped with elaborate finials, and hung

10 F. Schachermeyr, "Akrotiri-First Maritime Republic?" TAW 423-28 on the social dimensions of the Akrotiri

Shanks, "Two Israelite Cult Sites Now Questioned," Biblical Archaeology Review 14 (1988) 48-52. n As argued, inter alia, by N. Marinatos, "The West House at Akrotiri as Cult Center," AM 98 (1983) 1-19; Art and Religion; L 'iconographie minoenne 219-30; Foster (su-

settlement and their implications for private art and tastes;

M. Wiener, "Crete and the Cyclades in LM 1: The Tale of the Conical Cups," Minoan Tlw.lassocracy 17-25, dubbed the influence of Minoan culture abroad the "Versailles effect." Cf. E. Leach, "Landscape and the Prosperous Life: The Discrimination of Genre in Augustan Literature and Painting, • in R. Winkes ed., The Age of Augustus (Lou vain/

pra n. II).

14 Thera V, 17-20, 41-44; Thera VI, 19-34; Doumas, Thera 48-50. Michaelidou (supra n. 4) on the "clear prefer-

ence for imported ware" indicated in Room S, in contrast to the other rooms of the same house; Sakellariou, Fourth

Providence 1985) 189-95, esp. 191 on the "prosperous emu-

Cretological Congress 534, adduces the imported pottery as

lation of plutocratic style" and the "aggrandizement of social

status" through a deliberate variety of megalography; "Patrons, Painters and Patterns: The Anonymity of RomanoCampanian Painting and the Transition from the Second to the Third Style," in B. Gold ed., Literary and Artistic Patronage in Ancient Rome (Austin 1982) 135-73, esp. on Hellenizing wall decoration during the Augustan period as an expression of middle-class "fantasies of identification

additional evidence for the Minoan style of the frescoes, while N. Marinatos sees the Minoan imports as one of the room's ritual qualities.

15 The latrine was first called a "kitchen sink" (by S. Marinatos, Thera VI, 20, 24-29, figs. 2-3, pis. 58-63); more reasonably, Doumas, ArchEph 1974, 209-210, pl. 65b and Thera 54; Palyvou (supra n. 3) 184, 187 fig. 6; Barber (supra n. 3) 212-13. Inevitably, the toilet has been inter-

with a fabled stratum of society" (p. 153). 11 E.g., by A. Sakellariou, "The West House Miniature Frescoes," TAW 147-53 (Isidis Navigium); K. Foster, "Snakes and Lions: A New Reading of the West House Frescoes from Thera," AjA 92 (1988) 253 and Expedition 30:2 (1988) 10-20, argues for influence from the Egyptian

preted as an opening for ritual libations, and nearby objects, such as a clay rhyton in the shape of a lion's head, have been identified as "offerings" in support of a religious explana-

tion: N. Marinatos, AM 98 (1983) 14 and Art and Religion 48-49; contra R.L.N. Barber,JHS 107 (1987) 243. 16 Thera VI, 24-27, pis. 52-53, 58b and 59 (bowl of pigment still in situ; identified by N. Marinatos, Art and Religion 46, 49 as "body paint"). 17 Leach 1985 (supra n. 10) 190-91 on painted marble at

Jubilee Festival on the miniature frescoes.

12 C. Renfrew, The Archaeology of Cult (London 1985) 11-26,393-444. For a valuable survey of religion in Aegean archaeology, see J. Muhly's review of N. Marinatos, Art and Religion in Gnomon 59 (1987) 329-34. Near East: H.

Oplontis as an imitation of extravagant interiors.





with festoons. The ship frieze next door explains these panels as a sort of open cabin or portable shield for an important figure on board each ship, presumably the captain." The use of oxhides as temporary reinforcement of a wall under siege is attested by Homer (11. 12.263; below, n. 73) in a passage which suggests that a freestanding shield of the type depicted here would be more useful. The isolation and abbreviation in painting of this nautical or military equipment resemble the style of panels depicting figure-eight shields found at Mycenae, Tiryns, Thebes, and Knossos. 19 When first excavated, these ikria encouraged a biographical explanation of the paintings in the West House, making it the home of a sea-captain, the "House of the Admiral." This initial reading of the frescoes treated them as scenes derived from personal experience, with the Nilotic frieze setting the maritime adventure in Libya, hence the miniature fresco's early sobriquet, the "Libyan Fresco. " 20 While these panels do link Rooms 4 and 5 with a nautical theme, they contribute more to the general maritime decor of the West House rather than to any specific explanation of Room 4 as shrine or bedroom. The only figural decoration associated with Room 4 is a large panel of a young woman now restored on the jamb of the doorway leading to Room 5. 21 Her hairstyle, pose, and garment instantly dubbed her a young "priestess," one of those identifications that has been retained since discovery and encouraged religious interpretations. 22 She carries what is probably a brazier (resembling clay ones found at Thera) with red, glowing coals, on which she is sprinkling some substance as

if to release its fragrance when burned. The ritual explanation connects this figure with preparations, in Room 4, for a ceremony enacted in Room 5. 23 Context suggests a more practical explanation: she could be fumigating rooms near the lavatory, or applying cosmetics, rather than performing a ceremony." Admitting these rooms their domestic function in the life of Thera allows the "priestess" a more ordinary place in the daily life of the Theran household, in a genre scene related to the home. Suspending ritual associations between the frescoes of Rooms 4 and 5 allows an examination of the miniature frescoes without a more programmatic link to the Room 4 paintings than a nautical theme. The most famous room in the Aegean is a corner room ca. 4 x 4 m, found full of pottery but without much evidence for its function. The window sills were stacked with vessels, probably as a precaution in the final days of seismic unrest rather than in connection with any ritual. 25 The tripod table found on one of the north sills is a type often identified for offerings, but could as easily have served a more domestic function or could have been placed in the window for safety or to dry. 26 Imitation marble panels below the windows continue the decorative theme applied to the window in Room 4. Two large figural panels-two naked boys carrying fish-are restored at the outside ends of the two outer walls, flanking the windows. This theme is an Aegean favorite, familiar from the Middle Cycladic jar from Melos with a fisherman and his catch, and needs no assistance from ritual to account for its appearance on the walls of Thera. 27 Like the ikria pan-

18 Thera V, 41-42, fig. 6, pl. I; Thera VI, 20, 25-26, 34-35, 43, 49, 54. "L. Morgan, "The Ship Procession in the Miniature Fresco," TAW 639-40; M. Shaw, "Painted Ikria at Mycenae," A}A 84 (1980) 167-79, and "Ship-Cabins of the Bronze Age Aiu roV Eh]puuvn«oV Klvrpov rWv

Mv.~vwv, (Athens 1982) 54-63, 103, pis. 12-18; E. Davis, "The Iconography of the Ship Fresco from Thera," in W. Moon ed., Ancient Greek Art and Iconography (Madison 1983) 7; N. Marinatos, Art and Relzgzon 46-4 7; FOster (supra n. 11). 20 S. Marinatos, "The Libyan Fresco from Thera," AAA 7 (1974) 87-94; Thera VI, 34-60, called Room 4 the admiral's bedroom, with his portrait on board the largest ship on the south frieze of Room 5; cf. D. Page, "The Miniature Frescoes from Akrotiri, Thera," PraktAkAth 51 (1976) 135-52 and S. Stucchi, "II Giardino delle Esperidi e Ie tappe della conoscenza greca della costa cirenaica," QAL 8 (1976) 19-73. P. Haider, "Grundsiitzliches und Sachliches zur his-

13 identifies the red "coals" as cosmetic pigment.

" Thera V, 43-44; Hollinshead (supra n. 9). P. Muhly, Minoan Libation Tables (Diss. Bryn Mawr College, 1981) 278; N. Polychronakou-Sgouritsa, "Mv«1)vai"«h Tpl11'0lll«h Tpclwf{fS' 7rpocrf>opWP,"' ArchEph 1982, 26

20-33, cautions (p. 31) that such tables ar~ not exclusively ritual. 27 Melos jar: Atkinson et al., Excavations at Phylakopi zn Melos (London 1904) pl. XXII; Greek Art of the Aegean Islands. An Exhibition (New York 1979) 69 no. 23, pl. 27; Barber (supra n. 3) fig. 13. N. Marinatos, Art and Re/igzon 35-38 and L'iconographie minoenne 219-20 calls them "adorants" making a perpetual offering.

torischen Auswertung des bronzezeitlichen Miniaturfriests

auf Thera," Klio 61 (1979) 285-307 and Morgan, Miniature Wall Paintings 88-92 for critique of the "Libyan"





els from ships' cabins, these boys carrying fish abbreviate moments of island and seafaring life and bring the Cycladic world indoors. With its numerous windows, permanent paving of schist slabs, and imaginative painted decoration, Room 5 offered a comfortable and attractive environment for daily and frequent occupation, the closest to a "living room" of the many excavated at Thera. 28 Three sides of this room, on the north, east and south walls, carry miniature frescoes as an upper border ca. 40-43 em high, of which a length of ca. 7.50 m is preserved. The fourth (missing) wall on the west probably once bore a similar frieze, implied by a fragment found in the excavations. 29 This placement of friezes is traditional inside Minoan villas and elsewhere on Thera, but is also common for other ancient decorative painting inside private homes, most significantly in the house of the Odyssey Landscapes in Rome. 30 The three preserved walls present a landscape of ships and cities on the south facing military episodes on the north, linked by a Nilotic landscape on the east wall. As reviewed above (n. 20), Marinatos's original interpretation linked all three landscapes to the biography of the owner of the West House~an "Admiral" and veteran of a Libyan campaign. Both biographical interpretation and "Libyan" setting have ceded, in the last 10 years, to a more generic scenario within the Aegean. lt Minoan and Cycladic archaeologists continue to seek specific correlations to Aegean realities in the prosopography and topography of these frescoes. The landscapes have been compared to many a mainland coast and island in the Aegean, beginning with Thera itself, to find a setting in the contemporary Bronze Age world. Its costumes and physiognomies have similarly been identified with Mi-


Thera VI, 20, 28; Warren, "rvfiniature Fresco," 115;


noan, Mycenaean, Cycladic, and non-Aegean peoples in the reconstruction of plausible scenarios of the Late Bronze Age. The alternative to this historical approach invokes ritual as the source for the frieze's pictorial program. Morgan's identification of the scene as a seasonal ceremony has gained such wide acceptance that the scene is often labeled "procession of ships" or "nautical festival. " 32 I would like to reopen the debate on these frescoes and their pictorial relatives in Aegean art by examining their elements as part of a connected narrative transforming history into art, a visual counterpart to early epic poetry. The frieze is narrative in the most literal meaning, "erzahlend" or story-telling. No claim can be made to specific names or places in these paintings, just as there are few identifiable Greek myths illustrated in the Bronze Age. But the classification as "genre" dilutes the impact of these frescoes as an expression of the kind of narrative developed in epic poetry" Its images are cognates of epic motifs, without being illustrations of particular episodes. Nor does image defer to text, a view no longer popular in the study of art, where visual narrative is now accepted as independent of the text and even capable of generating stories derived from images. 34 The rich repertoire of visual formulae and motifs mirrors their proliferation not only in other visual arts but in poetic narratives now believed to be in circulation in the early Mycenaean age. THE SOUTH FRIEZE

The Ship Fresco on the south wall (figs. 1-3) is the best preserved and most debated picture in Aegean art, particularly enlightening for nautical archaeology'' A fleet of ships amidst dolphins spans a single narrative perspective between two towns on separate

the Eastern t\Jiediterranean in Ancient History and Prehis-

Page (supra n. 20) I 36 called the frescoes "the wall-paper of

tory. Studies Presented to Fritz Schachermeyr (Brrlin I 977)

suggests this room may have been "the gathering point for the occupants of the house" or even a dining-room. 29 Thera VI, 24, 34-38; Warren, "rvtiniature Fresco,"

Wall Paintings 88-92, I 46-54. 32 As in Barber (supra n. 3) 173 fig. 126, pp. 174-75; first argued by Morgan (supra n. 19) 629-44.

tributes a fragment with a "processional ship" to the west

"Theme in the West House Paintings at Thera," ArchEph 1983,85-105 and Miniature Wall Paintings 146-54.

a sitting-room in a private house"; t\-1ichaelidou (supra n. 4)

I I6- I8; Davis (supra n. I 9) 3-5 begins the narrative on the south wall. Morgan, Miniature Wall Pmntzngs I 62-63 at-

wall. 30 Davis (supra n. I 9) 5 n. to for parallels from Crete and Thera; Odyssey Landscapes: P. von Blankenhagen, "The Odyssey Frieze," RM 70 (1963) 100-46; J.J. Pollitt, The Art of the Hellenistic Age (Cambridge I 986) 185-209; E.W.

Leach, The Rheton-c of Space. Literary and Artistz'c Representations of LAndscape in Republican and Augustan Rome (Princeton I 988) ch. I. 31 S. Immerwahr, "Mycenaeans at Thera: Some Reflections on the Paintings from the West House," Greece and

173-91; Warren, "Miniature Fresco"; Morgan, Miniature




Fresco," 120-21; L. Morgan,

H See the essays in H. Kessler and M.S. Simpson eds., Pictorial Narratwe in Antiquity and the Middle Ages (Washington, D.C. I 983). 15 L. Basch, Le musie imaginaire de Ia marine antique (Athens 1987) 117-32. S. Marinatos, "Das Schiffsfresko von Akrotiri, Thera," in D. Grayed., Archaeologia Homen·ca G. Seewesen (Gottingen 1974) 141-51. 0. Hockmann, Die Schi.ffe des Minos. Schi.!Jbau und Seefahrt im a/ten Kreta. RGZ Arbeitsbericht (1982) 8-12; Morgan, Miniature Wall Paintings I 2 I -42; infra ns. 38-39.





Fig. 1. South frieze, "Ship Fresco," Room 5, West House, Akrotiri (Thera). (From Thera VI, pl. 9) landmasses which frame the fleet and whose figures focus on it. The town on the left (fig. 2) is more modest in size, architecture, population, and landmass, perhaps a small island, in comparison to the extensive promontory with many harbors and peaks which shelters the larger "city" on the right (fig. 3). The first town is encircled by a river which rises in a mountainous landscape and joins the sea abruptly, without a harbor. The rugged skyline beyond the town, made ferocious with a lion chasing deer, also runs above the pastoral world across the river from the town, on the viewer's left. A shepherd sits near a cluster of shelters, perhaps for flocks and their keepers, and seems in contact with the town through his conversation (?) across the stream with a standing townsman, in a similar long shaggy garment. This environment is far too typical of the Aegean to claim a specific identity: it illustrates universal patterns rather than a single locale. The detailed hierarchy of settlements represented throughout these frescoes-from pastoral shelter to small island town, larger city or palace, peak tower/sanctuary, animal enclosures, and stone springhouse-provides a contemporary witness to Bronze Age sites retrieved by archaeological surface surveys. For example, any Greek rural setting features tiny farmsteads or overnight shelters (67j/o'OVIts, «a6oi«la, or «a6toppau&s of the Shield of Herakles where dolphins define a harbor (207-15). The river encircles the town as Okeanos borders the shield of Achilles, or the world of Odysseus, an image always applied to circular shields or reconstructions of archaic cosmology (infra ns. 84-85). This prehistoric fresco suggests an alternative way to visualize this description from Iliad 18, a new source for poetic imagery that survives in Homer. Pictorial equivalents of Homeric formulae may have survived as vividly in the Greek visual imagination as in poetic memory, such that the description of the Shield of Achilles in Iliad 18 represents not so much an imaginary shield or a sophisticated cosmogony as a compound of narratives depicting nature and culture. The inhabitants of the smaller town face right: all eyes are on the stretch of open water with its ships, the primary focus of the fresco and its narrative action. In a maritime culture, the iconography of departure and arrival by sea dominates daily life as it does art, transcending the "seasonal" specificity argued by Morgan. The "fleet" consists of seven ships which float, almost leisurely, toward a second larger town whose occupants hail them in a friendly way (fig. 3). Only one ship is under sail, while the others are being rowed or paddled, hardly methods of propulsion for rapid travel. Together with their festive appearance, their pace has encouraged the view that they are not warships or a true military "fleet" under sail, but participate in a seaborne ceremony. 38 But every ship contains warriors, or at least their helmets, and each is commanded by a captain perched in one of the "cabins."" These military details make the fleet one of

the most "Mycenaean" elements of the fresco, as argued by proponents of "M ycenaeans at Thera. "' 0 This fleet is either departing on a military expedition or returning from one, arriving at a friendly port if not at its home. Whatever the specific scenario, this image of a fleet corresponds to the poetic tradition of a catalogue of ships, each vessel differentiated in passengers and emblems, just as leaders and cities characterize each entry in a poet's catalogue. The variety in locomotion and decoration could indicate heroic ranking according to lineage or privilege, the way shield emblems in Archaic art distinguish individuals without identifying them. Greek poetry offers epithets embellishing ships with flowers and visual decoration as a literary device, not a ceremonial one. In a context where poetry recall~ its own function,' ~indar c~al~enges anyone who cla&ms greater KTEava TE ""' Tt.p.'l among heroes who lived before Hieron (Pyth. 2.62-63}: EV· av8f.a a'{30.uop.a& O"T0,\ov clpET~ lrcEAaOfwv. Odysseus fetches Neoptolemos to Troy, after the death of Achilles, in a 1JO&K&AOITTC;A'!' 1171l (Soph. Phil. 343}, a ship whose gay decoration belied its grim mission, like the winning words of Odysseus himself. These epithets, and others, suggest that the decoration of ships represents an ornamental, narrative device, not a ceremonial function, in the fresco of the fleet. 41 What survives in the Homeric muster of ships (II. 2.494-877}, roster of heroes in battle (seen from the walls: 3.161-242), or in Hades (Od. 11.385-567) may be as old as panoramas like this one. 42 The thousands of vessels which sailed from Aulis far exceed the modest number on this fresco. But a fleet this small sufficed for the earlier expedition Herakles made to Troy, recalled in Iliad 5 (640-43, Lattimore trans.):

rinatos, AM 98 (1983) 9-10; Morgan 1983 (supra n. 33) 85-87 reviews the dagger and related scenes, without Homeric analogies. The first warrior in the Iliad imroduced with a simile is !\tenelaos, who is compared to just such a lion: Iliad 3.21-26. On Homeric similes and Mycenaean art, infra n. 85. 38 As argued by L. Casson, "Bronu Age Ships. The Evidence of the Thera Wall Paintings," !}VA 4 (1975) 3-10; cf. Sakellariou, TAW 149-52; 1\Jorgan (supra n. 19); S. Wachsmann, "The Thera Water-borne Procession Re-considered," IJN A 9 ( 1980) 287 -95; A. Raban, "The Thera Ships: Another Interpretation," l}NA 13 (1984) 11-19; Foster (supra n. 11). "A. Tilley and P. Johnstone, "A Minoan Naval Triumph," l}NA 5 (1976) 285-92; Warren, "Miniature Fresco" 121 n. 18; M.G. Prytulak, "Weapons on the Thera Ships'" l}NA 11 (1982) 3-6; Davis (supra n. 19) 9; J.

Ernstson, "The Ship Procession Fresco-The Pilots," l}NA 14 (1985) 315-20. 40 S. Iakovidis, "Thera and Mycenaean Greece," A}A 83 (1979) 101-102; Immerwahr (supra n. 31); R. Laffineur, "~fycenaeans at Thera: Further Evidence?" in lrlinoan Thalassocracy 133-39. Contra: J. Davis, "Mycenaeans at Thera: Another Look," A}A 85 (1981) 69-70, but see J. Crowley, "More on Mycenaeans at Thera," A}A 87 (1983) 83-85; J. Vanschoonwinckel, "Thera et Ia civilisation mycenienne," AntC/ 55 (1986) 5-41. 41 As suggested by R. Laffineur 1983 (infra n. 57) 113, who invokes Homeric similes and adjectives endowing warriors and their weapons with magical powers, as a parallel for the decorated ships. _. 2 M. Edwards, "The Structure of Homeric Catalogues," TAPA 110 (1980) 81-105, esp. 101-103 on catalogue form and narrative.

he came here on a time for the sake of Laomedon 's horses,





Fig. 2. South frieze, detail: first town. (From Thera VI, pl. 9) with six vessels only and the few men needed to man

I myself, however, kept my black ship on the outside, at the very end, making her fast to the cliff with a cable, and climbed to a rocky point of observation and stood there.

them, and widowed the streets of Ilion and sacked the city.

The pictorial details of arrival at the larger town (fig. 3) belong to the world of Aegean topography but also recall the poetic description of that world in Homer. The long rocky promontory running up to a hill with buildings, and the two small harbors sheltering ships suit several epic descriptions, beginning with lthaka (Od. 13.96-101):

The rocky promontory, small bays sheltering ships, and even the lookout point or ," Bryn Mawr Now. Supplement (Fall 1988) 7-11 on possible Cretan forerunners of the Odyssey.

Evans emphasizes the Minoan identity of the silver rhyton

and its kinship with epic city sieges, citing the Shields of Achilles and Herakles described by Homer and Hesiod, respectively. On the rehabilitation of the silver rhyton as a

51 S. lmmerwahr, "A Possible Influence of Egyptian Art in the Creation of Minoan Wall Painting," L 'iconographie mlnoenne 41-50 emphasizes the mosaic's formal and para tartic composition, closer to Egyptian conventions.

Mycenaean composition or at least commission, see W.S. Smith, lnlerconneclions in the Ancient Near East (New

52 M.C. Shaw, "The Miniature Frescoes of Tylissos Reconsidered," AA 1972, 171-88 for fragments which suggest

Haven 1965) 63-95; E. Vermeule, Greece in the Bronze Age (Chicago 1964) 100-105. E. Davis, The Vapheio Cups and Aegean Gold and Silver Ware (New York 1976)

a performance or games by male figures; Warren, "Minia-

ture Fresco," 127-28. "Davis (supra n. 19) II on Theran adoption of Minoan

227-30, no. 87, maintains Evans's argument for a Minoan craftsman but allows for a Mycenaean patron; however, she classifies silver vessds as Cretan, gold ones as Mycenaean

imagery, lacking scenes of "direct conflict," but she presumes lost military scenes. A. Sakellariou, "ldentite minoenne et identite mytenienne travers les compositions figuratives," L 'iconographie minoenne 292-309, stresses {pp.

(pp. 328-33). ., The inverse of his method has also been practiced: thus


Sakdlariou "de·militarius" the Thera scenes, denying the drowners and warriors in order to make the frescoc:s more

302-308) the un-Minoan nature of themes of conflict (hunt. battle). ,. K. Coleman, A Study of Painted Wall Plaster Frag-

"Minoan": Fourth Cretological Congress 534-35. "As Vermeule put it (supra n. 48) 102-103: "No one is





paring the spectacular new scaling from Chania with the Thera frescoes." The seal impression frames a single dominant and symbolic figure-priest or king ?-against a town designed to fit the lentoid shape. This image, however rich in its details of architecture, landscape, and symbolism, tells no familiar, heroic tale. Instead, it speaks a language idiomatic to Minoan iconography, one we may not be able to understand, and subordinates its landscape to a symbolic hierarchy dominated by this figure in a ritual or royal pose. Here the pursuit of narrative collides with the standard pursuit of style in Aegean prehistory, at a critical period in antiquity (the 15th century B.C.) and in modern scholarship (the last decade) focused on the distinction of Minoan from Mycenaean art. Ever since Matz, we have accepted, abused, and revised certain well-worn stylistic formulas about torsional, flowing Minoan art and the schematic, "emblematic" style of the Mycenaean artist. 56 The Thera excavations offered a new challenge to these conventions, for the chronology of the eruption coincides with the emergence of Mycenaean art, in a context dominated by Minoan imports and influenceY In conventional scholarship, Minoan style would be demonstrated in an artifact from the West House at Thera: a limestone tripod-table with a painted marine landscape, where the dolphins leap in a freeiy improvised setting of rocks and plants. 5 8 A typical Mycenaean comparison would invoke the ikria from the same house, where a single motif has been extracted and repeated for a

symbolic value of its own (above, n. I 9). Although these critical stylistic features in fact help distinguish the Mycenaean component of the miniature frescoes, they may have exaggerated Minoan images as mobile action, Mycenaean as arrested symbols. 59 One of the dangers of these definitions, even under recent modifications, is that they obscure the role of narrative. A more useful assumption might be that M yccnaean art tells a story familiar from Greek literature, while Minoan art accompanies a tradition whose language and meaning are lost. The miniature frescoes may be Minoan in their individual images, but these motifs contribute to a Mycenaean theme, as recognized by numerous scholars. 60 One would expect a narrative of the I 5th century B.C. to be represented in Minoan images, if not composed by Minoan artists. But their story could well belong to a Mycenaean, even Greek epic tradition, much as a Mediaeval illustration of the Trojan War might include a cathedral facade without converting epic heroes to Catholicism. This makes, for example, the horns of consecration in the large town an architectural convention, not necessarily an index of Minoan religion.•• Greek artists experimented in a similar manner in the early Archaic period, incorporating Oriental motifs into mythological scenes. 62 The emergence of Mycenaean art with the participation of Minoan artists, or even as a later phase of Minoan art, presents problems similar to the study of Greek art and artists in the production of early Roman art. 63 Once Minoan motifs and Myce-

ments from the Bronze Age Site of Ayia lrini in the Island of Kea (Diss. Columbia Univ. 1970} 200; see esp. 200-205 on

" E.g., in the work of Laffineur (supra n. 57} with his emphasis on the emblematic and symbolic value of M yce-

the kinship of the Kea frescoes with mainland rather than Cretan art. Cf. Sakellariou, TAW 148-49. "Hallager (supra n. 45}; E. Davis, "The Political Use of Art in the Aegean," A]A 90 ( 1986) 216. "Vermeule (supra n. 37; the Semple Lectures were delivered in 1971, before the discovery of the miniature frieze); J. Hurwit, "The Dendra Octopus Cup and the Problem of Style in the Fifteenth-Century Aegean," A]A 83 (1979) 413-26; G. Walberg, Tradition and Innovation in Minoan Art (Mainz 1986}. 57 As Immerwahr recognized (supra n. 31) 191 in calling the miniature frescoes "the pictorial equivalent of the early Mycenaean art of the Shaft Graves"; 0. Negbi, "The 'Miniature Fresco' from Thera and the Emergence of Mycenaean Art," TAW 645-55; R. Laffineur, "Early Mycenaean Art: Some Evidence from the West House in Thera," BICS30 (1983} I 11-22 and "Iconographic minoenneet iconographic mycenienne a l'epoque des tombes a fosse," in L'iconographie minoenne 245-66. 51 Thera V, 43-44, pis. 24b-25, C; Thera VI, 27; Greek Art (supra n. 27} 31, 79, no. 35.


naean images.

60 E.g., by Immerwahr (supra n. 31} 180: "a real narrative style, that might be termed Mycenaean"; Warren, "Minia-

ture Fresco," 129: "From Thera to Homer we may not need

the aid of the clew of thread"; Sakellariou (supra n. 45); Vermeule 1983 (infra n. 98) 142. 61 Thus Morgan's subtle analysis (supra n. 33) assumes that the arms of the woman depicted near the horns are

lifted in a "votive gesture," according to Minoan convention,

rather than in a gesture of greeting, appropriate to a story of

arrival or return, as maintained by Warren, "Miniature

Fresco," 119, Davis (supra n. 19} 8, and others. 62 As observed by Davis (supra n. 19} 6 who compares the artists of the Thera frescoes and their exploration of the Minoan pictorial repertoire to early Archaic vase-painters. "As observed by Hurwit (supra n. 56) 426 n. 2; cf. Pollitt

(supra n. 30} 150-63. In the Aemilius Paull us monument at Delphi, for example, style, subject and probably artist are Gre: RA 1971, 3-14; Fourlh Crelological Congress 532-38; Page (supra n. 20) 148, n. 1; Warren, "Miniature Fresco," 127 fig. 5. For an additional fragment showing a


swimming or drowning figure over a stylized marine background like that on the silver siege rhyton, see V. Lambri-





Fig. 6. North frieze, detail: landing and sea battle. (From Thera VI, pl. 7) tive (fig. 6). 72 Rarely discussed, this scrap of wall presumably lies too far from the missing city to fortify it, but it could protect the ships of an invading party, like the one built by the Greeks near Troy. The remains of such walls, or even of Early Bronze Age fortifications along the coast of Asia Minor, inspired legends about heroes invading, like the ni'xos 'Hpa•.\7/os remembered in the Iliad. 7 3 The wall built by the Greeks in the Iliad is most elaborately described in the moment of its destruction, like a warrior at the moment of death in epic poetry. Homer's wall features mysterious •pocrcrat, 7rvpyo crrparo?T(S~: 1.1 1.1) built earlier by the Greeks. The testimony of Thucydides contradicts Homer's account of a wall built in the lOth year, but agrees with that of Herodotus, who suggests a substantial structure built by the Greeks shortly after landing (2.118: lav .\a.\ovtrav: fr. 53 D = Plut. De Glor. Ath. 3). But the Kea frescoes, closest to the Thera miniatures in their narrative content, help establish tht Cyclades as one home of Bronze Age narrative cycles. 114 The Kea fragments indicate an architectural landscape with figures, including singing and dancing, a hunt and a feast, and a seascape with dolphins and boats; their epic potential is rich, although the scene is already being classified as a "festival."'" The frescoes from Thera suggest a Cycladic contribution to early epic poetry, just as discoveries from Lefkandi and other Iron Age sites have prompted a recent attribution of the Odyssey to Euboea. 116 Keos. Pan 1," Hesperia 42 (1973) 284-300; K. Abramowitz, "Frescoes from Ayia lrini, Keos. Parts 11-IV," H.sp.ria 49 (1980) 57-85; M. Cameron, "Theoretical Interrelations among Theran, Cretan and Mainland Frescoes," TAW 589; Warren, "Miniature Fresco" 127 suggests the "samt cyclt" as tht Thtran miniaturt frescoes (ptrhaps on tht thtmt of tht hunt, rathtr than battle: cf. Aspis 270-305). '"E.g., Barber (supra n. 3) 181; Morgan, MiniDture Wall Paintings 68. A ntw study of tht Kea frescoes by Elltn Davis and Lyvia Morgan is in progress: E. Schofield, A]A

Easton (supra n. 109) 189-90. G. Nagy, The Best of the Achaeans (Baltimore 1979) 140; cf. Hooker (supra n. 108) 7 and M. Jameson's portrayal, recalled by Vermeule 1983 (supra n. 98) 142, of "'tht Iliad as the rtrord of all the Myctnaean adventures ovtrseas', teltscoping and combining mtmorable tpisodes in epic song." 11 ' L. McCallum, "Frescoes from tht Throne Room at Pylos: A Ntw lnterprttation," A]A 91 (1987) 296. 1 " M. Lang, Minos 20-22 (1987) 333-41 for such a comparison of tht Pylos themes (battles, sacrifice, hunt, feasting) to Homtric ones. 11 ' Coltman (supra n. 54) and "Frescoes from Ayia Irini, 110 111

91 (1987) 242. 11 6


West (supra n. 92) 172.




any dimension at our peril. In philology, some training and research have recognized modern critical theory at the expense of visual evidence, with the danger that contemporary techniques of analysis can seem more essential than ancient realities. The plight of the miniature frescoes in modern scholarship illustrates a divergence in disciplines but also recommends


The "new chapter" opened by the Thera frescoes is written in several languages, few of which all can read. Hittite and Luwian omeiform, Mycenaean and Homeric Greek, Bronze Age ceramic chronology, Minoan fresco technique, Semitic languages and Canaanite poetry, and Indo- European linguistics and metrics are minimum requirements. The paintings are a sobering as well as inspiring reminder that an understanding of Aegean prehistory needs both statistics and style, pottery and poetry, and that we exclude

a reunion.



Errata The author of the the article, "A Tale of Two Cities: The Miniature Frescoes from Thera and the Origins of Greek Poetry", would like to correct errors reprinted here from the original. p. 520: In fact, Evans restored those very features which would provide a city under siege to match the Homeric passage, and certify a Minoan contribution to the prehistory of the Greek epic tradition. p. 532: Individual names and places suggested tantalizing ties to Homer: Piyamaradas (Prium?); Alaksandos (Alexandros, or Paris?); Tawagalawas (Eteokles); Taruisa (frmya > Tg o i 'YJ ?); and Wilusiya (I /.. to -si IFestsehr. E. Risch, Berlin & New York 19861 308 n. 2.0), has made it probable that

-~c.> -~a in the future and aorist of -3c.> verbs with dental stems is neither Aeolic nor Ionic but goes back to the 'Achaean' phase of the epic language. The 6p~~o­ lll" which I have mentioned is one example. The other verbs are {l~aAarr~c.>, TrTOA£J.1I3"', 1TWJ.IIl"', (~)wap­ i!"'• 6af300, KTEpEf3"', 6vcmaAi300; lyy\1aAi3"', CTTV~­ ).(lc.>, IAU.I3"', J.IEPJ.ITIPil.--rtearly all verbs of war. The last in the list suggests that the recurrent Homeric motif of r,mdering alternatives is an early feature. 7 So e.g. P. Cauer, Grun 1rp6s); Achaean and Aeolic discarded it at an early date in favour of its equally ancient rival1T6Tt ( > Myc. posi >Arc. 1r6s). Either it was preserved in the poetic language from the Middle Helladic era, as a useful metrical alternative, or it entered it in the first millennium from W. Greek. (v) eaaehm. A purely W. Greek type of future, apart from a limited group in Attic (1TAevaoV1JOI, etc.). 102 There is no longer any need to deny the W. Greek origin of these Homeric forms a priori. Other considerations have led us to establish a model which can accommodate them without difficulty. One important amplification of the Troy saga that was presumably due to the initiative of poets in or close to Euboea was an account of the gathering of the Achaean fleet at the narrowest point of the Euripus, between A ulis and Chalcis. That it gathered there was a matter oflocal pride, as we sense from Hesiod (if. above, p. 1so), and it is not likely to have been anything but a local invention. The invention included a catalogue of ships and contingents, a later version of which Homer sang and eventually adapted to a Trojan setting in his Iliad. The idea chat the catalogue goes back to some kind of Mycenaean 'document' is a wistful error. It is an Ionian composition, as the very first name in it suggests (nTJvet.ews, unmetrical in any other dialect) and as the repeated reliance on the form vees confirms; evevi)KovTa (6o2) points to Euboean Ionic, as was remarked earlier, and similarly EvCrrTJ (313) in a passage that also belongs to the Aulis narrative. And of all the twenty-nine contingents in the catalogue, none is so fully and distinctively characterized as the Euboean (536-44): IJEvEO 1TVeiOVTES •Aj30VTES ... 6ooi, om6ev KOIJOWVTES, a[XIJTJTOi IJEIJOWTES opEKTi)tcrtv IJEAlTjlatv 6wpT]KOS PTJ~EIV 5T]iwv a~cpi .v8E Ofj'Avs av~

(OJ. 12.369) (Od. 6.112),

about which he remarks: "diL..vOEv, employe pour decrire l'odeur du sacrifice se rcpandant dans l' air, convient aussi bien a decrire un son qui semble remplir I'air." I 1 11 ET 91; there is no implication of humorous or other word-play, or even suggestion that one phrase in any way echoes the other, although Parry does imply that there might be some chronological priority of one or the other. See below, pp. 2.86-88, and, for chronology of the v-movable, McLeod's review of Hoekstra (above, note 2.) 337-38. These are the only two usages of ap.


amJAO{'Y/aEOv ••••

On the other hand, some evidence seems to suggest that poetic signification can go hand in hand with diction in a curious way, bypassing denotative meaning. Various forms of yeywva occur five times in the Odyssey and eleven times in the Iliad unconnected with any 'Tbaaov . .. oaaov correlation to express the idea, "to make oneself heard" in various narrative situations.47 Only once is the verb of shouting a participial form of {Joaw (cf.Od. 8.305), and it is only in this line that a notion of difficult or urgent contact similar to that in some of 46 Cf. ,\aav aelpa'I10Eftva "veil," and alow~ "pride of chastity" from alow~ meaning roughly "pride of status." The former is alow~ as upheld by the women in wearing veils and bringing their attendants when they must go into mixed company; the latter is alow~ as upheld by men in battle, especially in defense of their native city. Thus, in the traditional language of the heroic poems, idea and diction are so closely linked that, as Parry often implied (c£ Studies I 126), the use of English as a descriptive tool may be self-defeating. All this does not mean, of course, that the poetic signification of the word Kp~OEftVOv is completely independent of its denotation in any and all narrative situations. At Od. 3·392, for example, where Kf>'I10EftVOV appears to denote the seal or stopper of a wine-pithos, any notion of a loss of alow~ must be quite inactive. One may easily




imagine, however, that if it were the suitors who were "broaching" (M£w) the Kp~SEf.LVOV of a wine-jar for their own feasting, and not Nestor for a libation to Athena, just this signification could emerge most forcefully.64 , One further group of examples must be considered in this complicated interaction of diction, meanings, and narrative situations. It would seem at first glance that the use of the word in question in line 2 of the Hymn to Aphrodite would again be a case in which no notion of sexual chastity or related type of alSw.mapoKp~OEp.vm

66 Cf. H. Vm. 5.82-90. For factitious explanation of these epithets, cf. Helbig (above, note 54) 218; Leaf and Bayfield, note on 406. In my view these natural explanations, which abound in the commentaries from Ewtathius onwards, usually complement poetical explanations of the same details. That is, the fact that real veils were probably made of linen and were really shiny in Homer's world by no means supersedes what we have said above with regard to the poetic function of this idea in the epic context. Homer's technique was not usually to depict the marvelous, but to transform the ordinary, through the medium of his "art language" (Whitman [above, note 29] Chap. 1). 67 Cypria, Frags. 5, 6; the same function is fulfilled by the ·npa' in H. Vm. 6: cf. esp. lines 8-10.


Vol. 98]



being. But it is still tenable as a general principle, applicable even to passages such as the Kp~SEfLvov, "stopper," of Od. 3·392, that poetic signification is always latent, if not active. It is inherent in the traditional Gestalt but not necessarily brought into play-into obvious resonance with the poetic context-each time that Gestalt is realized. We are thus enabled to formulate concisely one of the obviously crucial skills in the artistry of oral verse composition: when and how to bring into play the meanings inherent in the traditional diction. Whether or not this theory of the latency of poetic signification in some or all Gestalts of epic formulas be accepted, I think one must recognize that, on a statistical basis, meanings of various kinds seem to behave exactly like any other parameters of the Gestalt; to recur to the metaphors suggested earlier, they are like any other feature of the family, or any other fibre in the thread. Neither the denotation" veil" nor the poetic signification "chaste" is fully present in each and every appearance of the word Kp~SEfLVOv in Homer; by the same token, · " vel·1" (o'0OVTJ, Eavos-, r ~\ there are other ways o f denotmg KUJ\VfLfLa, 1


KaA07T'TpTJ, etc.) and there arc passages involving chastity in which no such object occurs. In these latter passages, finally, the idea may be channeled into some expression of" veiling" (KaA01T'TW; see above, note 6o) or simply remain implicit in the situation without giving rise to any diction whatevcr.68 In other words, poetic signification behaves like any other parameter of a Gestalt in this also, that it can appear independently. Just as the metrical parameter - v v I -- !-'II can be realized with no diction even remotely like 1rlovL STJILtP or meaning associable with the idea of" rich fat" or sacrifice in general (ayyEAos- ~A0Ev etc.), so these significations can appear without this rhythm or this diction; just as the rhythmical impulse v v - I - - v can appear without the words Amapa Kp~SEfLVa or any idea of chastity, so the idea can appear without support from rhythm or from diction. 68 For the latter group, to take only close parallels to the situation of Hecuba and Andromache, cf. II. 1.345-48 (note that Briseis follows the two heralds, not vice-versa), and Illas Parva, Frag. XIX. I have been assuming that the hymns and cyclic fragments, if not texts of purely oral performances, represent artistic practices close enough to be useful parallels. Sometimes they are especially useful to indicate the Fortleben of a phenomenon beyond Homer, and thus most likely a part of their common tradition.




At this point the objection could be raised that the foregoing discussion brgely depends not upon a "typical formula" (such as 8wK£ 8' ol, J,\y£• £0'Y]K£v, etc.) which merely advances the narrative, but upon a single, emotionally pregnant word, that is, a motif. A good deal more discussion would be required to give this objection the adequate treatment it deserves, but space forbids our undertaking this here. My own investigations of this question, though only preliminary ones, have led to the conclusion that the amount of poetically neutral diction in Homer-phrases which are mere space-fillers or which merely advance the narrative-is far less than we would ordinarily suppose. Despite its semantic modesty, even a unit like ap.a + dative by itself can, in the proper context, be powerfully evocative of an the honor that accrues to a person attended by chaperones or other followers; even a -rot can be pregnant with all the minatory signification of the formula (motif) Lit With these terms, four different sorts of runovers may be distinguished: free, pendant, embedded, and orphan. 2 I will briefly define free and pendant runovers, and then I will discuss in some detail embedded and orphan runovers. A free runover is bound neither to any particular dux nor to any particular comes. 3 Consider, for example, the runover M'I>JCo{, which occurs just twice in the epics: 1.

tv lie ot OJ.lC!XlAot ~aav &:hcom maJCo{, tv lie ~o\e K£V eilxroA.~v llptaJ.lq> Kal. Tproal. A.{!tOtJ.leY 'ApydTJv 'EA.EvTJv · ae6 l>' ba-tro ~tucret Cipovpa . . .

Il. 4.173-17 4


mo oe 1C£V E' llpt~XJ.lq> 1eal. Tproal. A.{!tot't£



In his article "The Distinctive Character of Enjambment in Homeric Verse," Milman Parry notes that the reader of the Diad and the Odyssey (the argument would apply only more strongly for the original audience) gradually forms "what may be called a sense of the formula. Meeting over and over the same group of words expressing the same idea, he comes to look on this group of words as a whole which has a fixed end" (258). Thus, Parry argues, the reader's responseto 11. 5.16 is conditioned by his memorl of Il. 5.66 and IJ. 17.49 = Il. 22.327 = Od. 22.16:



5 11. 3.335 II. 16.135 11. 19.372; II. 13.440; 11. 15.127; II. 18.371; 11. 22.286, II. 23.561; Od. 5.235 ahnost = Od. 22.80. 6 In our text, n. 5.16, with the enjambment, precedes any of the passages without enjambment, so the response of the reader might lead to the reverse expectation, the expectation of an enjambment where it does not occur. Parry, no doubt, would have argued that the listener's total experience of oral epic, rather than the experience of just this text in this order, is what determines the expectation. Even so, one might prefer to say that the listener or reader learns to expect that the enjambment may or may not occur.



98 5.

TuSdSero S' {mep ti>J.LOv apl.a'tepov fiA.uS' aJCroxi] Eyx£0c; is always bound to the comes oilo' £~uA.' uu"tov-but these slightly more complex relationships are only the result of the (near) identity of the two passages Il. 5.16-19 and Il. 16.477-481. Albert B. Lord has discussed such passages (which he calls "blocks of lines," but which I will call "formulaic molecules") in some detail, first in The Singer of Tales (58): 8 There are ... larger groups of lines which the singer is accustomed to use often, and through habit they are always found together. The repetition of these groups is sometimes word-for-word exact, sometimes not. Often enough the order of the lines is different. But these clusters of formulas or of lines, which are frequently associated together and are recurrent, also mark one of the characteristic signs of oral style.

He continues the discussion more extensively in Epic Singers and the Oral Tradition (75), particularly in the chapter "Homer as an Oral-Traditional Poet" (72-103). Lord notes the presence in both South Slavic epic and Homeric epic of "the repeated gnomic type of line or couplet": By the "couplet" in this case I mean two lines that are always (or almost always) found together . . . . The repeated couplets do not have to be gnomic in content, but may express any oft-repeated idea that can be expressed in two lines. Indeed the couplet is frequently expanded by a line or two. The main thing is that there be a more or less stable block of lines that is frequently repeated and plays an important role in oral-traditional composition.

Such formulaic molecules, which are fairly common in Homeric epic, may tend to protect a runover--::-here, both Eyx£0~ and the proper name in the following line-simply because the presence of the runover allows the molecule to continue. Runovers within molecules may be termed embedded. The function of EyxtC>c;, then, is to be part of the molecule, and in particular, to provide a metrical introduction for the phrase oM' £~aA.' au"tov. Once the 8 Lord's term block of lines may (unintentionally) tend to suggest that the repetitions are monolithic and unvarying, whereas they can be quite flexible; furthermore, they may be less than two full lines long; my term avoids these problems, and also indicates that the larger units are made of of smaller formulaic elements.




performer gets going m this particular molecule, the runover and comes naturally follow. 9 III.


So far I have been discussing runovers which are bound to dux and comes within a formulaic molecule. Many such runovers occur, and I will have occasion to discuss more as my argument proceeds. But another situation may arise: a runover word or short phrase can be tightly bound to a particular comes but not tightly bound to a particular dux. In the passages cited in example 6, the runover occurred twice along with the same dux and the same comes. In example 7, however, the runover occurs three times with the same comes, but with three different duces: Atav, Bci>po, xbtov, 7tEpl na"tpOKAo\0 6av6v-coc; Mvat' &A.~ 'Axa.tiJJv 1taA.A.ew1 .. .. Il. 16.140-141

h: ~~

apa. a\>pptYO MI!V'JlOt atix~ &v~piJJv

~pmCIN. oloiv te 1Coteooetat 6~pt!!OltcXtP'Jl.

Il. 5. 745-7 4 7


~' 5xea. q>Airyea. xooi ~~oeto, A&.~eto ~' tnO Ml!v'Jlat otix~ &v~piJJv ~pmoov. to\o\v te 1Cotrooetat 6PptJ10ltU"tp'Jl.

Il. 8.389-391

erA.eto ~I clA1ClJl.OV tnO PptSu I!Cra cmPa.p6v, tij> Ml!v'Jlot otix~ &v~piJJv ~pmoov. toioiv te 1CO'ttoaetat 6~ptJ10xa-cp'Jl.

Od. 1.99-101


1tav ~e oi. xdpooow cl'"fll 00AlX001CtaV £noc; PptSu ~a. cmPap6v 1CE1COpu6~ov. a.\rtap a1t 1 Oli!CIN &oxic; o\Jv-n:Aa.l!iJJvl XO.I!Q.i 1tOOE tEPI!t6oooa.. Il. 16.801-803

The runover mxUew (II. 16.142 and IJ. 19.389) occurs in just these two passages, embedded in a molecule. VI.


I believe that examining the behavior of runovers has a value in itself, as part of Homeric metrical and syntactic technique. But in addition there may be special applications to specific problems. In this section I touch briefly on three areas which may be illuminated by considering runovers.




A. The Iliad and the Odyssey: As the examples given above demonstrate, many enjambment formulas are identical in the Iliad and the Odyssey. Sometimes, however, the two poems can have different practices. In the next example we have a single runover which takes four different duces; in the Iliad it has one comes (times 2), and in the Odyssey it has a different comes (times 2): 26.

a\ 'tE xa'ta cr'taSJ,illv nmJ.lvfti:ov i)A&crxoucrw oopn Ev ciapwft CYcE 'tE 'YMl'YOpJlTJaet~ eMVTJcrEV oopn Ev eiapwft, CYcE 't''ta J.lalCPCx 1teAoV'ta~

Od. 22.300-301


The first, second and, fourth of these occur in similes, but not the third, which is spoken by Odysseus. 24 This is not a lone instance of different practice in the two poems. For example, J.laljf~O{ro~ is embedded in a couplet which occurs twice in the Iliad; on the first occasion Dione is comforting her daughter Aphrodite, who has been attacked by Diomedes; on the second occasion, Zeus is comforting Artemis, who has been attacked by Hera: 27.

"t\.c; vU cre ,;maS' epe!;e q>tAov 'tEKO~ OiJpavHnvcov Jla'VtS\~, ro~ d n xaxov pet,oucrav ev~;

II. 5.373-374; II. 21.509-510

But J.laljf~O{~ also occurs in a five-line molecule which appears twice in the Odyssey; on the first occasion, Telemachos is addressing the assembly in Ithaca; on the second occasion, Penelope is telling Eumaeus to summon the still-disguised Odysseus: 23 28.

oi. S' eic; '"tepov 1troAeUJlEVO~'ta nav'ta ~UnOL(a, evidentia, etc., as means to activate images in the minds of the listeners in order to produce butl.YJl;t~. the ultimate aim of any poetry(= performance). Among the best-known passages, and of particular interest for the present discussion, are the remarks in De Sub/. (c. 15) on q>aV"taa(a, Quintilian's discussion of the matter (10 6, 2, 29 ff.), and above all Plato's Jon (535b-e). Unfortunately, such discussions have traditionally been dealt with by scholarship concerned with the history of ancient rhetorical theory and literary criticism, while their direct relevance for ancient poetry itself has generally been neglected.



Discourse and Performance in Homeric Poetry


dynamics of the epic performance is no less an experience of the audience; and the involvement of the performer with his images is nothing other than the natural counterpart of the audience's involvement with their images, the natural consequence of their being "drawn" into the reality deployed by the performer. Epic narrative, in other words, is an intensification of the effect and purpose of ordinary oral narrative: it is concerned with reenactment; it recreates the past and makes it real in the here and now of the performance shared by the performer and his audience. In this process, the performer is a genuine interpreter, not merely in the cognitive ("private") sense discussed above, but in the literal sense of a mediator between the past and the present, a go-between effecting the realization of the past in the present. 49 When I say that the particle a v68o~ f..lEV ET)V. m)xa {)' ElQEl:~ emvw ioa vaLE JtOA\JXtll!!WV JtOA\, apyupE0\0\V emocj)up(o\S" apapu(as-. 8EU'TEpov au 8Wp1}Ka TlEp\ O'Ttl8EOO\V E8UVEV oto Kaa\yvl)'To\0 AuKaovos. 1\pJ.iooe 8' a\hws. QJ.Icj)\ 8' ap'


j3cXAE'TO ~(cj)os- apyUp01}AOV

XtlAKEOV, au'Tap ETlE\'Ta OclKOS" J.IE'Ya 'TE anj3apov Kpa'T\ 8 ' en' '(nnouplV.



8uvov 8e >..Oq,os Kaellnepeev E'veuev




Kuvl1}v e\huKTov E'81}KEv


o\ TlaAtXJ.I1}cj)\V apl)pEL

8' a\hws- Mevl>..aos apl}ios EV'rE E'8uvev. 0

First he put the greaves around his legs, 330 fine ones, fitted with silver ankle-pieces. Second he put on his breastplate about his chest, of his brother Lycaon; and fitted it to himself. And about his shoulders he threw his silver-studded sword 335 of bronze, and then his shield great and sturdy. And upon his mighty head he put a well-made helmet with horse-hair crest; and terribly did the plume nod from above. And he took a stout spear, which fitted his hands. And likewise warlike Menelaus donned his battle gear. ~40

gives the following (earlier lines extremely fragmentary):

EtAE('TO 8 aAK\J.Ia) 3oupE &J(w KEKOpU81JEVa XaAKW(


B au['Tws- Mevl>..aos) apTJta 'rEUXE E8uvev.

aom&x K[a\ TlllAll)ICa cj)anVTJ[v Ka\ 8uo 8oupe Kat Ka>.a[s- KVll)llt&xS' Emocj)(up\O\S apapu\asaiJ.cj)\ 8 a(p WJ.I0\0)\V j3cz>.E'r0 .et(cj>OS" apyup01}AOV37

338 339 339a 339b 339C

And he took two stout spears, tipped with bronze. 338 And likewise Menelaus donned his warlike armor, 339 His shield and shining helmet, and two spears 339a And fine greaves fitted with ankle-pieces, 339b And about his shoulders he threw his silver-studded sword. 339c 36 T. W. Allen, op. cit. Vol. 2.




S. West discusses the papyrus variant lines in the light of 'all the great arming scenes in Homer'.JS She lists the following: Iliad 5.735 ff (Athena), 11.16 ff (Agamemnon), 19.364 ff (Achilles), and Odyssey 22.122 ff (Odysseus); she further states that the order is always the same: 1) Greaves, 2) Cuirass (or breastplate), 3) Sword, 4) Shield, 5) Helmet, 6) Spear (or spears). She makes the comment that it was particularly important to put on the shield before the helmet, for reasons of convenience: the plume of the helmet would interfere with the shield-strap if the helmet were donned before the shield.39 In considering the papyrus' 'additional lines' 339abc, I observe firstly that the shield does still precede the helmet: indeed, items 4, 5, and 6 appear in order at the beginning, with greaves and sword being placed last. Now the preceding lines (328-338) have all dealt with the arming of Paris (using several lines repeated in other arming scenes), with that of Menelaus getting only the single line 339 in the 'vulgate.' The papyrus version gives Menelaus four Jines instead of one, and moreover none of the three extra lines is repeated from the earlier description, indeed one is 'unique' in Homer. Rather than Menelaus' arming being simply a repetition of Paris', it is more of a summary, with shield, helmet, and spears all mentioned in the same line. Thus one scholar's implication that the papyrus version 'has brought down the Homeric passage to the level of primitive epic poetry' by unartistic and tedious repetition, which 'only says that the armor of Menelaus was identical with that of Paris',4° is simply false, and 37 Iliad 3,339a-c: a (cf Iliad 6.322, 13.527, Odyssey 1.256, 12.228, 18.228, 22.101) b (=Iliad 18.459, cf 3.331) c (=Iliad 2.45 etc.) I here acknowledge my debt to the computerized database Homer in the Papyri, compiled by Dana F. Sutton, of the University of California, Irvine, 1990. 38 S. West, ibid. I notice that West omits the arming of Patroclus in Ililld 16.130 ff. Also all six elements are not always present: in the passages froJ:)l IliDd 5 and Odyssey 22 the first three items are missing.

39 Ibid. 40 M. van der Valk, op. cit. 545-6.




overlooks the obvious: the second description relates to the first by not repeating it, which would be rather 'tedious' (to use the same scholar's term), but by summarizing it, as mentioned above. In fact, the additional line 339a uses words for shield and helmet which are different from those used in Paris' arming. Thus the papyrus version still focuses upon Paris' arming, but also devotes some space to that of Menelaus: the two go together without any problems of 'tedious' or 'artless' repetition. (I note that, in contrast to van der Valk, Kirk suggests that a fuller description of Menelaus' arming would have underlined the unbalanced nature of the contest.41 This seems to me as unlikely, if not more so, than van der Valk's suggestions.) In this connection I note other 'abbreviated' scenes which mention lists of arms, such as Iliad 13.264-5 (which is admittedly not an actual arming scene), where spears and shields appear in one line, helmets and breastplates in the other.42 In looking at other examples of 'typical scenes,' in particular of sacrifices, I notice that there is considerable flexibility in retaining or omitting 'essential' elements.43 For instance, in the two sacrifice episodes of Iliad 1 and 2 (1.458-469 and 2.421-432), ten of the twelve lines are identical, and in the same order. However when we move to Odyssey 3 (3.447-73), we get only five of these same lines, and in a much longer passage overall. By way of contrast, in Odyssey 12 (12.359-365) we find a considerably shorter version, but still with six of these lines.44 In each of the Odyssey passages a flexibility of composition is exhibited which should lead us to treat with a more open mind passages, like that discussed above, where more than one version of an episode is preserved. In connection with the subject of 'typical scenes,' I should at least refer in passing to the substantial contributions which have been made in this area, beginning with Walter Arend,45 and including 41 C. S. Kirk, op. cit. 316. 42 Cf. also Iliad 23.457 ff. (Thetis requesting new armor for Achilles); 19.359 ff. 43 For these passages I refer to the OCT of Allen and Munro. 44 See also above, pp. 44-5. In these examples the elements are in the same relative order.




(among many others) significant work by Albert Lord,46 Bernard Fenik (on battle scenes).47 and G. S. Kirk.48 In a useful article summarizing scholarship on the subject, Mark Edwards points out that 'Use of type-scenes is probably a better test for orality, at least in Greek poetry, than use of formulae'.49 He also indicates that more work needs to be done on how type-scene structure relates to oral versus written The purpose of this brief (and ongoing) study has been to attempt to show that more than one variant reading in a particular passage can be considered 'authentic,' in the sense outlined I am trying to look at extant variants in the light of what we know about oral poetry; the examples discussed in this paper represent only a fraction of the variants preserved in the papyri and scholia. I refer again to the work of Lord, in particular his use of the term 'multiform' to describe poetry that is ever-changing, possessing no 'original' version. 52 In this connection I also mention the work of G. Nagy, who in an upcoming book develops the concept of mouvnnce; in addition he presents the idea that the oral tradition of 'Homer' must be explained as an evolutionary process, and he proposes a sequence of (at least) five successive periods of increasingly rigid forms of 'recomposition-in-performance.' Furthermore Nagy adduces relevant parallels from such other literatures as Old French, Proven\al, and Classical Arabic: he notes that, for example, three of the earliest 45 Die typischen Szenen bei Homer (Berlin, 1933); reviewed by M. Parry in MHV 404-7, first published in 1936. 46 In The Singer of Tales, e.g. 68-98 and 186-197. 47 Typical Battle Scenes in the Iliad (Hermes Einzelschriften 21, Wiesbaden, 1968). 48 Both in The Songs of Homer (Cambridge, 1962) and in Vols I and II of The Iliad: A Commentary (Cambridge, 1985 and 1990). 49 In Oral Tradition, 7/2 (1992) 284-330; the quote is from 289. Edwards observes that Parry had already noted this in 1933-35: MHV 451-452.

50 Ibid. 290. 51 See above, p. 40. 52 A. B. Lord, op. cit. 100.




manuscript versions of the Chanson de Roland possess not a single identical verse in common with each other.s3 I have chosen not to discuss here the somewhat irrelevant question of the literacy of 'Homer' himself. I do note the variations on the theory of the dictated text,54 including the hypothesis of B. Powell relating the origin of the Greek alphabet to the first writing down of the Homeric In spite of the attractions of the dictation view, it has been pointed out that little study has been done regarding the effects of dictation on a dictated text.56 In any case, this study has sought to show that the evidence points rather to a plurality of written texts, of equal 'authenticity' and authority, which do not all derive from a common archetype. By way of comparison I mention briefly some potentially parallel situations occurring in other disciplines, the first being the textual criticism of the Hebrew Bible. Some scholars have held the view that 'different pristine versions' of the various biblical books, each of equal status, existed, and that they reflected oral reformulations of these books; in some cases multiple variants in a passage have been accorded equal authenticity with each other, in much the same way that I have been suggesting for Homer.57 In addition the appearance of the Dead Sea Scrolls has led to new theories of the history of the biblical text, including the 'local texts' theory of F. M. 53 C. Nagy, Poetry as Perfomumce: Aucieut Greece aud Beyoud (forthcoming). 54 E.g. A. B. Lord, and more recently R. Janko.

55 Barry Powell, Homer and the Origin of the Greek Alphabet (Cambridge, 1991) (developing an idea first suggested by H. T. Wade-Cery, in The Poet of the Iliad, Cambridge, 1952). Powell imagines that Homer dictated his poetry to the inventor or 'adaptor' of the Creek alphabet- and suggests his name may have been Palamedes (236). The vexed question of the dating of the Creek alphabet I believe is correctly dealt with by the Semitic scholars F. M. Cross and J. Naveh: see especially the latter's 'Semitic Epigraphy and the Antiquity of the Creek Alphabet', in Ktldmos 31 (1992) 143-152. The proposed date given there is about 1100 B.C.

56 J. B. Hainsworth, in A. Heubeck, S. West, and}. B. Hainsworth, A Commentary on Homer's Odyssey, Volume I, Books i-viii (Oxford, 1988) 31.




The second comparable situation is the New Testament, specifically the book of Acts, whose text presents the textual critic with unusually thorny problems. Rather than there being one basic text with minor variants, there are two distinct forms of the text, called the Alexandrian and the Western. Both have early papyrus support, but the Western version is nearly ten percent longer. One of the theories advanced to account for this state of affairs contends that a perceived freedom to ' ... incorporate from oral tradition all kinds of additional details' led to a ' ... wild and uncontrolled growth of the text during the first and second centuries'.59 In addition I notice with approval the rejection of the methodology by which one or more manuscripts are compared to an external standard, and the replacement of this methodology by one in which manuscripts are first compared directly with each other.60 One might also think of how the existence of different versions of some episodes in the four Gospels can be thought of {at least in part) as surviving written records of one or more oral'performances.'61 Lastly I tum to the area of music, and refer once again to the article which motivated this paper. Hainsworth mentions the refuge sought by some scholars in musical analogies: Whitman in eighteenth century chamber music, and Havelock in jazz.62 In following up both references, I was somewhat disappointed to find only the briefest of descriptions. In this connection I can now cite a paper by Leo Treitler which argues that Gregorian plainchant melodies were composed and transmitted in a manner analogous to the composition and 57 E. Tov, Textual Criticism of the Hebrew Bible, (Minneapolis, 1992) 172-4, 288. 58 In Understanding the De11d Sell Scrolls, ed. H. Shanks (New York, 1992) ch. 11. 59 B. M. Metzger, A Textual Commentary on the Greek New Testament (London, 1971) 259, 264. 60 E. J. Epp &: C. D. Fee, Studies in the Theory and Metlwd of New Testament Textual Criticism (Grand Rapids, 1993) 62. 61 E.g. the two versions of the Lord's Prayer, in Matthew 6 and Luke 11. 62 Hainsworth, 'The Criticism of an Oral Homer' (see n.2) 94.




transmission of oral poetry.63 Treitler identifies musical'formulas' as well as a 'formulaic system,' along essentially the same lines as the Parry-Lord formulation. In addition I mention an unpublished PhD dissertation which contains a detailed analysis of the improvised performances of a jazz pianist, with specific reference and comparison to Homeric oral formulaic techniques.64 The jazz analogy is perhaps most pertinent when one bears in mind that no two performances are ever the same, and that none is more 'correct' than another; the most that can be said is that one is more 'inspired' (and inspiring) than another .65 As we continue to derive enjoyment and inspiration from Homer, although it usually has to be from a written text, let us endeavour to be mindful of that peculiar (to us) phenomenon of orality, along with all of its associated elements of tradition, freedom and innovation. Let us bear in mind that the poet had more than one way of singing a particular tale, and if fate has preserved evidence for us of this 'multiformity,' may it be a source of enrichment, rather than embarrassment, for our experience of Homer. Graeme D. Bird Harvard University 63 Leo Treitler, 'Homer and Gregory: The Transmission of Epic Poetry and Plainchant' Tire Music Quarterly 60(1974) 333-372. 64 Gregory E. Smith, Homer, Gregory, and Bill Evans? Tile Tl1eory of Formulaic Composition in the Context of Jazz Piano Improvisation (unpublished PhD diss. Harvard University 1983). When learning jazz piano myself, I can recall being told to transcribe and memorize the improvised solos of the masters (from tape recordings), and then to use them as a basis for my own original performances. 65 I also draw attention to the following quotes about jazz: 'The good musicianer, he's playing with it and he's playing after it. He's fmishing something. No matter what he's playing, it's the long song that started back there in the South.' 'Jazz ... has a rich, available tradition, and yet it thrives on freedom and innovation.' 'The blues had an active improvisational tradition.' And in reference to a song entitled 'Tiger Rag': 'There was no single composer: The music was still part of an aural tradition.' Quoted from L. Porter & M. Ullman, Jazz: From its Origins to the Present (Prentice Hall, 1993) S-6, 16, 31.


Acta Ant. Hung. 40, 2000. 479--488 MARTIN WEST


Zsigmond Rito6k has long been my friend, and even longer a friend of Homer. I hope that this short treatment of an old Homeric problem that is still highly relevant to modern controversies will give him as much pleasure to read as it has given me to write. ln Odyssey TJ 81 ff. we read how Odysseus approached Alcinous' palace. Before crossing the threshold he paused to take stock, ~ t£ yap iJ~::A.iou a'iyATJ nt/...£v itt o~::ldJVTJi) tpp{l;,romt

was nonsense. 'So konnte der dichter nur einen zeitgenossen des Alkinoos reden lassen, so konnte Odysseus der Penelope erziihlen, aber nimmermehr der rhapsode seinen zuhOrem.' 10 Friedlander pointed also to the oi without antecedent in 103 and to the fact that tv 'AI-xtv6ow in 132 referred back to the description of the palace interior, ignoring the orchard and vineyard, and would fit perfectly after I 02: 100 XPUOE\Ol 0. apa ICOUpot EOOJlJl!kilful among them and has left the most obvious traces. That so many critics have byen willing to overlook them does not speak well for the progress of Homeric criticism since 1851. All Souls College University of Oxford Oxford OXI 4AL Great Britain

Acla Ani. Hung. 40. 2000


Classical Quarterly 48 (i) 135-167 (1998) Printed in Great Britain

THE HOMERIC POEMS AS ORAL DICTA TED TEXTS In memory of Albert Lord The more I understand the Southslavic poetry and the nature of the unity of the oral poem, the dearer it seems to me that the Iliad and the Odyssey are very ellactly, as we have them, each one of them the rounded and finished work of a single singer.... I even figure to myself, just now, the moment when the author of the Odyssey sat and dictated his song, while another, with writing materials, wrote it down verse by verse, even in the way that our singers sit in the immobility of their thought, watching the motion of Nikola's hand across the empty page, when it will tell them it is the instant for them to speak the nellt verse.

So wrote Milman Parry late in his life. 1 His hypothesis about the origins of the Homeric text has influenced almost every line I have written about Homer. Writing a commentary on 3,000 verses of the Iliad only strengthened my view that the Iliad and Odyssey are texts orally composed in performance, written down by dictation from that same performance. What kind of text we are dealing with matters far more, for editing and interpreting Homer, than other questions, like whether the Iliad and Odyssey are the creations of one single poet (they are, in my view), or whether the poems were created in eighth-century Ionia. To digress for a moment on the latter point, the linguistic evidence offered by G. P. Edwards 2 and myself that these epics antedate Hesiod's three poems (including his Catalogue of Women) seems to me of paramount importance; I already disproved years ago 3 the claim, recently repeated, 4 that the choice of genre has any profound effect on the stage of linguistic evolution observable in any given text of the early epic tradition. I remain unconvinced by arguments for dating Homer to the seventh or sixth centuries. 5 On the contrary, the new arguments of C. J. Ruijgh 6 incline me to date the epics somewhat earlier than I used to, to c. 775-750 B.c. for the Iliad and slightly later for the Odyssey. Whether the Homeric texts are orally dictated compositions 7 is also an essential 1 M. Parry, The Making of Homeric Verse, ed. A. Parry (Oxford, 1971), p. 451 (written in January 1934). 2 The Language of Hesiod (Oxford, 1972). 3 Homer. Hesiod and the Hymns: Diachronic Denlopment in Epic Diction (Cambridge, 1982), pp. 15-16, 192, 220--1. 4 SeeM. L. West, 'The date of the Iliad', MH 52 (1995), 203-19, esp. pp. 204f. 5 For a survey see B. B. Powell, Homer and the Origin of the Greek Alphabet (Cambridge, 1991), pp. 186--220. I agree with West (n. 4), p. 207, that representations of the Iliad do not antedate c. 625; but those of the Odyssey go back to at least c. 660 (Powell, op. cit., p. 211). The arguments of H. van Wees (G&R 41 [1994), 1-18, 131-55) neglect the likelihood that vase-painters are representing heroic battles, with a millture of weaponry characteristic of different dates, and are therefore poor evidence for contemporary warfare. West's recent claim (pp. 211-19) that the destruction by flood of the Achaean wall at II. 12.17-33 is inspired by the destruction of Babylon by Sennacherib in 689 B.c. rests on an archaeological misconception: eighth-century Greeks were perfectly familiar with construction in mud-brick, often on a stone socle; this was normal and need not suggest borrowing from the Near East (p. 213). The effects of torrential rains on unprotected walls of mud-brick are apparent from most excavations of Bronze- and Iron-Age sites in Greece (at Ayios Stephanos in Laconia in my own ellperience), and would have been easily observable in the eighth century. 6 'D'Homere aull origines proto-myceniennes de Ia tradition epique', pp. 1-96 in J. P. Crielaard (ed.), Homeric Questions (Amsterdam, 1995), esp. pp. 21~. For similar datings see A. C. Cassio, 'KEINOI:, KAAAil:TEt/JANOl: e Ia circolazione dell'epica in area euboica', Annali di Archeologia e Storia Antica I (1994), 5~7. esp. p. 64, n. 66. 7 The theory was developed by A. B. Lord, 'Homer's originality: oral dictated tellts', TAPA 84




question because of widespread and persistent misunderstandings about the nature of orally composed epics. The approach of Milman Parry and Albert Lord has explained so much about the Homeric poems that it has all the simplicity and power of some masterly demonstration in physics, and to neglect it is to revert to the world of Newton from that of Einstein. Yet Parry's work, like Einstein's, raises the famous question of whether God plays dice: of what role there is for a creator, for the 'divine Homer' of the ancients, in the systematized and statistical cosmos of Parry's research. For over sixty years Homerists have been endeavouring to relate this new Homer to the Homer we knew before, that presence behind the text whose guiding intelligence is so apparent to any student reader. It is a testimony to the cogency of this approach that scholars have continued to work with it, vindicate it, expand it, and refine it in its own technical fields; but it is a measure of failure that there is still no agreed position on where it belongs in that wider spectrum of Homeric scholarship, without displacing certain other well-tried approaches. Many Homerists have been unaware that, in criticizing the oralist approach, they have been aiming at a moving target. Lord's work was far from over with The Singer of Tales. He next saw to the publication of the text and English translation of Avdo Mededovic's greatest epic, The Wedding of Smailagic Meho: 8 there is now no excuse to say that this poem is 'cited' as being of unusually high quality, or to refer only to very inferior poems from the South Slavic tradition. 9 I had the privilege of knowing him well for the last decade of his life. Just before he died in 1991, he published Epic Singers and Oral Tradition, 10 selected essays on Homer, the Kaleva/a, and South Slavic, Anglo-Saxon, Bulgarian, Central Asiatic and Medieval Greek epic poetry. Lord's last book, lightly edited by his widow, has recently appeared. In The Singer Resumes the Tale 11 he at last responds to the misunderstandings of his critics, an action which his friends had long urged him to do. His modest and gentle character prevented him from publishing any rebuttals during his lifetime. I can at last cite Lord's own words to show how some aspects of the work of Parry and Lord have since been modified, and others have been misunderstood. First, four modifications. First, oral composition is not the same as the composition of an oral epic. In The Singer of Tales the impression is given that oral epic is typical of oral poetry as a whole. Lord denies this explicitly in his later books. These include studies of English ballads and Latvian short songs as well as of the Serbo-Croatian ienske pesme or 'women's songs', which consitute a genre distinct from the jU!UlCke pesme or 'heroic songs'; the former are analogous to some genres of Greek lyric poetry, which share a similar origin. 12 Inevitably, a short poem admits a greater element of fixity in its (1953), 124-34. I have argued other aspects of this case in 'The Iliad and its editors: dictation and redaction', CA 9 (1990), 326-34 (published in Italian translation as 'L'lliade fra dettatura e redazione', SIFC 10 [1992], 833-43), and in a review of H. van Thiel, Homeri Odyssea, Gnomon 66 (1994), 289-95, as well as in The Iliad: A Commenrary. IV: Books 13-16 (Cambridge, 1992). pp. 37-8 (necessarily with extreme brevity there). Cf. similarly M. L. West, 'Archaische Heldendichtung: Singen und Schreiben', in W Kullmann and M. Reichel (edd.), Der Ubergang von der MUndlichkeir zur Lirera/ur bei den Griechen (Tiibingen, 1990), pp. 33-50; Powell (n. 5), pp 229-30; Ruijgh (n. 6), p. 26. 8 D. Bynum (ed.), with A. B. Lord, :tenidba hadji-Smailagina Sina, Kilzivao je Avdo Mededovic (Cambridge, MA, 1974); A. B. Lord (trans.), The Wedding of Smai/agit Meho, by Avdo MededoviC (Cambridge, MA, 1974). 9 So H. Lloyd-Jones, 'Becoming Homer', Nelf. York Review of Books (5 March 1992); cf. Rosalind Thomas, Lireracy and Orality in Ancient Greece (Cambridge, 1992), p. 49. 10 Epic Singers and Oral Tradition (Ithaca, NY, 1991 ). 11 A. B. Lord, ed. M. L. Lord, The Singer Reswnes the Tale (Ithaca, NY, 1995). 12 See M.L. Lord in Lord (n. I 1), pp. 62-8.




wording; but in fact even brief pieces exhibit far greater variation in wording and structure than one might expect. Secondly, Parry had originally hypothesized that all oral poetry consists of formulae, but The Singer of T11les showed that this does not apply to South Slavic epic, which consists largely of formular expressions. Although Homeric poetry contains a higher proportion of formulae than many other traditions, because of the fixity of the hexameter, if everything consisted of formulae, there would be no room for the kind of innovation, both within formulae and outside them, documented by the work of Hoekstra, Hainsworth, G. P. Edwards and myself in the Homeric poems and the post-Homeric tradition. Thirdly, Lord is often understood as having said that an oral tradition is killed off at once by the introduction of writing. He has recently offered a much more careful formulation: literacy carries the seeds of the eventual demise of oral traditional composition .... It is not, however, writing per se that brings about the change; traditional oral epic flourished in the Slavic Balkans for centuries in communities where significant portions of the population were literate. But gradually the epic came to be written down, and the concePt of a fixed text, and of the text, of a song came to be current. With that concept arose the need for memorization rather than recomposition as a means of transmission. 13

I now think that full oral composition in hexameters was still practised in Greece down to the middle of the sixth century, and indeed that it continued in remote rural areas even later, ultimately to inspire magical texts in hexameters and Theocritean · pastoral. Fourthly, Lord has also modified his early view that a text transitional between full oral composition and literary composition is impossible. 14 In one sense, even a composition of proven oral origin like Avdo Mededovic's huge poem The Wedding of SmailagiC Meho could be called a transitional text, because it was inspired by a written text. 15 Avdo, completely illiterate, had heard a version of this song, about 2,000 lines long, read to him from a published collection of songs; it had been taken down by dictation and published in 1886. When he sang it he maintained essentially the same plot, but so adorned the tale that his version was 12,000 lines in length, as well as much better. But his manner of composition is wholly oral-traditional. And there are other types of case. Lord has pointed to Petar Petrovic Njego§ II, a Montenegrin prince of the 1830s who was originally a traditional singer, then learned to write. At first he, as it were, dictated oral-traditional songs to himself, but as he became more educated he became more innovative, introducing rhyme and other non-traditional features into the poetry he wrote. 16 He was producing, in fact, written literature, just as ltalo Calvino does by modifying Italian dictated folk-tales. 17 But there is no reason to think that the Homeric poems are of this nature. Let me turn to five misunderstandings which have dogged Homeric studies as scholars have sought to digest the evidence, unwelcome to many, that Homer's magnificent compositions are the work of someone, as they might put it, so ignorant that he had not learned to use writing to compose them. On the contrary, I can only admire the intelligence and skill of a poet who could exploit his oral-traditional inheritance as well as Homer does. 13 15 16 17

14 Ibid., pp. 212-37. Ibid., p. 102. A. B. Lord (n. 10), pp. 62,68-71. A. B. Lord (n. II), pp. 23~. Ibid., pp. 213--20.




I. First, everyone has used the term 'Parry-Lord theory' or 'oral theory'. Lord has now protested against this: (T]he phrase 'oral theory' with regard to the investigations into South Slavic oral epic by Parry and me is a misnomer. These findings do not constitute a 'theory'; rather, they provide demonstrated facts concerning oral traditional poetry. . . . Where else but to a tradition continuing into modern times could a scholar go to look for clues to the nature of epics such as the Homeric poems and Beowulf, the method of whose composition is not documented and is subject to controversy? 18

This is correct; it is a hypothesis that the Iliad is an oral dictated text, quite a different thing to say this of the dictated texts in the Parry Collection, the genesis of which is meticulously recorded and documented (as anyone who goes to Harvard can verify). 2. The second misunderstanding is that oral poetry is 'improvised'. 19 This is true only in part: everything depends on what connotations we bring to the word 'improvise'. It is improvised in the sense that jazz is improvised-by using pre-existing blocks of material and skilfully putting them together or modifying them in new ways. Sometimes the effort shows: 'the poet makes a minor slip in the consistency of the plot, or he creates a line that is metrically faulty. Often, indeed, the best poets will be those who need to resort to prefabricated material with the least regularity. This must have been the difference noted by Callimachus 20 between Homer and the Cyclic poets, -rove 'av-rdp lTTt.oiccTJc refers to the Early Iron Age shield with its prominent central boss, 33 whereas dc7Tiboc d!L•f3p677Jc refers to the tower-shield seen on the dagger from Shaft Grave IV at Mycenae, and still reflects Proto-Indo-European syllabic r. 34 Moreover, such fossilized phrases as the Aeolic forms in·I>.wv i.p~v and adjectives like' EKn)pEOc (with -pE- for Ionic -pt-) show that the tale of Troy was already being told a couple of centuries before Homer, 35 and self-styled Aeneadae were certainly among the tale's early patrons. 36 In the texts themselves, Shipp plotted the incidence of relatively 'late· linguistic forms and found them clustered in, notably, the extended similes; these have long been conjectured to be one of the greatest glories of the 'monumental poet'. 37 But all this proves nothing about the quality of the poetry composed by Homer's predecessors: as it was oral and not written down, we do not have it. Since we have no texts earlier than the epics themselves (and certainly not the poems of Hesiod), we cannot directly examine the growth of the stories, as we can in certain other traditions, notably that of the Nibe/ungenlied. 38 But we can do so indirectly, through the approach called Neo-Analysis, if that approach is properly conceived: for I believe that the undoubtedly post-Homeric Epic Cycle preserves unchanged many story-patterns which Homer already knew and adapted in his epics. 39 The Homeric epics completely transcend the melodramatic soap-operas, replete with miracles and fantasy, which made up the Trojan and Theban Cycles, as well as their poverty of style. The purity of Homer's imaginative vision, no less than his compositional artistry, far outshone the quality of these works. As J. Griffin 40 has suggested, they were typical of the oral tradition from which the Homeric epics arose. But how can we say for certain that they were the best it could produce? From study of the diction, we can be sure that invention and innovation were possible-and widespread-in the prehistory of the tradition; and from Neo-Analysis we can be sure that Horner did change his tradition, not to say transcend it. Against the mass of evidence on which the researches of Parry and Lord rest, fine scholars in the Unitarian tradition, like Reinhardt and Griffin, have counterpoised the subtle and elaborate construction of the Homeric epics, their excellence in the depiction of character, and their 'literary' qualities, as clear evidence that Homer must have composed with the aid of writing. And there the matter stands, with no clear path of reconciliation between the two sides. The paradox of such great literature-for we cannot call it anything else-in an oral forrnular style has inevitably provoked the compromise, accepted by many, that, although Homer carne at the end of a long oral tradition, his poems are so good that they must have been composed with the aid of " R. Janko, The Iliad: A Commentary IV Books 13-16, p. 13. P Wathelet in Y Lebrun (ed.), Linguistic Research in Belgium (Wetteren, 1966), pp. 145-73. For further examples see Janko (n. 33), pp. 9-14; Ruijgh (n. 6), pp. 63--92. 35 Cf Janko (n. 33), pp. 15--19; Ruijgh (n. 6), pp. 53--{i3. ,. So Janko (n. 33), p. 19; cf. West (n. 4), p. 217, although there is no reason to think that the /lim/ was so named because of its place of composition rather than its content. 37 G. P. Shipp, Studies in the Language of Homer, 2nd edn (Cambridge, 1972). He was obviously wrong to suppose that they are interpolated. 31 Sec A. T. Hatto, The Nibe/ungenlied(London, 1965), pp. 37~95. 39 I ...~11 argue this case in my paper 'Homer and Neo-Analysis', to appear. "" 'The epic cycle and the uniqueness of Homer', J HS 97 (1977), 39-53. The superiority of Homeric epic to much of the tradition is vainly questioned by Nagy (n. I 8), p. 29. lo4




wntmg; they are 'oral-derived'. In fact this is not a reasonable and moderate compromise, but rests on an unexamined assumption, commonly made by literate and illiterate people alike, that written literature must be superior to oral literature. This is false: there is good and bad oral literature, just as there is good and bad written literature. As we will see, oral and written literature differ in certain respects, but we cannot assume that oral poets are incapable of managing large structures like the story-patterns of the Iliad and Odyssey, nor that they are incapable of achieving powerful effects in detail. 41 Other traditional poetries face the same problem. The Song of Roland is greatly superior in 'literary' qualities to the other chansons de geste, yet displays all the marks of oral composition and is older than the other chansons. 42 How do we explain Homer's superb quality, which is why all these questions matter? Is it owed to Tradition with a capital T, to the spirit of the Hellenic Volk, to an Editorial Board chaired by Pisistratus, to a poet's inspired pen, or to the genius of a single illiterate composer? Even if one believes, as I do, that the Iliad and Odyssey must be oral dictated texts, there is much to be said in favour of premeditation and writing as aids to Homeric composition. Lord showed that, once a poet is accustomed to the slower pace of dictation, he can take advantage of it to improve the quality of his performance, not only in terms of avoiding defective lines, but also in more careful premeditation of the story itsel( 43 But poets composing orally cannot go back and alter what they have compos~d. As Horace said in another context, nescit vox missa reverti. 44 If the Homeric poems were in fact composed with the help of writing, we would expect them to exhibit a much smoother surface and a much more self-conscious style than they have. But can we move from probable hypothesis to observing the process of dictation at work in the texts? I believe that we can, frequently. We can be certain that Homer did not use writing to improve his texts. A poet using writing or an editor altering his work would have done something about such incurably unmetrical verses as

On the morning when Odysseus is to slay the suitors, Zeus thunders 'from the clouds' as an omen to the hero, who is suitably gladdened ( Od 20.1 03-6): aUT{Ka


J{Jp0vrt]C£V d1r' alyA~EVTOC 'OAUp:trov,

EK v£tj>£wv·

y~BT}C£ 8€ 8ioc '08vcwlc. 8' E~ oiKoto yvv.j 1rpo£T}K£V aA£Tplc .,.>...,.,c{ov, lv8, ol J.'U>..a, £iaTo TTOr.p.Jv, AaWv.



Next, an old servant hears the thunder and exclaims to Zeus that 'there is no cloud anywhere' (Od 20.113-14): ~ pa J.LVATJV CT~caca i1roc tj>a TO, c~p.a avaKTI" 'Z£V 1TtiTEp, Oc TE 8£oic& Kal d.v8prd•TfYEpl-rrJc, throughout, when alternative phrasings were available. 52 Conversely, the less frequent a repetition, the more I think we should notice it. Take for example the famous couplet on the flight of a warrior's soul from hts booy: 1/Jvx~ ll' EK pE8iwv -rrTal-'ivrJ ,'11,/locllE {U{JT}KE•, Ov TrOT,.,.ov yoOwca., A"rUw tiv~poTJjTa Kat ~~'1V.

A common enough event in the Iliad, one might think; the distich ought to be rather frequent. Yet these ancient phrases are applied in the poem only to the two most important deaths, those of Patroclus and Hector. 53 This looks no less deliberate than Vergil's echo of these verses, which are likewise applied to the deaths of only two warriors--Camilla and Turnus. 54 These lines are used quite differently from the common descriptions of a warrior's demise, such as 'he fell with a crash, and his armour clattered about him'. In the latter case the context is not significant, beyond 49 Lord (n. II), p. 16. For other instances, see J. M. Foley, Traditional Oral Epic (Berkeley and Los Angeles, 1990). p. 28, n. 18. so Ibid., pp. 17-18. " Cf. (n. 33), pp. 22-5. Another tendency in the MSS of Homer is the standardization of repeated verses against each other: see my article 'The Iliad and its editors" (n. 7), esp. pp. 332-3 tSIFC 10 [1992), esp. pp. 840-2). ' 2 See my notes on Iliad 14.292-3, 310-12. 53 II. 16.856-7 = 22.362-3; see my note on 16.855-8, and cf. Taplin (n. 19), p. 246.

,. Aeneid 11.831





the fact that the warrior has been hit in the head or upper body. But in these verses on the deaths of Patroclus and Hector, I would advocate a mode of reading-the detection of long-range echoes-normally thought appropriate to Vergil, Dante, or other 'literary' texts. How can we reconcile these two ways of reading? 55 Can the new oral poetics-the approach of Parry and Lord-fit in with the old 'literary' poetics? Before answering this question I wish to consider briefly the use of 'themes'. W Arend 56 discovered, independently of Parry, that many Homeric scenes are 'typical', i.e. consist of a set of standardized elements; thus scenes of sacrificing. feasting, or voyaging by ship are made up of standard motifs, and often contain standard lines. It is in the combination of such themes into songs that the Bosnian singer has the greatest scope for originality57 In the use of themes and story-patterns we can best see an interaction between oralist and 'literary' considerations. For example, consider the journey of Poseidon to intervene in the battle, at the opening of Iliad 13. Zeus has momentarily looked away; Poseidon, seizing his chance to save the panic-stricken Greeks from losing their ships to Hector's onslaught, decides to intervene. He descends from the mountain-peak of Samothrace, whence he has been watching for his opportunity; the island quakes under his steps. But instead of going straight to Troy (as we might have expected), he proceeds to his house at Aigai. Wherever Aigai is, it is certainly out of his way. From Aigai he rides out in his chariot across the sea, which makes way for him, and all his sea-creatures gambol about as he crosses the Aegean. His progress is stately rather than swift; his chariot does not speed his arrival. When he is well off the Trojan coast, he goes to a submarine cave between Imbros and Tenedos, and parks his vehicle there. This cave is close to Samothrace, where he was to begin with. From there, by undisclosed means, he proceeds to the battlefield and appears to the Greeks, disguised as Calchas. Fortunately for the Greeks, who were losing, and showing due deference to Zielinski's law of Homeric narrative technique, the poet has stopped the clock, and Poseidon finds the battle exactly as he left it. Why does Poseidon travel to Troy by so roundabout a route, when the situation is so urgent? The usual reaction has been to excise either the whole passage, or at least the lines about Samothrace (so W Leaf), so that Poseidon does not cross and recross the Aegean. But when we combine 'literary' considerations with thematic analysis, the explanation soon appears. The description of Poseidon's journey stresses the importance of his arrival, gives us an impression of the god's awesome determination, builds suspense as we wonder about the fate of the Greeks, and vividly expresses Poseidon's three main characteristics, as earth-shaker, horse-god, and sea-god. To realize these aims Homer has combined two distinct themes. Theme I is that of a god intervening to counter the schemes of another. An unelaborated example of this is at Iliad 10.515-18: 58 otiS' d.AaocKontl}l' £[x'

d.pyvpDTo~oc )hr6M.wl', we (/)' :4.8T/''aL1)1' fLErov 'from on high, from out of the clouds [nephea]' in Odyssey xx 105, he is thundering from the sky, through the metonymy of the clouds. In this case, the ambivalence of clouds or sky is canceled only by the explicit statement, in the words of the singing woman at xx 114, that there is no nephos 'cloud' in the sky. In all other Homeric cases, the potential metonymic sense of nephos I nephea as 'sky' can remain in force. A particularly striking example of this metonymic sense of 'sky' is evident at Iliad XIII 523-524, where Zeus is pictured as sitting in grand isolation on the summit of Olympus, under a canopy of 'golden nephea' (a/J,: 6 yap