**Rethinking Orality II**

# **Transcodification: Arts, Languages and Media**

Edited by Simone Gozzano

**Volume 2**

# **Rethinking Orality II**

The Mechanisms of the Oral Communication System in the Case of the Archaic Epos

Edited by Andrea Ercolani and Laura Lulli

The series is funded with the grant of the Department of Excellence (2018-2022) awarded to the Department of Human Studies of the University of L'Aquila by the Italian Ministry of Education, University and Research (MIUR).

ISBN 978-3-11-075074-4 e-ISBN (PDF) 978-3-11-075196-3 e-ISBN (EPUB) 978-3-11-075205-2 ISSN 2702-7732 DOI https://doi.org/10.1515/9783110751963

This work is licensed under the Creative Commons Attribution 4.0 International License. For details go to https://creativecommons.org/licenses/by/4.0/.

Creative Commons license terms for re-use do not apply to any content (such as graphs, figures, photos, excerpts, etc.) not original to the Open Access publication and further permission may be required from the rights holder. The obligation to research and clear permission lies solely with the party re-using the material.

#### **Library of Congress Control Number: 2021951243**

#### **Bibliographic information published by the Deutsche Nationalbibliothek**

The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at http://dnb.dnb.de.

© 2022 with the authors, editing © 2022 Andrea Ercolani and Laura Lulli, published by Walter de Gruyter GmbH, Berlin/Boston The book is published open access at www.degruyter.com.

Printing and binding: CPI books GmbH, Leck

www.degruyter.com

### Table of Contents

Andrea Ercolani and Laura Lulli Introduction: Orality and Epic Poetry. Old Questions and New Perspectives VII

Elizabeth Minchin Mind-based Research Meets the Homeric Epics: Looking Again at Communicative Strategies in the Homeric Epics 1

Egbert J. Bakker Interformulaic Homer: Evidence from the "Wild" Papyri 19

Albio Cesare Cassio Two Chronological Extremes of the Homeric Language: πρόφρασσα and εἶπα 41

Chiara Bozzone Technologies of Orality: Formularity, Meter, and Kunstsprache in Homer 51

Giampiero Scafoglio A Cyclic Theme in the Odyssey: The Oresteia in Zeus' Speech (1, 28–43) 83

Jonathan S. Burgess Audiences of the Prophecy of Tiresias in Odyssey Book XI 103

Silvia Quadrelli Traces of Orality in the Histories: The Homeric 'Heritage' in Herodotean Battles and Speeches 117

Laura Lulli Some Reflections on Orality and Epic Poetry in Ancient Literary Criticism 145

Richard Hunter Homer and 'the Elegists': an Ancient Construction of Difference 165 VI Table of Contents

Sonja Zeman

Paradoxes of 'Orality': A Comparison between Homeric Oral Poetry and the Heroic and Courtly Epics in Middle High German 177

Franco Montanari Epos and Orality: Conclusive Remarks and Open Questions 207

Index Notable Things 213

Index Discussed Passages 217

### Andrea Ercolani and Laura Lulli Introduction: Orality and Epic Poetry. Old Questions and New Perspectives

The present volume is the arrival point of a long research path, which aimed to explore epic poetry as a case study to understand the communicative mechanisms of the ancient Greek culture, in the frame of a multifarious and multidisciplinary analysis of the codification, transcodification and trasmission of 'cultural messages'.

The previous step of the project¹ provided a series of reflections and indications about the modalities in which a cultural message can be produced, disseminated and transmitted, and this broad perspective can be usefully applied and/or verified in relation to one of the most complex cultural and literary phenomenon of the ancient Greek world. Indeed, thanks to its longue durée, epic poetry had the extraordinary ability to pass through different communicative systems, thus 'transcoding' its cultural messages and functions in a plurality of forms, even if exhibiting a well-defined profile in the frame of the 'rules' of a literary genre.²

Because of this, the attempt to understand the Homeric poems in an historical perspective cannot be limited to a traditional philological approach. Philology and textual criticism are a necessary condition to understand archaic epic, but, due to the special status of our extant evidence, there is a threshold which they cannot trespass without the help of other epistemological tools. So, a deeper understanding of the various 'dimensions' of the poems can be achieved, at first, if, starting from their 'textuality', we move to explain the interconnections with the original communicative conditions and contexts. This analysis will allow to highlight the gradual shifting of the epic poems from 'pure orality' to 'aurality', as well as their articulated relationships with the reading practices and the book culture, which, however, have never become a sheer alternative to spoken word.³ And drawing such picture requires the contribution of other research fields where texts are projected in a wider background.

See Ercolani/Lulli 2022.

For the concept of the "rules" or, better, the "laws" of the ancient Greek literary genres with their persistences and developments see the seminal work of Rossi 2020a.

For a broad overview of the different communicative systems in the Greek culture from the Archaic to the Imperial period, with an attention to the various and complex intersections between orality and writing, see at least Thomas 1992, Ong 2002, Sbardella 2006, 27–72.

In Homeric studies the necessity of explaining some textual peculiarities of the poems, going beyond the realm of philology, has been periodically risen since antiquity, as a sort of pendant or reaction to the Alexandrine perspective of the Omeron ex Omerou saphenizein. It would be an arduous task to offer not a comprehensive history, but even some mere glimpses of the multifaceted attemps to explore the epic "archipelago"⁴ under this specific point of view. From Murray 1934's ground-breaking definition of the Homeric poems as "book of culture", after an acute comparative analysis, to the merging of philology, ethnography and anthropology first elaborated by Milman Parry (Parry 1971)'s seminal essays (and systematically developed by his most brilliant student, Albert Lord, in Lord 1960), the 'new' Homeric philology looked for a dialogue with archaeology and history, opened to historical linguistics, semasiological analysis, history of religions and anthropological perspectives. And over the last decades a further improvement came from more refined ethnological comparisons,⁵ and finally from innovative attempts to explain the characteristics of the oral communication of poetry from the viewpoint of the cognitive sciences.⁶

But in front of so many works which offer innovative interpretative tools and methods, as well as groundbreaking visions of the Homeric poems, there is a question we would need to ask: have this variety of approaches and such instruments of analysis gained the right space in our school and university curricula? The limits of the itinera studiorum as well as the fences between the different disciplines are known. For this reason we try to offer to the readers at least some essential frames and sketchs of this complex picture: thanks to a dynamic and – hopefully – useful dialogue of various scholars, who try to propose and to renew the traditional hermeneutical tools, expanding their perspectives.

The result of such dialogue are the essays collected in this volume, which were originally discussed in an international conference held in the Department of Human Studies of the University of L'Aquila, on November 13–15, 2019. All of them have a common ground: the new paths they try to explore start from philology and textual analysis, and go beyond them with the use of new interpretative strategies which never betray extant texts; and all of them take into account the underlying communication processes that make the Homeric poems one of the most complex literary phenomena of ancient Greek culture, and that have

For this evocative definition see Rossi 2020b, 124.

See, in particular, the work of Burkert 1996.

See Minchin 2001 and the overview in Meineck/Short/Devereaux 2019, with previous bibliography. See also Giordano 2022.

contributed to the neverending success of the stories of the heroes of the Trojan war.

In such perspective, the choice of reconsidering the mechanisms of the oral communication system in relationship to the Archaic Greek epic poetry was not the consequence of the desire to add a further piece to an already multifarious debate that was preminent in the European cultural scene at least since the 17th century. Rather, the need to 'rethink the orality', considering the most complex evidence of orality from ancient literatures, has been all the more pressing and stimultating in the contemporary context. Indeed, the period we are living is strongly characterized by a powerful 'return to orality', i.e. by a massive use of the oral communication channel, especially in the new media, where it provides new forms of expressions in a dialectic relationship not only with the images, but also with writing and reading.

Therefore, the continuous resounding of oral communications that populate our daily life is a strong motivation to try to look back at the ancient roots of this medium, observing its effects and impact precisely on the epic and on the words of its most ancient heroes, which continue to be echoed in a post-modern world in constant search for new forms of expression.

### References


Lord, A. (1960), The Singer of Tales, New York.


Ong, W. J. (2002), Orality and Literacy, London.


Rossi, L. E. (2020b), "L'autore e il controllo del testo nel mondo antico", in: G. Colesanti/R. Nicolai (eds.), Luigi Enrico Rossi. Kηληθμῷ δ᾽ ἔσχοντο. Scritti editi e inediti. Vol. 3: Critica letteraria e storia degli studi, Berlin – Boston, 122–140 (= Rossi, L. E. (2000), "L'autore e il controllo del testo nel mondo antico", in: Seminari Romani di Cultura Greca, 3, 165–181).

Sbardella, L. (2006), Oralità: da Omero ai mass media, Rome.

Thomas, R. (1992), Literacy and Orality in Ancient Greece, Cambridge – New York.

## Elizabeth Minchin Mind-based Research Meets the Homeric Epics: Looking Again at Communicative Strategies in the Homeric Epics

Abstract: The publication of Milman Parry's and, later, Albert Lord's important work gave rise in the latter years of the 20th century to a number of questions about oral poetic composition: first, how did an oral poet prepare himself for composition in performance? second, what understanding did he, as a professional storyteller, have of memory and its functions? and, third, how did the poet exploit those functions of memory to assist him as he sang? These questions were at the time considered to be 'askable', because of a revived research interest in cognition, in how the mind stores information and how it retrieves it. Research into cognition in its many aspects has been a valuable resource for Homerists, not only for the insights it has provided into the processes of oral traditional composition. It has allowed us insight, too, into the poet's own understanding of the operations of memory and into the way he has built that understanding into his communication of both the motivations and the responses of his characters: the very impulses and reactions, that is, that drive the action of the poems. In this paper I offer an overview of how mind-based studies, when applied to the Homeric epics, can serve as tools for understanding. The two case studies that I include will illuminate, from different perspectives, the poet's compositional methods and communicative strategies, revealing something of the essential economy that underpins oral storytelling.

Keywords: Oral composition; cognitive scripts; shield of Achilles; land scape; s p ati a l memory; enactivist theory; motor resonance; anger of Achilles; emotion and memory.

## 1 After Parry and Lord: Cognitive Research and Oral Traditional Poetry

Milman Parry and Albert Lord's work resulted, over time, in significant revisions of Homeric scholarship.¹ Parry, following up on earlier European work on diction

Parry 1987; Lord 2000.

in the Homeric poems, had recognized their traditional and, indeed, their oral nature. Subsequently, with Lord, he made a close study of the then living tradition of South Slavic epic song. Their observations led them to argue that the Homeric epics too had not been composed in writing, nor were they memorized texts, but that they were products of a long tradition of oral composition.

Lord's great contribution to the study of the mechanisms of oral transmission was to bring the singer, as well as the song, into focus: he asked how singers acquired a repertoire of songs; how they built up a language rich in formulas; and how they acquired their so-called themes, or 'type-scenes' as we might describe them. From his observations of the South Slavic poetic tradition Lord had proposed that a long apprenticeship based on listening, learning, and the exercise of memory was the answer. But questions of a similar kind, which touched on the role of memory in the production of oral storytelling, were being explored also in experimental work on cognition.² And, in the case of Lord's 'themes', the answers suggested in that realm were somewhat different.

My own contribution to this enquiry – I make the point here that I am by no means alone in having adopted mind-based approaches – began with my identification of the Homeric type-scene with the cognitive script: that is, with knowledge that each of us has stored away in episodic memory, in many cases from early childhood.³ Stored knowledge of this kind comprises information about familiar real-world situations and the way they unfold. This insight into the relationship between type-scene and script has allowed us to develop a more plausible and more realistic view of a poet who composes not in writing but in performance. Sequences of actions that underpin, for example, the preparation of a barbecue meal, or of dressing, whether in armour or not, had been part of the poet's knowledge store, perhaps since his childhood.⁴ There was little or no need for the apprentice singer to undertake the kind of learning that Lord proposed. The appropriate pre-verbal script, drawn from episodic memory, would supply the poet with the required sequence of actions: his compositional task was to locate the words and phrases of his special poetic language in which to communicate them.⁵ Since the oral poet had no access to a written text to

For contributions from cognitive science, see, e. g., Schank/Abelson 1977; Dyer 1986.

Minchin 1992.

For the preparation of a feast, cf. the regular patterns underpinning Il. 1, 458–468; 7, 313–320; 9, 206–221; 24, 621–627. For the donning of armour, cf., for example, Il. 3, 330 –338; 11, 17–45; 16, 131–139; 19, 369–391. The donning of armour is a subgroup of dressing: Il. 10, 21–24, 29–31, 131–135.

Minchin 2001, 40–41.

prompt him, we can easily imagine that he standardized these sequences and rehearsed them in advance, in order to speed the connection between scripted action and the special language that expressed it at the very moment of performance.⁶

These insights have drawn attention in turn to the fruitful relationship that can be observed in the Homeric epics between other functions of memory and oral traditional practice. In addition to the advantages of episodic memory, as I have described it, auditory memory retains the rhythmic structure of traditional epic; it serves as a valuable constraining cue as singers search for the next phrase. Visual memory plays an important role in prompting the flow of narrative. Spatial memory assisted the poet in the disposition of his characters, in the tracking of the two armies as they move about on the plain, and in the composition and performance of that challenging exercise, the Catalogue of Ships.⁷

An evaluation of oral composition from a cognitive perspective encourages us to revise our ideas about the poet's communicative practice. It reveals the motives for his characteristically leisurely narration, especially of type-scenes, and his preference, for example, for extended similes. It explains his readiness to resort to repetition of words, of phrases, and of event sequences, speeches, and messages. All oral traditions have over time developed strategies and organizational 'rules' to support the poet in performance and to preserve the integrity of a song.⁸ No poet wants to break down, or to lose track, mid-performance.

### 2 Discourse Analysis and Related Mind-Based Studies

When cognitive approaches were introduced into Homeric studies, the rationale for a number of the poet's compositional habits became clear. What is essential for an oral poet, and indeed for the tradition in which he works, is a set of strategies that economize effort, with a view to making a difficult task possible. Just as the poet drew on everyday experience for his scripts, he drew on everyday conversational practice too, standardizing his presentation of a number of fre-

There is no doubt that poets practised their songs beforehand, just as speakers might 'run over' in their mind a talk they are about to give. Many of the poet's lists and catalogues, for example, would have required considerable practice.

For a useful introduction to research relating to these functions of memory, see experimental psychologist Rubin 1995; for application of this work to the Homeric epics, see Minchin 2001. Wallace/Rubin 1988; Rubin/Wallace/Houston 1993; Rubin 1995, 10; Minchin 2001, 70 –72.

quently-used speech forms such as rebukes, protests, and the refusal of invitations.⁹ When his characters tell stories to each other we observe that they follow an everyday storytelling format, familiar to us in the western world today. This format generates the prompts and repetitions that often create the appearance of what has been termed ring-composition.¹⁰ Or, again, as he reports question and answer sequences in the course of conversation, the poet adopts the practice of hysteron-prōteron, as this habit has been called by Homeric scholars. He deals with the second question immediately before him and commits to memory only the first – and thus he husbands his resources.¹¹ The solutions this poetic tradition has developed are economical for the poet and, equally, meaningful for his listeners. For audience members too face a challenge: unlike readers, who can adapt their reading pace to the text before them, audiences of oral performances have no such opportunity. The strategies that the performers have adopted to assist them in maintaining fluency – that is, the tradition's standardized, predictable forms – also assist the listening audience as they follow the tale. And so, as new explanations have been found for the characteristic phenomena of oral traditional epic, scholars have seen with clarity how poets in this oral epic tradition and their audiences have taken advantage of the resources of memory, and its capabilities, and worked with them.

### 3 Memory and the Individual

David Konstan has described cognitive science as "a mansion with many rooms".¹² I turn now to yet another 'room' within cognitive psychology, one in which we can explore how the workings of the mind affect or guide human behaviour.¹³ Drawing on this resource we observe how studies of autobiographical memory illuminate the behaviour of Nestor; how studies of the unwelcome persistence of memory allow us to understand the reactions of Achilles, in the Iliad, as he struggles to deal with the pain of Patroclus's death, or, turning now to the Odyssey, the pain that Ajax suffered on being denied the armour of Achilles after his death.¹⁴ An understanding of Theory of Mind, that ability that we all share to

Reber 1989; Coulthard 1992; Sacks 1992; Minchin 2007.

Minchin 2001, chapter 6.

Minchin 2007, chapter 4.

Konstan 2019, xv.

Neisser/Fivush 1994; Schacter 2001.

Minchin 2005; Minchin 2006.

varying degrees to attribute mental states – emotions, beliefs, knowledge – to others, allows us access to the mental resources that the poet draws on even as it illuminates the mind-based impulses and responses of his characters – impulses and responses that drive the action of the poems.¹⁵ And we observe how in that epic world, as in our own, the heroes attach stories to landmarks in the world they know and, in more intimate contexts, to precious objects in their possession. Our knowledge of the functioning of spatial and visual memory, and of their natural propensity to create vital links between location and other information, helps us understand the great wealth of story that emerges in both epics, prompted almost always by a visual cue.¹⁶

Mind-based research offers us the vocabulary that enables us to describe and explain specific features of the text in terms of their compositional and communicative significance; but it also offers interpretive capacity, allowing us to identify and account for the intentions and the motivations of the characters within the storyworld.

### 4 Where Can We Go From Here?

But have we – 'we' being scholars of oral traditional poetry – reached the limits of the insights that can be afforded to us by cognitive approaches? Can we take them further?

I suggest that we can. Now that we have the confidence to engage with psychological research and to work with it, I suggest that there is much more to be said in the areas that I set out above, as I shall demonstrate with two small case studies: first, in connection with the processes of composition, I focus on the Homeric poet's approach to landscape description; and, second, in dealing with an apparent hiatus in his account of motivation, I consider how mind-based research allows us to probe the poet's analysis of behaviour in the absence of the language that might describe it. I turn, in conclusion, to a discussion of the economy that underpins oral storytelling.

Herman 2003; Smith/Kosslyn 2007; Carlson/Koenig/Harms 2013; Zunshine 2015 (especially on Theory of Mind, supplying information to fill gaps in the narrative, and cognitive literary studies).

On imagery as a prompt to memory, see Rubin 1995, chapter 3.

### 5 Case Study 1: Communicating Landscape

The scenes on the great shield Hephaestus creates for Achilles, described in book XVIII of the Iliad, offer a remarkably sustained – and varied – series of landscapes and cityscapes. Many readers are charmed and delighted by these small scenes; and yet, as we quickly observe, the poet's account includes little precise spatial or descriptive language. So my question is: what mind-based resources does the poet draw on to evoke this range of landscapes so that his audiences also 'see' them and enjoy them?

I begin with the important insight framed by George Miller, one of the founders of cognitive psychology, who observes that descriptions of landscape scenery should never be minutely detailed.¹⁷ He explains that detailed descriptions are cognitively unrealistic, certainly for an oral storyteller and a listening audience. It is clear that the poet of the Iliad understands this, perhaps intuitively: he carefully avoids 'thick' description. Nevertheless, he succeeds again and again in his account of the landscapes of the Shield in conveying a sense of place. To account for his practice I have drawn on work in cognitive psychology by Barbara Tversky and Holly Taylor, who developed the notion of a 'spatial mental model' that can be derived from the topographical indicators that a speaker or writer may introduce.¹⁸ I consider too the non-topographical but elaborative information that may be included, such as epithets and discursive commentary. I use George Miller's further observation that, as we bring to mind a scene that is being described, we turn to our own memory store for generic information that will enhance our experience.¹⁹ Finally, I note that what makes any scene vivid, and arresting, is human presence.²⁰ So, now including one further element in this cognitive project, I turn to enactivist theories of cognition.

Drawing on experimental work in perception and in neuropsychology, enactivist theory offers valuable insights into the way in which storytellers create sequences of narrative-action, and into the way we audience members reconstruct these sequences in our minds as we follow them.²¹ And I argue that it is indeed

Miller 1993, 360.

Taylor/Tversky 1992.

Miller 1993, 358–360. Barbara Tversky's notion that landscape description often takes the form of a "cognitive collage" is relevant to this exercise: see Tversky 1993.

I have discussed the pleasure of audiences in the landscapes of the Shield of Achilles elsewhere: see Minchin 2021. My focus in the present paper, asIlook once more at the Shield of Achilles, is on the striking cognitive economy that underpins the poet's descriptive practice. See, e. g., Speer/Reynolds/Swallow/Zacks 2009; Kuzmičova 2012; Troscianko 2013. Whereas Grethlein/Huitink 2017, revisiting focalization, have discussed enactivist theories in relation to

because of the incorporation of human activity into each small scene that the audience is able to trace out and to visualize the landscapes in which those actions take place.

Recent work has shown that when we envisage a scene that depicts human action we do not achieve this by simply visualizing that action as passive thirdperson observers.²² Behavioural studies and experiments using neuroimaging have demonstrated that, as we process passages of this kind, we experience in our cerebral cortex sensory traces – called "motor resonance" – related both to the movement in question and to the object it is directed at.²³ Such sensorimotor responses, taken together, have been described as "enactment".²⁴ Furthermore, and this is critical for this project, when we encounter indicators of movement through space, regions involved in the navigation of spatial environments are also activated. Thus, as we process accounts of human action, we register a so-called "multimodal image" that represents the environment in which that movement occurs.²⁵ The spatial imagery that has been prompted may be fragmentary, instantaneous, and extremely short-lived, but it provides an intense but momentary sense of having physically entered a tangible environment – this is that sense of 'presence', of 'being there', that we all can recognize from our own experiences of listening and reading.²⁶ It is to those momentary experiences of enactment, along with the associated experience of 'presence', that we may attribute much of a narrative's intensity, its impact, and its memorability.²⁷

As he creates his landscapes the poet draws on all these elements. What we observe as we study the scenes that Hephaestus creates on the Shield is that, with the exception of the first (a landscape-seascape-skyscape view of his own world), the poet makes no effort to give us a comprehensive description. But in that exceptional first scene (Il. 18, 483–489) the poet gives us a vivid sense of the night sky as viewed from the earth below:

Homer's narrative passages, I propose that we can fruitfully apply these theories also to a study of how we 'read' the landscapes on the Shield.

Kuzmičova 2012, 24, 43–44.

Speer/Reynolds/Swallow/Zacks 2009, 995–998; on motor resonance, see Kuzmičova 2012, 26, 29.

For early use of this term, see Noë 2004; for a useful introduction to Noë's work see O'Regan/ Noë 2001, 939–1031.

Kuzmičova 2012, 35–36.

Kuzmičova 2012, 25, 26; on intensity, see Troscianko 2013, 189.

Kuzmičova 2012, 28–29, on "presence". When bodily actions are simulated fromafirst-person perspective, the reader (or listener) experiences the phenomenon of motor simulation: this becomes the vehicle for "presence and immersion".

#### 8 Elizabeth Minchin

ἐν μὲν γαῖαν ἔτευξ', ἐν δ' οὐρανόν, ἐν δὲ θάλασσαν, ἠέλιόν τ'ἀκάμαντα σελήνην τε πλήθουσαν, ἐν δὲ τὰ τείρεα πάντα, τά τ' οὐρανὸς ἐστεφάνωνται, 485 Πληϊάδας θ' Ὑάδας τε τό τε σθένος Ὠρίωνος Ἄρκτόν θ', ἣν καὶ Ἄμαξαν ἐπίκλησιν καλέουσιν, ἥ τ' αὐτοῦ στρέφεται καί τ' Ὠρίωνα δοκεύει, οἴη δ' ἄμμορός ἐστι λοετρῶν Ὠκεανοῖο.

He made the earth upon it, and the sky, and the sea's water, and the tireless sun, and the moon waxing into her fullness, and on it the constellations that festoon the heavens, the Pleiades and the Hyades and the strength of Orion and the Bear, whom men give also the name of the Wagon, who turns about in a fixed place and looks at Orion and she alone is never plunged in the wash of the Ocean.²⁸

Elsewhere the poet offers us little precise spatial terminology to help us envisage the landscapes and cityscapes that he includes: we find few spatial or directional pointers and little topographical detail. His landscapes are each presented sketchily, impressionistically. And yet, to take the next lines, the poet's first city-scene (Il. 18, 490 –496), as the focus for discussion, his representation of movement and action has an immediate effect on us audience members; it stirs motor resonance:


On it he wrought in all their beauty two cities of mortal men. And there were marriages in one, and festivals. They were leading the brides along the city from their maiden chambers under the flaring of torches, and the loud bride song was arising. The young men followed the circles of the dance, and among them the flutes and lyres kept up their clamour as in the meantime the women standing each at the door of her court admired them.

At 492–493 we leave home along with the brides and we move along the streets; we see the flaming torches and hear the wedding song: there is a suggestion of

I use Richmond Lattimore's peerless translation of the Shield (Lattimore 1951) and, with occasional variation, of the passages cited below.

excitement in the air. We dance (494) with the young men; and we hear the sound of flute and lyre. And then, changing places, we become the watchers, the women who watch from their doorways as the procession moves past (495–496).

In following the movement of the brides and their escorts, we make our way through a notional three-dimensional space, a generic space, in which all this action occurs. We don't consciously build up a picture in our heads, or even a spatial mental model, as Taylor and Tversky suggest we might; rather, through the poet's language cues, we gradually acquaint ourselves with a three-dimensional environment, which is built up from the scraps of spatial terminology and the fragments of narrative that the poet offers. These allow us to register a multimodal image, and, in the absence of further information from the poet, we supply relevant generic data from our own memory store: my own memory store, for example, provides some rough paving in the street and stone facades for the houses. Taken together, these elements contribute to the experience of 'presence' of 'being there', as we join the women standing in their doorways (495–496), enjoying the scene.

What I want to draw attention to here is the essential economy of this communicative strategy. The poet is sparing with spatial information, which would otherwise require cognitive effort to identify and to retain, both on his part and on that of his audience. The majority of the information he offers comes in vivid narrative form – always an easier option for him.²⁹ He judiciously offers also a certain amount of non-topographical information; and he leaves it to us, his listeners, to supply further detail as we require and to organize this assemblage of unrelated information into a more or less coherent cognitive structure.³⁰

It must be the case, therefore, that any one audience member's representation of the setting, and of the scene itself, will not be exactly the same as another's; nor will it be what the poet himself has envisaged. In this first cityscene on the Shield the street and the houses that line it are shadowy forms; but it is critical that they should be adequate for the purpose. What we listeners or readers bring to mind as we process this assemblage of information will be just enough to allow us to make sense of the action represented and to take pleasure in a scene of joyful celebration in a small community. My emphasis here is on the poet's strategic economy: he stretches his capacity only as far

See Lukacs 1962, 86–89; Linde 1981, 104–110; Minchin 2001, 100–131, especially 131.

This is not an entirely straightforward exercise, either for the poet or his audience. It could go wrong, as it does in the description of the second city, at 523–534, where, through a lack of necessary information, the poet leaves us in doubt as to whose cattle are being seized – and by whom.

as is necessary in order to encourage his audiences in their turn to bring to mind contextually appropriate – and pleasing – landscapes.

### 6 Case Study 2: Shaping the Story

My first example has illustrated successful communication. Like most scholars I am eager to show how communication in an oral context works. But it is instructive too to examine more problematic moments of communication. For these prompt us to locate and understand their cause. In this second example I bring our understanding of cognitive psychology and discourse analysis to bear on what is either a lapse in good storytelling practice or a test of an audience's willingness to collaborate.

Every storyteller's ambition is to tell an engaging story – one that holds the audience's interest – without labouring over excessive detail. And yet details there must be – details of cause, of motivation, and of effect – if the story is to be meaningful. The storyteller's solution is to share this burden with his or her listeners, obliging them to invest cognitive effort in the construction of the tale, supplying at various points of the narrative information about motivation and consequence.³¹ The poet of the Iliad on the whole manages this well: for example, when Hector in the course of his revealing exchange with Andromache in Il. 6 reaches out to take his son from his nurse's arms, Astyanax wails in alarm. His parents break into laughter, together (Il. 6, 466–471):

ὣς εἰπὼν οὗ παιδὸς ὀρέξατο φαίδιμος Ἕκτωρ ἂψ δ' ὃ πάϊς πρὸς κόλπον ἐϋζώνοιο τιθήνης ἐκλίνθη ἰάχων πατρὸς φίλου ὄψιν ἀτυχθεὶς ταρβήσας χαλκόν τε ἰδὲ λόφον ἱππιοχαίτην, δεινὸν ἀπ' ἀκροτάτης κόρυθος νεύοντα νοήσας. 470 ἐκ δ' ἐγέλασσε πατήρ τε φίλος καὶ πότνια μήτηρ […]

so speaking glorious Hektor held out his arms to his baby, who shrank back to his fair-girdled nurse's bosom screaming, and frightened at the aspect of his own father, terrified as he saw the bronze and the crest with its horse-hair, nodding dreadfully, as he thought, from the peak of the helmet. Then his beloved father laughed out, and his honoured mother […]

Carlson/Koenig/Harms 2013.

Although Homer describes the scene in sympathetic detail, he does not explain what underlies Hector and Andromache's response. He leaves it to us to supply the reason. The young parents' laughter is, in part, affectionate – for their son's fears are ill-founded. And Astyanax's problem, compared with the troubles that face them and their city, is an easy one to fix. In part too their laughter is a reaction to the tensions of the preceding minutes, now interrupted by their son and his needs.³² We are easily able to recognize what motivates their response, through reference to Theory of Mind. And, as we complete the essential chain of cause and effect that underpins this small scene, our success brings us pleasure and confirms our engagement with the poet and the story he tells.³³

There is, however, one notable occasion in the course of the Iliad-story when the poet appears reluctant to give his audience sufficient direction. He certainly invites, or indeed expects, his audience to supply motivation; but, although Theory of Mind is available to us, it is clear that we / audience members struggle to take advantage of it.

My subject in this exercise is the hero of the Iliad, Achilles, and his response in book I to his quarrel with Agamemnon. What Achilles takes away from this encounter has been a vivid recollection of the events as they occurred, as he spells them out to his sympathetic mother (365–392) and to others subsequently (9, 367–426; 16, 52–59). He recalls the circumstances in which Chryseis, Agamemnon's concubine, had been captured and how she was selected by the Achaeans for the king. He recalls the supplication of Chryses for the return of his daughter and Agamemnon's blunt refusal, despite the protests of the Achaean army. He recalls the subsequent coming of the plague and the seer Calchas's identification of the source. And he describes to Thetis, at 386–392, Agamemnon's angry response to his own efforts to reason with him and the unhappy outcome:³⁴

αὐτίκ᾽ ἐγὼ πρῶτος κελόμην θεὸν ἱλάσκεσθαι· ᾿Aτρεΐωνα δ᾽ ἔπειτα χόλος λάβεν, αἶψα δ᾽ ἀναστὰς ἠπείλησεν μῦθον ὃ δὴ τετελεσμένος ἐστί τὴν μὲν γὰρ σὺν νηῒ θοῇ ἑλίκωπες ᾿Aχαιοὶ ἐς Χρύσην πέμπουσιν, ἄγουσι δὲ δῶρα ἄνακτι· 390 τὴν δὲ νέον κλισίηθεν ἔβαν κήρυκες ἄγοντες κούρην Βρισῆος τήν μοι δόσαν υἷες ᾿Aχαιῶν.

Minchin 2008, 21–26.

The poet's practice is, I suggest, what Matthew Arnold describes as Homer's eminent rapid-

ity: Arnold 1896, 10.

And cf. also 1, 239–244, 295–296, 338–344.

It was I first of all urged then the god's appeasement; and the anger took hold of Atreus' son, and in speed standing he uttered his threat against me, and now it is a thing accomplished. For the girl the glancing-eyed Achaians are taking to Chryse in a fast ship, also carrying presents to lord (Apollo). But even now the heralds went away from my shelter leading Briseus' daughter, whom the sons of the Achaians gave me.

This memory fuels his continuing resentment over the coming days.

Withdrawing himself and his Myrmidons from the fighting, the hero is prepared to watch his Achaean comrades fall in battle. He asks his mother to call in a favour from Zeus so that he might intervene on the battlefield on his behalf (407–412):

τῶν νῦν μιν μνήσασα παρέζεο καὶ λαβὲ γούνων αἴ κέν πως ἐθέλῃσιν ἐπὶ Τρώεσσιν ἀρῆξαι, τοὺς δὲ κατὰ πρύμνας τε καὶ ἀμφ᾽ ἅλα ἔλσαι ᾿Aχαιοὺς κτεινομένους, ἵνα πάντες ἐπαύρωνται βασιλῆος, 410 γνῷ δὲ καὶ ᾿Aτρεΐδης εὐρὺ κρείων ᾿Aγαμέμνων ἣν ἄτην ὅ τ᾽ ἄριστον ᾿Aχαιῶν οὐδὲν ἔτισεν.

Sit beside him and take his knees and remind him of these things now, if perhaps he might be willing to help the Trojans, and pin the Achaians back against the ships and the water, dying, so that thus they may all have profit of their own king, that Atreus' son wide-ruling Agamemnon may recognize his madness, that he did no honour to the best of the Achaians.

What the external audience takes away from those early scenes is confusion: how is it possible that the individual who is introduced as the hero of this tale, and who had been so concerned and measured in his response to the plague that had beset the Achaean army (Il. 1, 52–67) – how could he abandon so readily not only the Achaean enterprise but also, so dramatically, his comrades? As Donald Lateiner observes, Achilles' early decision and his subsequent refusal to back away from it when the members of the Embassy (including his old tutor Phoenix) make a further appeal, have confounded critics across time.³⁵

The poet, I suggest, dodges the question. But I am interested, nevertheless, in what he shows us. I am interested in how the emotional impact of the quarrel, both during that great war of words with Agamemnon and subsequently in the Embassy, when memories of Agamemnon's high-handedness resurface, has af-

Lateiner 2004, 15, 22 n. 31, 27. On Phoenix and Achilles, see Cairns 2001, especially 211–212, in an important contribution to discussions of the regulation of quarrels in the world of epic.

fected Achilles' capacity to make important decisions. This is made clear in the hero's continuing inability both to set aside what was said in the course of his encounter with the king and to think beyond his own emotional response (Il. 9, 367–377):

γέρας δέ μοι, ὄς περ ἔδωκεν, αὖτις ἐφυβρίζων ἕλετο κρείων ᾿Aγαμέμνων ᾿Aτρεΐδης· τῶ πάντ' ἀγορευέμεν ὡς ἐπιτέλλω ἀμφαδόν, ὄφρα καὶ ἄλλοι ἐπισκύζωνται ᾿Aχαιοὶ 370 εἴ τινά που Δαναῶν ἔτι ἔλπεται ἐξαπατήσειν αἰὲν ἀναιδείην ἐπιειμένος· οὐδ' ἂν ἔμοιγε τετλαίη κύνεός περ ἐὼν εἰς ὦπα ἰδέσθαι· οὐδέ τί οἱ βουλὰς συμφράσσομαι, οὐδὲ μὲν ἔργον· ἐκ γὰρ δή μ' ἀπάτησε καὶ ἤλιτεν· οὐδ' ἂν ἔτ' αὖτις 375 ἐξαπάφοιτ' ἐπέεσσιν· ἅλις δέ οἱ· ἀλλὰ ἕκηλος ἐρρέτω· ἐκ γάρ εὺ φρένας εἵλετο μητίετα Ζεύς.

but my prize: he who gave it,

powerful Agamemnon, son of Atreus, has taken it back again outrageously. Go back and proclaim to him all that I tell you, openly, so other Achaians may turn against him in anger if he hopes yet one more time to swindle some other Danaan, wrapped as he is forever in shamelessness; yet he would not, bold as a dog though he be, dare to look in my face any longer. I will join with him in no counsel, and in no action. He cheated me and he did me hurt. Let him not beguile me with words again. This is enough for him. Let him of his own will be damned, since Zeus of the counsels has taken his wits away from him.

I suggest, however, that the uncertainty the poet has perhaps inadvertently created may respond to recent work on the impact of the emotions on memory.³⁶ Psychology tells us that when emotions are aroused, they are observed to have a positive impact on memory: that is, emotional content can enhance the vividness and the longevity of one's memory of an event.³⁷ This is especially so in connection with distressing events and, in particular, when people perceive changes in the status of their goals – that is, when fulfilment of those goals is favoured or when it is threatened.³⁸

Alexander/O'Hara 2009, 223.

For discussion with valuable references see Reisberg/Heuer 2004, 4–5. As Reisberg and Heuer argue, when our emotions have been engaged we tend, immediately afterwards, to mull over an event: because of its emotional nature, we revisit the event, both in our own thoughts and in conversation with others. Memory rehearsal of this kind promotes retention. Porter/Birt 2001; Schacter 2001, 161–183; Laney/Campbell/Heuer/Reisberg 2004, 1149.

But are our emotional memories, that is, our memories for events that have occurred in circumstances of high emotion, accurate? Research tells us that such memories may be vivid, but that they are also selective. In general, it appears, emotion promotes memory for an event's gist, its centre, but it does so at the expense of recollection for peripheral details: hence the terms "narrowing of attention" or "memory narrowing", in response to what is described as an "attention magnet".³⁹ The important corollary of this intense focus is that information that is peripheral (that is, not related to that emotion and the critical goals) is eclipsed, neglected.

This oral epic tradition that we associate with Homer's name was not aware of the psychological studies of memory narrowing that may go some way to explaining the hero's behaviour; but this tradition, or at least the poet of this Iliadstory, was sensitive to the psychology of the individual. Achilles' extreme behaviour, in the course of and as a consequence of the quarrel, is, certainly, not inconsistent with his exceptional origins. His anger is the anger of a hero whose mother is divine. But I suggest that it is also, in human terms, psychologically true.

The poet, however, has left it to us, the external audience, to supply the reason for what appears to be a callous disregard for the lives of others. I propose that we are able to find what we seek in these studies of cognition, in the links between emotion and memory, which offer us both a rationale for what we observe and the terminology in which to describe it. As I have noted, when individuals' high-order goals are threatened, they experience intense emotional responses related to those goals; their attention is tightly focussed on their preservation. Emotion enhances the subject's memory for what is central and puts out of his mind what is not. And thus it was that Achilles' response to Agamemnon's challenge to his honour resulted in a tremendous anger (Il. 1, 2–4)

ἣ μυρί' ᾿Aχαιοῖς ἄλγε' ἔθηκε, πολλὰς δ' ἰφθίμους ψυχὰς Ἄϊδι προΐαψεν ἡρώων, αὐτοὺς δὲ ἑλώρια τεῦχε κύνεσσιν οἰωνοῖσί τε πᾶσι […]

which put pains thousandfold upon the Achaians, hurled in their multitudes to the house of Hades strong souls of heroes, but gave their bodies to be the delicate feasting of dogs, of all birds […]

And all this time the hero himself felt no compunction.

Levine/Edelstein 2009, 834–836, 845–846.

The hero, we conclude, had set out to achieve payback by bringing Agamemnon to the realization of how essential he, Achilles, had been to the Achaean enterprise at Troy and by allowing the Achaean force to perceive how deficient Agamemnon had been as their leader ("so that thus they may all have profit of their own king", 410). At this point, therefore, Achilles viewed his comrades in the Achaean force instrumentally, as means to an end; he remained oblivious both at the time of the quarrel and its immediate aftermath and at his meeting with the Embassy, to the warmth of their friendship.⁴⁰

### 7 Conclusions

These two case studies reflect in different ways on the mechanisms of oral communication. My first example examines an oral poet's evocation of cityscapes and landscapes for a listening audience. Communicating just enough spatial information to suggest a particular environment, he supplements this with a narrative of movement within that environment; and he relies on his listeners to supply from their own memory store the details that enrich it. The economy of the process is striking – striking too in comparison with the unhurried descriptions we can find on the printed page.⁴¹ My second example considers the presentation of story and its relationship to the causal chain that underpins it. It too reflects on the economy of the storytelling process, but from a different perspective. Successful storytellers do not spell out every link in the chain of cause and effect. Their stories are, in a sense, incomplete; audiences must be prepared to supply motivational cues or causal links. These gaps are critical as a spur to engagement.When a storyteller presents a story in writing to a reading audience he or she can leave significant informational gaps that oblige readers to pause for thought, as they turn to Theory of Mind to supply the causal link that they need. And so, in today's world, as we work back and forward through the text of the Iliad, now on the page, and as we draw on observations from fields that throw light on human behaviour, we can make sense of Achilles' response to his treatment by Agamemnon, supplying at leisure the information that the poet omits. But when a story is told to a listening audience whose experience of the story is simultaneous with the telling, as were the first performances of

The hero's response to Ajax' words at 9, 624–642, especially at 630–632. Minchin, forthcoming develops the argument from Achille's point of view. The discussion focuses on the cognitive processes of the listener.

See, for example, Thomas Hardy's use of elaborated description in creating a setting for his novel Tess of the D'Urbervilles (Hardy 1983, 10–11).

the story of Achilles and the great quarrel, these gaps should be, or should have been, fillable in the moment. Otherwise the storyteller risks his audience's frustration, puzzlement, or even disengagement.

Both examples address the economy of storytelling.Whether we consider the descriptive or the narrative mode, we observe that the poet of the Iliad demands the contribution of his listeners to the project. This poet does not work in isolation; that would be uneconomical. This traditional storytelling genre is in a sense always collaborative. Even as we pay attention to the mind of the poet we must consider too the minds of his listeners.

### References

Alexander, K./O'Hara, K. (2009), "An Integrated Model of Emotional Memory: Dynamic Transactions in Development", in: J. Quas/R. Fivush (eds.), Emotion in Memory and Development: Biological, Cognitive, and Social Considerations, New York, 221–255.

Arnold, M. (1896), On Translating Homer, London.


Lattimore, R. (1951), The Iliad of Homer, Chicago.


Zunshine, L. (ed.) (2015), The Oxford Handbook of Cognitive Literary Studies, Oxford.

### Egbert J. Bakker Interformulaic Homer: Evidence from the "Wild" Papyri

Abstract: Oral-formulaic theory saw in the Homeric formulas a system enabling singers to compose the epic tale while performing before an audience. Everything other than this situation was dismissed as "literate", hence not relevant for the study of Homer as oral poetry. Today we are able to appreciate that "literacy" is not a monolithic concept, but culturally determined and embedded. The existence of texts changes, but does not end, the nature of Homer as oral poetry. In this paper I explore a number of ways in which the formulaic system is cognitively important in the presence of written texts of the epic. First, formula as formulas are a powerful aid in memorization and recall, which makes them indispensable not only for oral composition, but also for the performance of a written Homer. Second, the Homeric poems begin also to constituteameta-language when phrases are used with full awareness of the context(s) in which they occurred previously. The conscious use of formulas as reference to other instances of the formula can be called "interformularity".

Keywords: Intraformulaic; oral-formulaic theory; Homeric text; mnemotechnic.

This chapter contributes to two separate strands of Homeric scholarship, formulaic analysis and textual criticism. These two subfields are not naturally or commonly viewed together, but I will argue that they are interrelated in ways that increase our understanding of both the Homeric formula and the Homeric text. Textual criticism and formulaic analysis have not been natural bedfellows in the history of Homeric scholarship. The formula, in the classic conception authoritatively developed in the work of Milman Parry and Albert Lord, is an adaptive response to the unavailability of writing in the composition of the poems: the idea of formula and formulaic composition presupposes the absence of writing. And text is what comes 'after' the formula, the stage of the transmission of the Homeric transmission in which writing has eliminated the need for the oral formula, since the tale, now written, can be read, eliminating the need for the formula and the support it provides.¹

The idea of textuality as "post-oral" – the ideological opposite of the idea of formularity as "pre-textual" – has been outdated in its classic form for many

On the mutual incompatibility of orality and literacy, see Lord 1960, 129.

OpenAccess. © 2022 Egbert J. Bakker, published by De Gruyter. This work is licensed under the Creative Commons Attribution 4.0 International License. https://doi.org/10.1515/9783110751963-003

years, although it is not entirely dead.² More conciliatory is the conception of an 'oral-derived' Homer, the idea that oral habits do not immediately disappear with the advent of writing, so that Homer can continue to be studied as an oral poet without further ado by scholars who do acknowledge the presence of a text, or texts.³ The trouble with such conceptions is that "formula" and "text", "oral" and "literate", remain opposite notions whose unmodified discreteness is anachronistic. Advances in the study of textuality in attested historical contexts, not to mention the growing understanding of formularity in language in general, allow us to bring formula and written text together in a way that was inconceivable to the pioneers of the theory of oral-formulaic composition.

It is true that in order to get the nature of oral-formulaic composition in sharp focus, Homeric poetry had to be construed as textless in its composition, performance, and transmission. But the fact remains that in studying an oral Homer we cannot ignore the central presence and existential importance of what brought us Homer in the first place: the Homeric text, the written Homer that was available to readers, writers, and performers of the poems through the (late?) archaic and classical periods and beyond. This contribution will therefore address what seems at first sight a paradox: the importance of the formula for a textual criticism of an oral Homer. As we will see, neither "formula" nor "text" will emerge unchanged from this exercise.

### 1 Memorizing Homer

In Xenophon's Symposium we meet Niceratos, the son of the famous Athenian general Nicias,⁴ who informs his fellow symposiasts, in a conversation about what each of them considers "good" (3, 3), that his father had him learn all of the Iliad and the Odyssey by heart, as part of his education. Even now, he claims, he can still recite either poem in its entirety:

Xen., Symp. 3, 5 ὁ πατὴρ ὁ ἐπιμελούμενος ὅπως ἀνὴρ ἀγαθὸς γενοίμην ἠνάγκασέ με πάντα τὰ Ὁμήρου ἔπη μαθεῖν· καὶ νῦν δυναίμην ἂν Ἰλιάδα ὅλην καὶ Ὀδύσσειαν ἀπὸ στόματος εἰπεῖν.

My father, who took great care that I would become a good man, had me memorize all of the verses of Homer: even now I could recite the entire Iliad and Odyssey by heart.

See, e. g., Friedrich 2019.

See, e. g., Foley 1990, 5–10. Related is the idea of the "oral-dictated text" first introduced by Lord 1953; see Lord 1960, 126–128.

Plat., Resp. 327c 2; see Davies 1971, 405.

How could Niceratos pull off this remarkable feat of memory? He was wealthy and may well have been in possession of a personal copy of the poems. We may wonder what this text was like, as we will be doing throughout this chapter. We may also wonder whether a written copy of Homer was the only source of Niceratos' achievement. Classical Athens was a time when performance was still a powerful factor in the transmission and reception of Homeric poetry. Performance was the domain of the guild of the rhapsodes, the professionals of the Homeric tradition, whose shows Niceratos says he attended on a nearly daily basis (3, 6). He may have learned a great deal of Homer there; but as a knowledgeable member of the audience he must also have contributed to the agonistic quality of the performance: the rhapsode would know that due to the presence of listeners like Niceratos his audience was appreciative, but critical; he would need to please and impress them by putting up a good Homer if he wanted to win the prize in the competition.

Rhapsodes were not held in high esteem by the Socratics,⁵ nor by many modern scholars. In the orthodoxy of oral poetry scholarship (admittedly now superseded) the rhapsode comes to be seen as the reproducing reciter of fixed textual material, as opposed to the oral bard, the ἀοιδός, who re-creates the tale in performance, aided by his proficiency in the system of formulas. But the earliest usage of ῥαψῳδός and ἀοιδός does not make such a clear distinction between the two terms; and the idea of "rhapsode" is associated with "stitching," not with texts and reciting.⁶ The conception of the rhapsode as the reciter of the fixed Homeric text, then, is just as untenable as the rigid opposition of "orality" to "literacy." The idea of fixed texts, with any differences between them being due to scribal error (as in the medieval textual transmission), is inappropriate for classical Athens, as we see in more detail shortly.⁷

But even if we assume, for the sake of the argument, that Niceratos and the rhapsodes did have such texts, the recitation of them from memory would still not be the uncreative, mechanical reproduction thought to be the work of the rhapsode in some modern scholarship. Memorizing a text as long as the Iliad,

In Xenophon's text Antisthenes calls the rhapsodes "foolish" (ἠλιθιώτερον) after Niceratos admits that he attends their performances (3, 6).

Creative vs. reproductive: Parry 1971, 337, 356, 443; Kirk 1985, 15–16. Ancient use of the terms: Nagy 1990, 21–28; Graziosi 2002, 21–40; Skafte Jensen 2011, 145–153; González 2013. In the often-cited opening phrase of Pindar's Second Nemean Ode the two ideas are conjoined: Ὁμηρίδαι | ῥαπτῶν ἐπέων […] ἀοιδοί, "the Homeridae, singers of stitched verses" (Pind., Nem. 2, 1–2).

See Nagy 2009, 445 on the divergence of Aeschines' (Against Timarchus) Homer quotations from the medieval manuscripts.

even if the wording of the text is fixed and uncontroversial, is not a passive retrieval of narrative statements to be reproduced from some memory base; it is a dynamic cognitive operation: remembering the Iliad is reenacting the tale, indeed, performing it, if only mentally. The cognitive scientist David Rubin has analyzed the mnemotechnical aspects of oral traditions in terms of interlocking cognitive constraints, the systematic elimination of freedom, or the guidance of choice.⁸ The Homeric formulas and the intricate metrical patterns with which they interact are obvious examples of such cognitive constraints, as are the visual and spatial aspects of the narrative.⁹ Rubin's analysis of meter and formula can be combined with new developments in linguistics to provide an adequate account of Homeric poetry as a linguistic system, a language in which singers were fluent.¹⁰ But the important thing here is that fluency in this language is just as essential for Niceratos when he memorizes and recites Homer as it is for the textless "singer of tales". No amount of rote memorization of a text can lead to a convincing rendering of the Iliad; the "reciter" (if that is the right word) needs the same cognitive patterning as the "singer". Both are, to various degrees, speakers of the same language, the special language of the Greek epic tradition; both are subjected to the same set of cognitive constraints.

### 2 Adorning Homer

But memorizing Homer was not a one-way street from a fixed text to its recitation – in fact, the nature of the recitation itself as a dynamic operation would loop back into the very texts on which it was based. A good place to start here is looking at a contemporary of Xenophon's Niceratos, namely Ion the rhapsode as represented in Plato's dialogue of the same name. Throughout the dialogue we can glean important insights on the art of "doing Homer" even if the dialogue as such is skeptical of, if not hostile to, the art of the Homeric rhapsode. I quote here from the very beginning, where Socrates makes Ion a series of ironic compliments, to which Ion gives self-pleased replies:

Rubin 1995, 90–121.

Rubin 1995, 39–63.

See Bozzone 2010, 34–37, drawing a parallel between the way in which an apprentice singer acquires mastery of the formulaic systems (Lord 1960, 21–29) and the way in which children learn their language. Bozzone analyzes Homeric formulas analogously to "constructions" in ordinary language understood as "learned pairings of form and function" (e. g., Goldberg 2006, 5); see Bakker 2019, 86–88.

#### Plat., Ion 530b5–c6

καὶ μὴν πολλάκις γε ἐζήλωσα ὑμᾶς τοὺς ῥαψῳδούς, ὦ Ἴων, τῆς τέχνης· τὸ γὰρ ἅμα μὲν τὸ σῶμα κεκοσμῆσθαι ἀεὶ πρέπον ὑμῶν εἶναι τῇ τέχνῃ καὶ ὡς καλλίστοις φαίνεσθαι, ἅμα δὲ ἀναγκαῖον εἶναι ἔν τε ἄλλοις ποιηταῖς διατρίβειν πολλοῖς καὶ ἀγαθοῖς καὶ δὴ καὶ μάλιστα ἐν Ὁμήρῳ, τῷ ἀρίστῳ καὶ θειοτάτῳ τῶν ποιητῶν, καὶ τὴν τούτου διάνοιαν ἐκμανθάνειν, μὴ μόνον τὰ ἔπη, ζηλωτόν ἐστιν. οὐ γὰρ ἂν γένοιτό ποτε ἀγαθὸς ῥαψῳδός, εἰ μὴ συνείη τὰ λεγόμενα ὑπὸ τοῦ ποιητοῦ. τὸν γὰρ ῥαψῳδὸν ἑρμηνέα δεῖ τοῦ ποιητοῦ τῆς διανοίας γίγνεσθαι τοῖς ἀκούουσι· τοῦτο δὲ καλῶς ποιεῖν μὴ γιγνώσκοντα ὅτι λέγει ὁ ποιητὴς ἀδύνατον. ταῦτα οὖν πάντα ἄξια ζηλοῦσθαι.

I have certainly many times envied you rhapsodes, Ion, for your art, because it requires you to deck yourselves out and appear as beautiful as possible, and because it forces you to spend time with many excellent poets, Homer in particular, the best and most divine of all poets, and to acquire a thorough understanding of his thought, not just of his verses – all that is a matter for envy. Because the thing is that you can never be a good rhapsode without understanding what the poet says: the rhapsode needs to become an interpreter of the poet's thought to his audience; and if you do not know what the poets wants to say, this is impossible. In sum, all that is worthy of envy.

#### Plat., Ion 530d6–8

καὶ μὴν ἄξιόν γε ἀκοῦσαι, ὦ Σώκρατες, ὡς εὖ κεκόσμηκα τὸν Ὅμηρον· ὥστε οἶμαι ὑπὸ Ὁμηριδῶν ἄξιος εἶναι χρυσῷ στεφάνῳ στεφανωθῆναι.

And, yes, it is worth hearing, Socrates, how beautifullyIhave adorned Homer; so that I deserve, I think, to be crowned with a golden crown by the Homeridae.

In spite of the irony we can take away some interesting points from Socrates' "praise" and Ion's reaction to it: the rhapsode goes beyond learning the Iliad and Odyssey by heart; he does not stay at the surface of the poems' epē and strives to arrive at an in-depth understanding of Homer's "thought" (τὴν διάνοιαν ἐκμανθάνειν, μὴ μόνον τὰ ἔπη);¹¹ he is an "interpreter" of Homer (a ἑρμηνεύς), which will later in the dialogue (533d-e) be refined as the idea of a mediator between the poet and the audience; and, finally, Ion has "adorned" Homer, using perfect tense, κεκόσμηκα, that presents this act as some kind of lasting achievement.

Part of what this "adornment" means is surely Ion's histrionic performance, since the act of κοσμεῖν is also said explicitly to involve his appearance on the stage (τὸ σῶμα κεκοσμῆσθαι; cf. 535d2). Ion's art is certainly a matter of stage acting, because later in the dialogue Ion testifies to a particularly emotional style of performance, when he tells Socrates that he intensely relives the emotions of the

At Xen., Symp. 3, 6 we read, without the irony, that the rhapsodes do not understand the meaning of the texts (τὰς ὑπονοίας οὐκ ἐπίστανται, Socrates speaking). Cf. the interrogation of Hippias on Homeric subjects in Plato's Hippias Minor.

tale (535c4–8), and is intent on ensuring that his audience has the same experience. Ion's well-known line is that when the spectators weep, he will laugh, because he will likely win the prize, and when the spectators laugh, he will weep, because he probably won't (535e5–6).

But Ion's "adornment" of Homer is also a matter of the tale, the text, itself in its verbal articulation, because the idea of kosmos is pertinent to the telling of the story. The noun κόσμος denotes "order," "adornment"; applied to a song or a tale it would designate the best possible way in which you can present it, the most convincing version of it you can think of.¹² The noun itself does not occur in Ion, but it is regularly used in Homer, most often in the prepositional phrase κατὰ κόσμον "in the right, proper way", which can be said of the (singing of) song itself: "you sing of the plight of the Achaeans in the right way" (λίην γὰρ κατὰ κόσμον ᾿Aχαιῶν οἶτον ἀείδεις, Od. 8, 489), says Odysseus to the Phaeacian bard Demodocus, which in the mouth of Odysseus the eyewitness must apply no less to the accuracy of the events themselves than to the way in which the tale is told.¹³

Now if Ion were reciting a completely fixed text of the poems, it would not be easy to think of something that "adorning Homer" could mean apart from theatrics: the text would restrain him. But as already indicated, positing such a fixed and stable Homer in Classical Athens would be anachronistic; it would fly in the face of all that we know now about texts in attested historical contexts. From a wide range of cultures and periods we (now) know that "early" textuality does not produce stable manuscript traditions of poetic works.¹⁴ There is considerable variability, more than when transmission is purely textual and differences between manuscripts predominantly due to scribal error. Under these conditions the conceptual boundaries between speaking and writing, reading and hearing, become blurred: scribes are performers, even when the text they are producing is actually meant to be a text, rather than a transcript dependent on an actual per-

See Elmer 2010, on "ornamentation" in Greek and Bosniac as applying both to events and their representation in the speech-event of the epic performance. Kosmos (or kita in Bosniac) is "narrating the song as fully as possible" (Elmer 2010, 287) as matter of a given singer's personal style.

A little later Odysseus may be using the term κόσμος for the song itself, when he mentions the ἵππου κόσμον (8, 492), which possibly means "the song of the Wooden Horse" rather than "the construction of the horse." See Elmer 2010, 296 (with more references); González 2013, 195–198.

The primary impulse for this perspective on variance in early textuality ("scribe as performer") comes from medieval studies, e. g., O'Brien O'Keeffe 1987, Doane 1994, Cerquiglini 1999; overview of literature on other cultures and periods in Ready 2019, 192–215. See Bakker 1997, 23–27; Bakker 2019, 85.

formance. Their writing is not a mechanical act of copying. They know by heart what they write (and read); their writing is guided by memory, even when they write in order to copy an existing text.

We can call the texts produced in such a milieu "multitexts", manuscripts belonging to one and the same textual tradition showing variance that is not caused by scribal copying error. Multitexts are different from multiforms; in the latter case variance is potentially much greater – multiforms, if written down, do not necessarily belong to the same textual tradition – and can be caused by divergences between two or more independent performances.¹⁵

We might wonder what Ion's "adorning Homer" would mean in such a situation. Would κατὰ κόσμον "according to the true nature of the tale" not mean more freedom to "tweak" the text, to "adorn" the tale in order to produce a better Homer? Under these conditions the art of the rhapsode exists in the tension between on the one hand following where his memory leads him in his desire to reproduce Homer faithfully, and on the other the desire to "do a good Homer" in the competitive agonistic sphere of the public festival that served as décor to the performance of the epics.¹⁶

Doing Homer κατὰ κόσμον under these conditions, in the absence of a monolithic standard text, would mean that you would not consciously or willfully suppress lines or scenes; you would also not deliberately change lines that came to you out of your memory, although it would happen frequently that your memory ran out—again, memorizing a text as long as the Iliad or Odyssey was not a matter of rote learning—but you were able, because of your general competence in Homeric formulaic Greek, to substitute a similar phrase for the missing one.

### 3 Augmenting Homer

But you could very well – and surely many rhapsodes actually did – add lines out of a desire to make Homer "even better"; for example, you might be interested in elucidating Homer, in removing possible ambiguity, in enhancing the emotional impact of a given scene, or in underlining similarity you saw between two

See Ready 2019, 202.

Elmer 2010, 286–289 quotes a suggestive interview in which the Bosnian master-singer Avdo Mededović conceives of kita "ornamentation" as doing justice to the epic events themselves: if other singers omitted descriptive detail, they would have 'omitted' them from the (ideal version of) the tale.

given scenes – all by adding lines that you knew from memory, lines that typically are used elsewhere in Homer.

The transmission of Homer in antiquity, then, is subject to a process of what we could call "immanent expansion". This phenomenon has long been recognized, but mostly as a matter of creeping interpolation, of which Homer needs to be purged by the editor, whose task it is to recover the "authentic" Homeric text, an editorial stance and practice, pioneered in antiquity by the Alexandrian scholar Zenodotus, that continues to the present day.¹⁷ The stance taken in this contribution is that "immanent expansion" is not something that obscures the view of the 'real' Homer; it is Homer, 'Homer' being defined as the poetry viewed under its diachronic aspect: Homer is the process of its own transmission, and that includes both oral transmission and textual transmission.¹⁸

The best window on the reality of this "augmentation" of Homer – and on the general culture of scribal performance described earlier – comes from Ptolemaic Egypt, the time and place that has supplied the oldest physical witnesses to the Homeric text. The surviving papyri from this early (4th-3rd century BC) period not only display "horizontal" variance with respect to the medieval manuscripts (i.e., line-internal variants in word choice or prosody), but also "vertical" variance: they have lines that do not occur in the medieval manuscripts or omit lines that occur in them, thus upsetting the numerus versuum on which all modern Homeric textual criticism is built.

These early papyri are often called "wild," or "eccentric," as if the variation they display takes place in a "wild" or "eccentric" way with respect to an established standard text. The view taken here is that the "wild" features they display are a sign precisely of the absence of such a text. They are produced in the milieu described earlier, in which the activity of copying has to be seen as a kind of performance.¹⁹ The Ptolemaic papyri provide valuable evidence for the fact that the existence of written texts does not stop the linguistic innovation that for Milman

Bolling 1925; Apthorp 1980. See the overview in Bird 2010, 49–56. The latest editor of the Homeric poems, Martin L. West, isastrong proponent of the perceived need to purge the text from interpolated versus iterati.

On Homer as the personified and retrojected primordial poet who is reenacted in performance, see Nagy 1996, 59–86; Homer as myth turns into "transcendental" what in reality is the immanent principle of stability of the tradition.

See Ready 2019, 197–203, who points out (201) that the divergence between papyri and the medieval standard text (as well as between papyri) is very unlikely to be due to the fact that different papyri are different transcriptions of actual performances. Bird 2010, 71 reminds us that papyri confirm the accuracy of the Homeric quotations in Athenian texts such as Plato, Hipp. Minor and Aesch., In Timarch: these are not quotation "errors" resulting fromafaulty memory.

Parry was inherent in living oral traditions.²⁰ But the emphasis in this chapter is on the most distinctive feature of the Ptolemaic papyri, the "vertical" variation they display with respect to the standard medieval text.

There are plus-verses and minus-verses: cases where a papyrus reads a line that does not occur in the manuscripts, and cases where, conversely, a papyrus skips a line that occurs in most manuscripts. Plus-verses and minus-verses seem at first sight to be opposites, but I will argue below that they are in fact two sides of the same coin. Plus- and minus-verses are treated asymmetrically by scholars who think in terms of an authentic text: the former tend to be seen as evidence for interpolation to be 'purged' by the textual critic, whereas the latter are seen as windows on a less augmented, "purer" Homer.

Some plus-verses are unique and not found anywhere else in Homer, such as Il. 3, 302a (on which see below) and Od. 9, 537a, which provides detail about the second rock the Cyclops hurls at Odysseus' boat.²¹ More frequently they occur elsewhere in Homer and are hence called versus iterati. Likewise, the phenomenon of minus-verses concerns lines that occur elsewhere in the epics. We are here concerned with ways in which such versus iterati and non iterati can play a role in understanding the process of "adorning Homer" as practiced by Ion.

Consider now the case of Il. 3, 302–305 and Iliad pap. 40, a document covering books II and III of the poem. In the left column is the manuscript text, in the right-hand column the text of the papyrus. The context is as follows: preparations are being made for a duel between Paris and Menelaus, with at stake the fate of Helen and the final outcome of the war. Agamemnon has invoked Zeus, Helios, and Gaia as witnesses for an oath by which the two parties will honor the outcome of the fight, and he has performed an oath sacrifice. An anonymous Achaean or Trojan has uttered a prayer that whoever will be the first to violate the arrangement made under oath will die along with his family and descendants. This is where the fragment takes off:

 ὣς ἔφαν, οὐδ᾽ ἄρα πώ σφιν ἐπεκραίαινε Κρονίων. [ὣς ἔφαν εὐχό]μενοι, μέγα δ᾽ ἔκτυπε μητίετα Ζεύς a [Ἴδης ἐκ κορυ]φων̣ ἐ̣π̣ὶ̣ δὲ στεροπὴν ἐφέηκεν·

Parry 1971, 332. In the papyri there is, e.g., a preference for augmented forms, e. g., Il. 2, 205 ἔδωκε (see S. West 1967, 46), Od. 9, 560 ἐφάνη (S. West 1967, 247); for modernized morphology (e. g., Od. 9, 520 ἐθέλῃ for ἐθέλῃσ᾽) and pronunciation (e. g., Od. 9, 546 νεῖα for νῆα). See Ready 2019, 192–193, providing a veritable index to S. West 1967.

<sup>ἐ</sup>ξ αὐλῆ<sup>ς</sup> <sup>ἀ</sup>νελὼ<sup>ν</sup> <sup>ὅ</sup>θι ο<sup>ἱ</sup> θυρεὸ<sup>ς</sup> ̣π̣α̣ . οσ ε[.] . [, for which <sup>π</sup>άρος <sup>ἔ</sup>σκεν and λα[ε]ος <sup>ἔ</sup>πλεν have been proposed as possible restorations.


First we notice that in 303–304 there are minor differences with our text along with one plus-verse: in the manuscripts Priam speaks among (μετά) the Trojans and the Achaeans, whereas in the papyrus he speaks to (προτί) his own people, the Trojans and their allies. This is the kind of horizontal variation one can expect under the relatively fluid conditions of early textuality. The plus-verse 304a does not add much to the passage, but it can be seen in principle as triggered by the formulaic conditions: all of the ten attestations of the line in Homer occur after a line beginning with κέκλυτε, in which a speaker addresses a collective.²² But we can be more specific: a cluster of these are found in book VII, where Trojan leaders (Hector, 7, 68, Antenor, 7, 349, and Priam himself, 7, 369) address both armies (Hector) or only the Trojans (Antenor and Priam) under conditions that bear similarity to the present situation: a duel between an Achaean and a Trojan is involved, with much at stake concerning the possible end of the war and the terms under which this might take place.

At 302 we see that the restrained narratorial comment "and Kronos's son (we now know) did not yet bring this to fulfillment for them" is expanded into a fourline description of Zeus showing through thunder and lightning that he is not (yet) willing to punish any oath breakers: the war will not be terminated by this duel and Achaeans and Trojans will have to endure further heavy losses and suffering.We may wonder what motivated this expansion. In our transmitted text we notice the particle ἄρα, which can be described as "evidential": it suggests that there is evidence from which the narrator concludes that Zeus does not grant the wish.²³ The scribe of the papyrus may have wanted to make this implicit evidence explicit, as if the narrator "knew" that there was thunder and lightning in the narrative, but didn't say so. The performer in this way, "interprets" Homer for his audience (remember Ion's use of the term ἑρμηνεύς), with the intention to deliver a Homer κατὰ κόσμον.

Now verse 302 itself as it stands in our text is also (with adjustment for grammatical number) found at 2, 419:

Il. 2, 419–420 ὣς ἔφατ᾽, οὐδ᾽ ἄρα πώ οἱ ἐπεκραίαινε Κρονίων ἀλλ᾽ ὅ γε δέκτο μὲν ἱρά, πόνον δ᾽ ἀμέγαρτον ὄφελλεν.

this is how he spoke, but Kronos's son did not yet bring this to fulfillment for him; no, he did accept the sacrifice; but he was increasing wretched toil.

The situation is similar: a character (Agamemnon) makes a sacrifice and a prayer that amounts to wishful thinking – it pertains to what he hopes will be the course of events in the plot. The narrator informs us that Zeus did not grant his wish.

But the corresponding line 302 in the papyrus is modeled on a verse that comes much later in the poem, in book XV at the height of the Battle of the Ships:

Il. 7, 68, 349, 369; 8, 6; 19, 102; Od. 7, 187; 8, 27; 17, 469; 18, 352; 21, 276.

Bakker 2005, 98.

Il. 15, 377–378 ὣς ἔφατ' εὐχόμενος, μέγα δ' ἔκτυπε μητίετα Ζεύς, ἀράων ἀΐων Νηληϊάδαο γέροντος.

Thus he spoke praying; and greatly he thundered, Zeus of the counsels, answering the prayer of the old man, Neleus' son.

Again a character (Nestor) makes a prayer, this time not a case of wishful thinking that is at odds with the course of the plot, but a real, urgent prayer for salvation, reminding Zeus of earlier sacrifices and his promise to grant the sacrificer a safe nostos, and imploring him now to ward off imminent destruction for the Achaeans at the hands of the Trojans. The god thunders in response, at first sight as a sign that he answers Nestor's prayer; but the sign is ambiguous, because the Trojans hear the thunder too and double the intensity of their attack, increasing Nestor's plight.

The first actual plus-verse (302a) is nowhere else found in Homer;²⁴ but 302b-c are: they are found at the beginning of book II, when Agamemnon wakes up after the dream sent by Zeus. No public prayer here, nor any celestial phenomena; just the mistaken belief – again – that Troy will be taken this very day. The narrator intrudes here as well and announces that this is not going to be the course of events: much more battle and suffering is in store for both Trojans and Danaans.

Il. 2, 37–40 φῆ γὰρ ὅ γ' αἱρήσειν Πριάμου πόλιν ἤματι κείνῳ, νήπιος, οὐδὲ τὰ ᾔδη ἅ ῥα Ζεὺς μήδετο ἔργα· θήσειν γὰρ ἔτ' ἔμελλεν ἐπ' ἄλγεά τε στοναχάς τε Τρωσί τε καὶ Δαναοῖσι διὰ κρατερὰς ὑσμίνας.

he asserted that he was going to take Priam's city that same day, fool: he had no idea of those actions Zeus was contemplating: for, yes, he was still going to put further woes and suffering for Trojans and Danaans alike through strong battles.

The papyrus lines 3, 302b-c are so appropriate in their context that some scholars believe that they are original here and interpolated into the beginning of book II as 2, 39–40.²⁵ The idea of primacy implied here – i.e., the distinction between

The reading Ἴδης ἐκ κορυφῶν is plausible; the phrase is not attested in Homer, but the equivalent dative expression (Ἴδης ἐν κορυφῃσι) is (Il. 11, 183; 14, 332; 15, 5; 22, 171); see S. West 1967, 52.

Details in S.West 1967, 53; by contrast, 302d is found suspect, since typically the act of oathtaking is mentioned in the previous context (e. g., Il. 14, 280, Od. 10, 346). The rationale for the

an "authentic" line and its repetition as an interpolation elsewhere – is problematic; but the idea of similarity between contexts has great potential, as will be further developed below.

We observe, then, that the oath sacrifice as preparation for the duel between Paris and Menelaus partakes in a network of contexts (Agamemnon's dream, his sacrifice, the duel between Hector and Ajax) revolving around false hopes and mistaken ideas about the course of the war. Phraseology pertaining to this common theme circulates among the various contexts involved. Before we pursue this further, let us pause for a moment to reflect on the importance of context for the study of Homeric formulas.

### 4 Context and Randomness

Just as in language at large, words, formulas, in Homer are inseparable from their contexts.Words are associated with specific contexts in ways that make language infinitely more complex, cognitively, than mere grammatical rules. And if this is true for language in general, it is doubly true for the special language of Homer. Formulaic phrases are so strongly associated with the contexts in which they are uttered that they come to stand for those contexts. Uttering a formula is in this way to evoke the formula's context, to state that the present context is like the context(s) with which the formula is associated.²⁶

In this way the plus-verses are much more than the mere repetition of lines. They evoke the contexts with which these lines are indissolubly associated: the inane wishful thinking in the face of what the narrator knows to be the course of events in the war, and the inherent ambiguity of signs from Zeus, such as the thunder.

It is important to realize that the reference of the repeated formulas in the plus verses is to the Iliad specifically, and not to the epic tradition at large, to any generic prayers that do not come to fulfillment or to any generic signs of Zeus that are ambiguous. And this leads to the idea of randomness in the distribution of Homeric formulas.

It is often assumed that the sum total of formulas we find in Homer is a subset of a much larger set of oral poetry that is now lost. An instructive quote in this regard is from Milman Parry in a passage on repetition:

line, however, can be that after Agamemnon's oath (3, 276–291) and the oath sacrifice he performs (3, 292–296) a matching oath on the Trojan side is needed.

See Bakker 2017, 59.

It is important […] to remember that the formula in Homer is not necessarily a repetition, just as the repetitions in tragedy are not necessarily formulas. It is the nature of an expression which makes of it a formula, whereas its use a second time in Homer depends largely upon the hazard, which led a poet, or a group of poets, to use it more than once in two given poems of a limited length. We are taking up the problem of the Homeric formulas from the side of the repetitions, but only because it is easier to recognize a formula if we find it used regularly, and that it helps the poet in his verse-making. (Parry 1971, 304 = Parry 1930, 122 – emphasis added –)

In order to qualify for being a formula, a phrase doesn't have to occur in Homer twice or more, because it can be a phrase that happened to occur only once, while it is frequent in poems that are now lost. For Parry, the Homeric poems are no more than "two given poems of a limited length" which contain far fewer formulas than the sum total of all the formulas in all early epic poetry, much of which is now lost or has never been committed to writing. Similar is the definition of the influential concept of "traditional referentiality" by John Miles Foley:

Traditional referentiality […] entails the invoking of a context that is enormously larger and more echoic than the text or work itself, that brings the lifeblood of generations of poems and performance to the individual performance or text. (Foley 1991, 6 – emphasis added –)

This comes close to the idea of formulas being associated with given contexts; but the crucial difference is the idea that the associated contexts are seen as matter of 'the tradition' (at large), of which the Homeric poems are no more than a single and partial instantiation.

It is of course undeniably true that there is formulaic material in Homer that is also found elsewhere. The line ὄφρ᾽ εἴπω τά με θυμὸς ἐνὶ στήθεσσι κελεύει / ἀνώγει (line 304a in the papyrus) is also found in Hesiod (Theog. 645; fr. 75, 14 M.–W.). But it cannot be true that the set of formulaic expressions that is Homer is a chance, random, affair. If it were, the survival of the Homeric poems would be a chance survival, and this is obviously not the case.

### 5 Interformulaic Homer

The Homeric poems have been preserved for no other reason than that they are the Homeric poems. We have to assume, then, that in many cases formulas in Homer are repeated for no other reason than that they are formulas in Homer. The repetition is done to signal a link or similarity between two contexts within Homer, sometimes across the boundary of the two poems. I have elsewhere called this principle "interformularity".²⁷ This term is modeled on "intertextuality," of course, but I am less interested in the use of this term by literary scholars, when they talk about the pervasive allusive nature of, say, Latin poetry; rather, I feel affinity with the idea of intertextuality in ordinary language, as discussed in the work of Boris Gasparov:

Whatever one chooses to make out of a quantum of language matter, one cannot abstract it from its original association with a certain experiential landscape out of which it has been drawn by memory. (Gasparov 2010, 3–4)

In the case of Homer the universe in which language is imbued with the meaning and the contexts with which it is associated is much more constrained and turned on itself than the real world. And the "experiential landscape" that is part of the meaning of formulas is not social; it pertains to the progression of the epic tale that everyone knows so well. The art of Homeric repetition, or at least part of it, is to create a dense network of associations that holds the poem together, thematically and structurally.

Reusing lines, whether or not modified, are of central importance in this regard. From the point of view of the practicing rhapsode they are attempt at doing Homer κατὰ κόσμον; they are where the rhapsode's memory led him, either because he had in fact heard (or seen) the plus-verses earlier in a given particular context, in the performance of someone else and/or in a text he had seen – or he composed (or rather re-composed) them himself, potentially setting up a model for others to follow.

This technique of weaving the text together by means of the texture of the formulas is not confined to the "wild" papyri and their plus-verses. It equally involves the minus verses in those papyri; and it is ubiquitous in the standard text that came out of the pipeline of Hellenistic textual scholarship.²⁸

To stay with the duel between Paris and Menelaus, when the fighting begins and the combatants enter the delimited area in between (ἐς μέσσον) the two armies, we see the following formulaic sequence:

Il. 3, 340 –343²⁹ οἱ δ᾽ ἐπεὶ οὖν ἑκάτερθεν ὁμίλου θωρήχθησαν, ἐς μέσσον Τρώων καὶ ᾿Aχαιῶν ἐστιχόωντο δεινὸν δερκόμενοι· θάμβος δ᾽ ἔχεν εἰσορόωντας Τρῶάς θ᾽ ἱπποδάμους καὶ ἐϋκνήμιδας ᾿Aχαιούς.

Bakker 2013, 157–169.

See also Schein 1999; Bakker 2017.

Cf. Il. 23, 813–815.

And they when on either side each with their people they had armed themselves, you could see them striding into the middle space in between Trojans and Achaeans they strode,

eyes frightening; and amazement held them as they were watching, the horse-taming Trojans and the well-greaved Achaeans.

This is apparently a formulaic sequence specifically associated with duels conceived as spectacle; vv. 340–342, modified, are also is used in the Funeral Games for Patroclus to introduce the fight between Diomedes and Ajax.³⁰ Verse 343 is strictly speaking redundant after Τρώων καὶ ᾿Aχαιῶν in 341, and is hence a candidate for interpolation in the eyes of the scholars seeking to purge the Homeric text of redundancies. It is, in fact, a minus-verse in an unpublished Oxyrhynchus papyrus used by West in his Iliad edition. Accordingly, from the standpoint of this papyrus and its scribe the transmitted 3, 343 is a plus-verse. The scribe who put it there very probably had in mind a scene not very far ahead, when Athena comes down from Olympus like a celestial phenomenon:

Il. 4, 79–80 κὰδ δ᾽ ἔθορ᾽ ἐς μέσσον· θάμβος δ᾽ ἔχεν εἰσορόωντας Τρῶάς θ᾽ ἱπποδάμους καὶ ἐϋκνήμιδας ᾿Aχαιούς. And down she jumped into the middle space; and amazement held them as they were watching,

the horse-taming Trojans and the well-greaved Achaeans.

The line in question (now 4, 80) is here not informationally redundant, and it is not disputed by any wayward papyrus. Its reuse at 3, 343 is due, I propose,³¹ to the desire to exploit and enhance the mutual attraction between the two scenes, the beginning of the "real" fighting (after Athena's persuading Pandaros to shoot an arrow at Menelaus, 4, 93–104) and the staged spectacle that was the duel between Paris and Menelaus. Just as the duel is a stylized, staged version of the "real" fight, so that real fight comes now to be seen, conversely, as aligned with a staged spectacle, attended not only by the direct combatants, but also by the gods, whose privileged perspective is shared with the poem's audience.³²

Il. 23, 813–815 οἱ δ᾽ ἐπεί […] θωρήχθησαν / ἐς μέσον ἀμφοτέρω συνίτην μεμαῶτε μάχεσθαι, / δεινὸν δερκομένω· θάμβος ἔχε πάντας ᾿Aχαιούς.

The idea of 're'-use for a line that occurs later in the poem is paradoxical only in a reductively linear reading of the poem, as if the reader-performer encounters each line successively as new.

See Myers 2019, 80–81.

Textual critics seeking to purge the Homeric text of perceived redundancies in order to recover a more authentic Homer tend to dismiss plus-verses as redundant, while accept minus-verses as evidence of a purer, more authentic text. The latest editor of the Odyssey, Martin West, is a clear example of this practice. He suppresses Od. 9, 489 in Odysseus' description of the escape from the Cyclops:

Od. 9, 487–490 αὐτὰρ ἐγὼ χείρεσσι λαβὼν περιμήκεα κοντὸν ὦσα παρέξ· ἑτάροισι δ' ἐποτρύνας ἐκέλευσα [ἐμβαλέειν κώπηισ', ἵν' ὑπὲκ κακότητα φύγοιμεν], κρατὶ κατανεύων· οἱ δὲ προπεσόντες ἔρεσσον.

But I took a long punting-pole in my hands and thrust alongside; my companions I urgently ordered to lean on the oars, so that we could escape from danger, nodding with my head. And they fell forward on the oars and rowed and rowed.

The line is not found in a number of manuscripts, but it is presumably its absence in two papyri that makes West decide to keep it out of his text, considering it an interpolation of Od. 10, 129, in the passage describing the escape from the Laestrygonians. It may have been absent in some renditions (although after ἐκέλευσα in the previous line we expect an infinitive³³), but its presence in others is very natural, considering the mutual attraction the separate adventures of Odysseus' wanderings have to one another, and this applies to the Cyclops episode and the Laestrygonians in particular.³⁴

A second example is more complicated. In book X, when Odysseus has learned from Circe that he has to travel to Hades to consult the soul of Teiresias, he reacts as follows:

Od. 10, 496–500; 4, 538–541 ὣς ἔφατ᾽, αὐτὰρ ἐμοί γε κατεκλάσθη φίλον ἦτορ· {κλαῖον δ᾽ ἐν λεχέεσσι καθήμενος, οὐδέ νύ μοι κῆρ ἤθελ᾽ ἔτι ζώειν καὶ ὁρᾶν φάος ἠελίοιο. αὐτὰρ ἐπεὶ κλαίων τε κυλινδόμενος τε κορέσθην,} καὶ τότε δή μιν ἔπεσσιν ἀμειβόμενος προσέειπον·

Thus she spoke, but for me my dear heart broke: I wept sitting in the bed, and my heart No longer wanted to be alive and see the light of the son. But when I had had my fill of crying and rolling about I then in answer addressed her in reply.

See Heubeck 1989, 38.

Bakker 2013, 24–25; 74–75.

The first four lines also occur in book IV, in Menelaus's account of his travels, when the hero learns from Proteus the Old Man of the Sea that his brother Agamemnon has been killed by Aegisthus. They are not found in Papyrus 431, an unpublished Oxyrhynchus papyrus that contains this part of the Circe episode. This makes West decide that they have been interpolated into book X from book IV. In an article accompanying his edition, he says that there is "poetic gain" from deleting the lines in book X.³⁵

There is no reason to expunge the lines from the Circe context. On the contrary, its details, in particular the idea expressed in v. 498 (no longer wanting to be alive and see the light of the sun), are very pertinent to the themes of the Circe episode, with its emphasis on death and rebirth: the voyage to Hades is a symbolic death – when they return, Circe addresses them as δισθανέες, "those who died twice."³⁶ But beyond being fitting in their context, the lines gain significance in their resonance with Menelaus' reaction to the news from a divine informant. It is the mutual attraction between the two scenes that counts, not the question whether either of them is "primary" or "interpolated".³⁷

Suppressing in our text the minus-verses of papyri does not give us an authentic Homer any more than does restoring digammas and other archaic features beyond the evidence of the manuscripts.³⁸ A minus-verse in a given papyrus may well be evidence for the transmitted verse in the manuscripts as a standardized plus-verse.³⁹ The oxymoronic nature of that expression means that repetition is not mere redundancy or an epiphenomenon of transmission; it is baked into the very texture of Homeric poetry. And repetition is not a matter of replicating mere words or lines, but a means to draw attention to what the poet, or rhapsode, or scribe, perceives as similarity between two or more scenes; repetition is in fact what makes Homeric phrases mean what they mean.

West 2017b, 21.

Cf. 10, 174–175. One of West's arguments is that 10, 500 in the reading he prefers, καί μιν φωνήσας ἔπεα πτερόεντα προσηύδων, fits better immediately after 496. But presumably for this reason most manuscripts have καὶ τότε δή μιν ἔπεσσιν ἀμειβόμενος προσέειπον, the apodotic continuation καὶ τότε (δή) being more acceptable after a preceding subclause starting with αὐτὰρ ἐπεί (cf. Od. 9, 345, 363).

In spite of the criticism of West's editorial practice I wish to emphasize that his editions are invaluable for their exhaustive documentation of the papyrological record.

See Graziosi/Haubold 2019 on the risk of constructing a pre-Homeric (and therefore un-Homeric) text.

Cf. Il. 9, 311 ὡς μή μοι τρύζητε παρήμενοι ἄλλοθεν ἄλλος at the beginning of Achilles' great speech, a line that does not occur in the quotation of the passage by Plato's Hippias (Plat., Hipp. Min. 365a1–b2).

### 6 The Epic Cosmos

On this basis we can assert that the Iliad and Odyssey are a language; not a language in the usual sense of "Homeric language", the curious mix of dialects, of archaic and modernizing features, and of metrically determined formulas, which comprise its primary features. Homer is also a narrative meta-language; a language in which many phrases convey meaning, not merely through their common linguistic features (phonetic, syntactic, semantic, pragmatic), but also through their place and role in the narrative: they mean what they mean because of the contextual associations attached to them.

And like every language, this meta-language is nothing without its speakers and hearers, who are members of a language community for whom utterances mean much more than their surface linguistic meaning. The contexts with which the communicative units of this meta-language (part-line formulas, whole lines, blocks of lines) are "impregnated" are not social, physical, or psychological; they are bound up with the tale. The universe of discourse in which both performers and audiences, scribes and readers, of Homer operate is much more closed and focused than the universe of discourse that speakers of a language, or groups of speakers of an ordinary language, share; it is the tale and its constitutive features.

This is what a contemporary of Niceratos and Ion, the philosopher Democritus, may have meant when he calls Homer an ἐπέων κόσμος,

Democr. B21 D.–K. Ὅμηρος φύσεως λαχὼν θεαζούσης επέων κόσμον ἐτεκτήνατο παντοίων.

Homer, whose share was a divine nature built a cosmos of all kinds of different epea.

ἐπέων κόσμος "orderly arrangement of epea" is a remarkable phrase, which combines our modern idea of a "universe of discourse" with the idea of the poem itself as universe, a universe that is "epic" in a double sense: it constitutes the epic tale and it has epos, the epic verse, as its building material, a universe made up of formulas and interformulaic connections.

Being able to build and propel this cosmos is being able to, and being interested in, adding verses that strengthen the internal structure of the narrative universe by highlighting similarities between two or more scenes. This is part of what Ion may have meant when he talks about "adorning Homer". When he says κεκόσμηκα τὸν Ὅμηρον, who knows how many interformulaic repetitions he has brought to our text? Repetitions for which we have no one to praise – or to blame – but Homer himself. And the fact that there are always hearers like Niceratos in the audience indicates that this operation is not only a matter of creation and supply, but also of appreciation and demand. People who know Homer well can also tell a good Homer when they hear one.

### References


Labarbe, J. 1949. L'Homère de Platon. Liège.


## Albio Cesare Cassio Two Chronological Extremes of the Homeric Language: πρόφρασσα and εἶπα

Abstract: This contribution falls into two parts, each dealing with modifications undergone by the Homeric text at very different chronological stages. The first part aims at showing that the archaic feminine pseudo-participle πρόφρασσα "strong-willed / benevolent" is not just a linguistic curiosity of limited interest, but its obsolete morphology combined with its exclusive use for goddesses, and only when they are addressed by male individuals, betrays the existence of a pre-Homeric 'typical scene' unnoticed by scholars so far. In the second part I hope to show that, contrary to ancient and modern opinions, the Homeric inflection of (ἔ)ειπον was not purely thematic, since the only transmitted form of the 2nd pers. plur. imperative is εἴπατε, while ἔειπας / εἶπας is often well attested in Homeric manuscripts and papyri. Interestingly enough, these -α- forms, ubiquitous in many spoken dialects, prompted reciters and scribes of Homer to engage in 'bold' innovations, like imperative εἶπον found in a 'wild' Hellenistic papyrus; they, however, were discarded fairly soon and are unknown to our medieval manuscripts.

Keywords: Homer; oral literature; Greek dialects; Indo-European reconstruction; linguistic innovations; written fixation.

It is commonly accepted that in the Greek-speaking world oral texts recounting heroic deeds go back to extremely remote antiquity, i.e. Mycenaean times, if not earlier; as a consequence it is hardly surprising that our Homeric text, which took its definitive shape in the 8th/7th century BC, appears as a conglomerate of a number of linguistic stages and local variations, attested in the innumerable variants found in our manuscripts.

From the last decades of the 19th century our knowledge of this matter has enormously been enlarged by the discovery of numerous Homeric papyri. In some cases they have revealed that certain lines had originally a more archaic shape than the one found in Medieval manuscripts, thus confirming conjectures made by modern linguists; in others they have showed that in the Hellenistic age the text had undergone remarkable modifications, and some forms had been 'modernized' to an extent unknown to medieval manuscripts. Yet, no apparatus can tell us what happened before the earliest written text took shape. The key to the very earliest compositional phases can only be found in the text itself, as Richard Bentley discovered more than three centuries ago.

In what follows I shall deal briefly with two case studies which are also two chronological 'extremes': the former will take us back to remote times when some athematic forms were remarkably different from those we are familiar with; the latter will touch on an innovation found only in Ptolemaic papyri and happily surviving in Modern Greek – although the roots of the innovation are much older than Hellenistic times.

### 1 ΠΡΟΦΡΩΝ / ΠΡΟΦΡΑΣΣΑ

The adjective πρόφρων is amply attested in Greek poetry from Homer onwards, and is often rendered with "gentle" or "friendly", but these meanings are in fact secondary developments. The basic meaning is "of one's free will",¹ or "ready for action";² not surprisingly, in Homer the adjective is often applied to the gods, especially when they bestow favours on mortals (e. g. Il. 8, 175–176 γιγνώσκω δ' ὅτι μοι πρόφρων κατένευσε Κρονίων / νίκην καὶ μέγα κῦδος). However, it should be noted that divine benevolence is far from guaranteed, since at times a πρόφρων god can lead a mortal astray.³

Πρόφρων belongs to the well-known archaic category of two-ending adjectives, hence the same form is used for males and females, e. g. Od. 13, 359 αἴ κεν ἐᾷ πρόφρων με Διὸς θυγάτηρ ἀγελείη. In the V book of the Odyssey when Calypso is obliged by Zeus to set Odysseus free, she says with superior nonchalance: "I, πρόφρων, will show him, without concealing anything, the means of reaching his fatherland unscathed" (5, 143–144 αὐτάρ οἱ πρόφρων ὑποθήσομαι οὐδ' ἐπικεύσω / ὥς κε μάλ' ἀσκηθὴς ἣν πατρίδα γαῖαν ἵκηται). For this reason one is slightly taken aback on discovering, a few lines later, that the same Calypso addressing Odysseus (161) describes herself as πρόφρασσα: "do not despair: I, πρόφρασσα, will let you go", ἤδη γάρ σε μάλα πρόφρασσ' ἀποπέμψω.

So a special feminine of πρόφρων did exist. Yet its very existence is surprising, since it is completely unexpected for this class of adjectives – a cheerful Greek woman could be described as εὔφρων, certainly not \*εὔφρασσα or \*εὔφραττα. Moreover, at Od. 5, 161 πρόφρων would have created no metrical problem: ἤδη γάρ σε μάλα \*πρόφρων ἀποπέμψω.

See LSJ 1940, s.v.

See Montanari 2015, s.v.

Theogn. 403–404 ἀνὴρ κέρδος διζήμενος, ὅντινα δαίμων / πρόφρων εἰς μεγάλην ἀμπλακίην παράγει.

Πρόφρασσα is not confined to that passage; yet it is extremely rare in Homer (five occurrences in all, with no textual variants), and other surprises are not lacking. Besides Calypso' words in Od. 5, 161 just quoted, πρόφρασσα appears in the following four lines:

Il. 10, 290: Diomedes addresses Athena, recalling how helpful she was to his father Tydeus, and asks for her protection and help.

μάλα μέρμερα μήσατο ἔργα σὺν σοὶ δῖα θεά, ὅτε οἱ πρόφρασσα παρέστης. ὣς νῦν μοι ἐθέλουσα παρίσταο καί με φύλασσε. Note ἐθέλουσα, which conveys the same meaning as πρόφρασσα.

Il. 21, 500: Hermes addresses Latona and grants her victory without fighting: ἀλλὰ μάλα πρόφρασσα μετ' ἀθανάτοισι θεοῖσιν εὔχεσθαι ἐμὲ νικῆσαι κρατερῆφι βίηφιν.

Od. 10, 386: Odysseus addresses Kirke, asking her to to return his companions to human shape:

ἀλλ' εἰ δὴ πρόφρασσα πιεῖν φαγέμεν τε κελεύεις, λῦσον, […]

Od. 13, 391: Odysseus addresses Athena (389 ff.), recalling how he was helped by the goddess during the Trojan war:

αἴ κέ μοι ὣς μεμαυῖα παρασταίης, γλαυκῶπι,

καί κε τριηκοσίοισιν ἐγὼν ἄνδρεσσι μαχοίμην

σὺν σοί, πότνα θεά, ὅτε μοι πρόφρασσ' ἐπαρήγοις.

Both concept and wording bear a strong resemblance to the πρόφρασσα παρέστης of Il. 10, 290.

In his commentary on Od. 5, 161 Hainsworth⁴ explained πρόφρασσα as follows: "fem. of πρόφρων < \*προ-φρn̥τi̯α, after the analogy of -ντ- stems, type χαρίεντfem. χαρίεσσα, with -ε- after the masc. for the expected α < nσ", an explanation repeated in his commentary on Il. 10, 290.⁵

Unfortunately, Hainsworth's words are doubly embarrassing. Firstly, one cannot see the arcane reason why a perfectly normal -n- declension – πρόφρων πρόφρονος – should inconvenience a -nt- declension in order to createafeminine, and one which is completely unnecessary. Secondly, Hainsworth seems to be unaware of the (right) explanation independently offered by Wackernagel (1914, 41–42 = Wackernagel 1955, 1143–1144) and Meillet.⁶

See Hainsworth in Heubeck/West/Hainsworth 1988, 269.

Hainsworth 1993, ad Il. 10, 290 "πρόφρασσα, also at 21, 500 and 3 x Od., is an epic feminine of πρόφρων (which also serves as a feminine, e. g. at 244). The formation is analogical after archaic feminines in -ασσα of ντ-stems".

See Wackernagel 1914, 41–42 = Wackernagel 1955–1979, 1143–1144; Meillet 1920.

As a matter of fact, our πρόφρασσα presents several peculiarities. For a start, while πρόφρων in Homer is also attested in cases different from the nominative singular (Il. 8, 39–40 πρόφρονι θυμῷ; 22, 303 πρόφρονες εἰρύατο) πρόφρασσα is found in the nominative singular only. And, as Wackernagel noted,⁷ in the two Odyssean lines quoted above (13, 359 αἴ κεν ἐᾷ πρόφρων με Διὸς θυγάτηρ ἀγελείη, and 5, 144 αὐτάρ οἱ πρόφρων ὑποθήσομαι οὐδ' ἐπικεύσω) πρόφρων is "im Prädikats-Nominativ, quasi-adverbiell", and "nur in dieser letzter Funktion kann dafür in Homer auch πρόφρασσα eintreten".

Then Wackernagel wentastep further and noted that the meaning of πρόφρων "strong- willed" was virtually the same as that of ἑκών (earlier ϝεκών) "wittingly, purposely":⁸ the latter is in fact an old participle, the remnant of a verb built on the I. E. root \*u̯eḱ- (Vedic váṣṭi) "desire, want", whose full inflection disappeared in Greek at an early stage; the original femin. participle with the zero grade of the suffix was ϝέκασσα (/ϝέκαττα) < u̯eḱn̥ti̯a, a grade still kept in Cyren. ἕκασσα, Cretan ϝέκαθθα⁹ (later ἑκοῦσα¹⁰). The virtual identity of meaning favoured the creation of feminine πρόφρασσα by analogy to ϝέκασσα (proportion ϝεκών : ϝέκασσα = πρόφρων : πρόφρασσα).¹¹ The same explanation was independently offered by Meillet 1920, although his discussion was far less detailed and insightful than Wackernagel's. Since ϝεκών / (ϝ)έκασσα started its career as a participle, its use in predicative function must have been inherited,¹² which made it the perfect model for the creation of a specific feminine 'suffix' for a functionally feminine predicative πρόφρων.¹³

Πρόφρασσα seems to have always been confined to the limbo of Homeric linguistic eccentricities, otherwise not especially interesting to specialists of litera-

Wackernagel 1914, 42 = Wackernagel 1955, 1144.

See n. 6.

See LSJ 1940, s.v. 2.

See Buck 1955, par. 163.8a.

The ancient zero grade of the suffix was later replaced by a full grade in our Homeric text and in Classical Greek.

See also Risch 1974, 136 par. 50 a: "ἑκοῦσα zu ἑκών; älter ist schwundstufiges ἕκασσα < \*ϝεκn̥τi̯α (Kyrene, SEG IX 72, 87) welche Form das ihm nachgebildete πρόφρασσα zu πρόφρων (‐ν- Stamm!) auch für die epische Sprache voraussetzt". Ablaut problems of the root in the Greek forms are discussed by Harðarson 1993, 62.

It is noticeable that both ϝεκόντ- and its Vedic counterpart uśánt- are the participles of a verb meaning "want", "desire", but share the peculiarity of lacking what is most obvious in transitive verbs, namely the specification of an object, which however appears in Vedic finite forms of váṣṭi; in practice they are used as (predicative) adjectives: see Lowe 2014, 179–180. As such ϝεκών ~ ϝέκασσα was very close to Wackernagel's "Prädikats-Nominativ, quasi-adverbiell" πρόφρων. Something similar is found in Latin volens, which not surprisingly appears as a separate entry in Latin dictionaries.

ture. Yet, things stand differently. As we have seen, Wackernagel had already rightly emphasized that πρόφρασσα was always used within very narrow limits, only in the nominative and only in predicative function (πρόφρασσα παρέστης, inherited from his participial model ϝέκασσα). But there is more to it than that.

In theory πρόφρασσα could be used for any woman, but in fact is used for minor or major goddesses only: Calypso, Kirke, Athena, Leto/Latona. And πρόφρασσα is never used in a narrative. All the occurrences are found in dialogues that are in many ways 'typical' scenes, conversations between a male hero and a goddess, who is obviously more powerful than the hero, and his 'helper'. Odysseus and Diomedes invoke a πρόφρασσα Athena, Odysseus speaks to a πρόφρασσα Kirke, Calypso addresses Odysseus defining herself as πρόφρασσα.

The basic pattern – an exchange of words between a 'minor' male individual and a powerful goddess – is evident even in the less typical scene, that of Iliad XXI, where the minor god Hermes gives up fighting with Leto, the mother of Apollo and Artemis, and tells her that she can go to Olympus and boast πρόφρασσα before the other gods about his victory over Hermes.

As we have seen πρόφρασσα was created at a very ancient stage, when ϝέκασσα was still perceived as the femine participle of an athematic verb meaning "will"; at that chronological level other feminine participles must have kept their inherited zero grade of the suffix, like ἔασσα for the verb "to be" (Mycenaean a-pe-a-sa "absent women" and some 1st millennium dialects, later ἐοῦσα); ἴασσα for the verb "to go", later ἰοῦσα;¹⁴ the change did not entail any metrical problem. So at some point ϝέκασσα became (ϝ)εκοῦσα, along with ἐοῦσα and ἰοῦσα, since it was still perceived as the partiple of a verb, but πρόφρασσα never became \*προφροῦσα because was categorized as the strange and old feminine of an adjective and was never modified.

At the compositional level, πρόφρασσα is likely to have been devised for a specific dialogue situation {inferior male [mortal/god] – superior goddess}, and in this frame a special morphological change was enacted in order to emphasize the feminine nature of the powerful goddess. As a matter of fact the breach of morphological rules in the formation of personal feminine names is amply attested in Ancient Greek from Mycenaean times onwards. For instance the same two-ending adjective can be used regularly when applied to feminine common nouns (τὰν θηροφόνον λογχίδα in Lycophronid. PMG 2, 3) but can be provided with a recognizably feminine ending when used for women or feminine deities (Theogn. 1, 11 Ἄρτεμι Θηροφόνη); see also the two-ending θεόδοτος

Buck 1955, par. 163.8, 8a.

(Bacchyl. 8, 27 θεοδότους εὐχάς), but Θεοδότη pers. name; mycen. a-ti-ke-ne-ja = ᾿Aντιγένεια, Homeric ᾽Ιφιγένεια and ἠριγένεια, all feminine modifications of twoending adjectives in -γενής (Eur., Hippol. 26 εὐγενὴς δάμαρ); also note θεά and θέαινα for a feminine θεός.¹⁵ Homeric ἠριγένεια seems to have enjoyedamidway status between adjective (ἦμος δ' ἠριγένεια φάνη ῥοδοδάκτυλος Ἠώς, Il. 1, 477 and passim) and personal name proper (Od. 23, 347–348 αὐτίκ' ἀπ' Ὠκεανοῦ χρυσόθρονον Ἠριγένειαν / ὦρσεν, the epithet χρυσόθρονος being regularly used for such goddesses as Hera or Artemis).

In conclusion, from both the linguistic and the literary viewpoint πρόφρασσα is far more interesting than it looks at first sight. Although it was theoretically obvious that the epic language was old enough to have used at an early stage such feminine athematic participles with zero grade of the suffix as (ἀ)ϝέκασσα or ἔασσα, the advantage of πρόφρασσα is that it provides indirect proof of this state of affairs. Besides, at the literary level, the odds are that πρόφρασσα was created for a special 'dialogue frame', and one of high antiquity, as revealed by its being modelled on obsolete morphology. We get a glimpse of a small 'typical scene' much older than our Homer, and we are reminded once more that the study of Homeric language is not 'other' than the study of Homer as literature.¹⁶

### 2 ΕΙΠΑΤΕ ~ ΕΙΠΑ

In what follows – my second case study – we move from a rare and poetic adjective to a 'normal' verbal form, amply used in Homer and in Classical Greek, and still happily alive in Modern Greek: the aorist εἶπον / εἶπα "I said", modern Greek είπα. In Homer and elsewhere numerous forms of thematic (ἔ)ειπον < (ἔ) ϝειπον < \*(e)weik<sup>w</sup> -¹⁷ are attested, and those who read Martin West's edition of the two poems systematically find the expected thematic forms, e.g. Od. 9, 171 καὶ τότ' ἐγὼν ἀγορὴν θέμενος μετὰ πᾶσιν ἔειπον, Il. 1. 286 ναὶ δὴ ταῦτά γε πάντα γέρον κατὰ μοῖραν ἔειπες.

Yet things are more complicated than this. In Homer the first person ἔειπον is found in that line and elsewhere with no variant, but the second person is a

Wackernagel 1928, 25 ff., Morpurgo Davies 2000, 17, Risch 1974, par. 50c.

Remarkably enough πρόφρασσα is completely unknown outside Homer, including Hesiod and the Homeric Hymns; and in spite of the Hellenistic and late obsession with rare Homeric coinages, Apollonius Rhodius and the authors of epic poems of Imperial age gave πρόφρασσα a cold shoulder.

In its turn dissimilated from \*(e)wewk<sup>w</sup> -.

different story: ἔειπας is often well attested in various manuscripts and papyri in both the Iliad and the Odyssey,¹⁸ and at Hymn to Hermes 261 ἀπηνέα μῦθον ἔειπας is attested "in both branches of the tradition" with only two minor variants.¹⁹ The same goes for contracted εἶπες / εἶπας. At Il. 1, 106 μάντι κακῶν οὐ πώ ποτέ μοι τὸ κρήγυον εἶπες (printed by West 1998) the traditon is seriously split between εἶπες and εἶπας, the latter accepted by Aristarchus, in his turn reproached by a later grammarian who wanted a 'regular' thematic declension.²⁰

The problem occurs again, and in a more virulent manner, in the case of 2. pers. plur. imperatives: Od. 3, 427: εἴπατε δ' εἴσω / δμῳῇσιν κατὰ δώματ' ἀγακλυτὰ δαῖτα πένεσθαι, with εἴπατε in all ms. and accepted by Aristophanes of Byzantion;²¹ again a grammarian, Erotianus, dismissed the transmitted text of this line and required 'regular' εἴπετε by invoking the analogy of completely thematic δραμεῖν in the Homeric text.²² The same goes for Od. 21, 198: εἴπαθ' ὅπως ὑμέας κραδίη θυμός τε κελεύει, with εἴπαθ' in all mss. except one (εἴπεθ' ὅπως U²³); and at Hes. Theog. 108 εἴπατε δ' ὡς τὰ πρῶτα θεοὶ καὶ γαῖα γένοντο there are no variants. So in the 2nd persons we have far more -α- forms than Chantraine would let us believe ("quelques traces"²⁴). In all these cases Martin West systematically accepted the objections of the ancient grammarians and systematically printed the thematic forms (εἶπες, εἴπετε).

There is little doubt that the earliest thinkable declension type of (ἔ)ειπον was thematic, like e. g. ἔλιπον, ἔλιπες, ἔλιπε, but Homer is no handbook of early Greek morphology. Different verbs have different stories.

It is no accident that the -α- forms securely attested in the Homeric tradition are found only in the 2nd person. Today nobody seriously thinks that (ἔ)ειπας and (ἐ)είπατε belong to an old athematic aorist.²⁵

It is infinitely likelier that εἶπας was created by analogy to the 2nd and 3rd person endings of the indic. of sigmatic aorists, especially those where [s] had dis-

E. g. Il. 24, 379 ἔειπας in pap. 14 C R G (West 2000, ad loc.); Od. 8, 166 ἔειπας 397<sup>s</sup> GHPM (West 2017, ad loc.).

See Vergados 2013, 423, ad loc.

Schol. ad loc. τὸ δὲ εἶπες "εἶπας" ᾿Aρίσταρχος γράφει, κακῶς· εἰπών γὰρ ἀεὶ καὶ εἴποιμι λέγομεν. b (BC).

Schol. ad Od. 3. 427e (Pontani 2010, 149) εἴπατε δ' εἴσω] ᾿Aριστοφάνης τὸ εἴπατε δ᾽ εἴσω ἀντὶ τοῦ εἰσαγγείλατε; Slater 1986, 195.

Schol. ad loc. (Pontani 2010, 150 ad Od. 3 427–428) εἴπατε] Ἐρω(τιανὸς) "εἴπετε"· ὡς ἀπὸ τῆ (ς) "δραμὼν" μετοχῆς "δρα[μέτην"λέ]γει, οὕ(τως) ἀπὸ τῆς ["εἰπὼν" οὐκ "εἴπατε"] ἀ̣λλ' εἴπετε".

καὶ οὕ(τως) χρῆται Ὅμηρος· διὸ οὕ(τως) γραπτέον.

Monac. gr. 519 B, saec. XIII-XIV.

See Chantraine 1958, 386.

A once-widespread idea, see e.g. Risch 1974, par. 88 i.

appeared, e. g. 3rd ἔνειμε – 2 nd ἔνειμας / 3rd εἶπε – 2 nd εἶπας: for this reasons 2nd persons were the first to be affected.²⁶

This makes it easy to explain why in Homer we often find a variant (ἔ)ειπας in the 2nd person but only (ἔ)ειπον in the 1st one, only εἴπατε in the 2nd person plural but only εἴπομεν in the 1st one, since the 1st person was the last one to be reached by the process which had started in the 3rd one. In short, the morphology of this aorist as attested in Homer seems to belong to a sort of middle stage, with the -α- forms creeping into the 2nd persons but not enjoying the almost complete success later attested in Ionic and in other dialects.

As a matter of fact, many more -α- forms materialized in post-Homeric Greek, especially in East Ionic and the Doric dialects (to a lesser extent in Attic). Imperative plural εἴπατε easily lead to the creation of imperative singular εἶπον: Pind. Ol. 6, 92 εἶπον δὲ μεμνᾶσθαι Συρακοσσᾶν; the 'Orphic' lamella found at Hipponion (5th century BC – hexameter! –)²⁷ has εἶπον "say!" at line 10; Aristoph. eccl. 255 τούτῳ μὲν εἶπον [imper.!] εἰς κυνὸς πυγὴν ὁρᾶν. Such imperatives as λάβον and ἄνελον are attested for Syracusan Doric.²⁸

In Herodotus εἶπα, εἴπαντες, εἶπαι etc. are amply attested in the ms. tradition²⁹ and accepted by modern editors. The oldest Attic 1st person εἶπα is found in Solon, fr. 34, 6 W. οὐ χρεών· ἃ μὲν γὰρ εἶπα, σὺν θεοῖσιν ἤνυσα (trochaics; Ionic influence?), and Sicilian εἶπα was poetically arrayed as ἔειπα by Empedocl., fr. 17, 24 D.-K. ὡς γὰρ καὶ πρὶν ἔειπα πιφαύσκων πείρατα μύθων; later εἶπα and other -α- forms became the most widespread ones in Hellenistic Greek, and are amply attested in our mss. of the Septuaginta and the New Testament. Medieval and modern Greek inherited some -α- forms while others remained thematic: hence the modern 'mixed' inflection είπα, είπαμεν, είπατε, but είπες, είπε, and imperative πές.

Obviously enough, the pressure of the real, spoken language made itself felt fairly soon in Homeric recitations and consequently in written texts: the old fight between conservation and innovation that characterized the epic language at the time of its final definition in the 8th-7th century BC was resumed in a rather disorderly manner in Hellenistic times. In 1925 Octave Guéraud published one of the longest and most impressive Ptolemaic papyri of Homer, one going back to the 3 rd century BC and containing more than the half of the IX book of the Odyssey and about one hundred lines of the X. It is one of the so called 'wild' papyri, cer-

See Willi 2018, 79–80, with previous bibliography.

See Sacco 2001.

See Kassel/Austin 2001, 309; Glossar. Italiot. no. 49; cf. late and modern Greek ἔλαβα.

Powell 1938, 105–106, s.v. εἶπα.

tainly very wild if compared to a modern printed text of the Odyssey: some lines are added, others are absent, others are completely different from those in the medieval mss., hiatuses tend to be eliminated and morphology is often modernized.³⁰

And it is precisely in the Guéraud papyrus that we find two interesting 'modern' forms of our aorist:


Both imperative εἶπον and 1st person (ἔ)ειπα are never found elsewhere in Homer.

The Homeric text is so full of archaic, artificial or otherwise puzzling forms, that one is prone to forget that without the support of a good deal of 'normal' vocabulary and morphology the text itself would have been incomprehensible. When new forms were metrically compatible with the old ones, old morphology was most exposed to variation over time. In this case, too, we are before extremes. The 'modern' forms of the Guéraud papyrus are clearly unacceptable for a Homer of, say, the 7th century BC, but honestly, what of Martin West's 'original' εἴπετε, accepted against εἴπατε transmitted without variants? It runs the risk of being in fact pseudo-archaic – in practice a homage not to Homer but to the narrow analogical reasoning of ancient grammarians.

In conclusion, 1st person indicative εἶπα and imperative εἶπον attested in the Guéraud papyrus are nothing but the Hellenistic continuation of a process set in motion many centuries before. These forms cannot be accepted as Homeric but at the same time they prompt us to be more cautious in our evaluation of the -αforms attested in the Medieval manuscripts of Homer, especially imperative εἴπατε and 2nd person indicative εἶπας, which should not be backtranslated into εἶπες and εἴπετε, but should be accepted (as Aristarchus and Aristophanes of Byzantium did) as the initial instances of an evolution bound to be successful until the present day.

P. Sorb. inv. 2245. See Guéraud 1925; S. West 1967, 223–256.

### References


Wackernagel, J. (1914), "Akzentstudien II", in: Nachrichten von der Gesellschaft der Wissenschaften zu Göttingen, Philologisch-Historische Klasse, 20–51 = Wackernagel, J. (1955–1979), Kleine Schriften, Göttingen, 1122–1153.


### Chiara Bozzone Technologies of Orality: Formularity, Meter, and Kunstsprache in Homer

Abstract: This paper examines three formal features of Homer's poetry (formularity, meter, and Kunstsprache) in light of recent advances in the fields of linguistics and cognitive studies. Each feature is compared to analogous phenomena in the contemporary world in order to understand how it sustains the poet's performance and impacts the audience's experience. Specifically:

1. Homer's reliance on prefabricated linguistic sequences (i.e., formularity) is compared to similar findings in the field of corpus linguistics. Given the high rate of prefabricated sequences in spoken and written modern language corpora (over 50%, according to some studies), is it correct to say that Homer is really no more 'formulaic' than ordinary speech? New data on collocational measures in modern and ancient corpora is discussed.

2. Homer's meter is compared to the regularization of intonation units found in the speech of professional sportscasters, which correlates with decreased error rates (Kuiper 1996). This comparison suggests that meter should be regarded as an adaptive response to performance pressure, which enhances fluency by limiting the poet's choices.

3. Homer's composite Kunstsprache (marked by the coexistence of diachronic and diatopic variants) is compared to mixed performance dialects in contemporary popular music. While traditional thinking is that dialectal and artificial variants are maintained in the language of epic in order to satisfy the meter, comparison with pop music suggests that more complex sociolinguistic motivations are at work.

Keywords: Formularity; chunking; collocation; meter; intonation unit; Kunstsprache; dialect.

### 1 Introduction

Formularity, meter, and Kunstsprache (i.e., the reliance on a mixture of linguistic features of different diachronic and diatopic origins) are by far the most notable formal features of Homer's language. One cannot read a line of Homer without encountering the powerful (and interconnected) effect that these three forces exert on the poet's diction. The relationship between these three features has often been conceived as illustrated below, with meter effectively 'causing' the other two.

Fig. 1: The traditional view of the relationship between meter, formularity, and Kunstsprache.

This is Witte's famous position that Homer's language is "ein Gebilde des Epischen Verses",¹ a quote that summarizes a line of work on Homeric diction that became prominent in the first decades of the last century,² and that would ultimately culminate in Parry's own investigations of formularity (and, not to forget, Kunstsprache as well).

According to this line of thought, poets are 'forced' to resort to odd linguistic features (i.e., Kunstsprache) and prefabricated linguistic sequences (i.e., formularity) in order to conquer the formidable adversary that is the hexameter – a very demanding verse, made even more demanding (this is Parry's contribution) by the oral improvisatory nature of the poetry. I believe (and I am not certainly the first one to suggest this) that this view of meter as primary is really more of a product of the history of the field than a persuasive insight into the workings of the poet's art. Meter was always the most noticeable and clearest formal feature of Homer's poetry, and the careful study of meter led to the discovery of the other two features. But order of discovery is not always order of generation, and focusing too much on meter might cause us to neglect some other important factors at play.

In this paper (which reproduces an early and condensed version of selected topics now developed in Bozzone 2021), I attempt to turn this traditional conception on its head: I argue that meter, formularity, and Kunstsprache all emerge from the circumstances of oral performance, and that they all serve (in their own way) to support the poets and their audience in achieving their goals: they are adaptive responses to the cognitive challenges of oral performance. They are all, in other words, technologies of orality.While this argument has previously been (at least partially) made for formularity and meter, I believe this

Witte 1913, 2214.

See Witte 1913, Witte 1972, Meister 1921. These studies rested on an earlier scholarly tradition, see Janko 1992, 8–19.

perspective is new for the epic Kunstsprache. In what follows, we will first look at formularity, meter, and Kunstsprache separately, and we shall conclude by discussing some of the ways in which they interact.

### 2 Formularity

Formularity has held pride of place in discussion of Homer's language for much of the 20th century, as the field grappled with oral-formulaic theory and its impact on our experience of the poems. Formularity was so closely studied, in part, because it held the promise of setting the work of Homer (and other oral traditional poetry) apart from ordinary written poetry, thereby carving out a clear line of separation between the oral and post-oral world.³ Thus the famous quote by Albert Lord:

there are ways of determining whether a style is oral or not, and I believe that quantitative formula analysis is one of them, perhaps the most reliable. (Lord 1968, 16).

Some quantitative work on ancient epic seemed to confirm Lord's belief,⁴ though ultimately this line of research has been pushed to the margins in recent decades. The recent realization that formularity in a more general sense (i.e. the reliance on prefabricated linguistic sequences) is a core feature of human language tout court,⁵ and that rates of formularity in spoken and written corpora of modern languages are remarkably high,⁶ and perhaps as high as those in Homer, has complicated this position. On the one hand, we now have a much better understanding of formularity as a linguistic and cognitive phenomenon (more on this below). On the other hand, we are left to ask: is Homer then not that special? Is quantitative formula analysis the wrong tool to try to establish the orality of a text?⁷

In order to answer this question, we should first step back and ask: what is formularity's adaptive advantage? How exactly does it contribute to the poet's (or speaker's) performance? As reviewed in Bozzone 2016, the emergence of formu-

This line of thinking, which sees an unbridgeable divide between the oral and post-oral world, was championed for ancient Greek by the work of Havelock 1963.

See for instance Cantilena 1982 on the Homeric Hymns.

See Bozzone 2010, Bozzone 2014.

See Erman/Warren 2000; see further below.

Note that quantitative formula analysis was never the only proposed measure by which to assess the orality of a text. The most extensive discussion of this issue for the Greek tradition is to be found in Peabody 1975, who himself proposes several additional tests.

larity has to do with the limitations of our working memory, and the strategies we can employ to bypass them. It is really a specific case of a general psychological phenomenon known as chunking.⁸ Let us set up a brief psychological experiment: imagine we are given a few seconds to memorize a long number:

### 1304205630051968

Most people will succeed at retaining only the first few digits, famously reflecting the diminutive size of our working memory.⁹ Yet, some people will be able to remember the entire sequence. Rather than being gifted with a larger working memory, these people are most likely using a better strategy (a memory trick, so to say): they are chunking the information, i.e., grouping individual numbers into larger chunks. This is what such a chunking strategy could look like:¹⁰

### 1304 2056 3005 1968

Rather than 16 individual items, we now have only four to remember – and moreover, these now look like years, a data format which is more memorable than meaningless digits.

In this brief experiment, chunks are created instantly, through conscious analysis and effort, and retained for a brief period of time. A more effective, long-term strategy for chunking involves repetition. This is what we see at work in formularity in natural language (as well as any other complex behavior that we learn to master through repetition and training). As speakers encounter the same linguistic sequence over and over again (e.g., the fixed greeting formula "How are you?"),¹¹ they start to treat it as a single unit, and to retrieve it as such from their long-term memory instead of generating it anew, piecemeal, every time they need it. The same is true of musicians learning to play scales, or swimmers learning to combine the individual movements needed for a backstroke into a single, smooth routine. In all endeavors, chunking results in a dramatic lightening of the load of working memory, which makes any task at hand much easier to accomplish. Similarly, oral poets in training who repeatedly encounter the sequence πόδας + ὠκὺς + ᾿Aχιλλεύς will end up chunking it into the single unit πόδας ὠκὺς ᾿Aχιλλεύς: now, instead of managing three words, their

For a recent study on how chunking helps working memory, see Thalmann/Souza/Oberauer 2019.

The classic study is Miller 1956; see now also Cowan 2004. For a discussion of the structure of working memory (according to one recent model), see Baddeley/Eysenck/Anderson 2015, 44–58.

This should also remind us of how credit card numbers and phone numbers are usually written out, again for ease of processing.

For formulas in everyday language, see Wray 2002 and Kuiper 2009.

working memory can deal with a single (larger) 'word',¹² thereby freeing up space for longer stretches of language production.

When it comes to language production, whether everyday or oral-poetic, working memory is not the only factor that pushes us to rely on chunks (i.e., formulas, prefabricated linguistic sequences, etc.). Often enough, and certainly more often than we realize, chunks are not just a convenient way to say things: they are the only idiomatic (i.e., correct, natural, non-odd sounding) way to say things.¹³ When translating between languages, chunks are often revealed by the failure of word-by-word translation: so English "How are you?" cannot be translated in Italian as "Come sei?", and viceversa Italian "Come stai?" cannot be translated into English as "How do you stay?".¹⁴ Such errors in translation, which are really failures of idiomaticity, are often what set native speakers apart from even accomplished non-native speakers of a given language.

Similarly, πόδας ὠκὺς ᾿Aχιλλεύς is not just the most convenient way to realize the referent Achilles in a given position in the line: we have to assume, based on our poems, that it is also the only natural, traditional, appropriate, or idiomatic way to do so – the only one that achieves the correct traditional referentiality.¹⁵ Oral-traditional poets who compose in performance are then reliant on formulas for at least two reasons: in order to lighten the load of their working memory, and in order to sound like they 'fit' into the tradition in which they are performing and are conversant with its conventions. A correct formulaic usage signals to their audience that they are 'native speakers' of the oral epic idiom.

Let us now return to our original question. Formularity (i.e., chunking) is widespread in natural language. Some recent measures point to a little over 50% of formularity (prefabs or prefabricated expressions, i.e., repeated linguistic sequences that are conventionalized) in spoken and written corpora of contemporary English, with little difference between the two mediums.¹⁶ The reported formular density in Homer is around 60%.¹⁷ Does this mean that Homer is not

For the concept of formulas as "big words", see Foley 2002, 11–21.

This has been discussed by Pawley and Syder as the puzzle of native-like selection (Pawley/ Syder 1983, 193).

For this reason, Statistical Machine Translation (i.e., matching words, phrases and sentences in parallel-language corpora), had advantages over purely Rule-Based Machine Translation (Chitu 2007). Google Translate used Statistical Machine Translation until 2016, when it switched to Neural Machine Translation (Turovsky 2016). For Statistical Machine Translation, see Kohen 2009. For a history of Machine Translation, see Hutchins 2005.

See Foley 1991.

Erman/Warren 2000; see discussion in Bozzone 2010.

Note that the units of measure here are different. By formular density, Pavese/Boschetti 2003 mean the percentage of lines in Homer that are mostly formulaic. This is different from another

really that different from natural language in general? Can quantitative formular analysis really tell us whether a text was composed orally or not?

I have tried to answer this question in the past (essentially by explaining how meter, under different conditions and within different poetic traditions, can sometimes encourage reliance on formularity and sometimes discourage it),¹⁸ but it has not ceased to puzzle me. After all, to a modern reader, Homer feels significantly more repetitive than natural spoken or written language, and more repetitive than most ancient authors we know. Are we just too habituated to 'ordinary' repetitiveness to notice it? And why does Homer's repetitiveness stick out in comparison?

As it turns out, the size of chunks might have something to do with this. The large majority of prefabricated expressions (i.e. chunks) that we use in everyday life are relatively short. In English, this amounts to 2–3 words.We can gauge the size of these chunks by looking at the most frequent (2-, 3-, 4, and 5- word) collocations¹⁹ occurring in natural language corpora and their distribution. We use collocations as a shortcut here, since collocations can be detected automatically, unlike prefabs (in Erman and Warren's definition), which require native-speaker judgments. The table below shows the 10-most frequent 2-word, 3-word, 4-word and 5-word collocations in the Lancaster-Oslo-Bergen (LOB) corpus:


Tab. 1: The 10-most frequent 2-word, 3-word, 4-word and 5-word collocations in the LOB corpus.

measure given by Pavese and Boschetti, formulaic percentage, which is the percentage of lines in Homer that contain at least one formula (the value here is 84%). Both measures are different from the percentage of prefabricated expressions in a corpus used in Erman/Warren 2000, which is a word-based measure (i.e., what percentage of words in a corpus is chosen "freely" vs. the percentage that is part of prefabricated linguistic sequences).

Bozzone 2010, 33–34.

 Collocations are text-based units formed by 2+ words which tend to recur together in a given corpus. Hoey 2005, 5 defines them as follows: "<collocation> is a psychological association between words (rather than lemmas) up to four words apart and is evidenced by their occurrence together in corpora more often than is explicable in terms of random distribution."


Tab. : The -most frequent -word, -word, -word and -word collocations in the LOB corpus. (Continued)

As we can see, most of these are appreciably shorter than what we would consider to be a formula, as only few of them constitute complete semantic units. Nonetheless (and somewhat simplifying), their high frequency speaks to the likelihood of them being treated as units for the purposes of language processing. What interests us more specifically is the distribution of these collocations in our corpora with respect to their length: as a rule, shorter collocations are repeated much more frequently than longer ones. From our table above, the most-frequent 2-word collocation type, of the, is repeated 9009 times in the LOB corpus; the most frequent 5-word collocation type at the end of the is repeated only 28 times. Fig. 2 below illustrates this difference by plotting types and token counts of our collocations in the LOB corpus.

In Fig. 2, the token and types lines touch when there is only one example instance of repetition for a given sequence (i.e. when a given collocation is a singulum iteratum). For 2-word collocations, about 93% of all tokens are repeated more than twice, i.e. only 7% of 2-word collocations in our corpus are singula iterata. For 5-word collocations, only 16% of all tokens are repeated more than twice, i.e., 83% of 5-word collocations are singula iterata.

This tendency holds true (as far as I know) in all natural language corpora. It also holds true, for instance, for Herodotus [Fig. 3], which, in terms of corpus size and language provides perhaps our best comparandum for Homer. Here too, 2-word collocations are repeated very often, but longer collocations are repeated increasingly less. At 5-word collocations, 90% are singula iterata (so the tendency to not rely on long collocations is even stronger than in the LOB corpus).

When we look at Homer [Fig. 4], the picture is quite different. While the general tendency holds, the token line declines a lot less steeply in Homer than in the corpora we have seen so far. In fact, in Fig. 4 below, the lines do not come close to touching at all. At 5-word collocations, only 75% are singula iterata.

Fig. 2: Types and Token Counts of 2-, 3-, 4-, and 5-word collocations in the LOB corpus of written English.

Fig. 3: Types and Token Counts of 2-, 3-, 4-, and 5-word collocations in Herodotus.

In other words, Homer relies on longer collocations a lot more frequently than other corpora do (and this is probably why we perceive Homer as more repetitive).

Fig. 4: Types and Token Counts of 2-, 3-, 4-, and 5-word collocations in Homer.

In fact, the data presented here might barely scratch the surface on the extent of chunking in Homer. Some famous formulaic runs in our poems go well beyond the 5-word count: they take up multiple verses. We can think for instance of type scenes, such as arming, sacrifice, and banqueting, where the same formulaic material is repeated, with minor variations, over several lines.²⁰ Chunks of comparable size are not to be found (or exceedingly rare) in natural language corpora.

So why do we find these long chunks in Homer? As noted above, chunking is a strategy that helps us maximize the functioning of our working memory. From this perspective, small chunks (e. g., four groups of four integers, each group counting as a single unit for mnemonic purposes) are better than no chunks at all (e.g., sixteen separate integers counting as sixteen separate units). But large chunks are of course even better (e. g., two groups of eight integers, each counting as a single unit), in that they help us stretch the capacity of our working memory even further.

Classic psychological studies on the memory of expert vs. novice chess players show that expert players are able to handle information chunks that are

Structures of this kind have been recently discussed as multiforms by Frog forthcoming (building upon a proposal by L. Honko/A. Honko 1995, L. Honko/A. Honko 1998).

much larger than those that novices can handle.²¹ Thanks to these larger chunks, expert chess players can remember the position of pieces on a board much more accurately, even after short exposure, and they can plan ahead more easily during a match, because their working memory is not overwhelmed by the task at hand. Oral traditional poets (e. g., Homer) are then arguably akin to expert chess players: through long training, they have learned to handle word-chunks (formulas) as well as story-chunks (themes) of a larger and larger size. This helps them remember (i.e., recreate) songs they have heard before (even after a single exposure),²² as well as to plan ahead more effectively during their performance. Because their working memory is not overwhelmed, poets have more cognitive resources left over for ornamentation and characterization, for instance, as well as for all of the attention to detail that makes an oral tale great.

Based on the preliminary data above, can we then say that we have discovered, under the guise of collocational length, a better quantitative test for orality? Not quite. What collocational length measures is not orality per se, but just the level of optimization of working memory. Optimization usually results from extended training (in the form of repetition of complex tasks and the chunking that results), and extended training is usually a response to a particularly cognitively challenging task. On the ancient Greek landscape, oral composition in performance, to my mind, is the only scenario that would justify the need for such stark optimization strategies. But the proof, if anything, is indirect, and other scenarios cannot be excluded a priori.

### 3 Meter

While we are used to discussing meter as a (sometimes defining) feature of poetry,²³ contemporary linguistic theory tells us that metrical organization (i.e., metricality, whereby syllables are arranged into metrical feet) exists as part of the phonology of many (and in the view of some phonologists, perhaps all)

See Chase/Simon 1973.

This is the famous experiment conducted by Parry with Avdo Međedović, who could recreate (and significantly expand) a song he had heard only once before (in particular, this was The Song of Bećiragić Meho, which Međedović learned from Mumin Vlahovljak. The two versions are provided in full in Lord 2000, Appendix I).

Though Aristotle famously disagreed: εἴη γὰρ ἂν τὰ Ἡροδότου εἰς μέτρα τεθῆναι καὶ οὐδὲν ἧττον ἂν εἴη ἱστορία τις μετὰ μέτρου ἢ ἄνευ μέτρων (Arist., Poet. 9, 1451b1–4).

human languages.²⁴ For instance, metricality can play a role in the assignment of word accent, or in morphological derivation.

This is the case for ancient Greek, where phonologists have argued that syllables are grouped into moraic trochees, i.e., a sequence made up of either one heavy syllable (–) or two light syllables (⏑⏑).²⁵ Moraic trochees can be seen at work in the assignment of the Greek recessive accent,²⁶ where the accent (high tone) is assigned immediately to the left of the rightmost moraic trochee (i.e., the rightmost metrical foot). In this framework, the procedure for assigning the recessive accent is as follows:


For instance, in the recessively accented verbal form ἐδυνάμεθα, "we were able", from the right-edge of the word, we can form two moraic trochees, each consisting of two light syllables: (με.θα) and (δυ.να). The remaining syllable ἐ is left unfooted. The accent is then assigned immediately to the left of the rightmost foot (μεθα), thus landing on the second syllable of the foot (δυνα): ἐδυνάμεθα.²⁷

Moraic trochees also play a role also in some processes of nominal derivation.²⁸ For instance, Greek neuter nouns in -μα(τ)- have the option of picking either a full-grade root or a zero-grade root as their base. So, for the verb χέω "pour", we have the form χεῦμα "that which is poured or flows" (Iliad+), formed on the full grade of the root, next to χύμα "id." (Aristotle+), formed on the zero grade of the root. While the second pattern is, from the chronological point of view, an innovation, its emergence can be explained through considerations of foot parsing: specifically, χύμα can be neatly parsed into a moraic trochee

See generally Hayes 1995, Hammond 2011, Hyman 2006, Hyman 2009.

For an introduction to metrical feet and their designations, see Hayes 1995. For an in-depth discussion of ancient Greek foot structure (and the data used to establish it), see Devine/Stevens 1994, 102–117, and, from a theoretical perspective, Golston 1990. Devine/Stevens 1994 is also the most thorough treatment of ancient Greek prosody to date.

See Sauzet 1989, Golston 1990.

Golston 1990, Gunkel 2010, Sandell 2015, 166–170. However, alternative analyses of the assignment of the recessive accent without reference to foot structure, such as Itô/Mester 2017, have also been proposed.

Gunkel 2010, chapter 1, Gunkel 2011.

(χύ.μα), while χεῦμα results in one unparsed syllable: (χεῦ).μα. The first form is thus phonologically preferable to the second. Gunkel analyses several innovative -μα(τ)- formations that reflect this preference. In short, metrical considerations (in this case, the preference for words that are easy to foot in moraic trochees) can impact word-formation patterns.

But the role of metricality in Ancient Greek does not end here: Golston and Riad have argued that moraic trochees can be used to describe all of Greek lyric meters, and that Greek lyric meters reflect and exploit the natural phonological preferences of the language (in that phonological constraints that are violated in lyric meter are low-ranking in Greek phonology overall).²⁹ In other words, poetic meter in Greek appears to be a special configuration of phonological meter. Whether one agrees with this specific analysis of lyric meters or not, this overall principle is not a new idea.

Sapir (himself a linguist and a poet) had put this rather poetically in his Language: "[verse] is merely the language itself, running in its natural groves."³⁰ Within Greek philology, a similar thread had been picked up by Nagy, in describing the origin of the hexameter as it relates to formularity (and arguing that formularity came first):

at first […], traditional phraseology simply contains built-in rhythms. Later, the factor of tradition leads to the preference of phrases with some rhythms over phrases with other rhythms. Still later, the preferred rhythms have their own dynamics and become regulators of any incoming phraseology (Nagy 1974, 145).

The most recent incarnation of such a theory of the origin the Homeric hexameter is to be found in Bakker,³¹ who also argued that the regularization of meter was a process which continued (and was amplified by) the textualization of our poems:

See Golston/Riad 2005. Talking about constraints and ranking thereof reflects the framework of Optimality Theory (Prince/Smolensky 1993), within which Golston and Riad operate. Sapir 1921, chapter XI.

Of course, the origin of the Greek hexameter has been greatly debated. A rich recent discussion within the framework of generative metrics is Kiparsky 2018. For a brief review of earlier theories (proposed by classical philologists or Indo-Europeanists) see Hackstein 2010, 413– 414, with references. Note that the specific historical question as to how a particular verse form arose (whether, for instance, the components for the hexameter were inherited from PIE prehistory and how far back we can go in our reconstructions) is different from the more general question posed here (addressed by Nagy and Bakker above), which is how poetic meters in general emerge from speech, serving which goals, and under which conditions.

Meter is not something superimposed on language, a form that exists independently of it; meter emerges from language as part of the process by which special speech emerges from speech. (Bakker 1997, 146).

While drawing a line connecting ordinary speech and special speech can be intuitively satisfying, the question of how exactly this process is supposed to unfold has not, so far, been given a precise answer.What are the intermediate steps leading from the unplanned, clumsy, variable intonation units of natural speech³² to the tightly regulated structures of poetry? And why does this happen? How does this technology serve the poets and their audience? Diachronic studies of existing meters have shown how prosodic regulations can become stricter over time, and how syntactic preferences can turn into metrical preferences,³³ but no study so far has observed the crystallization of regular speech into metrical units as it unfolds.

A few hints as to how poetic meter (or a rough version thereof) might arise from speech come from a markedly non-poetic context: the speech of sportscasters, i.e., professionals (one might say oral performers) trained in the art of describing events quickly as they unfold. These and other smooth talkers have been extensively studied by Koenraad Kuiper in his eponymous 1996 book, which brought to light the extent to which such extraordinary feats of fluency are sustained by an increased reliance on formulaic speech.

But formularity is not all that happens: Kuiper observed that there are clear audible correlates to formulaic behaviors while performing a cognitively-demanding task. Specifically, various prosodic aspects of the speech (which we shall see below) become regularized. These regularizing effects appear more clearly and reliably the harder the cognitive task at hand – and therefore the higher the need to relieve the speaker's working memory of unnecessary choices. When comparing play-by-play commentary of a slow-moving sport, like cricket, to play-by-play commentary of a fast-moving sport, like horse-racing, the differences are stark.

Among cricket commentators (a famously slow-moving sport), Kuiper was able to detect formulaic behaviors in the form of discourse structure rules as well as the reliance on some formulas, but while the speech was fluent, it still contained pauses, had an uneven articulation rate, and followed a variety of natural intonational patterns – i.e., it sounded like normal speech. Horse racing commentators

For an introduction to the properties of unplanned, natural speech (and that they arguably reflect the workings of human consciousness), see Chafe 1994. Chafe's work served as the basis for Bakker's study of speech-like properties in Homer (Bakker 1997).

Studies in this direction include Fränkel 1926, Fränkel 1955 and Cantilena 1995.

present a very different picture:³⁴ as soon the race calling portion of their comment starts, they follow strict discourse structure rules and rely almost exclusively on formulas; they display abnormal levels of fluency, with no hesitations and shorter pauses, and an even articulation rate. Moreover, their intonation becomes droned or chanted (i.e., articulated in a monotone, as opposed to the varied intonational patterns of natural speech). Their speech, in other words, is a lot more 'metrical' (i.e., prosodically regulated) than that of cricket commentators. If we imagine having a dial that can increase or decrease cognitive load during speech, then prosodic regulation ('meter') is something that seems to naturally start to appear when the dial is all the way up (and perhaps when other circumstances, like extensive training in the task at hand, are in place).

But how, one might ask, is the speech of horserace commentators akin to poetic meter? We can observe at least four types of similarities.

1. Length: the even articulation rate, as well as the regularization of the length of intonation units,³⁵ can be seen as analogous to the creation of metrical cola or lines of regular length. In English, which is a stress-timed language,³⁶ this regularization will likely result in restrictions on the number of main word-stresses within an intonation unit. This can be seen in Kuiper's data, where we observe multi-line stretches showing the same (or almost the same) number of main word-stresses per unit:³⁷

Kuiper 1996, 10–21.

While this effect is not remarked upon by Kuiper, it is observable in the transcript he provides (Kuiper 1996, 11–15). While not all intonation units in the race-calling portion of the commentary are of identical duration, there are several stretches in which consecutive intonation units form 'blocks' of relatively homogenous length. Note that, since Kuiper does not provide time measures for his transcript, the relative length of intonation units here has to be inferred based on general facts of English prosody (and the crucial piece of information, provided by Kuiper, that articulation rate is even).

In stress-timed languages, stressed syllables (which may in themselves be of different length) are produced at approximately regular intervals, and unstressed syllables adapt to fit in between those intervals (see Nespor/Shukla/Mehler 2011 on the typology on prosodic timing). In these languages, counting main word stresses is then a good way to estimate the length of an utterance (for a more precise measurement, the length of each stressed syllable should also be taken into account).

The text is from Kuiper 1996, 12. I have added solid underlining to signal words carrying a main stress; broken underlining indicates that the stress status of the word is uncertain. The number in parentheses represents the number of main stresses in each intonation unit. Since a recording of Kuiper's data is not available, I relied on the judgement of a native speaker of North American English to reconstruct the likely placement of stresses in each line.

and El-Red the leader by two lengths (4) from Speedy Cheval and Race Ruler (4 or 5) On the rails to Twilight Time (4) Belvedere is going up three wide (4) Florlis Fella in the center. (3) False Image attacking around three wide. (5) In between them is Little River (4) Lone Eagle is up three wide (4)

A cursory examination of Kuiper's transcript suggests that several areas have undergone this type of regularization, and that this regularization precisely does not happen in the non-race-calling portions of the commentary (i.e., the introduction, before the horses go off, and the end, after the winner is announced). 2. Rhythm: a closer look at Kuiper's data reveals that not just word stresses are being regularized, but that several lines display recurring rhythmic patterns. Perhaps the most frequent of these patterns is a line consisting of four prosodic words, each with a main stress, and each corresponding to at least one bimoraic foot (i.e., either a heavy syllable or two lights), and in which the odd-numbered stresses appear to be the stronger ones, yielding a trochaic rhythm:³⁸

Note that the lines 'breaking' the pattern in the middle seem to constitute a kind of enjambement – in that taking the first word of the second line ("False")

In the text below, above each line, a "/" is used to mark strong word main-word stresses (experienced as beats) and a "\" to mark weak main-word stresses (experienced as offbeats). Lines conforming to the trochaic pattern just described are marked with a checkmark.

and attaching it to the end of the first line would yield a regular pattern for both lines. And while the pattern just described does not hold true for all the lines of the commentary, a cursory examination of the transcript indicates that more regularities of this type could be found in the data.

What is interesting here for our purposes, is that by combining prosodic regulations affecting length and rhythm, we obtain something that resembles an English meter, which typically specifies (a) how many stressed syllables (ictic syllables) there should be within a line and (b) which type of metrical feet (e. g. iambs vs. trochees) should be employed to build the line (e.g. an iambic pentameter builds lines out of five iambic feet, allowing for five ictic syllables).³⁹ Our pattern here, while certainly not as regulated (in that it does not specify rules for unstressed words or syllables), nonetheless specifies (a) how many stressed words there should be within a line and (b) that these stressed words should be arranged in a trochaic rhythm (effectively forming two trochees).

3. Intonation: the adoption of a droned or chanted intonation, with fewer variation in pitch in comparison to natural speech, is something we know from our modern experience of poetry (though we normally do not consider it to be part of the meter itself). Fabb reports:

In the recorded performances of English poetry that I have examined, the intonation contour is often more limited in pitch range and variation, a type of what Ladd 1978 calls stylized intonation […] Byers (1979, 371) similarly notes that poetry is generally spoken at a lower pitch and with a narrower pitch range than conversation or prose, and with relatively little variation from line to line. (Fabb 2015, 44–45)

As far as sportscasters are concerned, this type of intonation has been described in a privative way (i.e., in that it shows less range than natural speech).⁴⁰ But there might be some further regularization happening, which Kuiper describes only cursorily. In natural speech, each phonological phrase is assigned a specific intonational pattern (or tune), which usually conveys a specific meaning.⁴¹ It ap-

For a generative analysis of some metrical structures of English, see Hayes 1989, 221–244. Kuiper also reports that, in race-calling, there is a tendency for the absolute pitch to increase stepwise as horses near the end of the race, likely as a way to convey urgency and emphasis (Kuiper 1996, 19): "the intonational note usually rises in semitones to a high point at the finishing post and then gradually comes down as the commentator moves thought the last cycle." In the terminology of Hayes 2009, a tune is the specific pitch pattern with which words in a given phrase or utterance are said; in intonation languages like English, tunes "convey abstract meanings of their own, usually related to the information structure of the utterance" (Hayes 2009, 292). Examples of English tunes are, for instance, the Declarative tune (used in answering an information-seeking question like "Where do you live?"), and the Emphatic question tune (used for asking a question like "Where do you live?"). Each tune is composed by a sequence

pears that some of Kuiper's commentators have developed specific tunes that they use to mark particular moments in their performance (i.e. with meanings that are specialized for the task at hand):

Commentators also have intonational ornaments of various kinds. For example, Reon Murtha uses a fall tune, that is, a slight drop in pitch on the last stressed syllable of formulae which are in turn at the end of particular sections of the discourse, for instance, the completion of a cycle. (Kuiper 1996, 21)

Going back to poetry, it might be the case that poets reading their poetry out loud using the same intonational contour (i.e. tune) for each line might simply have gone further in this process of prosodic regularization (i.e. they are relying on a stereotyped 'poetry' tune).

One should note that the specific prosodic features of a given language will likely determine what type of prosodic regularization takes place. The regularization of word stresses makes sense for a language like English, which is a stresstimed, stress-accent language. Ancient Greek is a mora-timed language, in which the primary acoustic correlate of accent (or stress) is high pitch:⁴² the prediction here is that the regularization would affect in primis moras (which it does), though pitch (i.e., word accent) might also be involved. As far as the Homeric hexameter is concerned, Danek and Hagel report (summarizing Hagel 1994):

There is statistical proof that end-accented words are avoided at caesura, but favored at metrical bridges. The melody of the 'typical' hexameter fell at the middle caesura as well as at the end of the verse. The 'typical' melodic contour consisted of a double, sometimes triple rise and fall. (Danek/Hagel 2002)

Following this line of reasoning, the Homeric hexameter, too, had a stylized intonation contour or tune (or perhaps a few such tunes). While intonational features are not usually part of how we describe Greek meters, the case can be made that standardized tunes also belonged to the poet's metrical grammar.

of tones: for instance, the Declarative tune is composed by the sequence MID – HIGH\* – LOW, while the Emphatic question tune goes MID – LOW\* – HIGH. Each tone within a tune has to be anchored to specific syllables within the phonological phrase; tones marked with a star (starred tones) need to be anchored to the main word-stress of a phrase or utterance.

For this reason, ancient Greek is often described as a "pitch accent" language (e. g., Probert 2003, 3, Gunkel 2014, 7), but the appropriateness of this designation in general (see Hyman 2009) and in the case of ancient Greek in particular (see Sandell 2019), has been questioned. For ancient Greek prosody, see again Devine/Stevens 1994.

4. Pauses: finally, the elimination or shortening of natural pauses (and elimination of hesitations) can also be seen as similar to poetry, where normal units of speech production (e. g. syntactic or intonational phrases) become subordinate to the non-linguistic unit that is the poetic line.⁴³ In other words, the pace and rhythm of delivery is no longer dictated by fluctuating processing concerns, but a pre-established pattern takes over.

Of course, these phenomena of prosodic regularization are only incipient in Kuiper's data. Nevertheless, we can observe their emergence, coupled with an increased reliance on formularity, as an adaptive response to the high cognitive strain. In situations where the working memory could become overwhelmed, formularity and meter-like qualities help speakers sustain their language production by artificially narrowing down their choices. Once these two 'emergency' technologies are put into place (and both naturally require training in order to be mastered), speakers can achieve an abnormal level of fluency. At least for an oral poet, then, meter is not a straitjacket to fight against: it is a powerful technology (one might conceive of it as extra gear), which streamlines language production and maximizes cognitive resources.⁴⁴

### 4 Kunstsprache

We are all familiar with the phenomenon of ancient Greek Kunstsprachen,⁴⁵ i.e., as summarized by Morpurgo-Davies (numbering and emphasis mine):⁴⁶

(2) In a number of instances the choice of dialect is independent of the origin of the author; Pindar was from Thebes but did not write in Boeotian. Hesiod was also from Boeotia but composed in epic language, i.e. in a composite form of Ionic. We have Ionic prose, Doric prose and Attic prose, but, for instance, the Hippocratic corpus is written in Ionic, though Hippocrates himself was from Cos,aDoric place.

(3) The literary dialects are no perfect match for the epigraphical dialects: the Doric of Attic choruses is far less Doric than that of, e. g., Peloponnesian inscriptions.

<sup>(1)</sup> an interesting pattern of dialect or language switching tied to the view that some linguistic forms are more suitable than others for certain linguistic genres. Epic verse is written in some form of Ionic. Attic tragedy is written in Attic except for the choruses which are in a modified form of Doric. Lyric poetry can be in Aeolic; literary prose cannot.

See Fabb 2015, chapter 3.

The view of meter as something directly related to cognition (and to the limitations of working memory specifically) has been recently championed by Fabb 2015, chapter 7. For more in this direction, see Bozzone 2021.

See Cassio 2016 for individual treatments.

See Morpurgo-Davies 2002, 157.

Of all Kunstsprachen, the language of Homer is perhaps the most complex and most debated.⁴⁷ Complex, since most (though not all) scholars recognize at least four dialectal components in our poems, namely:


Debated, because there are important open questions about each of these layers or components. A necessarily insufficient list will include the following: whether the Mycenaean component is the oldest, or yet older elements can be detected, whether the Aeolic component was there at all, whether the Ionic component represents East or West Ionic, and finally, whether the Attic component is just a matter of written tradition, or whether it also played a role in the final stretches of oral transmission and performance.

A few examples will suffice to illustrate the thorough mixing of diachronic and diatopic variants in Homer. In the following lines, two different variants of the same features occur within the same verse:⁴⁸


In 1., the poet combines two diachronic variants: the (conservative) uncontracted allomorph of the 2nd sg. pres. imp. ending –εε with its (later) contracted counterpart –ει. In 2., we have two diatopic variants: the Aeolic and Ionic forms respectively of the 1st person plural personal pronoun. 3. represents a possible case of combined diachronic and diatopic variation: while the variant –οιο for the genitive singular of the thematic declension is etymologically older than the con-

Forarecent overview of Homeric diction and its linguistic features, see Passa 2016. Phase theory, i.e., the idea that the epic diction acquired its heterogenous features (linguistic, but in some versions also mythological) through the movement of the tradition from one geographical center to another (likely over several centuries) originates with Ritschl 1838. Many reformulations followed, importantly Parry 1932, who adapted phase theory to an oral conception of Homer. Again, see Passa 2016 for more references.

See Hackstein 2010, 407–408.

tracted –ου common in Attic, it also taken as a distinctly continental Aeolic feature by some.⁴⁹

All of the questions above concerning the different components of the epic diction are intertwined with the Homerica quaestio itself: in Homer's odd dialect, scholars have seen an opportunity to garner information about (depending on one's predilections) either the poet himself or the poetic tradition to which he belongs. In this direction, the coexistence of these different components in Homer's text has been explained by (a) the poet's biography⁵⁰ (b) the history of the tradition⁵¹ or (c) a combination of both.⁵²

What is particularly interesting for us is that, in all modern accounts of Homer's Kunstsprache (and its origins), meter has played a key role. While the archaic, dialectal, or artificial features themselves might betray something about the history of the poet or the poetic tradition, their presence in the poems is purely instrumental: they help satisfy the stringent metrical requirements of the hexameter. The preference of the poets, is assumed, would be to use their native idiom as much as possible when composing (and thus precisely not to rely on an artificial, mixed idiom). So for instance Parry, discussing what happens when poets "borrow" from a linguistically foreign poetic tradition:

they [the poets] make the foreign poetry fit their spoken language in so far as they can do so without any great loss. The new poems thus take on straightaway a local color (Parry 1971, 337–338)

It is only when using the local idioms is impossible that poets default to the foreign (or archaic, or artificial) option. So for instance Nagy's current theory of the Aeolic default, which assumes (as most scholars do) that the poets responsible for our epic tradition were speakers of Ionic, who inherited (or were deeply influ-

This is phase theory, as introduced above.

See Haug 2002, 106, 146, 160.

This vein of explanation begin already in antiquity with the many Lives of Homer. An extensive review of the ancient evidence on Homer's life and whereabouts (including early mentions of Homer's name, direct and indirect literary citations, and the biographic tradition) is available in Latacz 2011. A recent attempt to contextualize the contradictory ancient accounts has been made by Nagy 2009/2010 (summarized in Nagy 2011), who sees them as reflecting the gradual Athenian appropriation of Homer, and preserving the memory of an even earlier power struggle between the Ionian and Aeolian dodecapoleis.

This, for instance, is Wachter 2007, who speaks of an Ionic Homer, native of Smyrna, who single-handedly imports the Aeolic tradition of epic, translating it into his native dialect, while leaving in some Aeolic features that would have been easily understood by the audience in his native city (which was exposed to Aeolic influences).

enced by) an Aeolic epic tradition,⁵³ and who only defaulted to Aeolic forms in their compositions when no other alternative was available:

In Homeric diction, if an Ionic form is available to fit into a metrical position that is already occupied by an Aeolic form, then the Aeolic form is replaced by the corresponding Ionic form, but the Aeolic form is preserved wherever no metrically equivalent Ionic form is available. Homeric diction defaults to Aeolic forms when it has no metrically equivalent forms in Ionic (Nagy 2011, 144).

Nevertheless, these complex questions may be, in a sense, premature. While many scholars have focused on dissecting ancient Greek Kunstsprachen, very few have asked whether there are any contemporary parallels to these odd linguistic objects.⁵⁴ Are these artificial languages exclusive creations of the ancient Greek literary landscape? Are they common to oral traditions in general? Are they a purely literary phenomenon? If modern Kunstsprachen could be found, we could study their historical development and their sociolinguistic context with much a much greater depth and breadth of detail than the ancient record allows. We could get a better idea of how these Kunstsprachen relate to the spoken dialects of the poets who use them, and we could see whether meter (or other similar formal requirements) is the only factor at play in their adoption. In the end, this data would help us return to the Homerica quaestio with much sharper tools. But where should we go to look for modern Kunstsprachen? As it turns out, really not so far. Modern Kunstsprachen are literally all around us, just not in places we previously thought to look.

The idea of an Aeolic phase has come under close scrutiny in recent years. In the first place, several scholars have criticized the idea of Aeolic as a unitary dialect group, both on linguistic and archeological grounds (see Parker 2008 and Rose 2008). Even before then, scholars have argued that what counts as Aeolic in Homer should simply be interpreted either as 'archaic' (the classic treatment in this sense is Strunk 1957), or as trivial parallel innovations in the language of epic. The most recent attempt to advocate for the legitimacy of the Aeolian traits in Homer is Nagy 2011, which contains a discussion of all of the features just mentioned. For an opposite view, see Miller 2014, chapter 25.

A contribution in this direction is Hackstein 2002, chapter 3, who investigated parallels to Homer's mixed dialect in literary prose, particularly with reference to the coexistence (and intermixing) of diachronic variants. As it turns out, this type of variability is well-documented in the history of many Indo-European languages, from standard German (whether in Martin Luther or in Standard Modern High German), to 18th-century Russian, to Tocharian.

### 4a Adele and the Brit Pop Pronunciation

In February 2012, when the British singer Adele won six Grammy Awards, The Telegraph published an article titled "Grammy Awards: Americans baffled by Adele's accent" (The Telegraph, 2–13–2012). The reason for this bafflement was clear: Adele was born and raised in Tottenham, England, and is a native speaker of Cockney English, as one can easily notice in interviews, as well in as award acceptance speeches. However, notable Cockney features are completely absent from her singing:


Moreover, Adele shows the same type of dialect mixing that we have just observed for Homer in the three examples above:

But it don't matter, it clearly doesn't tear you apart anymore (Hello, 25, 2015).

Here, in the same line, we see a non-standard English morphological variant, it don't, immediately followed by the standard variant it […] doesn't. What is peculiar here is that the non-standard variant does not seem to come from Adele's own spoken dialect, but rather (as we shall see), from much further afield. Why does this happen?

As it turns out, Adele is not alone in this type of linguistic shift. She is actually part of a long tradition of British singers which seem to almost completely lose their native accent during performance. This phenomenon was first studied in a classic contribution by Trudgill,⁵⁷ who coined the term "Brit Pop pronunciation" to capture this behavior in popular music the 1960s and 1970s. Some of the most notable features of this pronunciation are as follows:

Here and below, I conventionally use the symbol [r] for the English post-alveolar approximant. The standard IPA symbol for this sound would be [ɹ] (in IPA, [r] represents an alveolar trill instead).

See Mott 2012.

See Trudgill 1983.


In fact, these changes are not limited to pronunciation (phonology) alone, but they spread to grammatical and lexical choices. For instance, Brit-Pop singers tend to avoid lexical items that are stereotypically British, like mate and fancy, and tend to opt for the more neutral friend and like. Of these and the features above, Trudgill observed:

No single British variety has all these features, and the vast majority of singers who use these forms do not do so when speaking. There can be no doubt that the singers are modifying their linguistic behavior for the purposes of singing (Trudgill 1983, 252).

But why would this singers feel the need to carry out such modification? It seems like the main goal of the Brit-Pop pronunciation was for British singers to sound more 'American' in their performances. Since most of the popular music up to the 1960s and 1970s was produced in the US, British singers who wished to insert themselves in that tradition were modifying their dialect to conform to what they perceived was the standard dialect of the genre. They were trying, in other words, to 'pass themselves off' as legitimate members of that performance tradition.

To an extent, their linguistic modification was successful: while the features listed above are largely extraneous to British dialects, they are indeed present in different varieties of the English spoken in North America. Adele's American fans would have readily accepted her as a speaker of US English, and were surprised to find out that she was not.

What is fascinating, however, is that no single, clearly identifiable, American dialect actually encompasses all of the features above. While (a) is a systematic trait in North America, (d) is typical of dialects spoken in the South US, which however normally (and at least still in the 1960s and 1070s) did not preserve coda /r/ as in (c). Similarly, (b) represents a common trait in North America, but the vowel /æ/ is subject to diphthongization to [eə] prenasally in many areas of the US and Canada (e. g., somewhat simplifying, New Jersey, Florida, Canada, New England). In southern dialects, the same sequence is often realized with the triphthong [æjə]. So, while a speaker of Boston English would say [deə̃ns] (while dropping coda /r/, much like speakers of British English or from the South of the US), a speaker from the South could say [dæjə̃ns].⁵⁸ British singers, in other words, were adopting an odd, artificial mix of dialectal features for the purposes of singing.

If this sounds familiar, it is because it is indeed very similar to what we have seen at the beginning of section 4 above concerning ancient Greek Kunstsprachen:


We might ask why this phenomenon is apparently so common, and how we might explain it. Following Trudgill, we can seek some answers in the realm of sociolinguistics. First, this is not simply linguistic accommodation,⁵⁹ whereby the speaker (consciously or unconsciously) tries to match the linguistic features of his audience. After all, British singers use these features regardless of whether they are performing for a domestic or international audience. As Trudgill suggests, a more appropriate model is Le Page's theory of linguistic behavior, originally developed to describe individual linguistic choices in multilingual communities and the creation of Creoles:⁶⁰

The individual creates for himself [/herself] the patterns of his [/her] linguistic behavior so as to resemble those of the groups with which from time to time he [/she] wishes to be identified or so as to be unlike those from whom he [/she] wishes to be distinguished. (Le Page/ Tabouret-Keller 1985, 181).

In other words, the singers are not trying to adopt the language of their audience, rather, they are trying to adopt the language typical of the performance tradition in which they want to fit. For Adele, this would be the performance dialect of rhythm and blues, which uses features of the English spoken in the South of the US – this, for instance, is where the non-standard agreement in the verse

Forasystematic treatment of the regional features of North American English, see Labov/ Ash/Boberg 2006.

See Giles/Smith 1979.

See Le Page 1978, Le Page/Tabouret-Keller 1985.

above comes from. This linguistic shift has then to do with self-presentation,⁶¹ and it can be impacted as well by the topic of discourse and other situational considerations.⁶²

What is even more fascinating, is that in some cases the motivation to identify with a specific group or tradition seems to override the need of being understood: this is true of genres, like opera, where singers perform in a language which is often not controlled by most of their audience, as well as in the extreme cases of songs performed entirely in Gibberish.⁶³

But if the goal of these singers is to adopt the dialect of another existing group of performers within their genre, why does their shift result in a mixed "artificial" dialect, and not simply in consistent adherence to the linguistic rules of the model group? When it comes to adult speakers trying to imitate another linguistic variety (just like adult speakers trying to learn a second language), the results are never perfect. As Trudgill remarked, "the end-product of this language modification […] is by no means entirely successful."⁶⁴ Many variables are at play here, but in practice many factors (riders) limit the capacity of speakers to perfectly approximate another linguistic variety:⁶⁵


For an introduction to the concept of style (i.e., how speakers employ language variation to express social meanings) see Coupland 2007.

For instance, Simpson 1999, 353 analyzes the unusual density of specifically New-York English features in the song Money for Nothing by the British band Dire Straits, who otherwise do not use conspicuously New-York English features in their repertoire. Here the topic of the song itself is the reason for the linguistic shift: the singer is voicing a character heard speaking ataNew York bar, and he is thus picking up some of his most notable dialectal features (in this case, they happen to be largely lexical). In Greek poetry (especially monodic lyric), we can think of several situations where the poet is effectively lending his voice to a character, and thus potentially facing a similar motivation for a linguistic shift.

An example is the Gibberish English song Prisencolinensinainciusol by the Italian singer Adriano Celentano.

See Trudgill 1983, 254.

See La Page 1978.

These riders give rise to the linguistic variation (or dialect mixing) discussed above. As for 1. and 2., British singers only had partial access to their model group (defined vaguely as "American singers"), and ended up not identifying it clearly: as a result, they selected features from different dialectal groups within North America.⁶⁶ Moreover, singers are not always consistent in their linguistic choices, either because they are unable to always make the necessary changes in their behavior 3., or because they do not always feel equally motivated to make those changes 4.

What seems to play no role, at least in contemporary music, in driving the singers' choice of dialect, is what has been hailed as Homer's main motivation for keeping extraneous dialectal material in his performance language: namely, meter. These singers decide to adopt 'foreign' linguistic features in order to convey a sociolinguistic message, not out of metrical necessity. And while some contemporary genres have rhyme, and singers can be shown to use dialectal variants opportunistically when it comes to satisfying the rhyme scheme,⁶⁷ purely formal considerations of this sort seem to be neither necessary nor sufficient to explain the existence of these mixed performance languages. It seems that Tolkien was then close to the mark when he observed, about the Kunstsprache of Beowulf:

This sort of thing – the building up of a poetic language out of words and forms archaic and dialectal or used in special senses – may be regretted or disliked. There is nonetheless a case for it: the development of a form of language familiar in meaning and yet free from trivial associations, and filled with the memory of good and evil, is an achievement, and its possessors are richer than those who have no such tradition. (Tolkien 2006, 55)

The Brit Pop pronunciation is, of course, just one case of a contemporary Kunstsprache. Many more are there to be studied, like the dialect of Hip-Hop,⁶⁸ or the so-called Punk Rock Pronunciation, which presents one of the best contemporary parallels to phase theory.⁶⁹ I am convinced that a more exten-

In this direction, most British singers choose to restore coda /r/ in their singing, since the trait sounds generally "American" to them, even though coda /r/ is actually dropped in many sociolects of the South of the US (just like it is in Britain), and specifically in those dialects associated with rhythm and blues and country music.

For instance, the British band the Arctic Monkeys, usually known for the abundance of local phonological features (Sheffield English) in their performance dialect, appear to default to the received/international pronunciation of the sequence [aʊ], instead of local [aː], for the words around and ground, which carry the rhyme, in their song Mardy Bum. See Beal 2009, 11–12. Hip-Hop, at minimum, combines features of African American Vernacular English and Standard American English, though it is particularly receptive to further additions. For an introduction to the sociolinguistics of Hip-Hop, see Cutler 2007.

See Bozzone 2021.

sive examination of such contemporary Kunstsprachen holds the key to some of our most complicated questions about Homer's language.

Finally, to return to the main theme of this paper, we should ask: what does this technology (i.e. a linguistically mixed Kunstsprache) do? How does it emerge from the conditions of oral performance, and how does it support the performers and their audiences? Here, the main motivation appears to be sociolinguistic (one might say aesthetic) rather than cognitive in nature. Kunstsprachen do not help sustain the effort of working memory, but they do serve as a link between the performer, the tradition (i.e., previous performers) and the audience. They provide an efficient way for performers to signal their artistic roots and affiliations, and they prime the audience to expect a given type of performance (they help, in other words, to signal genre). They have, in Foley's terms, traditional referentiality, much like formularity and (arguably) meter also do.

### 5 The Relationship Between Meter, Formularity, and Kunstsprache Revisited

This paper has argued that meter, formularity, and Kunstsprache should be viewed as distinct adaptive technologies that emerge in response to the challenges and conditions of oral poetic performance. Of course, this does not mean that these three technologies do not interact with each other in fundamental ways. As is well known, formulaic sequences are (mostly) made to fit metrical cola, and odd linguistic forms (part of the Kunstsprache) seem more likely to be employed if they are metrically advantageous. Conversely, following formulaic patterns might lead the poet to commit some metrical imperfections (especially when combining formulas), and some notable features of Kunstsprache might be embedded in old formulas (more so than in new expressions). The goal of this paper is not to do justice to these complex interactions, and it is certainly not to deny their existence. The point made in this paper is that to simply claim that formularity or Kunstsprache only exist to satisfy the meter, or vice versa⁷⁰ is to miss the point of what these technologies actually afford the poets and their audiences.

Meter, formularity, and Kunstsprache, as technologies of orality, serve (at least in part), to make a linguistic text more memorable. The adjective "memorable" here is intended in multiple senses:

I.e., any approach that claims the primacy of one formal feature over another, or that tries to reduce one feature to the other.

1. meter and formularity appear to lighten the load of working memory, as they segment a text into chunks that can be easily processed. This advantage holds primarily for the poets, but it is arguably important for the audience as well, making a performance easier to take in, and thus more pleasant.

2. Meter, formularity, and Kunstsprache also seem to help recall from longterm memory. Note that recall here can work in two ways: as investigated by Rubin,⁷¹ additional formal constraints added onto a linguistic text can improve the accuracy with which it is remembered (which is always a creative process). Thus, if a poet is trying to create or recreate a tale out of traditional components, these technologies will help; they will also make the performance more easy to remember for the audience; but recall can also mean the power to evoke other experiences that are stored in the audience's long-term memory. In this sense, these three technologies have the power to make oral texts evocative of previous performances: they can index both individual performances and the tradition as a whole, thus contributing to a text's referentiality – to its richness of meaning and to its aesthetic effect. That these poetic technologies are strictly dependent on the workings and limitations of human memory should not surprise us: the Muses are daughters of Mnemosyne, after all.

### References

Baddeley, A./Eysenck, M.W./Anderson, M. C. (2015), Memory, London – New York.

Bakker, E. (1997), Poetry in Speech, Ithaca, NY – London.


Byers, P. P. (1979), "A Formula for Poetic Intonation", in: Poetics 8, 367–380.

Cantilena, M. (1982), Ricerche sulla dizione epica I. Per uno studio della formularita` degli Inni Omerici, Rome.

See Rubin 1996.


Sandell, R. P. (2019), "The Place of Attic-Ionic Greek in Word-Level Prosodic Typology". Presentation, in: New Ways of Analyzing Ancient Greek 1, 13–14 December 2019, Göttingen.

Sapir, E. (1921) Language: An Introduction to the Study of Speech, New York.

Sauzet, P. (1989), "L'Accent du grec ancien et les relations entre structure métrique et représentation autosegmentale", in: Langages 24, 81–113.


Thalmann, M./Souza, A. S./Oberauer, K. (2019), "How does Chunking Help Working Memory?", in: Journal of Experimental Psychology: Learning, Memory, and Cognition, 45, 37–55.

Tolkien, J. R. R. (2006), The Monsters and the Critics and Other Essays, London.

Trudgill, P. (1983), "Acts of Conflicting Identity: The Sociolinguistics of British Pop-Song Pronunciation", in: P. Trudgill (ed.), On Dialect: Social and Geographical Perspectives, Oxford, 141–160 [repr.: Trudgill, P. (1997), "Acts of Conflicting Identity: The Sociolinguistics of British Pop- Song Pronunciation", in: N. Coupland/A. Jaworski (eds.), Sociolinguistics: A Reader and Coursebook, Houndmills, 251–266].


Wray, A. (2002), Formulaic Language and the Lexicon, Cambridge.

### Giampiero Scafoglio A Cyclic Theme in the Odyssey: The Oresteia in Zeus' Speech (1, 28–43)

Abstract: This paper focuses on the first mention of the Oresteia-theme (that is the murder of Agamemnon by Aegisthus and the revenge by Orestes) in the Odyssey, notably in the speech of Zeus in the council of gods (1, 28–43). I try to prove that this theme, coming from the epic cycle, is a holdover of the oral tradition, and an evidence of the origin and belonging of the Odyssey to this cultural background. The structure of the oral tradition and the rhapsodic practice allowed and even promoted exchanges and cross-references between different stories. Then, the Oresteia-theme is not an example introduced to demonstrate a moral issue (as scholars usually interpret it); on the contrary, the moral issue is deduced a posteriori from the mythological reference that is a part of, and a hint to, the big picture to which the Odyssey belongs.

Keywords: Odyssey; Oresteia-theme; epic cycle; oral tradition; neo-analysis; oral theory.

### 1 The Homeric Poems and the Epic Cycle

Some preliminary clarifications.When I talk about the epic cycle, I mean the oral tradition, the whole of mythological tales and songs passed on orally from generation to generation by the aeds and rhapsods.¹ Moreover, when I talk about the cyclic poems, I mean the later written texts, the literary works (such as the Cypria and Little Iliad) that are now lost and that we know from ancient evidence and fragments.² If the epic cycle as oral poetry is earlier than the Iliad and the Odyssey, the individual poems arising from the epic cycle came later, from the 7th century onwards.³ This is why we can say (with an only apparent paradox) that the Homeric poems and the epic cycle influenced each other mutually.⁴ The Iliad

For an effective definition of oral tradition as a "constellation" or "a loosely related consortium of flexible narratives", see Foley/Arft 2015. See Foley 1991, 1–59; Nagy 2015a.

For fragments and evidence of the cyclic poems on the Trojan subject, see Bernabé 1987, 36– 165; Davies 1988, 27–76; West 2003, 64–171. On the formation of these individual poems see West 2015.

See Davies 1989; Davies 2001, 2–5.

See Burgess 2001, 149–157 and passim; Scafoglio 2004b; Nagy 2015b.

OpenAccess. © 2022 Giampiero Scafoglio, published by De Gruyter. This work is licensed under the Creative Commons Attribution 4.0 International License. https://doi.org/10.1515/9783110751963-006

and the Odyssey indeed arose from the same cultural substrate as the cyclic poems, which in turn have been influenced by the Homeric poems.

I cannot leave aside the issue of whether the Iliad and the Odyssey are themselves cyclic poems. Most scholars (from Jasper Griffin to Margalit Finkelberg)⁵ answer they aren't, without any doubt. It is true that the Homeric epics have specific characteristics that distinguish them from the cyclic poems (as everyone knows, from Aristotle onwards). Yet I must confess I have some doubts: maybe the insistence on the specificity of the Homeric poems (specificity that cannot be denied, of course) risks making us lose the perception of the common grounds, what I would call the underlying congeniality of all the archaic epics, as well as the sense of the diversity between the Iliad and the Odyssey. The problem is further complicated by the editorial process carried on over centuries, from Pisistratus to the Alexandrine grammarians, with the result to fix up and regularize the text and structure of the Homeric poems, making them similar to each other and different from the cyclic poems. Moreover, our judgment is affected by the condemnation without appeal expressed by Aristotle on the cyclic poems⁶ and by our very limited knowledge of them.

Anyway, the Iliad and the Odyssey incorporate themes and episodes from the epic cycle. The crossover between the Homeric poems and other parts of the cultural tradition is not at all surprising, if we consider the latter as a long story that includes many other stories, something like a world history that develops through several stages, starting from the birth of the universe. This is exactly the image of the epic cycle we find in the Library of the Byzantine patriarch Photius, who outlines a summary of the Chrestomathy by Proclus: a source believed to be reliable, maybe the most important evidence on the cyclic poems.

Photius, Bibl. 239, 17–19⁷

διαλαμβάνει δὲ καὶ περὶ τοῦ λεγομένου ἐπικοῦ κύκλου, ὃς ἄρχεται μὲν ἐκ τῆς Οὐρανοῦ καὶ Γῆς μυθολογουμένης μίξεως, ἐξ ἧς αὐτῷ καὶ τρεῖς παῖδας ἑκατοντάχειρας καὶ τρεῖς γεννῶσι Κύκλωπας. διαπορεύεται δὲ τά τε ἄλλως περὶ θεῶν τοῖς Ἕλλησι μυθολογούμενα καὶ εἴ πού τι καὶ πρὸς ἱστορίαν ἐξαληθίζεται. καὶ περατοῦται ὁ ἐπικὸς κύκλος ἐκ διαφόρων ποιητῶν συμπληρούμενος, μέχρι τῆς ἀποβάσεως Ὀδυσσέως τῆς εἰς Ἰθάκην, ἐν ᾗ ὑπὸ τοῦ παιδὸς Τηλεγόνου ἀγνοοῦντος κτείνεται.

The Homeric poems indulge much less in fantastic and horrid elements, according to Griffin 1977. They are far more Panhellenic than the cyclic poems: Nagy 2010, 3–28. The 'special status' of the Iliad and the Odyssey within epic tradition can be understood as 'meta-cyclic': Finkelberg 2011; Finkelberg 2015; see also Tsagalis 2011.

Cf. Arist., Poet. 1459a37–b16, with Stroud/Robertson 1996; Scafoglio 2007; Fantuzzi 2015.

Text according to Severyns 1938. Translation is mine.

Proclus talks about what is called the epic cycle, which starts from the mythological union between Heaven and Mother Earth. This union generated three children with a hundred hands each and three Cyclops. Proclus examines the legends of the Greek people about the gods and their relationship with the historical truth. The epic cycle, made up by several poets, ends with the arrival of Odysseus in Ithaca, where he dies, killed by his son Telegonus, who does not know him.

In view of the reliability of his source (Proclus' Chrestomathy), we can trust Photius' statement about the comprehensiveness of the epic cycle.⁸ We find a confirmation in the remains of a cyclic theogony (or more than one) that constituted the beginning of the endless chain of events leading to the Trojan war and proceeding further with the return journeys of the heroes. Actually, not all scholars agree on the existence of the cyclic theogony, but there are several clues pointing to that direction: Alberto Bernabé collected the evidence on this subject in a dedicated section of his edition of the archaic epic fragments, while Malcom Davies refers to a cyclic titanomachy that in all likelihood was a part of (or a poem linked to) a lost theogony.⁹ Livio Sbardella found traces of a cyclic theogony within the text of Hesiod's Theogony and put it in relation with rhapsodic competitions.¹⁰

Furthermore, the summaries of the cyclic poems outlined by Proclus, which we know from a direct tradition and which Albert Severyns has carefully restored, give an impression of continuity and even close connection between the individual poems, as if they were the single images forming the big picture of the epic cycle. This is, at least partly, a construction realized by Proclus himself: we know that he removed overlaps and duplications from his summaries of the individual poems, in order to make their plots fit with each other.¹¹ For instance, he eliminated the tale of the fall of Troy and the subsequent events (in particular, the killing of Polyxena and Astyanax) at the end of the Little Iliad,¹² to avoid overlapping with the beginning of the Ilioupersis. Nevertheless, he was able to do such an operation, and he was motivated to do it, just because of the underlying consistency between the cyclic poems: we might say that Proclus envisioned and restored the continuity of the epic cycle through targeted intervention on the individual poems. Therefore, the overview of the single poems he outlines is not fully reliable, but for that very reason it is closer to the original configuration of the epic cycle, which was a systematic and comprehensive re-

See Holmberg 1998, who speaks of the oral tradition as "one large, universalizing text".

Bernabé 1987, 8–10 (Theogony), but see also 11–16 (Titanomachy); Davies 1988, 16–20.

Sbardella 2015. See West 1983, 121–126; D'Alessio 2015.

See Scafoglio 2004a.

Cf. schol. ad Lycophr. Alex. 1268 (360, 4 Scheer) = Ἰλιὰς μικρά, fr. 21 Bernabé (= 20 Davies).

view of legends pivoting on the Trojan war, but going far beyond this central point. The aeds and rhapsods drew the subjects and contents and, to some extent, even the structure and style of their songs from this mythological repertoire.¹³ Each time, they had the opportunity to choose individual episodes according to the convenience and the chronological limits imposed by the circumstances.

We must rely on this background to view not only the cyclic poems, but also the Iliad and the Odyssey.¹⁴ Relying on it, the relationship between the Homeric poems and the cyclic themes appears indeed in the right light, in a perspective of fluid communication and interchange reflecting the original consistency and continuity of the oral tradition.

In the light of these considerations, it is not surprising that the Iliad and the Odyssey are full of references to the epic cycle.¹⁵ I chose to study a single theme connecting the Odyssey with the wider background of the oral tradition: the theme that I would call Oresteia, going from the return home of Agamemnon and his murder up to the revenge by Orestes.¹⁶ This theme is inserted at the beginning of the Odyssey and it is strongly present in the Telemachy. I will focus on its first appearance, opening the way to many other references that will come later, in several parts of the poem.

### 2 The Oresteia in Zeus' Speech

The plot of the Odyssey starts with a council of the gods, just after the proem: it is precisely here that the Oresteia-theme makes its appearance, in an inconsistent and problematic way that (in my view) has been underestimated and misunderstood by most scholars.¹⁷ The first topic introduced in the debate of the gods, by

For the rhapsodic performance as the original form of circulation of the epic cycle see Burgess 2004.

See Edwards 1992; Ercolani 2006, 63–81; Foley 2007.

See Ercolani 2006, 113–117 and 123; Kullmann 2015; Finkelberg 2015; Scafoglio 2015.

As de Jong 2001, 12, puts it: "the Oresteia story is an embedded story, which is referred to repeatedly in the Odyssey, by different characters, to different addressees, and for different reasons". Form a moral perspective (usually recognized by scholars), Agamemnon provides some parallels with Odysseus, Clytemnestra with Penelope, Orestes with Telemachus, and Aegisthus with the suitors. See Olson 1995, 24–42; Marks 2008, 17–35; Alden 2017, 77–100.

See e.g. S. West 1983, 190: "la vendetta di Oreste è l'ultimo avvenimento importanteenaturalmente costituisce, in cielo come in terra, argomento di conversazione"; so, in the prominent position occupied by this theme at the beginning of the Odyssey "non c'è nulla di artificioso o macchinoso".

the mouth of Zeus, is the murder of Agamemnon by Aegisthus and the revenge by Orestes.

#### Od. , –¹⁸

τοῖσι δὲ μύθων ἦρχε πατὴρ ἀνδρῶν τε θεῶν τε· μνήσατο γὰρ κατὰ θυμὸν ἀμύμονος Αἰγίσθοιο, τόν ῥ᾽ ᾿Aγαμεμνονίδης τηλεκλυτὸς ἔκταν᾽ Ὀρέστης· τοῦ ὅ γ᾽ ἐπιμνησθεὶς ἔπε᾽ ἀθανάτοισι μετηύδα· "ὢ πόποι, οἷον δή νυ θεοὺς βροτοὶ αἰτιόωνται· ἐξ ἡμέων γάρ φασι κάκ᾽ ἔμμεναι, οἱ δὲ καὶ αὐτοὶ σφῇσιν ἀτασθαλίῃσιν ὑπὲρ μόρον ἄλγε᾽ ἔχουσιν, ὡς καὶ νῦν Αἴγισθος ὑπὲρ μόρον ᾿Aτρεΐδαο γῆμ᾽ ἄλοχον μνηστήν, τὸν δ᾽ ἔκτανε νοστήσαντα, εἰδὼς αἰπὺν ὄλεθρον, ἐπεὶ πρό οἱ εἴπομεν ἡμεῖς, Ἑρμείαν πέμψαντες, ἐύσκοπον ἀργεϊφόντην, μήτ᾽ αὐτὸν κτείνειν μήτε μνάασθαι ἄκοιτιν· 'ἐκ γὰρ Ὀρέσταο τίσις ἔσσεται ᾿Aτρεΐδαο, ὁππότ᾽ ἂν ἡβήσῃ τε καὶ ἧς ἱμείρεται αἴης'. ὣς ἔφαθ᾽ Ἑρμείας, ἀλλ᾽ οὐ φρένας Αἰγίσθοιο πεῖθ᾽ ἀγαθὰ φρονέων· νῦν δ᾽ ἁθρόα πάντ᾽ ἀπέτισεν".

Among them, the father of gods and men was first to speak, since in his heart he thought of blameless Aegisthus, whom far-famed Orestes, Agamemnon's son, had slain. Thinking of him, Zeus spoke among the immortals and said: "See, now, how ready mortals are to blame the gods. They say, indeed, that evils come from us, but they cause their own suffering, overstepping their bounds through their blind madness, as happened to Aegisthus. He took to himself the wedded wife of the son of Atreus and killed him on his return, overstepping his bounds, even though he knew of the death looming over him, since we had warned him, sending Hermes, the keen-sighted Argeiphontes, not to kill the man nor to take his wife. 'The son of Atreus, in fact, shall be avenged by Orestes, when he has grown up and wanted to return home'. This is what Hermes told him, but he could not prevail upon the heart of Aegisthus, while speaking for his sake; and now he has paid the full price of all".

In the Odyssey, at the very beginning of the plot, in the pivotal scene that sets events in motion, attention is initially focused on Aegisthus and the revenge of Orestes: this is the first word of Zeus in the debate that will lead to the departure of Odysseus from Calypso's island, starting the action of the poem. It seems a rather strange beginning. Why does Zeus talk about Aegisthus' crime and its consequences at the opening of the council of gods? Actually, an explication is provided right after by Athena, who takes the opportunity to introduce a

Text according to Van Thiel 1991. Translation is mine.

topic that is close to her heart: the captivity of Odysseus. She tells her father that Aegisthus deserved such a wretched death, while Odysseus was affected by unmerited suffering, as a prisoner on Calypso's island, far from home and family (1, 45–62). Therefore, the goddess takes advantage of Zeus' reference to Aegisthus' crime, in order to plead for her protégé; but this is not enough to explain the reason of that reference in Zeus' speech, especially at such a key moment.

Modern scholars point out the analogy, on the one side, between Aegisthus' impiety and the behavior of the suitors in Odysseus' house, on the other side, between Orestes' revenge and the killing of the suitors by Odysseus on his return.¹⁹ In this outlook, Zeus' reference to Aegisthus would be a touchstone for judging the behavior of the suitors, as well as an anticipation of the revenge of Odysseus on his enemies. Thus, the beginning of the poem would announce (albeit implicitly) the end of the story, which accordingly would take on a sense of consistency and completeness in both narrative structure and moral meaning. Aristotle and most Alexandrine scholars would have appreciated this fascinating explanation, which (I guess) works very well a posteriori, but does not give account for the choice of the Oresteia-theme as a term of comparison for the story of the Odyssey. The analogy between the former and the latter is in fact generic on the moral side, and it is of limited relevance with regard to the narrative structure.

In my opinion, the right explanation may rather be found in the relationship of the Odyssey with the epic cycle, which encompassed the tales of the return journeys of the Greek heroes after the conquest of Troy, including the story of the murder of Agamemnon by Aegisthus and Clytemnestra, and Orestes' revenge. This story was fixed in writing (at a later time) in a cyclic poem, the Returns (Νόστοι), as we learn from Proclus' summary.

Procl., Chrest. 301–302²⁰

ἔπειτα ᾿Aγαμέμνονος ὑπὸ Αἰγίσθου καὶ Κλυταιμήστρας ἀναιρεθέντος ὑπ᾽ Ὀρέστου καὶ Πυλάδου τιμωρία.

S.West 1983, 190: there is "una stretta analogia" between Aegisthus and the suitors, and this analogy "contribuisce in modo significativo alla presentazione di questi ultimi" (but she admits: "sebbene ciò non regga al vaglio della loggica", sic). Marks 2008, 17: the reference to Aegisthus and in general to Orestes' myth issues "a programmatic assertion of the god's own role in the Odyssey", in the sense that "just as Orestes, who acts with Zeus' approval, suffers no retribution for killing Aegisthus, so Zeus will intervene at the end of the Odyssey to ensure that Odysseus will not suffer for killing the suitors". Alden 2017, 77: the audience "will easily see that if the gods were prepared to send Hermes to warn Aegisthus not to follow through on his bad intentions, they should be all the more willing to send Hermes to help Odysseus".

Text cited (here and infra) according to Severyns 1963. Translation is mine.

Then, after the murder of Agamemnon by Aegisthus and Clytemnestra, there is the revenge by Orestes and Pylades.²¹

In an ideal reconstruction of the big picture of the epic cycle, that we can recognize in approximate form in the summaries of the cyclic poems outlined by Proclus, the Odyssey came just after the Returns: μετὰ ταῦτά ἐστιν Ὀμήρου Ὀδύσσεια (Chrest. 306). In this perspective of continuity and complementarity,²² the borders separating the Oresteia-theme and the story of Odysseus were not so clear-cut as it seems today, when the Odyssey is a literary work in its own rights, detached from the epic cycle that does not exist anymore.

I do not mean that the story of Odysseus was the same as that of Agamemnon, Aegisthus and Orestes, of course; the former was distinct from the latter, just as the latter was distinct from the other stories included in the Returns, such as the return journeys of Neoptolemus and Menelaus. But all these tales were related to each other, and even interpenetrated, because of several factors:


Furthermore, the structure of the oral tradition and rhapsodic practice allowed and even promoted exchanges and cross-references between different stories, since the aeds and rhapsods passed easily from one subject to another, to meet the requests of their public. We can imagine indeed that a singer could insert, within his song,areference to a different subject in order to recall a song he performed before, to announce the theme of his next song, or to suggest a possible topic to the audience, drawing their attention and arousing their curiosity.²⁴

"The murder and avenging of Agamemnon formed the main heroic subject matter of the narrative" in the Returns, while "the returns of other heroes were accommodated within this frame", according to West 2015, 100.

See the epic cycle as "one large, universalizing text" in Holmberg's view (Holmberg 1998, 456).

Cf. Procl., Chrest. 296–298.

On rhapsodic performances cf. the evidence given by the Odyssey itself (especially the songs of Demodocus in the book VIII) and by Plato, Ion 531a2–9, Rep. 600d5–6, Laws 658d6–8; with

Taking a step further, I would add that such cross-references are an integral part of the formulaic technique, conceived in the broader sense: not limited to linguistic segments and stylistic features, but extended to themes and sections of content, as proposed by Gregory Nagy.²⁵ I agree with him that, "from the standpoint of oral poetics, each occurrence of a theme (on the level of content) or a formula (on the level of form), in a given composition-in-performance, refers not only to its immediate context, but also to all other analogous contexts remembered by the performer or by any member of the audience".²⁶ This is even truer, I guess, when it comes to themes that are not consistent with the context in which they are located, such as the references to cyclic characters and episodes in the Homeric poems. Accordingly, this kind of cross-references can be interpreted as a result of the oral origins of the Iliad and the Odyssey, and as a mark of their contiguity and congeniality with the Epic Cycle.

In this perspective, the current understanding of Zeus' speech on Aegisthus and Orestes should be overturned, in order to restore the right relation between cause and effect. The Oresteia-theme is not an example introduced to demonstrate a moral issue that is central to the Odyssey; on the contrary, the moral issue is deduced a posteriori from the mythological reference that is a part of, and a hint to, the big picture to which the Odyssey belongs. To put it in other words, the Oresteia-theme is the cause and not (as it is usually understood) the consequence of the moral issue. This is not to devaluate the ethical meaning of Zeus' words on Aegisthus' guilt, which is an important point. I would rather frame this cross-reference (as well as many others) in the dynamic and diachronic dimension, taking account of the origins and evolution of the Odyssey as oral poetry belonging to a broader and richer tradition that substantially affects its written form. This gets even more evident if we look more closely at the speech of Zeus.

### 3 Traces of the Oral Tradition

In the sentence that introduces the direct quotation, Aegisthus is called ἀμύμων (l. 29). This epithet, almost universally rendered by "blameless" (from μῶμος

Nagy 1979, 15–65; Burgess 2004; Tsagalis 2018. On the "interdependence between performance and text" see Bakker 1993.

See Nagy 1990, 18–35.

See Nagy 1996, chapter 2 and, in particular, p. 50.

with the privative alpha),²⁷ is not consistent with the context and does not fit the character as described by Zeus: the behavior of the murderer of Agamemnon and seducer of his wife is anything but blameless.²⁸

Faced with this difficulty, scholars suggest broader and vaguer meanings, such as "noble" and "illustrious";²⁹ there is who gets to specify that ἀμύμων is often used "to mark personal appearance or nobility of birth, and sometimes without regard to moral excellence".³⁰ This is evidently an attempt to justify the awkward presence of the epithet in that context by weakening its meaning. Even worse, the radical followers of Milman Parry refuse to recognize any meaning to this epithet, considering it as "empty" and purely ornamental.³¹

Anna Amory Parry has taken a different path, analyzing all Homeric passages where we find that epithet and coming to the conclusion that it means "handsome", concerning physical beauty and charm (the qualities of a seducer).³² But this intriguing proposition is far from being demonstrated, affected as it is by subjective impressions. After all, in the Homeric world, physical excellence is not independent from moral judgment. Moreover, keeping in mind the myth as a whole (especially the events related to Iphigenia, but later also Cassandra's role at the side of Agamemnon), we have reason to believe that it is not with handsomeness that Aegisthus seduced Clytemnestra.

Alfred Heubeck, in turn, proposed a different (but quite improbable) etymology and meaning of the epithet ἀμύμων, which would be connected with the verb ἀμεύομαι ("to surpass") and would express excellence by birth or other qualities, and/or capacity to overcome the enemies, without moral implications.³³ Conversely, according to some scholars, it is perfectly understandable that Zeus calls Aegisthus "blameless" because, by killing Agamemnon, he avenged his fa-

Cf. Hesychius, s.v. μῶμαρ· αἶσχος, φόβος, ψόγος. For the scholia on the Homeric passage see Pontani 2007, 31.

At the time, Kirchhoff 1879, 166–167, considered Od. 1, 1–87, as an interpolation, taking that "inappropriate epithet" as strong evidence of the lateness of the whole passage, added by an "incompetent poet".

E. g. Cunliffe 1963, 26, s.v. ἀμύμων: "an epithet of persons of general commendation, noble, illustrious, goodly, or the like", applied to Aegisthus "as a merely conventional epithet".

Autenrieth/Flagg 1958, 21, s.v. ἀμύμων.

E. g. Stanford 1959, XIX. The metrical argument is rightly refuted by Lowenstam 1993, 48, since the epithet ἀμύμονος has no "unique metrical value", as proved by the metrical alternative of III, 310, ἀνάλκιδος Αἰγίσθοιο.

Parry 1973. Contra, Combellack 1977.

Heubeck 1987. Contra, Pulleyn 2000, 150.

ther (just like Orestes, who is praised for his revenge passim in the poem; and like Telemachus himself should do).³⁴

Recently, David Elmer argued that the poet deliberately refers to Aegisthus such ambiguous epithet to question the ethical dimension of the whole epic (including Odysseus' final revenge on the suitors).³⁵ On this basis, Ruobing Xian claims a closer connection between Aegisthus and Odysseus as evoked by the epithet "blameless", which is often applied to the latter in the poem: Zeus' reference to Aegisthus in his first speech is to be understood as "a kind of coded signal to Athena, by which she immediately takes up the hint and changes the subject to Odysseus".³⁶

This difficulty of interpretation gives us the opportunity to return briefly to the broader issue of the formulaic epithets. It is undeniable that they fulfil a lexical and metrical function rather than providing important semantic integrations to the context. Indeed, they are not completely meaningless, as the radical followers of the oral theory believe it: such misconception is not a mistake made by Milman Parry; it is rather a bad use of his good ideas. The formulaic epithets indeed find their raison d'être, their motivation, in a permanent feature or in a potential skill of a character or of a thing: Achilles is "fleet-footed" (πόδας ὠκύς ᾿Αχιλλεύς) whether he is standing or sitting, or even sleeping; but we know that he is very fast, if and when he runs. The fact that the epithet ἀμύμων is referred to several characters in both Homeric poems confirms that it is a formulaic element, but this does not exempt us from finding an explication for its attribution to Aegisthus, who is not blameless at all, and even less so in the situation recalled by Zeus.

We may think, therefore, that Aegisthus was blameless in another moment (in all likelihood, an earlier time), which is not included in the speech of Zeus, who focuses on the recent and turbulent events leading to his death. Then, this epithet referred to Aegisthus comes from the past: the individual past of the character as well as the collective past of the epics: it comes from the oral tradition, just like the whole Oresteia-theme evoked by Zeus. The same of course can be said (in an admittedly non-problematic way) of the patronymic ᾿Aγαμεμνονίδης that accompanies the name of Orestes (l. 30).

There is another awkward point that, if properly understood, can confirm my interpretative proposition. The speech of Zeus includes the direct quotation of the words addressed by Hermes to Aegisthus to warn him: ἐκ γὰρ Ὀρέσταο

See Combellack 1982; Danek 1998, 40.

Elmer 2015, 176–183.

According to Xian 2020, the apparently inconsistent connection between Odysseus and Aegisthus "helps the audience to reflect upon the epic hero's fate and deeds".

τίσις ἔσσεται ᾿Aτρεΐδαο / ὁππότ᾽ ἂν ἡβήσῃ τε καὶ ἧς ἱμείρεται αἴης (ll. 40 –41). The passage from the oratio obliqua to the oratio recta is unexpected and abrupt, as scholars did not fail to notice:³⁷ Stephanie West admits that it is "very brusque", but she still finds an explanation, claiming that the direct speech serves to "underline the importance" of the concept expressed by Hermes.³⁸

I think, instead, that the insertion of the quotation depends on the origin of the Oresteia-theme, coming from an oral song, to which those lines formerly belonged. The speech of Zeus is, at first, an overview of the events, bearing some traces of orality in its language (such as the apparently inconsistent epithet ἀμύμων); but then it incorporates even a formulaic segment that is distinctly recognizable because of the direct quotation and its abrupt insertion. In the lost oral song, those words were pronounced by Hermes to Aegisthus, but they could have been spoken mutatis mutandis also by other characters, for instance by Zeus ordering Hermes to warn Aegisthus, and maybe by the latter reporting Hermes' warning to Clytemnestra or to someone else. Of course, we cannot know who pronounced those words, apart from Hermes; but we recognize the formulaic nature of the direct quotation, and we can lead it back to that lost song.

I would still add a minor point that works ad abundantiam, to support my statement. Zeus introduces the case of Aegisthus with the introductory formula ὡς καὶ νῦν, which presents the events as very recent and fresh in memory. We know indeed that the murder of Agamemnon happens in the period in which Odysseus is being held on Calypso's island: we can easily think (even within the uncertain chronological framework of the legend) that it happens a short time before the council of gods. This is the evident meaning of that introductory formula. However, if we bear in mind that the time of the events is represented as a linear succession of tales in the oral tradition, that expression should indicate not only the chronological proximity, but also the contiguity, or better the continuity, between two stages of the long and varied story that is the epic cycle. This impression is enhanced by the repetition of the adverb νῦν at the end of the speech (l. 43).

On the transition from oratio obliqua to oratio recta in Homeric language as a trace of oral tradition see Rajić 2008.

S. West 1983, 194.

### 4 The Moral Meaning of the Oresteia-Theme

Now we come to the moral meaning of Zeus' words, a meaning that is not the reason why the case of Aegisthus is introduced (as I tried to prove), but which is evidently present and not irrelevant. The guilt of Aegisthus is summarized with a short expression that is nevertheless meaningful and problematic: ὑπὲρ μόρον. Zeus states that mortals cause their own suffering ὑπὲρ μόρον (l. 34) and that Aegisthus acted ὑπὲρ μόρον, taking the wife of Agamemnon and then killing him (l. 35). Scholars do not agree on the meaning to be attributed to this expression: if we consider μόρος as a synonym of μοῖρα and we translate the latter as "fate", "destiny", the phrase ὑπὲρ μόρον would mean that mortals sometimes act against their fate and, in particular, Aegisthus committed that crime going against the destiny.³⁹ However, in the Homeric world that is a mirror of the Greek archaic way of thinking, the concept of μοῖρα does not mean a fate that controls all aspects of life: it is rather the duration (literally the "portion") of the existence granted to men and communities.⁴⁰ It is the superior and inflexible determination of two crucial moments, the start and the end, in both individual and collective life: the birth and death of men, as well as the foundation and fall of cities and kingdoms. Then, what does ὑπὲρ μόρον mean in Zeus' speech?

I think it means "over the limit", indicating the overcoming of a boundary that delimits and defines a condition of normality, a state of balance and harmony, rather than a moral law stricto sensu. Each man is given joys and sorrows, to some extent, by the gods (as Achilles says to Priam in the book XXIV of the Iliad); but here Zeus states that sometimes "mortals increase their own suffering over the limits", that is "beyond the measure" established by the gods: in this case, they should not accuse the gods, because they are responsible for their exceeding sorrows. This is what Aegisthus made: he procured his own death by taking Agamemnon's wife and then killing him, overcoming the limits, upsetting and disrupting the current status of things, breaking the balance and harmony. The revenge of Orestes is seen as a necessary consequence of Aegisthus' crime, a blood price paid by the guilty:alife for a life, without any problematization. This

It would be "un paradosso teologico", according to S. West 1983, 191–192. Aegisthus acted "against the ordinance of his fate", in the view of Pucci 2018, 149–152. Finkelberg 2020, 234– 240, speaks of "anti-apology of Aegisthus" as opposed to "Agamemnon's apology".

See Pirenne-Delforge/Pironti 2011. I cannot share the all-encompassing definition by Schadewaldt 1966, 108: "als die allgemeine regelnde Macht im Grundbestand all dessen, was ist, ist Moira göttlich und eine Gottheit, aber in ihrer Allgemeinheit eben kein Gott von umrissenem Charakter". For a full treatment of all occurrences of μοῖρα and the terms having the same root in the Homeric epics see Sarischoulis 2008, 27–128.

is already a moral concept in a broad sense, but we cannot yet speak of a mature ethical vision: it is rather the starting point in a path of cultural and moral evolution that will lead over time to such a vision.

In the Odyssey, we can find the different stages of this evolution (corresponding to different episodes and moments), due to the progressive and stratified composition of the poem.⁴¹ In this light, the moral meaning of Zeus' speech is not inconsistent with the whole of the Odyssey, but it reflects the older stratum of the evolutionary process: it is what we may consider to be the meeting point, or rather the intersection, between the epic cycle and the Homeric world.⁴²

In Zeus' speech, the character of Orestes has no moral connotation: he does not embodies a heroic ideal and does not pose any ethical problem; the attention is concentrated on his revenge. In the Telemachy, Orestes will be mentioned on several occasions (by the goddess Athena, by Nestor and by Menelaus) as a positive example and even a moral model for Telemachus, on the account of the great glory he obtained by revenging his father's death.⁴³ Here is a list of all references:

Od. 1, 298–300: Athena speaks to Telemachus; Od. 3, 103 ff. praes. 193–200: Nestor speaks to Telemachus; Od. 3, 234–235: Athena speaks to Telemachus; Od. 3, 254–312: Nestor speaks to Telemachus; Od. 4, 512–547: Proteus' speech, reported by Menelaus to Telemachus. Cf. also Od. 1, 325 ff. (Phemius sings about ᾿Aχαιῶν νόστον […] λυγρόν).

Undertaking a systematic analysis of all these references to Agamemnon, Aegisthus and Orestes in the Telemachy, we may follow the gradual integration of the

I fully agree with Segal 1994, 195–196: "there is little doubt that the Odyssey incorporates older strata of beliefs" in the moral and theological field.

Allan 2006 advocates a substantial homogeneity in moral matters between the Homeric epics and the epic cycle, but without a gradual evolution. This last point seems to me most improbable. Ultimately, I would call it dynamic congeniality rather than (static) homogeneity.

"The story of Agamemnon's death is not the same each time it is told", because "individual narrators routinely adapt it to their own purposes, while the characters who listen to these tales understand and interpret them in accord to their own plan and preoccupations" (Olson 1995, 27). Orestes becomes thus a reference point in the moral growth of Telemachus; but the choice of this example is determined, in my opinion, by the previous reminder of the legend made by Zeus in his first speech, where the tale of Aegisthus' crime and punishment has not yet such a moral purpose.

Oresteia-theme in the moral evolution achieved in the Odyssey;⁴⁴ but we cannot yet see it in Zeus' speech.

### 5 What about Clytemnestra?

However, a striking point in Zeus speech (whichIhave deliberately left for last) is the silence on Clytemnestra's role in the murder of Agamemnon: she is presented as the "wedded wife" (ἄλοχον μνηστήν) of the king, taken⁴⁵ by Aegisthus, who stands out as the only responsible. Moreover, she is not even mentioned as a victim of Orestes' revenge.⁴⁶

It is possible that, in the original version of the legend passed down in the epic cycle, she did not have a pivotal role in the murder, in line with the marginal condition of women in Greek archaic society; but what is sure is that, in no version of the legend, she could escape the revenge of her son. Yet there is no mention of this revenge in Zeus' speech, and not even in the many other references to the Oresteia-theme spread in the whole of the Odyssey.

Actually, the responsibility of Clytemnestra in the murder of Agamemnon is clearly stated in the speech that the latter addresses to Odysseus in the underworld (11, 404–434): here Aegisthus is still the leading man of the crime, but Clytemnestra shares the action and the blame as the second main character. She is at Aegisthus' side while he murders Agamemnon (ll. 409–411);⁴⁷ she is the one who kills Cassandra (ll. 421–423), and maybe she even strikes the blow of grace to her husband (ll. 423–425, but admittedly the development of the events is not

Although the Orestes-theme in the Telemachy and also in the whole Odyssey has been the focus of some studies (listed supra, n. 16), this specific research perspective has never been adopted, as far as I know.

The verb γαμέω (line 36) is a uox media that means "to marry", "to take to wife", as well as "to take for a lover" or "for a concubine" (for different kinds of relationships, and even mere sexual intercourse).

Actually, there is a short and indirect hint to the matricide, notably 3, 309–310, referring to the funeral fest that Orestes gave for Aegisthus and Clytemnestra, but without explicit mention of the killing of the latter. Aristarchus took these lines to signify that both Aegisthus and Clytemnestra were murdered by Orestes (schol. MQRT ad Od. 3, 309–310). Not by chance, this passage is omitted in some ancient editions.

Agamemnon's words: ἀλλά μοι Αἴγισθος τεύξας θάνατόν τε μόρον τε / ἔκτα σὺν οὐλομένῃ ἀλόχῳ, οἶκόνδε καλέσσας, / δειπνίσσας, ὥς τίς τε κατέκτανε βοῦν ἐπὶ φάτνῃ, "but Aegisthus, after preparing a destiny of death and inviting me home to feast, with my wretched wife killed me, just like people slaughter an ox at a manger".

plainly explained).⁴⁸ Little later, in Agamemnon's second speech, Clytemnestra seems to be the very responsible of the crime (notably ll. 452–453).⁴⁹ In the dialogue between Odysseus and Agamemnon in the underworld, the mention of Clytemnestra serves a comparison with Penelope and becomes part (obviously ex contrario) of the portrait of the latter as the ideal wife.⁵⁰ This is a confirmation that Clytemnestra's role in the murder of Agamemnon was known to the oral tradition, from which it emerges in certain points (and not in others) of the Odyssey. Anyway, Orestes' revenge against his mother is completely (or at least almost completely) ignored in the poem. Why?

It is evident that this part of the legend has been removed, intentionally deleted, by the poet.⁵¹ It is not an accident that here I start speaking of "the poet", while until now I only talked about "the poem". The removal of this part of the legend must be traced back to moral grounds, in line with the elimination of other turbid and scabrous stories in the whole of the Homeric epics.

This selective approach applies, for instance, to the burning issue of human sacrifice, which was well present in the epic cycle (mainly with the episodes of Polyxena and Iphigenia), while it is almost completely deleted from the Iliad⁵²

[…] αὐτὰρ ἐγὼ ποτὶ γαίῃ χεῖρας ἀείρων / βάλλον ἀποθνήσκων περὶ φασγάνῳ∙ ἡ δὲ κυνῶπις / νοσφίσατε κτλ., "but I lay dying upon the earth, raising my hands, with the sword in my body, while that bitch went away". See the freer translation by Butler/Power/Nagy: "I raised my hands to kill the slut of a murderess, but she slipped away from me".

ἡ δ᾽ ἐμὴ οὐδέ περ υἷος ἐνιπλησθῆναι ἄκοιτις / ὀφθαλμοῖσιν ἔασε∙ πάρος δέ με πέφνε καὶ αὐτόν, "my wife did not let me satiate my eyes with the sight of my son: she killed me before I could".

See Katz 1991, 29–53. Clytemnestra, "the archetypal evil woman", is set up throughout the poem "as the obvious inverse parallel to Penelope", as Wohl 1993, 35, puts it.

However, I cannot agree with S. West 1983, LXXIX, that Orestes' matricide is ignored in the Odyssey because it "did not fit in the pattern", that is the double parallelism between Aegisthus and the suitors, Agamemnon's son and Telemachus, as well as the opposition between Clytemnestra and Penelope. Nor can I share the more sophisticated interpretation by Marks 2008, 25–27, who recalls the alternative legend of an unfaithful Penelope that is emarginated from the Odyssey, but which at the same time is alluded to in the poem: "the potentially unfaithful Penelope is then an analog of Clytemnestra, so that the story proceeds in part as if Odysseus and Telemachus are preparing to confrontaconniving and homicidal wife and mother"; it follows that the suppression of Clytemnestra in Zeus' first speech is somehow "analogous to the suppression of this 'other Penelope' by the Odyssey itself".

The only exception is the sacrifice of twelve young Trojans made by Achilles at Patroclus' funeral pyre (Il. 18, 336–337; 21, 26–32; 23, 22–23; in particular 23, 175–177). Schol. T ad Il. 23, 174–176, interprets this sacrifice as illustrating the brutality of Achilles's nature, which Patroclus' death made even more savage. See Hughes 1991, 49–56. In the poem, however, there is no retrospective recall of the most infamous sacrifice of Iphigeneia: in fact, it never happened, if she is to be identified with the Ἰφιάνασσα mentioned at Il. 9, 145 and 287 as Agamemnon's

and disappears altogether from the Odyssey. The removal of the killing of Clytemnestra by Orestes seems perfectly consistent with this particular setting of the Homeric epics:⁵³ I take it as evidence of a poetic individuality, consciously working on the literary material passed down in the oral tradition.⁵⁴

This setback affords me the opportunity to conclude with a methodological reflection. I just tried to correct a point that I consider as a common mistake in the interpretation of the Oresteia-theme in Zeus's speech, that is seen as the result of an authorial work aimed at achieving a literary and even a moral purpose: a position shared by neo-unitarian and neo-analytic scholars. Using the tools of oral theory, I tried to demonstrate that Zeus' reference to Aegisthus and Orestes is rather the outcome ofadiachronic and more complex process that has benefited from the contribution of one or more poets, but which cannot be explained just as a creative operation.

On closer inspection, this approach proved to be effective, but not yet exhaustive. In my view, it worked very well to a certain extent, but it could not account for all aspects of the Oresteia-theme in Zeus's speech: this theme cannot be considered as the result of an authorial work driven by a moral intention (as I just stated), but it bears not negligible traces of an authorial rework underlying a moral concern (particularly as regards the removal of Clytemnestra's involvement in both Aegisthus' crime and Orestes' revenge, a removal that I see as the clear imprint of a poetic individuality).

The standpoint of neo-analysis, which has turned out to be inadequate at the preceding stage, comes back into play here and provides useful input. Finally I would say, with a pun that sums up the complementary relationship just tested between the oral theory and neo-analysis, that the former fixes the mistakes committed by latter, while the latter achieves the process left undone by the for-

daughter who is alive and well. But even if Iphigenia and Iphianassa are different people, the fact remains that Homeric epics avoid mentioning the sacrifice. See Burgess 2001, 150 –151; Currie 2015, 291–292.

In the Iliad, women are enslaved, but they are never killed (Griffin 1977, 45). Moreover, Homeric epics avoid reference to killing within the family (Alden 2017, 84–85, n. 34). If Orestes is to be retained as a constructive paradigm for Telemachus on the moral level, his behavior has to be without spot or blemish: this is why his matricide is played down.

As for poetic individuality in Homeric epics see Edwards 1980, who takes Iliad I as a case study. A balanced assessment on the negation of poetic individuality in Homeric poetry by the "Parry-Lord thesis" (the hard core of the oral theory), as well as on the opposite "aesthetic reaction", is made by Thomas 1992, 29–36.

mer.⁵⁵ The two approaches can help each other, if adequately managed: each of them will contribute in its own way to the understanding of the Homeric poems, as well as to the knowledge of the epic cycle.⁵⁶

### References


Davies, M. (2001), The Greek Epic Cycle, London.

Not casually, there have been some noteworthy attempts to reconcile the two orientations over the past thirty years, e. g. Slatkin 1991; Danek 1998; Burgess 2001 and Burgess 2009. See the theoretical appraisals by Finkelberg 2003, Burgess 2006 and Montanari 2012.

I tested this methodological approach to Greek archaic epics in my book: see Scafoglio 2017, mainly the methodical premises, 3–8, with Norton 2019 and Stelow 2019.


Thomas, R. (1992), Literacy and Orality in Ancient Greece, Cambridge.


### Jonathan S. Burgess Audiences of the Prophecy of Tiresias in Odyssey Book XI

Abstract: With attention to non-Homeric narratives of continuing adventures of Odysseus after the slaughter of the suitors, I explore the potential meanings of the prophecy of Teiresias in book XI of the Odyssey. In particular I am concerned with the various meanings possible for audiences with in the poem (Odysseus listening to Teiresias, and Penelope hearing her husband's report of the prophecy) and external audiences, notably those informed of nonHomeric narratives. It is recognized that audiences might variously respond to the prophecy, or even hold disparate reactions in mind as they listen.

Keywords: Odyssey; Teiresias; Odysseus; Penelope; Telegonus; Telegony; exile.

In book XI of the Odyssey Tiresias tells Odysseus that after a successful return to Ithaca he must undertake an inland journey before returning home to a death "from the sea". This sequence of post-return events – inland journey, return home, death – roughly approximates the cyclic Telegony, in which a sojourn on the Greek mainland is followed by a return to Ithaca and death at the hands of his son Telegonus. I have previously argued for a textless intertextuality between the two epics, mediated by oral, epichoric legends about Odysseus. According to this argument, the words of Tiresias allude to non-Homeric narratives, with the allusion embracing a wide range of significance, including reference, variation, agonism, and suppression. I am concerned here with how internal characters and external audiences might understand the words of Tiresias.

As with my previous discussions,¹ my argument assumes the oral performance of a multiplicity of early Panhellenic epics and epichoric analogues. I emphasize that my intertextual concerns are with oral traditions, not necessarily specific epics like the cyclic Telegony. Besides updating and extending my approach to the prophecy of Odysseus' death "from the sea", I wish to put to one side the question of how ancient and modern scholars interpret the words of Tiresias. I first ask what the prophecy might mean its internal audience, Odysseus, as well as to his own internal audience, Penelope, when the hero repeats his version of the prophecy to his wife. My focus is on what the words of Teiresias

See Burgess 2014; Burgess 2017a.

OpenAccess. © 2022 Jonathan S. Burgess, published by De Gruyter. This work is licensed under the Creative Commons Attribution 4.0 International License. https://doi.org/10.1515/9783110751963-007

could mean to an ancient listening audience informed by nonHomeric narratives when listening to a performance of the Odyssey.

Ancient audiences of the Odyssey would need to negotiate between nonHomeric material and the Odyssey's telling of the return of Odysseus. Post-return narratives do not present a uniform story about what happens after the death of the suitors. They vary in nature and detail and they contradict the Odyssey's closure. The Telegony, for example, would seem to challenge the very ideology of the Homeric poem, not simply add on further adventures.² It is also apparent that the Odyssey itself is aware that the consequences of the slaughter of the suitors are usually more complicated than the epic's ending would have us suppose.³ The prophecy of Teiresias in book XI itself looks beyond the boundaries of a poem that is insistent on closure. None of this, however, means that an ancient audience would need to choose allegiance between the Homeric epic and apparently competing analogues. At its most complex, reception of the Odyssey may have involved both recognition of alternative traditions and a simultaneous surrender, however momentarily, to the Odyssey's expectations of its ideal audience.

When Odysseus seeks out Tiresias in the underworld, Tiresias reports generally about the sequence of events that might lead to his return to Ithaca and vanquishing of the suitors (Od. 11, 100 –118). He then mandates a so-called "inland journey" after their death (Od. 11, 119–137; cf. 23, 281 ff.):


For recent discussions of the Telegony, see West 2013, 288–315, Tsagalis 2015, Lulli 2018. For intertextuality between the Odyssey and the Telegony (or its story), see Ballabriga 1989, 297–301; Tsagalis 2008, 63–90; Alden 2017, 107, 109, 175 n. 9; Arft 2019.

ἀβληχρὸς μάλα τοῖος ἐλεύσεται, ὅς κέ σε πέφνῃ 135 γήρᾳ ὕπο λιπαρῷ ἀρημένον· ἀμφὶ δὲ λαοὶ ὄλβιοι ἔσσονται. τὰ δέ τοι νημερτέα εἴρω.

But when in your halls you kill the suitors, whether by deceit or openly with sharp bronze, go forth then, taking a well-balanced oar, until you reach those men who do not know the sea or eat food mixed with salt or know of red-cheeked ships or well-balanced oars, which are wings for ships. And I will tell you a very clear sign, and it won't escape your notice: when another traveler meeting you says that you have a chaff-destroyer on your shoulder, indeed, after sticking the well-balanced oar in the earth and offering fine sacrifices to lord Poseidon, a ram, bull and a mounting boar, return home and offer holy hecatombs to the immortal gods who hold wide heaven, all in order. And death to you from the sea, yourself, harmless – quite suchlike – will come, which will slay you, by sleek old age worn. And about the people will be prosperous. I speak these things as true. (All translations my own)

"It's tough to make predictions, especially about the future", reportedly said Yogi Berra.⁴ Teiresias does not actually provide the hero with directions for the way home, as Circe had suggested (Od. 10, 539–540). He only vaguely outlines alternative scenarios in the return to Ithaca. Successful return is contingent upon particular choices in a certain sequence of events, well described as a "logic tree" by Peradotto.⁵ The "inland journey" will occur only if Odysseus avoids participating in the slaughter of the cows of Helius, successfully returns home, and kills the suitors.⁶ And then the "inland journey" must occur, though no explicit motivation is provided for it. The contingent nature of the prophecy, with alternative scenarios and opaque causality, might remind us of variation in accounts of the further adventures of Odysseus. Even if the "inland journey" does not allude specifically to non-Homeric narratives, its very manner of presentation suggests multiformity about the return to Ithaca and its consequences.⁷

According to Teiresias, Odysseus must travel inland with an oar on his shoulder until he meets someone who does not know what it is. The motivation for the journey is not made explicit, but many find it significant that Odysseus must sacrifice to Poseidon far from the sea. Apollodorus (Epit. 7, 34) specifies that Odysseus propitiated Poseidon through sacrifice during a mainland journey

The professional baseball player and coach was celebrated for malapropisms. Source: https:// en.wikiquote.org/wiki/Yogi\_Berra, accessed 4/10/2019, which references another site that finds a Danish origin for variations on this proverbial statement: https://quoteinvestigator.com/2013/10/ 20/no-predict/ .

See Peradotto 1990, 63–93.

Bakker 2013, 14–134 interestingly explores connections between Teiresias' prophecy and Poseidon's role in the plot structure of the Odyssey, and beyond; see in particular Bakker 2020. See West 2014, 216 for the view that the prophecy represents a pre-Homeric poem featuring a prominent wrath of Poseidon.

to Thesprotia, which also occurs in the cyclic Telegony.⁸ Tiresias further states that Odysseus is to return to Ithaca, where he will die. This return to Ithaca is notable, since in a comparable tale type the protagonist desires to leave a climate or place (like the sea) and permanently settle away from it.⁹ Settlement in a new home happens in the Telegony, but only for a time. According to the brief summary by Proclus, the hero marries the Thesprotian queen and produces a child before eventually returning to Ithaca. Both the "inland journey" mandated by Teiresias and the mainland journey narrated by the Telegony and Apollodorus thus resemble the "The Sailor and the Oar" tale type (to use the title in Hansen 2002),¹⁰ but they are consonant with the Odyssey in that the hero returns to Ithaca, where he dies.

Tiresias' specific description of the death of Odysseus on Ithaca (11, 134–137; cf. 23, 281 ff.) has inspired much controversy:

θάνατος δέ τοι ἐξ ἁλὸς αὐτῷ ἀβληχρὸς μάλα τοῖος ἐλεύσεται, ὅς κέ σε πέφνῃ γήρᾳ ὕπο λιπαρῷ ἀρημένον· ἀμφὶ δὲ λαοὶ ὄλβιοι ἔσσονται.

And death to you from the sea, yourself, harmless – quite suchlike – will come, which will slay you, by sleek old age worn. And the people about will be prosperous.

My translation reflects the halting ambiguity of the syntax and words, and my interpretation of the potential meaning of these words will be discussed below. The most obvious meaning of the prophecy is that Odysseus will die apart from the sea, in his old age, with his people prosperous. The question remains whether the apparent meaning is the actual one. Odysseus is reporting the words of Tiresias to the Phaeacians, and so it is Odysseus who provides his own internal response to Teiresias. It isacurt and uninformative reply, affirming his acceptance that the contents of the prophecy must be fated (Τειρεσίη, τὰ μὲν ἄρ που ἐπέκλωσαν θεοὶ αὐτοί, "Teiresias, these things then I suppose the gods themselves have ordained", Od. 11, 139). If Odysseus understands the words of Teiresias to mean that Odysseus will not perish at sea, this of course is what the hero who has wandered long at sea would want to hear. We do not know

See especially Hartmann 1917, 216–217; Cook 1995, 96; Cerri 2002, 167–168.

Hansen 1977; Hansen 2002, 371–78; Hansen 2014. See the inverse journey of a 12th-century Turkish army ordered by the Sultan to march to the coast and not return without "sea-water, sand, and an oar" (Seal 2012, 249).

See Hansen 2002, 371–378.

what the Phaeacians think of the prophecy, beyond that they are generally impressed by his tales, even if Alcinous ambiguously raises the question of truthfulness during the intermezzo (Od. 11, 363–369).

There is another internal exchange between speaking character and listening character in book XXIII when the main narrator describes Odysseus reporting his prophecy to Penelope. After Odysseus is recognized by Penelope, he tells her that he has learned from Teiresias there are more trouble for him:

Od. 23, 248–250 ὦ γύναι, οὐ γάρ πω πάντων ἐπὶ πείρατ' ἀέθλων ἤλθομεν, ἀλλ' ἔτ' ὄπισθεν ἀμέτρητος πόνος ἔσται πολλὸς καὶ χαλεπός, τὸν ἐμὲ χρὴ πάντα τελέσσαι.

oh wife, since not at all have we come to the limits of trials, yet still there will be unlimited trouble hereafter, much and difficult trouble, all of which I must see through.

Odysseus' statement is noticeably vague in terms of quantity (ἀμέτρητος, πολλὸς) and time (ὄπισθεν).¹¹ Though the hero is understandably frustrated that his challenges are not over, his words seem to exaggerate the relatively simple mandate of Teiresias to go far inland, sacrifice, and return home. One might see here the main narrator speaking through his character with a meta-poetic awareness of numerous alternative post-return adventures that are more complex and durative than an "inland journey." When in response to Penelope's insistence that Odysseus reveal the coming "trial" (ἄεθλον, 261), the hero's version of the prophecy surprisingly introduces the further complexity that he will need to visit many cities (μάλα πολλὰ βροτῶν ἐπὶ ἄστε' ἄνωγεν / ἐλθεῖν, 267–268). This detail has reminded scholars of the third line of the Odyssey (πολλῶν δ' ἀνθρώπων ἴδεν ἄστεα καὶ νόον ἔγνω, "and he learned of the cities of many peoples and their mind"), which itself does not seem easily harmonize with the wanderings of Odysseus reported in books IX–XII.¹² The Odyssey thus seem on occasion to acknowledge alternative narratives to which it does not subscribe.

It also is hard not to be reminded by Odysseus' reference to πείρατα that the scholia to Od. 23, 296 use the terms πέρας as well as τέλος in reporting arguments by Aristophanes and Aristarchus about the true ending or climax of the poem.¹³ That report has been central to arguments against the authenticity of the received ending of the Odyssey after Od. 23, 296, when Odysseus and Penelope

See Buchan 2004, 234–236; Tsagalis 2008, 70 –74 for creative intepretation of this exchange between Odysseus and Penelope.

This does not mean that Apologos is non-traditional: Burgess 2017b.

See Heubeck 1992, 342–345, with Tsagalis 2008, 70; Purves 2010, 74–75.

go to bed. My position is not an analytical rejection of the ending of the Odyssey, but I also distinguish myself from "naive unitarians" (to use the conventional phrase without pejorative intent) who fully subscribe to the Odyssey's efforts at closure. Odysseus' wording in Od. 23, 248–250, as well as the difficulty with which the epic justifies its own closure, may testify to the existence of non-Homeric narratives about the further adventures of Odysseus.

After Odysseus recounts the "inland journey" to Penelope, he repeats Tiresias' description of his death (Od. 23, 267–284). Though Odysseus predicted that Penelope would find the prophecy unpleasant (266), Penelope ignores the "inland journey" and focuses on the predicted old age of her husband at his death (εἰ μὲν δὴ γῆράς γε θεοὶ τελέουσιν ἄρειον / ἐλπωρή τοι ἔπειτα κακῶν ὑπάλυξιν ἔσεσθαι, "if indeed the gods will bring about a better old age at least, then there is hope that there will be for you an escape from evils", 286– 287). Peradotto has argued that Penelope's response should guide our interpretation of Tiresias' words: "In fact, the surest guide we possess upon which to base our own response to the prophecy is the response of Penelope when she hears it […] and that, as we shall see, simply reinforces the ambivalence of the prophecy with its own ambivalence and guarded conditionality".¹⁴

I largely agree with the characterization of the words of Teiresias and Penelope as ambivalent, and Penelope's words may display 'guarded conditionality'. One might wonder what Penelope is actually saying and how she says it (Tsagalis suggests that she speaks with "bitter irony"¹⁵). But Penelope clearly understands the words of Teiresias to predict a peaceful death at home in old age. As an internal reaction to the prophecy, this need not guide our response as an external audience. And in general I do not share Peradotto's pessimism about employing non-Homeric narratives to explore what an external audience, both ancient and modern, might make of her words, as well as those of Teiresias. The internal responses of Odysseus to Teiresias and Penelope to Odysseus may well coincide with the most natural and most common understanding of the prophecy, but this is not the only possible interpretation.

Let us remember that the Odysseus' report of the prophecy of Teiresias is actually provided by the main narrator. How does the main narrator expect an external audience to interpret an obscure prophecy spoken by one character to another, as reported by the second character to an internal audience? The narratological box-within-boxes arrangement would encourage an external audience to distinguish its reaction to the prophecy from that of internal audiences

See Peradotto 1990, 73.

See Tsagalis 2008, 71.

like Odysseus and Penelope. Just as an external audience might recognize that Odysseus is adjusting his story for his internal audience, the Phaeacians,¹⁶ it would also be able to separate its understanding of the prophecy from the hero's.

Knowledge of different accounts of the return of Odysseus and its aftermath would certainly complicate transmission of meaning between main narrator and the external audience.¹⁷ Many have long connected the words of Teiresias with the story of the Telegony, in which the son of Odysseus and Circe, Telegonus, arrives at Ithaca and unwittingly slays his own father (as in the summary by Proclus; cf. Apollod., Epit. 7, 36). With this story in mind, we might rather think that Teiresias refers not to death removed from sea, but death arriving out of the sea (ἐξ ἁλός, Od. 11, 134). Telegonus arrives from the sea, after all, and other sources, notably Apollodorus, indicate that the spear of Odysseus was tipped with the spine of a stingray – a creature of the sea.¹⁸ An ancient audience aware of a traditional story of death ἐξ ἁλός might conclude that "death from the sea" is the real meaning of the words, and that Odysseus (not to mention Penelope) has been fooled by the apparent meaning of death "removed from the sea".¹⁹

The Telegony's version of the hero's death is usually judged a ridiculous misinterpretation, or perverse manipulation, of Tiresias's prophecy. But the Greek actually favors the Telegony's interpretation. The preposition ἐκ with the genitive usually means "out of" not "apart from" with verbs of motion, and especially out of the sea in Homeric usage. That is why there was ancient controversy over the meaning of the words, as scholia to this passage indicate. The controversy persists in modern times, despite a tendency to ignore it. Merry and Riddell concluded that the seer's words can only mean "death out of the sea"²⁰ (and therefore an interpolation based on the post-Homeric Telegony). M. L.West concluded that the Odyssey was alluding to a pre-Homeric story of "death from the sea". For him, this is not the Telegonus story, but rather an excrementitious story employed by Aeschylus: a seabird's droppings containing a barb fall onto the hero's head.With the whimsy that we now miss,West deemed this tale "the incontinent

See Most 1989.

Marks 2008, especially chapters 2–4, well explains what seem to us as textual cruxes in the Odyssey as choice of alternative performance.

Actually this is a justified emendation of the text; the detail is often found in most surviving versions of the story. See Thompson 1947, 279– 281 for the main ancient sources about the stingray and its role in the Telegonus story.

See Cerri 2002, 153–159.

See Merry/Riddell 1886, 455. Pace Tsagalis 2015, 390 n. 80, the seeming counter example at Od. 15, 272 inherently contains the notion of moving "out of" before being "apart".

heron".²¹ Other scholars have also pursued alternative interpretations of Teiresias' words.²²

I myself do not argue that the Telegony predates the Odyssey, or predates a later crystalized form of the Odyssey, as Ballabriga²³ and some oralists would argue. Instead I think that the Telegonus story, or one very much like it, is pre-Homeric. West's preference for the Aeschylean version over the Telegonian version is unnecessary, but it could be that there were multiform versions of "death from the sea". Perhaps there was a folkloric motif, or riddle, about "death from the sea," that both the Telegony and Aeschylus variously employed.

The question remains whether the Odyssey employs this riddling motif also. There have been reasonable objections to the possibility of Teiresias' words implying "death out of the sea" as opposed to "death away from the sea." Teiresias' description of the death as "gentle" (ἀβληχρός, 135) does not seem appropriate for death by spear-point.²⁴ But if ἐξ ἁλός is best interpreted as "out of from the sea," despite our first impression, we should wonder if ἀβληχρός is also misleading. I have previously suggested that the poisonous nature of the spine of the stingray, well-known in antiquity, might lead to a slow or "peaceful" death.²⁵ But now I favor a more lexical solution. The word ἀβληχρός has a biform βληχρός; they both mean "weak, gentle."²⁶ Odysseus would understand the word to mean "gentle," but audience informed of the Telegonus story could hear it as "not-gentle," interpreting the alpha as privative.²⁷ Presumably the true meaning is to be realized after the fact, as often with prophecies. A belated realization, of course, is not narrated by the Odyssey, but an informed audience might consider it crucial to the meaning of Teiresias' words.What is absent becomes all the more proleptically significant.

Teiresias also states that the people about will be ὄλβιοι when Odysseus dies. If this means that the hero's people are "happy" while by the corpse, that suggests a gentle death,²⁸ but I take the phrase to refer to the prosperous state of the people ruled by Odysseus at the time of his death. Nagy sees the peo-

 The alpha-prefix of ἀβληχρός probably results from a preceding word's final alpha becoming attached to a preexisting βληχρός. See Reece 2009, 122–132.

See West 2013, 307–315 ("Excursus: The Death of Odysseus"), West 2014, 14 ("The Incontinent Heron"). West argues that the Telegony ascribes to the story found at Aesch., Psychagogoi, TrGF fr. 275.

See Burgess 2017a, 34 n. 28, with further bibliography.

See Ballabriga 1989, 297–301.

Hartmann 1917, 74 n. 69, 221.

Burgess 1995, 234 n. 70; Burgess 2001, 153–54. Cf. Ael., Nat. Hist. 1, 56; 2, 36, 50.

See Cerri 2002, 156.

So, Tsagalis 2015, 393 n. 89.

ple as "blessed" because of cult worship of Odysseus,²⁹ but this strains the significance ὄλβιοι and there is no evidence of burial of Odysseus on Ithaca.³⁰

Our concern is with ancient audiences, and in particular, audiences informed by non-Homeric accounts.³¹ Historical Ithacans of the Classical age, on the evidence testimony of the lost Aristotelian Constitution of the Ithacans, believed that Odysseus was exiled to Italy after the murder of the suitors.³² The judicial decision is made by Neoptolemus, king of northwest Greece, seeking wider control over the region. Apollodorus (Epit. 7, 40) provides a similar story, though here Odysseus travels to Aetolia, where he marries, has a child, and dies – in old age – apart from Ithaca. Some want to change Italy to Aetolia in our source, so as to harmonize the two similar tales, but Odysseus was often associated with Italy, and not just in the wanderings.³³ Ancient reports reference the burial and oracle of Odysseus in both northwest Greece and in Italy.³⁴ In the Telegony, the corpse of Odysseus is conveyed to Circe's island (Proclus), where apparently he is buried (so Hyginus, Fab. 127). It may be that the Telegony's Aeaea was the one localized with Monte Circeo.³⁵

Exile for Odysseus is non-Homeric, but in the Odyssey Odysseus asserts to Telegonus that he expects to be exiled, since that is the normal consequence of domestic murder (Od. 23, 117–122). Other passages are relevant: self-exile is pondered by Odysseus at Od. 20, 41–43 when he asks Athena where he should go in flight after the death of the suitors (Athena is evasive in her reply). At Od. 24, 430 –431 Eupeithes, father of the slain Antinous, suspects that Odysseus will flee to Pylos or Elis. At Od. 22, 55–58 Eurymachus proposes the type of compensation that is the counterpart of exile in the Aristotelian Constitution of the Ithacans. At Od. 24, 354–365 Odysseus tells Penelope that he will go into a kind of mini-exile at Laertes' farm, in expectation that report of the murders will spread quickly. At Od. 24, 352–355 Laertes expresses fear that the Ithacans will soon arrive after having sent messages to cities in broader Kephallenia (the region of the Ionian Islands in Homeric epic). Such passages assume consequences for the murders, including exile or self-exile; there is no expectation of the

Nagy 2013, 1. 11, especially section 40.

See Malkin 1998, 107–108.

See Lulli 2014 on cyclic poetry as "submerged" by its epichoric and non-canoncial nature, with good discussion of the concept of a "cycle".

Plut., Quaest. Gr. 14 (Mor. 294c–d) = Arist. fr. 507 Rose; 511, 1 Gigon.

Phillips 1953; see also Malkin 1998, 178–209.

Cf. schol. Lycoph. 799=fr. 508 Rose; 513 Gigon; ps.-Arist., Peplos 12, 13 lemma; Hyg., Fab. 127, with Burgess 2019.

See Burgess 2019; Burgess 2020.

supernaturally imposed peace at the end of the Odyssey. Some might conclude that the passages are ad hoc details of no importance that became mined in a textualist manner by the Telegony for an inventive narrative. It is more efficient, in my view, to conclude that these passages acknowledge the type of consequence that typically occurs in narratives of Odysseus' return to Ithaca.

There are therefore various possible reactions to Tiresias' prophecy. The most natural one is that Odysseus will die peacefully, in old age, on Ithaca, safely away from the sea. This is the meaning apparently accepted by Odysseus and Penelope, and it is also harmonious with the epic's ideology and desire for closure. Tiresias' apparent meaning is entirely consonant with the hero's centripetal quest for home, family, and rule. But an audience informed of the further adventures of Odysseus would think that Tiresias means something other than "death apart from the sea". The successful return of Odysseus to Ithaca is a strong climax, and it may be that continuations of the story were rather free to innovate, especially in respect of local interests. Certainly the Telegony and the epichoric stories do not agree with one another. But Teiresias' words would seem to suggest that the Telegonus story (not the Telegony itself) in particular precedes the Odyssey. The myth of Telegonus certainly exists outside of the cyclic poem. Sophocles' Ὀδυσσεὺς ἀκανθοπλήξ, "Odysseus struck by a spine," is one example; more recently Telegonus is featured alongside Homeric material in Madeline Miller's Circe (2018).³⁶

Yet audiences can be variously informed. Teiresias does not actually specify where Odysseus will die. Local audiences might instead have been pleased to think that the seer alludes to death of Odysseus at other locations. The small island of Ithaca, after all, is not the best place to die safely removed from the ocean. And it is not especially efficient to have Odysseus leave Ithaca and travel inland only then to return once again to Ithaca in order to die. The narratives of permanent exile could fulfill the apparent meaning of death apart from the sea. Exile and death in Aetolia, for example, well matches "The Sailor and the Oar" tale type that is seen to underlie Tiresias' "inland journey".

It is interesting that the cyclic Telegony – sometimes described as being based on epichoric tales of Northwest Greece – specifies that Odysseus does return again to Ithaca. This is despite divergent characterizations of the hero in the two epics: the Homeric Odysseus is a centripetal hero, always longing for home, whereas the cyclic Odysseus is a centrifugal hero, who leaves home again and makes a new home. Yet the Telegonus story remains centrifugal in its insistence that the corpse of Odysseus is removed to Aeaea. This detail at the end of the

On the Telegonus story, see Grossardt 2003.

Telegony is not as odd as it may seem to us, since there is no evidence that Odysseus was buried at Ithaca, as mentioned above. The Odyssey may seem to want us to think that Odysseus will be buried on Ithaca, but perhaps it could not make this claim explicitly.

Margalit Finkelberg has employed the term "meta-epic" to portray Homer as an agonistic poet who incorporates and subsumes other poetic traditions.³⁷ In reply, I have used the term "meta-cyclic" to argue that Homeric poetry often seeks to work "with" other contemporaneous traditions, such as those found in the epic cycle, rather than suppress them. It is in that spirit that I am exploring the possible variety of meaning in Tiresias' words. Justin Arft has recently well articulated this type of intertextuality in the Odyssey, stating that it "is able to invoke cyclic nodes […] as a matter of resonance or dissonance. When the dissonance invites contradictory outcomes, as it does with references to the death of Odysseus, the epic can either ignore, subordinate, or make use of the dissonant element(s)".³⁸

My focus is on aural reception of the poem, not the implicit intentions of the poem. I think that an ancient audience could make its own choices about whether to "ignore, subordinate, or make use of dissonant elements". And I would argue further that ignoring, subordinating, or using alternatives need not be mutually exclusive. An audience with alternative knowledge of Odyssean myth – say, the Telegonus story, or an exile narrative – should nonetheless be able to subscribe also to the Homeric portrayal of the hero, and to the Homeric arrangement of the story. Odysseus well describes the pleasure that arises in listeners of a bard (Od. 9, 1–11), and Eumaeus testifies to the enchantment produced by the bard-like story-telling of the disguised Odysseus (Od. 17, 514– 521). A contract always exists between a narrator and its audience, a certain willingness to suspend disbelief. We can easily imagine spellbound listeners of the Odyssey, even those informed of alternative legends, accepting to some degree the Odyssey's positioning of its closure. They might recognize the ominous and allusive ambiguity in Tiresias' words, yet still be happy, for a time, to hear the description of Odysseus' death as Odysseus does: a peaceful passing away in old age at home in Ithaca, safely away from the sea. Instead of advocating for a single solution to the riddle of Teiresias' prophecy, we might prefer to entertain the possibility of such a subtle response.

One might think here of Ruth Scodel's Listening to Homer, which on the basis of comparative anthropology argues that Homeric audiences would have readily

See Finkelberg 2015.

See Arft 2019, 172.

accepted the poet's ad-hoc presentation of "tradition".³⁹ Her argument tends to favor invention more than I consider likely within early epic, but it well describes the negotiation between bards and listening audiences. As modern scholars, we recognize with abhorrence the socio-political problems of the Homeric world, rooted as it is in warfare and a slave economy, but nonetheless employing a double consciousness we can also allow ourselves to fall under the spell of Homeric poetry. We should therefore be able to hypothesize different audiences of the Odyssey, of different place and time, each having their own type of reception. They could conceivably be enchanted by a performance of the Odyssey while aware of Odyssean myth that is inharmonious with the Odyssey. If Homeric audiences could sense a non-Homeric undercurrent beneath the words of Tiresias, they could still go with the tide of the great Homeric epic.

### References

Alden, M. (2017), Para-narratives in the Odyssey. Stories in the Frame, Oxford.


Buchan, M. (2004), The Limits of Heroism: Homer and the Ethics of Reading, Ann Arbor.


Scodel 2002. See now Hutchinson 2017 on the effect of Homeric verse on an audience.


## Silvia Quadrelli Traces of Orality in the Histories: The Homeric 'Heritage' in Herodotean Battles and Speeches

Abstract: The paper focuses on a specific aspect of the 'Homeric heritage' in Herodotus' work, and it is based on a 'functional' approach. Some Homeric elements, which depend on the cognitive and oral communication system typical of epos, can be found even in Herodotus and their presence could clarify certain aspects related to the composition and publication of the Histories. The Histories are a written work, but they still show characteristics of oral communication. These characteristics may depend both on the oral sources, used by Herodotus, and on a probable oral publication of the Histories. One element, which is related to oral communication, is the typical scene. In particular, the analysis focuses on the battle scenes: to describe the battles of the Histories, Herodotus uses recurring situations and elements, which can be reused in different contexts and combined in various ways. Another element related to oral communication is the function of Herodotus' speeches. The speeches of the Histories, like the Homeric ones, are not a rhetorical and literary exercise, but explain, motivate and generate the events: they are the 'driving force' of the action. This functionality of speeches reflects a still oral culture, in which spoken words have the same – and sometimes even greater – importance than the action. In conclusion, the comparison with the Homeric model highlights some 'traces' of orality, which influenced the composition of the Histories and could be connected to the oral context of publication, through performance, of Herodotus' work.

Keywords: Herodotus; typical scenes; battle; speech; orality; performance.

### 1 Introduction

Herodotus plays a leading role in the relationship between orality and writing in archaic Greece. Indeed, the 5th century is a period of cultural transition in which there is a dynamic coexistence between orality and writing, with mutual interactions that are often hard to retrace.

By this period, writing had already become part of the mechanisms of composition and transmission of texts. The length of the Histories, as well as the frequency of the verb γράφω, implies that the work was composed with the help of

OpenAccess. © 2022 Silvia Quadrelli, published by De Gruyter. This work is licensed under the Creative Commons Attribution 4.0 International License. https://doi.org/10.1515/9783110751963-008

writing. But the Histories still show distinctive features of oral communication, like ring composition, parataxis and digressions.¹ These characteristics may depend on the oral sources used by Herodotus² and on the probable oral publication of Herodotus' work.³ In the 5th century, orality was still involved in the diffusion of works. Although poetic compositions were more frequently developed in writing, they continued to be orally performed, as were the sophists' speeches.⁴

A comparison with archaic epic poetry can be helpful to clarify how the Histories were composed and published. It is important to bear in mind that historiography and epos shared the same audience: a heterogeneous one in terms of social background and, often, Panhellenic.⁵ Specifically, it is fundamental to draw a comparison between Herodotus and Homer. Homeric influences are recognizable in the narrative techniques as well as in the thematic and lexical choices of the Histories.⁶ In this work I focus on a specific aspect of the 'Homeric heritage' in Herodotus, namely the 'functional' approach. What is at issue is the hypothesis that some Homeric elements, which depend on the cognitive and oral communication system typical of epos, can be found even in Herodotus. Their presence could clarify certain aspects related to the composition and publication of the Histories.

See Lang 1984a, in particular 69, Rossi 1992, 99, n. 68, Bettalli 2009. On the influence of oral communication in historiography in general see Sbardella 2006, 56–58.

Murray 1987, Murray 2001 and Bettalli 2009, 51. On oral sources in Herodotus see Momigliano 1961–1962, Darbo-Peschanski 1985, Lang 1984b, Murray 1987, Asheri 1988, XXVIII–XXXIV, Evans 1991, Dorati 2000, 39. Fehling 1989 instead argues that Herodotus' sources are unreliable and represent a literary fiction.

In support of the idea of oral publication via public performances see in particular Jacoby 1913, Parke 1946, Canfora 1971, 658–660, Momigliano 1978, Lang 1984a, Evans 1991, Canfora/ Corcella 1992, 441–448, Thomas 1992, 123–127, Thomas 2000, Hornblower 1996, 25–28, Dorati 2000, 17–52, Bakker 2002, 28–29, Moles 2002, 33–34, Slings 2002, 63, Raaflaub 2002, 150 – 153, Nicolai 2004, Evans 2009, Oliver 2017. Powell 1939, Beltrametti 1986, Johnson 1994, Rösler 2002 do not agree with this hypothesis.

Oliver 2017 speaks of "wisdom performance". For the juxtaposition between Herodotus, sophists, Hippocratic writers and natural philosophers see Thomas 1992, Thomas 1993 and Thomas 2000. Nagy 1987, 177, on the other hand, places Herodotus among the logioi and considers him a prose aoidos.

On the indebtedness of historiography to epic poetry see Jacoby 1913, 502–504, Huber 1965, Strasburger 1972, Griffin 1990, Rengakos 2006a, Montanari 2006, Marincola 2007, Pelling 2006a and Rutherford 2012. On the audience of historiography and epos see especially Sbardella 2006, 56–58 and Porciani 2009, 25.

Among the numerous studies on Homeric influences on Herodotus see Caskey 1942, Asheri 1988, LX–LXI, Huxley 1989, Boedeker 2002, Montanari 2006 and Baragwanath 2012.

The first element that will be analysing is typical scenes – specifically, the battle scene. Typical scenes have been deeply studied over the last century, but the debate is still open.⁷ However, most scholars have noted the connection between typical scenes and the oral context of composition and publication of archaic epic poetry.⁸ Indeed, the epic singer could resort to a repertoire of themes, formulas and typical scenes that could be put together in different ways to create an extemporaneous composition.⁹ As Ercolani highlights, a performer would not always improvise: sometimes he would recite a poem by heart. According to this scholar's reconstruction, a series of episodes – a sort of 'narrative plot' – would first be memorized, and then recomposed through the use of formulaic techniques, with the addition of new elements and the omission of others, mainly based on the performance context and the audience's expectations.¹⁰

A typical scene, therefore, provides an aid for the aoidos in different narrative and performative contexts. Consequently, we should regard it not as a rigid structure but as a very flexible one. A typical scene can be adapted through expansions, digressions and reductions, which ensure a better link with the context, avoiding the impression that the scene is an extraneous 'block'. Its structure is so flexible, in fact, that it can even be adapted to a completely different context such as that of the Histories. However, we should keep in mind that the Histories are a prose work and hence that it is rather difficult to find word-by-word repetitions in them, of the sort which are instead quite common in verse compositions. In relation to this work, I use the expression 'typical scene' to refer to a set of repeated actions and situations that tend to be de-

Among the numerous studies on typical scenes see Arend 1933, Parry 1936, Bowra 1950, Lord 1951, Armstrong 1958, Lord 1960, Kirk 1962, 72–80, Dimock 1963, 50–57, Russo 1968, Fenik 1968, Gunn 1970, Nagler 1974, Richardson 1974, Edwards 1975, Edwards 1980, 1–3, Kirk 1985–1993 (in particular Kirk 1985, IX, Kirk 1990, 15–27, Edwards 1991, 11–23, Richardson 1992, 14–15), Edwards 1992 (who provides a summary of studies on the subject and an extensive bibliography), Reece 1993, Giordano 1999, 227–237, Gainsford 2001, Minchin 2001, 2–8 and 32–72, Friedrich 2002, Clark 2004 and Bakker 2004.

However not all scholars agree on the connection between typical scenes and orality: see, for example, Dimock 1963 and Hoekstra 1969.

Lord 1962, Lord 1985, Gentili 1984, Friedrich 2002, Ercolani 2006, 65–66; on the link between memorization techniques and typical scenes, catalogues and other compositional structures see Minchin 2001. On oral composition in general see Finnegan 1977, Sale 1996.

Ercolani 2006, 67. Ercolani points out the importance for the audience of a formulaic style, which facilitated learning, met the public's expectations, and increased the effectiveness of communication (Ercolani 2006, 165). On the performance and audience of Homeric poems see also Martin 1989, Martin 1997 and Bakker 1997.

scribed with similar words. These scenes are arranged on the basis of recurring content elements that provide a sort of traditional repertoire, and which can often be inserted or omitted without significantly altering the episode.

### 2 Typical Battle Scenes in Herodotus

The Histories present several typical scenes,¹¹ but the most frequent is undoubtedly the battle scene. The description of battles is quite frequent: in addition to clashes between Greeks and Persians, Herodotus also dwells on the wars that marked Persian expansion during the reigns of Cyrus and Darius. Like many other aspects of the Histories, the structure of battle scenes is influenced by the Homeric model.¹² Right from the proem of the Histories, Herodotus sets the war between the Greeks and the Persians in continuity with the Trojan war: it is a conflict between East and West, between Greek civilization and barbarians. Homeric battles, however, differ in certain respects from the historical reality of the Persian wars, in which hoplite tactics and great mass fighting prevail. So, Herodotus adopts a new kind of prose description for battles, to represent a more complex type of war than the one portrayed in epic. Herodotus focuses on an overview of the events occurring during each battle and omits certain details that are typical of epic narrative: the bloody details of Homeric death scenes, for example, are replaced by a technical description of what is happening, with references to the type of weapon and the parts of the body affected. Nevertheless, there are Homeric echoes in the language, narrative structure and descriptions of Herodotean battles.

In addition to the Homeric model, the structure of battle scenes is deeply influenced by the nature of Herodotus' sources, which is to say – as Tritle points out¹³ – accounts by veterans. Indeed, oral traditions are generally characterized by recurring structures, anecdotes and typical elements, which are also important elements in Herodotean battle scenes.

I would argue that Herodotus describes most battles using a typical structure made up of recurring situations and elements, which can be reused in different contexts and combined in various ways. Within this structure, which can be defined as a battle scene, there are certain Homeric elements, such as lexical or thematic references, or even typical Homeric situations, such as catalogues, an-

See Quadrelli 2017 and Quadrelli 2019 concerning the presence of seduction, dressing and arming scenes and dream scenes in the Histories.

More recently see Marincola 2018.

Tritle 2006, 214.

droktasiai and fights over a corpse. In addition, in Herodotean battles there are recurring motifs¹⁴ that do not have a Homeric antecedent, but which are typical in the Histories.

To highlight the frequency and the different possible combinations of recurring elements, some motifs have been collected in the following table:¹⁵

The term "motif" in this work will be used as a synonym for typical or recurring element. It has not been possible to include all battle scenes and recurring elements from the Histories in the table. I have chosen to analyse, in addition to the most important battles between Greeks and Persians, the episodes with most recurring elements.


Tab. 1:

Recurring

 elements

 and

situations

 in Herodotean

 battles.


:Recurring elements andsituations in Herodotean battles.(Continued)

Tab.

Traces of Orality in the Histories 123

In the table it is evident that the most important battles (Marathon, Thermopylae, Artemisium, Salamis and Plataea) have the highest number of recurrent elements. Similarly, certain motifs, such as the holding of a council or androktasiai, recur in almost every episode analysed. To describe battles, Herodotus uses a series of more or less frequent elements and modifies their order according to the narrative context. This mechanism of composition is similar to that of typical battle scenes in Homer.¹⁶

As one would expect, all Homeric battle scenes are in the Iliad; in order to avoid the monotony of excessive repetitiveness, it is necessary to describe battles in a typical yet, at the same time, varied way. In this respect, Bernard Fenik has pointed out that the typical battle scene differs from other typical scenes, such as arming, departure or arrival scenes, which consist of repeated actions essentially described with the same words: battle scenes are characterized by typical details but differ from one another.¹⁷ Therefore, according to Fenik's interpretation, the bard had a series of elements at his disposal, such as formulas, verses, typical details and recurring situations, to create battle scenes. Despite this apparently mechanical composition, the episodes in question also present non‐typical details and different combinations of single elements which help to avoid excessive repetitiveness.

It is possible, therefore, to analyse some recurrent elements in the Histories and – at least in certain cases – to point to their relation to the corresponding typical situation in Homer. Specifically, this kind of analysis focuses on catalogues, androktasiai and aristeiai.

Herodotus often dwells on the description of armies about to fight and their formation.¹⁸ The model here is the Homeric catalogue and especially the Catalogue of Ships (Il. 2, 484–770), which can be considered a typical scene for its structure and recurring language.¹⁹ The information provided for each army, following Powell's schematization, can be organized into three different narrative patterns: the names of the nation and its generals, its place of origin and the

On typical battle scenes in Homer see Kirk 1962, 75–80, Beye 1964, Hainsworth 1966, Fenik 1968, Armstrong 1969, Latacz 1977, Tsagarakis 1982, 104–133, Thornton 1984, 86–92, Kirk 1985– 1993, Niens 1987, Van Wees 1986, Van Wees 1988, Edwards 1987, 241–244, Kirk 1990, 21–26, Van Wees 1997, Hellman 2000, Mueller 2011, van Wees 2011.

Fenik 1968.

It is difficult to distinguish between catalogues and formations in the Histories, which is why the two are grouped together in the table above.

For some studies on Homeric catalogues see Beye 1964, Powell 1978, Edwards 1980 and Kirk 1990, 168–177.

number of ships.²⁰ Upon initial analysis, the catalogue of ships seems to display a rigid and repetitive structure. However, if we examine the scene in its individual elements, it emerges that not all of the information is provided for each army and, above all, that the order of the basic elements (nation, place and generals) varies significantly. Therefore, here too the typical scene is not a collection of verses that are repeated verbatim or with minimal variations, but rather a recurrent structure with typical elements that can be combined in various ways.²¹

In the Histories too there are several catalogues.²² The most famous and bestdeveloped one is the review of Xerxes' troops in Doriscus, at the beginning of the expedition against Greece (7, 60 –99).²³ For each contingent Herodotus gives some information about the soldiers' nation, equipment (including costumes and weaponry) and the commanders' names, along with some minor details, such as the origin of the nation or of its name, often in the form of short anecdotes. All these elements feature simultaneously in the description of the Persian contingent, which is the first to be presented (7, 61); for the other contingents Herodotus provides only partial information, but he always highlights the name of the Persian commanders. The most interesting aspect of the comparison between Herodotus' work and the Homeric poems is the presence of a recurring structure in all catalogues, and especially in the two most important army lists. Although there are no specific lexical references in Il. 2, 484–770 and in Hdt. 7, 60 –99 the information about nations and commanders is organized according to the same narrative pattern and the description is often enriched with anecdotes. The structure is therefore the same, but the amount of information and its arrangement differ, both between the two works and across different catalogues within the same work. The hypothesis that Herodotus may be consciously drawing upon Homeric typical scenes is reinforced by the details that the historian provides with regard to contingents. Indeed, some troops are described as having equipment more suitable for a military parade than a real expedition. As Vannicelli

Powell 1978 also focuses on the lexical aspect of the scene, further highlighting its typicality. In Homeric poems there are other catalogues (Il. 3, 161–244, 16, 168–197) in addition to the Catalogue of Ships, and there are also non-military catalogues (Il. 5, 382–404, 14, 315–328; Od. 11, 225–329).

As in the Homeric poems, in the Histories there are also some non-military catalogues: the catalogue of Darius' districts (3, 89–97) and the list of peoples and tribes ruled by Darius (5, 49, 5–7).

On Herodotean catalogues in general and on the catalogue of Xerxes' army in particular, see Erbse 1992, 125–127 and Vannicelli 2013. See especially Armayor 1978, Vannicelli 2017, Nicolai/ Vannicelli 2019 for a comparison between the catalogue of Xerxes' army and the Catalogue of Ships.

points out,²⁴ visual evidence seems to suggest that Persian soldiers were not actually equipped with bow, spear and shield at the same time. The Herodotean catalogue, therefore, turns out to be a typical scene designed to meet the expectations of the audience, rather than a real description of troops.

In the Histories there are other, shorter catalogues that include elements typical of this kind of scene: the number of men or ships (for single contingents and in total), the names of the most important generals (usually Athenians and Spartans), and anecdotes.²⁵ In the table above, catalogue and formation are listed together because the two types of scene have similar characteristics and are often woven together. Formation scenes, like catalogues, are characterized by elements that can be combined in various ways:adescription of the setting (Herodotus often refers to a river), anecdotes and the position of a particular contingent, with an indication of the enemy contingent against which it is being deployed.²⁶ Some passages, moreover, can be considered a mixed kind of scene since they have elements typical of both formation descriptions and catalogues.²⁷

In battle descriptions Herodotus often refers to heroic actions of a sort that is typical for the Histories. One element already present in Homeric battles scenes and which recurs in the Histories too is 'few against many'.²⁸ Herodotus highlights how a few Greeks will often fight against a much larger army, but this motif is present even before the account of the Persian wars, for example in the Battle of Pteria (1, 77) and in the battle between Lycians and Persians (1, 176, 1). The Greeks differ from the barbarians not only in number – they are consistently less numerous – but also for the way they fight: another recurring element in the Histories is the disorder of barbarians.²⁹ These two motifs play a secondary role in the Homeric poems but they often recur in the Histories; consequently, it is possible to point out that while some typical situations have a Homeric antecedent, Herodotus develops them independently to create new topoi in the Histories.

Two significant recurring elements are illustrated in the table above using the terms for the corresponding Homeric typical scenes: androktasiai and

Vanicelli 2013. See also Armayor 1978.

Hdt. 7, 202–204; 8, 1–2; 8, 43–48.

Hdt. 1, 80; 5, 110; 6, 111; 8, 85; 9, 46–48.

In Hdt. 6, 8, for example, the author describes the soldiers' formation and mentions the number of ships; at 9, 28–32 the scene is more complex and we also find the names of leaders and some anecdotes.

Cf. Il. 2, 119–133; see also Nicolai/Vannicelli 2019.

Cf. Il. 3, 1–9; 4, 427–438.

aristeiai. At the end of almost all battles, Herodotus lists the names of valiant men – sometimes these are not individual warriors, but entire nations – and also names some illustrious dead combatants. The most important battles end with a reference to those who stood out, according to the spirit of epic aristeiai.³⁰ Here too Herodotus 'inherits' a typical Homeric situation, but he develops it in a different way because he must adapt the scene to a new context. Indeed, as one would expect, in the Herodotean work the aristeia is less developed than in the corresponding Homeric scene and usually it consists of a simple list of names, sometimes associated with an account of some memorable actions. Unlike in the Histories, in the Iliad the aristeia is one of the most extensively developed typical situations,³¹ since it focuses on the glorious deeds of a hero, which are the central theme of epic poetry. In short, Homeric aristeiai can be divided into five stages: arming, fights preceding the hero's wounding, wounding and recovery of the hero, duel and epilogue.

In the Histories this scene is reworked and often consists of a simple list of names, because it must be adapted to a new context: Herodotus' account is no longer centred on the brave actions of a single hero, but on the effort of many warriors fighting for their freedom and their people. Nevertheless, the pattern of the aristeia is recognizable in a key episode of the Histories that many scholars consider to be the most Homeric of all: the Spartan feat at Thermopylae.³² The passage could be divided as follows:


In some passages (7, 227; 9, 105) Herodotus actually uses the verb ἀριστεύω.

On Homeric aristeia see, in particular, Krischer 1971 and Thornton 1984.

Gainsford defines the Spartans' feat at Thermopylae as aristeia and he points out that, as in the Homeric scene, the action begins with the arming and preparation of the Spartans and ends with a defeat. In addition, in the episode there are many references to the epic theme of κλέος (Gainsford 2013). For other Homeric influences on the description of the battle of Thermopylae see Dillery 1996, Pelling 2006a, Foster 2012, Carey 2016, Marincola 2016, Vannicelli 2017.

The battle of Thermopylae can therefore be considered a real aristeia, unfolding from the arming of hero to his death. Unlike the Homeric scene, the episode concerns not a single hero, but the whole Spartan contingent. The adaptation of the scene reflects a change in the actual way of conceiving war: the outcome of the war no longer depends on the glorious deeds of the kings and heroes who are the protagonists of epos, but on armies of citizens.

The account of Thermopylae continues, like that of other battles, with what has been defined as an androktasia, that is a list of fallen combatants who are worthy of remembrance. This typical situation is similar to the aristeia – a list and some anecdotes – and it is characterized by elements already present in Homeric androktasiai.

In the Iliad, the androktasiai follow a pattern that, according to Charles Beye,³³ we might call the ABC schema, in which A stands for basic information, such as the names of the slain man and the slayer, B for the anecdote and C for the contextual information, such as the description of how the man is slain or what happens after his death. As for the catalogue, of which the androktasiai could be considered a subtype, the three recurrent elements (ABC) are not present in every scene and always in the same order; rather, there are a great variety of solutions, depending on the narrative context.

For an analysis of Herodotus' androktasiai, anecdotes – that is, element B in Beye's scheme – are particularly noteworthy. Beye detects some recurring themes in the anecdotes, such as the social rank and wealth of the hero, his birth, place of origin, marriage, flight to avoid blood vengeance and a seer's prophecy.³⁴ However, there are androktasiai in which slain men are listed exclusively by proving the name of their father and their place of origin. So, once again, a simple list conceals the possibility of combining recurring elements in various ways, so as to make each scene different.

In Herodotus' catalogues of the illustrious dead it is possible to recognize both the structure identified by Beye and many themes characteristic of Homeric anecdotes. One example is the androktasia following the battle of Thermopylae (7, 224, 227), characterized by several elements typical of Homeric androktasiai: the slaying of two brothers, their father's name and (at 7, 224) a brief anecdote about marriage, even though Herodotus is referring to their father Darius' marriage. It is relevant that Herodotus uses the verb ἀριστεύω at the beginning of the list of men worthy of being remembered (μετὰ δὲ τοῦτον ἀριστεῦσαι λέγονται Λακεδαιμόνιοι δύο ἀδελφεοί). At Thermopylae it is impossible to distinguish the

Beye 1964, 347. Beye classifies androktasiai as a type of catalogue.

Beye 1964, 358.

best warriors from the illustrious dead, aristeiai from androktasiai, since all Spartans were killed. This detail highlights the similarity between the two typical situations in the Histories:³⁵ in both cases, Herodotus lists names in a typical way and often adds the name of the father, the place of origin and an anecdote. Finally, androktasiai sometimes consist of a simple reference to the number of fallen soldiers, with an essential and stylized reduction of the whole to one of its component parts. This element, however, is particularly important because in the Iliad androktasiai are a type of catalogue, and numbers play a key role in catalogues.

In the Histories there are several typical situations that recall the Homeric model and Herodotus often develops them in a new way. For example, we may examine the passages concerning the deaths of Leonidas (7, 224–225) and Masistius (9, 22–23). In both cases the armies fight to retrieve the bodies of their leaders and both episodes recall the struggle over Patroclus' corpse (Il. 17, 1; 18, 238). The Herodotean scenes are less developed, but the Homeric influence is evident, especially if we analyse the contents of the episodes. Herodotus may be describing a fight over Masistius' corpse that really happened, but the struggle over the body of Leonidas is implausible in the historical context of the battle of Thermopylae. The Three Hundred were aware that they could not defeat the Persian army and that, at the end of the battle, Leonidas' corpse would fall into enemy hands.³⁶ The fight over the body, however, makes sense if it is considered a typical situation, as is also suggested by the Homeric elements that recur throughout the episode.³⁷ This motif inherited from epic poetry becomes typical in the Histories, as evidenced by the struggle over Masistius' body, a passage that, like the previous one, is deeply influenced by the Homeric model.³⁸

So in the Histories there are many recurring elements, but some have no Homeric antecedent.³⁹ Herodotus inherited certain elements from Homeric battle scenes and used the same mechanism of composition to create new recurring motifs that became typical in the Histories. One element that seems specifically

In the table above the passages considered to be aristeiai and androktasiai occur within a short space from one another and sometimes coincide.

Tritle 2006 instead contends that the struggle over the body of Leonidas is not a literary fiction.

The fight over Leonidas' corpse is part of an androktasia, Leonidas and the Three Hundred pursue kleos (cf. 7, 220, 2) and, finally, the Spartans drive back the Persians from Leonidas' body four times (7, 225, 1), just as Hector attempts to wrest Patroclus' corpse away from the Achaeans three times (Il. 18, 155–156). For other Homeric elements in the struggle over Leonidas' body see Flower 1998, Pelling 2006a, Foster 2012, Gainsford 2013, Carey 2016.

For Homeric influences in this episode see Masaracchia 1978, 162.

Some non-Homeric elements can be found in the second part of the table above.

'Herodotean' is the number 'three hundred'.⁴⁰ The Three Hundred are by definition the Spartans led by Leonidas, but this number occurs in several episodes in which acts of bravery are accomplished by a small group of warriors:

1. 1, 82: in order to settle the outcome of the war, at Thyrea three hundred Spartans fight in a duel against three hundred Argives. In the end only three warriors survive, two Argives and a Spartan, and the two Argives, believing themselves to be the victors, return to Argos, whereas the Spartan despoils the enemy corpses and wait at his position. The next day, both armies claim the victory, there is another battle and the Spartans win.

2. 9, 21–23: during the battle of Asopo three hundred Athenians accept the call for help of Megarians and these three hundred picked warriors fight over the body of Masistius.

3. 9, 64: Herodotus recounts an anecdote about Mardonius' slayer, the Spartan Aeimnestus: during the third Messenian war, he leads three hundred warriors and dies in battle with them.

4. 9, 67: at Plataea, after the Persian retreat, the Thebans continue to fight and the best three hundred of them are killed by the Athenians.

There is a considerable amount of evidence regarding the existence of units of three hundred Spartans,⁴¹ which would explain the number of men at Thermopylae, Thyrea and in the Messenian war. As far as the three hundred Thebans at Plataea are concerned, we know of units of three hundred men at Thebes, but probably after the Persian war. In the case of the Asopo episode, the issue is more complex because the existence of units of three hundred hoplites in Athens is an unproven hypothesis.⁴² Probably based on the existence of an actual Spartan unit, Herodotus uses the number three hundred as a typical element: whenever some valiant warriors stand out, often dying in battle, their number becomes three hundred. The typicality of this number is evident especially when it comes to the battle of Asopo. Indeed, we have no information about this unit of men in Athens and, furthermore, the scene culminates in the fight over Masistius' body and recalls the same typical situation as Termopylae.

So, in order to describe the many battles of the Histories, Herodotus uses recurring situations and elements that can be reused in different contexts and combined in various ways. The Herodotean battles are all different, yet at the same time similar to one another. Events are placed within a structure inherited from epic tradition, and which becomes typical in the Histories: a structure con-

For the number 300 in Herodotus and in historiography in general see Ruffing 2013.

In addition to the Herodotean evidence (1, 82, 3; 7, 205, 2; 7, 224, 1; 8, 124, 3; 9, 64, 2), see Daverio Rocchi 1990 and Vannicelli 2017, 557–558.

See Asheri 2006, 200 –202.

sisting of Homeric elements, Homeric typical scenes and new motifs introduced by Herodotus.

In conclusion, if we accept the extension of the meaning of the concept of typical scene proposed by Fenik,⁴³ we can affirm that the Histories feature typical battle scenes. Indeed, Herodotus adapts certain elements of the epic tradition to the new context of prose historiography and uses the same mechanism of composition to create new recurring motifs that become typical in the Histories.

These mechanisms of composition are related to the oral composition and publication of epic poetry: their presence in the Histories may reflect a close relationship between Herodotean historiography and orality, especially when it comes to the publication of Herodotus' work. In any oral performance, the relationship between the speaker and the audience is fundamental. Above all, it is important for the speaker to satisfy what has been called the "horizon of expectations" of the listeners. An audience used to epic performances and typical scenes will expects some episodes to be narrated in a specific way: typical elements and situations are a sort of signal used by the performer to orient his audience and establish a reassuring and effective mode of communication. The presence of typical scenes in Herodotus' work is not merely the echo of a literary model; the use of composition mechanisms that take the listeners' horizon of expectations into account suggests an oral publication of the Histories.

If we acknowledge that Herodotean historiography is connected to the oral communication system, we can also better understand the function of Herodotean speeches.

### 3 Speeches in Herodotus: Words as the 'Driving Force' of the Action

In the 1980s, Mabel Lang published an interesting work on Herodotean speeches. She analysed them in relation to Homeric ones, and pointed out that they have three main functions: to motivate the following action, to explain how and why events occur and to prefigure something that will happen.⁴⁴ For Lang, these functions are related to the oral style that derives from the context of composition of the Histories and from the oral nature of Herodotus' sources. Speeches introduce new narrative material in a 'natural' and realistic way, and they are functional to the narration of events like actions, as Pelling states:

Fenik 1968.

Lang 1984a. I used the same terms ("motivate", "explain" and "prefigure") used by Lang.

"for an important sense speeches are action. They play their part – often initiating often responsive – in a chain of events".⁴⁵

This aspect reveals a close link with the function of Homeric speeches: indeed, several scholars have dwelled on the relationship between speech and action in the Iliad and Odyssey. In particular, Bark argues that in the words-action combination both components are equally important, whereas Larrain, examining the first eight books of the Odyssey, acknowledges that the main function of speeches is to recall what has happened or to anticipate what will happen. For Richardson speeches are an integral part of the action and not merely a support. Finally Dentice di Accadia Ammone argues that in Homeric poems the spoken word seems to be more important than the action and that speeches contribute to the plot development.⁴⁶

In the case of Herodotus, as in that of Homer, the analysis of speeches can be centred on the relationship between words and action: the origin of this link, as Lang points out, may lie in a context of oral communication.⁴⁷

In oral cultures, the spoken word has a much more concrete function than in our own, text-based culture; it has the power to evoke action and, at the same time, to make it visible in the listener's mind.⁴⁸ In a sense, words do not describe the action but create it through a magical-creative power. They have an almost sacred role: oaths, prayers to the gods, oracles and supplications are gestures based on words but they invariably generate an action. An oath requires a given behaviour, a prayer to the gods never remains unfulfilled and generates an action in the divine world, supplication implies a reaction, and oracles – whether they be interpreted correctly or not – influence the actions of those

Pelling 2006b, 103.

Barck 1976, Larrain 1987, Richardson 1990, 70–88, Dentice di Accadia Ammone 2012, in particular 300–301.

Zali 2015 argues that "Homeric influence on Herodotus in the extensive use of direct speech and the diversity of speeches goes without saying, and much of the material presented in the Histories belongs to the archaic period, largely dependent on oral traditions which give the Histories its distinctly conversational colour". Pelling 2006b, 103 instead draws a relation between the frequency of direct speech and the probable oral fruition of the Histories.

For Zumthor 1984, 35, 62 writing involves the separation between thought and action, whereas the voice, in its primary function before the influence of writing, does not describe but acts. See also Canfora 1971, 653 and Aloni 1998. On the role of the spoken word in ancient theatre see Giovannelli 2015, 45–46: "come in ogni forma di racconto orale, nel teatro greco era la parola a condurre lo spettatore attraverso il mutare dei luoghi e delle condizioni atmosferiche. […] ciò che viene nominato assume esistenza senza difficoltà". See also Waters 1966, 157, who highlights the link between Herodotus' speeches and the epic tradition in terms of their dramatic aspect.

who hear them. So, in all these cases the spoken word generates an action but, at the same time, is an action itself.⁴⁹

In an oral performance, moreover, when the singer gives the floor to his characters, there is an identification of the performer with the protagonist of the narration;⁵⁰ an identification that in the Odyssey becomes a narrative device in Odysseus' account of his wanderings,⁵¹ increasing the audience's involvement.⁵² At the same time, the words spoken by the character/performer are functional to the plot development. Kirk even affirms that Homeric poems are "narrative expressed as drama", since action – just as in theatre – is evoked in conversations between characters rather than in the account provided by the poet-narrator.⁵³ The spoken word explains, motivates, invites to act and, in some way, constitutes the 'driving force' of the action, because it pushes it forward. If we try to analyse Homeric poems from the point of view of the function of speeches, we can see that the main themes of the Iliad and the Odyssey, that is the wrath of Achilles and the return of Odysseus, unfold mainly through speeches. The words uttered by the characters are the 'driving force' of the actions related to the specific mythical episodes of the Homeric poems. In other words, the most important events for plot development are not narrated in the third person by the omniscient narrator, but are recounted directly by the characters, who – in their dialogues – explain the situation and motivate the following action.

In order to understand the function of speeches, we can analyse the case of the Odyssey, in which speeches are most important, especially considering that from book IX to book XII Odysseus himself describes his wandering to Alcinous in the first person. The dialogues during the two councils of the gods (Od. 1, 28–95 and Od. 5, 1–43) explain Odysseus and Telemachus' situation: the former is a prisoner on Ogygia, the latter is in danger of an imminent ambush by Penelope's suitors. The words spoken by the gods at the two councils trigger the departures of Odysseus and Telemachus and foreshadow the hero's arrival in the

For Zumthor 1984, 333, in an oral culture the voice traces the action that follows and, consequently, it is the action itself. Ong 1986 also argues that in oral cultures the spoken word is a means of action and is inseparable from the event it evokes.

Richardson 1990, 88: "Homer's storytelling was a live performance. The performer was a narrator of events and an imitator of characters, and he held those two roles distinct when rendering speech in his discourse". For Cantilena 2002, 33 "il poeta epico non si limita a 'dire' i discorsi dei suoi personaggi, ma presta loro voce, espressività, carattere".

Martin 1989.

See Ong 2002, 45, who considers this identification typical of the empathetic and participatory nature of oral thought.

Kirk 1990, 28–35.

land of Phaeacians. After Odysseus' stay at Alcinous' court and his arrival on Ithaca, the dialogue between him and Athena motivates and generates the hero's choice to disguise himself as a beggar and his arrival at the palace. Finally, the speeches between Odysseus and Penelope also play a central role: in the first encounter (Od. 19, 535–553) there is a description of the dream that prefigures the slaughter of the suitors, in the second (Od. 19, 570 –581) Penelope's words motivate and generate the bow contest, and in the third (Od. 23, 163– 230) the recognition between Penelope and Odysseus is realized through the well-known 'bed test', entirely described by characters themselves.

So, Homeric speeches mainly have the function of explaining facts and of motivating and anticipating, or even generating, events that have yet to occur.

Herodotus' speeches have the same function. Some of them explain not only what is happening, but also how and why it is happening, and they can be considered the real 'driving force' of the action: they give essential information for the development of the story and, in this way, the action unfolds. These plot developments are not narrated in a third-person narrative; consequently, the words spoken by the characters are necessary to understand the sequence of events.

Some important events concerning the war against the Persians are explained, motivated and generated by speeches: in 3, 134, for example, the words that Atossa addresses to Darius motivate and generate the Persian expedition against Greece. The dialogue between the two characters also induces Darius to send observers to Greece, for a reconnaissance that anticipates and prepares the invasion. Similarly, Themistocles' words to Eurybiades in the assembly (8, 59–62) and his following speech, reported by the messenger Sicinnus (8, 75), generate the battle of Salamis. Finally, the dialogue between Xerxes, Mardonius and Artemisia (8, 100 –102) motivates and generates the king's return to Persia and the consequent election of Mardonius as leader of the army.

The same function can be detected in some speeches related to another of the main themes of the Histories: μεταβολή. One example is the episode of Gyges and Candaules (1, 8–12), which is emblematic of μεταβολή: Candaules' words explain the origin of the whole story (1, 8, 2) and the plot through which the bodyguard is able to see his wife naked (1, 9). Candaules' wife's revenge too is motivated and generated by her speech to Gyges (1, 11, 2–5), which – as 'driving force' of the action – causes the murder of Candaules and the following rise to power of Mermnadae in place of Heraclidae. Another important episode related to the theme of μεταβολή is the Persians' takeover, with Cyrus, to the detriment of the Medes. Speeches play a central role in the account of Cyrus' childhood, which is spared a certain death when he is still an infant: Astyages' speech to Harpagus (1, 108, 4) foreshadows the ruin of the Persian noble and his future alliance with Cyrus; Harpagus' words to his wife explain why he does not intend to kill the child (1, 109, 3) and his speech to the cowherd describes how the child should be exposed and killed (1, 110, 3); the cowherd, talking to his wife, tells her how the child was given to him and how he discovered his true identity (1, 111, 2–5); finally, the woman's reply generates the most important action of the episode: the stratagem to save Cyrus (1, 112, 2–3).

Speeches also generate the main action of the Histories: the Persian invasion of Greece. These speeches occur during the famous Persian council (7, 8–11): following Mardonius' advice (7, 5), Xerxes decides to invade Greece and holds an assembly of the noblest Persians to inform them of his decision. The king explains the reason for the expedition, namely the need to increase honour and power (7, 8α) and to punish the Athenians (7, 8β). Artabanus' speech focuses on the recurring theme of μεταβολή and highlights the risks of the expedition, foreshadowing Xerxes' defeat (7, 10). Finally, the king's reply generates the invasion of Greece.

The words spoken by Herodotus' characters, therefore, look to the past (explaining the causes of events), to the present (motivating the action) and to the future (prefiguring what will happen), according to a framework reminiscent of epic poetry.⁵⁴ The function of Herodotean speeches could reveal a link with the Homeric poems and, more generally, with archaic poetry, closely related to a context of oral performance. Consequently, speeches could be considered a component of the orality that probably characterized the publication of the Histories: their role as a 'driving force' of the action is not a stylistic 'legacy' of the archaic poetic tradition, but an aspect that is functional to the oral performance context. For an audience used to epic recitals and theatre, the words spoken by characters highlight important narrative passages and push the action forward. Therefore, the use of speeches in the Histories is designed to meet the expectations of an audience of listeners.

### 4 Conclusions: A Comparison With Thucydides

The presence of typical scenes and the function of speeches are clues to reconstruct the mechanisms of composition and publication of the Histories. Herodotus' work is certainly a literary composition, reflecting the mechanisms of written texts, but it was conceived for oral performances, that is for public readings. Therefore, the composition may have been influenced by the cognitive system of oral communication.

Since he is inspired by Muses, Hesiod knows τά τ' ἐσσόμενα πρό τ' ἐόντα (Hes., Theog. 32).

In support of this argument, a comparison with Thucydides may be useful. In the proem the Athenian historian announcesadifferent destination for his work (Thuc. 1, 22, 4):⁵⁵ he states that his work will be ἐς μὲν ἀκρόασιν ἀτερπής; consequently, the elements that we considered 'traces' of orality in Herodotus should be absent in Thucydides' work.⁵⁶

Thucydides' battle scenes are different: there is a tendency to avoid the detailed narration of military events and the description usually consists of a list of expeditions and cities besieged or conquered. According to Ferrucci,⁵⁷ some episodes – such as the siege of Plataea and the civil war in Corcyra – have a universal significance and serve as paradigms. The following episodes similar to the paradigmatic ones are not described in as much detail: the mechanisms of Thucydides' narrative are opposite to the 'search for what is typical' in Homer and Herodotus.⁵⁸ For example, we find no typical catalogues, with names of leaders, numbers, places of origin and anecdotes. These are replaced with partial catalogues which are often integrated into the narrative, meaning that they do not appear as lists.⁵⁹ Sometimes Thucydides mentions only a commander, referring to the other leaders as "his colleagues", and reports either only the numbers of certain contingents or only the total number of men, without distinguishing between contingents. Any catalogues found are coherent with the intentions declared in the proem: Thucydides avoids the mythodes, in the sense of epic heritage, necessary to make the listening experience more pleasurable, and prefers what is useful for the historical narration. So, Thucydidean catalogues are not typical, because they do not have to meet the horizon of expectations of the audience.

The same goes for speeches and their function. Thucydidean speeches have been extensively studied, but few studies dwell on their relationship with narrative sections.⁶⁰ As regards Herodotean speeches, I have highlighted the relation-

See in particular Canfora 1971, Ferrucci 2009, 76, but the debate is still open.

The subject is treated here in summary, to provide an overview, but it should be examined in depth. The bibliographical references are a selection of the numerous studies on the relationship between Herodotus and Thucydides and on Thucydides' speeches.

Ferrucci 2009, 88.

Nevertheless, it is possible to underline the presence of an element already present in Homer and then expanded by Herodotus: the 'few against many' motif (Thuc. 2, 76, 3; 2, 100, 5; 4, 36). The macrostructure 'exhortation-catalog/formation-battle', instead, is to be considered a topos of historiography since its individual parts are not typical.

Thuc. 1, 29, 4; 1, 46, 1–2; 1, 60; 1, 61, 1; 1, 94; 1, 107, 5; 1, 117, 2; 2, 9; 2, 31; 2, 56, 2; 2, 66; 2, 79; 2, 80, 5–7; 2, 98; 3, 17, 1–2; 3, 100, 2; 4, 42, 1; 4, 53, 1; 4, 70, 1; 4, 93, 3–4; 4, 124, 1; 4, 129, 2; 5, 2, 1; 5, 6, 4–5; 5, 57; 5, 84, 1; 6, 31, 2; 6, 43; 6, 98, 1; 7, 19, 3–4; 7, 20, 2; 7, 33, 1; 7, 57–58; 8, 25, 1. De Romilly 1956, Hunter 1973, Morrison 2006, Rengakos 2006b.

ship between words and action that derives from the archaic Greek culture, still closely related to the mechanisms of oral communication. In Thucydides, speeches do not have the function of explaining, motivating and anticipating the action, and even when they explain, motivate and anticipate things, they do not replace the narrative sections. Indeed, much of the information contained in speeches is anticipated or repeated in Thucydides' third-person narration. Sometimes, the account of events even contradicts what has been said in speeches. With regard to speeches, in the proem Thucydides specifies that he will report what characters may have said in given circumstances (1, 22, 1–2). There is no identification between narrator and character; rather, the author highlights the verisimilitude of his speeches.⁶¹ The spoken word loses the objectivity and concreteness that characterize it in an oral culture. Speeches have a different function in Thucydides: characters' words are not the 'driving force' of the action, they do not replace it, but they provide an interpretation of it. Evidence of this is the presence of pairs of speeches that give opposite interpretations of the same situation.⁶² Two opposite speeches cannot push the action forward: on the contrary, the action stops, and the focus becomes the comparison between two different points of view and two different interpretations of facts. Thucydides makes rhetorical use of speech, which becomes an interpretative instrument; therefore, speech is accessory to praxeis and does not coincide with action.

So, there is no close link between words and action and, although they start from a particular and concrete event, Thucydides' speeches often concern universal themes. Even in the Histories there are universal and philosophical speeches, such as the constitutional debate and Artabanus' speeches. Herodotus' speeches are not always the 'driving force' of the action and are also influenced by rhetoric, which is already present in Homer and later becomes prevalent in Thucydides.

In conclusion, we can affirm that Herodotus is a transitional figure: he marks the passage from the archaic age to the Classical age, from epic poetry to Thucydidean historiography and, specifically, from the use of speeches as the 'driving force' of action, typical of a still oral culture, to universal speeches, influenced by rhetoric, a key feature of 5th-century Athens.

Canfora 1971 points out that Thucydides' speeches are introduced by τοιάδε.

For example, Thuc. 3, 37–40; 3, 42–48; 6.9–14; 6, 16–18; 6, 33–34; 6, 36–40; 6, 82–87; 6, 89–92.

### References


Canfora, L./Corcella, A. (1992), "La letteratura politica e la storiografia", in: G. Cambiano/L. Canfora/D. Lanza (eds.), Lo spazio letterario della Grecia antica, 1, Rome, 433–472.

Cantilena, M. (2002), Sul discorso diretto in Omero, in: F. Montanari (ed.), Omero tremila anni dopo. Atti del Convegno di Genova, 6–8 luglio 2000, Rome, 21–40.

Carey, C. (2016), "Homer and Epic in Herodotus' Book 7", in: A. Efstathiou/I. Karamanou (eds.), Homeric Receptions Across Generic and Cultural Contexts, Berlin, 71–90.

Caskey, J. L. (1942), "Herodotus and Homer", in: ClWeekly 35, 267.

Clark, M. (2004), "Formulas, Metre and Type-Scene", in: R. Fowler (ed.), The Cambridge Companion to Homer, Cambridge, 117–138.


de Romilly, J. (1956), Histoire et raison chez Thucydide, Paris.

	- Festgabe für Wolfgang Schadewaldt zum 15. März 1965, Pfullingen, 29–52.

Ong, W. J. (1986), "Text as Interpretation. Mark and After", in: J. M. Foley (ed.), Oral Tradition in Literature. Interpretation in Context, Columbia, 147–169.

Ong, W. J. (2002), Orality and Literacy. The Technologizing of the Word, London – New York.


Rutherford, R. B. (2012), "Structure and Meaning in Epic and Historiography", in: E. Foster/D. Lateiner (eds.) (2012), Thucydides and Herodotus, Oxford, 13–38.

Sale, M. (1996), "In Defence of Milman Parry: Renewing the Oral Theory", in: Oral Tradition 11, 374–417.

Sbardella, L. (2006), Oralità. Da Omero ai mass media, Rome.

Slings, S. R. (2002), "Oral Strategies in the Language of Herodotus", in: E. J. Bakker/I. J. F. de Jong/H. van Wees (eds.) (2002), Brill's Companion to Herodotus, Leiden – Boston – Köln, 53–77.

Strasburger, H. (1972), Homer und die Geschichtsschreibung, Heidelberg.

Thomas, R. (1992), Literacy and Orality in Ancient Greece, Cambridge.

Thomas, R. (1993), "Performance and Written Publication in Herodotus and the Sophistic Generation", in: W. Kullmann/J. Althoff (eds.), Vermittlung und Tradierung von Wissen in der griechischen Kultur, Tübingen, 225–244.

Thomas, R. (2000), Herodotus in Context. Ethnography, Science and the Art of Persuasion, Cambridge – New York.

Thornton, A. (1984), Homer's Iliad: Its Composition and the Motif of Supplication, Göttingen.


Van Wees, H. (1986), "Leaders of Men? Military Organization in the Iliad", in: Classical Quarterly 36, 285–303.


## Laura Lulli Some Reflections on Orality and Epic Poetry in Ancient Literary Criticism

Abstract: The paper examines some relevant cases of the views of the orality which appear in ancient literary criticism. Starting from a reconsideration of the oral system in the broader communication context of the Hellenistic and Imperial period, the analysis will underline the specific functions attributed to orality in extant works of literary criticism, with a special focus on the perspective followed by the author of the treatise On the Sublime. Indeed, from the survey it appears clear that the exploration of oral dimension becomes, on one hand, a lens for the study of the previous literary tradition, in particular of the Homeric poetry, and, on the other, a crucial experience in the rhetorical paideia of the future ruling classes.

Keywords: Homeric poems; ancient literary criticism; orality.

When we think about the analytical method that ancient literary criticism adopts for the study of Homeric texts, we usually envisage it as a complex and multilayered system.¹ This system was developed at first by the rhapsodes themselves and later, over the centuries, by poets and philosophers,² but also by grammarians (most notably, yet not exclusively, Alexandrian ones),³ rhetoricians and var-

A synthetic and comprehensive outline of the different figures involved in the complex process of the Homeric ancient exegesis is in Pontani 2005, 23–103.

For a view of both the role of the rhapsodes in the first phases of the commentary and interpretation of the Homeric text, and the beginnings of the Homeric scholarship between the 6th century BC and the 4th century BC, until Aristotle, see Cassio 2002, 118–132, with further bibliography. The starting point of a professional scholarly activity regarding the poetical texts of the Greek tradition (i.e., above all, Homer and the lyric poetry) in the Platonic reflections, in particular in Plato's Ion, is underlined by Hunter 2011. An overview about the function of the Hellenistic poets in the construction of the Homeric scholarship is in Rengakos 2002, with a synthesis of the previous critical debate.

In the huge bibliography regarding Homeric scholarship see, at least, the reconstruction of the work on the Homeric poems by the Alexandrian grammarians provided by Montanari 1998, with a special focus on the processes of the ekdosis,West 2001, in particular 33–45, about Zenodotus, Nünlist 2011, and Schironi 2018, in particular 47–90, about Aristarchus. For an overview of the philological practices in the Homeric scholia see Schmidt 2002. An attempt to underline the development of the literary criticism in the context of the scholia is made by Nünlist 2009. On this topic see also Richardson 2006.

OpenAccess. © 2022 Laura Lulli, published by De Gruyter. This work is licensed under the Creative Commons Attribution 4.0 International License. https://doi.org/10.1515/9783110751963-009

ious literary critics.⁴ These were among the most prominent intellectual figures in charge of the cultural interactions and interconnections between the Greek world and the new order of the Roman empire. When we turn to consider the reflections developed within such a system, we find ourselves dealing with a corpus of analytical works, all based on the close examination of well-defined texts, stabilized in written form very close to that which has reached us (leaving aside the well-known Alexandrian discussions on the end of the Iliad, or idiosyncratic ideas, luckily of limited appeal, such as Zenodotos' view that the description of Achilles' shield was not Homeric enough⁵).

In this perspective, then, we might wonder whether the original and/or constitutive orality of Homeric epic poetry was considered by ancient critics, in some way, and thus discussed, and generally received by them, in their different cultural contexts and in their various works. In other words, was there any perception of the ancient oral system of communication through which archaic epic poetry had been composed and transmitted? Was there any conscious or even indirect awareness of the traces of the oral mechanism behind the construction of the archaic epic poems which continued to be visible even in fixed written sets of songs, which had clearly been transmitted and regarded as unitary poems? Did the oral component of epic poetry continue to maintain a role as a specific characteristic of the literary genre even in the different critical perspectives of later interpreters of such poetic works? Finally, were the ancient literary critics conscious of the oral aspects of the epic poems and, if so, how did they analyze them?

The attempt to answer these questions is not easy and probably it can only be tentative. In any case, a broad and deep examination of the mechanisms of the oral communicative system would probably benefit from a survey – with no claim to completeness – of the 'views of orality' which were developed, or even just hinted at, by ancient critics and intellectuals interested in early Greek epic poetry.

In order to cast some light on ancient views on orality, it is worth setting out not from theoreticians, but from a poet, albeit one imbued with philology, namely Theocritus: a fine experimenter of new solutions and literary genres thanks to

For a broad perspective on the position of the Homeric poetry in the ancient criticism, with a focus on the rhetorical criticism, see Hunter 2015.

The athetesis and the diorthosis practiced by Zenodotus and Aristarchus in their ekdoseis, with the creation of a specific corpus of semeia, are analyzed by Montanari 2015, in particular 653– 660, with previous bibliography. About the influence which such Alexandrian philological work on the Homeric text had even on the Alexandrian literary production, above all on Apollonius Rhodius, see Rossi 2020a.

his deep knowledge and reassessment of the previous literary tradition. In Idyll 16, composed to celebrate Hiero's exploits, Theocritus begins by voicing his disapproval of those patrons who, driven by their desire for wealth,⁶ tend to avoid paying poets and fail to consider the crucial function of poetry as a powerful tool to perpetuate their memory. In his attempt to rebuild eulogistic poetry and to adapt it to the needs of Hellenistic patrons,⁷ Theocritus refers to Simonides and alludes to Pindar,⁸ while also mentioning Homer.⁹ However, before dealing with these exemplary cases, at the beginning of the Idyll, Theocritus adopts the point of view of an hypothetical patron and states:

Theocr., Id. 16, 16–21 πᾶς δ' ὑπὸ κόλπου χεῖρας ἔχων πόθεν οἴσεται ἀθρεῖ ἄργυρον, οὐδέ κεν ἰὸν ἀποτρίψας τινὶ δοίη, ἀλλ' εὐθὺς μυθεῖται 'ἀπωτέρω ἢ γόνυ κνάμα· αὐτῷ μοί τι γένοιτο.' 'θεοὶ τιμῶσιν ἀοιδούς.' 'τίς δέ κεν ἄλλου ἀκούσαι; ἅλις πάντεσσιν Ὅμηρος.' 20 'οὗτος ἀοιδῶν λῷστος, ὃς ἐξ ἐμεῦ οἴσεται οὐδέν.'

and each, his hand within his purse-fold, looks to see whence he may win money and will not rub the very rust therefrom to give another, straight answering rather, 'The knee is closer than the shank; may somewhat befall me myself', or 'Heaven rewards the poet', 'And who would listen to another? Homer is enough for all', 'He is the best of poets who shall get naught of me'. (transl. A. S. F. Gow)

Cf. Theocr., Id. 16, 14–15 οὐκ οἶδ᾽· οὐ γὰρ ἔτ᾽ ἄνδρες ἐπ᾽ ἔργμασιν ὡς πάρος ἐσθλοῖς / αἰνεῖσθαι σπεύδοντι, νενίκηνται δ' ὑπὸ κερδέων; for an analysis of Pind., Nem. 7, 17 as a close parallel for the topic of the kerdos in Theocritus see Sbardella 2004a, 68. The motive of the gain is furthermore developed in vv. 22–33, with a special accent on the relationship with the role of the poets, who are defined Μοισάων […] ἱεροὺς ὑποφήτας (v. 29), sacred interpreters or even priests of the Muses. For the echoes of the Homeric, as well as the Pindaric visions of the poet's role in this Theocritean syntagm see Gow 1952, 311; for a comparation of the Theocritean passage with Pind., Isthm. 1, 67–69 see Sbardella 2004a, 71–72.

The complex pattern of the poetical perspective presented in Idyll 16 about the patronage of Theocritus is analyzed by Hunter 1996, 77–90, who highlights the fine mixture with elements of the Pindaric tradition and the new needs of the rulers in the Ptolemaic period.

Cf. Theocr. Id. 16, 34–47. For a comprehensive analysis of the allusive references to Pindar and Simonides in the Idyll 16 see Hunter 1996, 82–90, 97–109.

Cf. Theocr. Id. 16, 48–57. The role of Homer in Idyll 16, as well as in the other Idylls of the Corpus Theocriteum is examined by Hunter 1996, 90 –97.

The utmost consideration for poetry, expressed by a patron who is completely enslaved by the desire for profit, deals with Homer:¹⁰ he is enough for all – "who would listen to another?" (τίς δέ κεν ἄλλου ἀκούσαι). In this rhetorical question, designed to underline the absolute centrality of Homeric poetry on many levels in the cultural context of the Hellenistic period, the key word is the verb ἀκούω, which stresses the importance of the oral dimension of epic poetry,¹¹ and alludes to the possibility for the public to enjoy such compositions aurally.¹² This is all the more remarkable if we consider that, in the Hellenistic context of the dissemination of literature, the glory of the great authors of the past was generally committed to books, which preserved their verses and expressions: Dorica and Charassos have long turned to dust, writes Posidippos (122 A.–B. = XVII G.–P.), but the bright singing columns (v. 6 αἱ λευκαὶ φθεγγόμεναι σελίδες) of the bookroll allow Sappho's poetry to survive.¹³ Similar state-

 The oral dimension of the epic poetry during the Hellenistic period, especially for its diffusion in various festivals, as well as for its presence in the many epideixeis, "recitations", and akroaseis, "recitations" (the term underlines the role of the audiences) of wandering poets, is a deeply rooted phenomenon. See, at least, Fantuzzi 1988, XXXVI–XLII, who, however, outlines a difference between the "epici mestieranti", professionals engaged in the public festivals, and the "poeti 'colti'", whose literary production was intended to be read or performed in the epideixeis at court. A reassessment of the oral aspect of such Hellenistic and Imperial epic poetry, in view of the theory of the so-called "oral dictated text" (Jensen 1980, 10 and passim) was formulated by Jensen 1980, 125–127. On the varietas and the considerable amout of epic production in the Hellenistic period, with considerations about the status of the epic poets in the context of the Hellenistic courts, see Ziegler 1988, in particular 15–25. The phenomenon of the wandering poets in the Byzantine period has been illuminated by the study of Cameron 2015, according to whom "attention has been drawn to the exceptional mobility, social and geographical, of professors and teachers of rhetoric, but it has not been remarked that professional poets traveld from city to city every bit as much as an ambitious professor of rhetoric like Libanius or the young St. Augustine" (15). For an overview on the complex mixture between festivals, spectacles, and elite cultural circles as occasions for the performance, and/or for the readings of epic poetry in Egypt in the 3th-6th century AD see Miguélez Cavero 2008, 194–198.

 The epigram is cited by Athen. 13 p. 596c. For such a sofisticated image, with a personification of a constitutive part of the bookroll see Bing 1988, 33, who mentions also the case of a "silenced" papyrus column in Aceratus (AP 7 138, 3–4 = FGE p. 3 σοῦ δὲ θανόντος, / Ἕκτορ, ἐσιγήθη καὶ σελὶς Ἰλιάδος). For other similar images see Page in FGE p. 342 ad v. 1198.

Regarding the relationship between Theocritus and Homer, Hunter 1996, 94 underlines that the centrality given to Homer in the Theocritean poetry is part of the poet's broader strategy of presenting himself as the new Homer.

In this perspective it is worth considering the reflection of Sbardella 2004b, 88, who underlines that "[…] l'uso del verbo ἀκούω non è né casuale né desemantizzato, ma rimanda al modo tradizionale attraverso cui nei secoli il patrimonio di canti attribuito ad Omero era stato fruito e continuava in parte ad esserlo ancora, per mezzo della viva voce dei cantori".

ments in praise of books can be found in many other epigrammatists.¹⁴ But Theocritus' focus on the acoustic dimension of poetry¹⁵ implies that the Homeric poems continued to be perceived as a kind of literature mainly addressed to an audience of listeners and, therefore, able to maintain a close link with an oral communication medium, even in a period characterized by the increasing use of the written communication system at every stage in the life of a literary text.

As is well known by now, this phenomenon could be traced back to the broader, actual circumstances of the diffusion of epic poetry in the Hellenistic age: poetic performances by groups of rhapsodes and wandering poets are uninterruptedly attested in the epigraphic and papyrological evidence up to the Imperial age and beyond.¹⁶ To remain in Theocritus' Egypt, Zeno's archive – our largest source of knowledge on daily life in mid-3rd-century Egypt – mention at least one set of readings of (or conferences on) Homeric poetry,¹⁷ but references to live performances by poets, along with various forms of musical entertainment and dances, are recorded for almost every local eorté.¹⁸ Listening to poetry was

For an overview of the images of the bookroll, with various functions, in the Hellenistic epigrammatic production see Bing 1988, 29–33.

It is notworthy to add that such sensibility for an acoustic dimension of the poetry in Theocritus is not limited to the reflection on and the reception of the Homeric poetry, but it is also a constitutive aspect of his own poetical practices, since many of his poems seem to have been aimed to performative occasions, both in the context of the Ptolemaic court and in more private frame, such as the symposia. For the performative destination of the Theocritean poetry see, at least, Zanker 1987, 79, who considered the "city audience" the main public of the Theocritean Idylls; see also Pretagostini 2007, Pretagostini 2009, 3, 14–29.

For an overview of the epigraphic evidences of the wandering poets in the Hellenistic period see Guarducci 1929, and Pallone 1984. Chaniotis 2009, in particular 253–269 outlines the role of the historians, the orators and the envoys as wondering intellectuals who contributed with their performances to build up the collective and the cultural memory of the various poleis in Hellenistic period and until the Roman time. See Manieri 2009, 50–53 for the role of the contests of epic poetry in the agonal occasions in Beotia; for similar presence of the epic poetry in the Spartan agonals festivals see Massaro 2018, 178–179, 182–183, 189–190. A useful review of the papyrological evidence from Oxyrhynchus about the agonal occasions and the poetical program performed there, including the epic poetry, from the Hellenistic to the Imperial period, see Remijsen 2014.

The main evidence of such intellectual activity is provided by a letter sent to Zenon (P. Cair. Zen. 4. 59603, middle 3th century BC), which, although its fragmentary condition, appears to deal with a lecture περὶ τοῦ ποιητοῦ (l. 3), organized in Philadelphia by Mnesitheos; moreover, Zenon's correspondent should regard this event quite important, since he asked Zenon to collect as many pupils as possible. The editor princeps, Edgar C. Campbell, suggested that the expression περὶ τοῦ ποιητοῦ alludes to Homer, as "a living poet would surely have been named" (P. Cair. Zen. 4, 55).

See above, 148, n. 12.

common in Hellenistic Egypt. Moreover, reading poetry from a book also entailed a strong oral component, as reading aloud was the main reading practice until Late Antiquity.¹⁹ Grammarians such as Dionysius Thrax even give instructions about how to read Homer aloud (εὐτόνως, "with due energy") and the performative reading of epic was taught in school.²⁰ These elements – i.e. continuous performative experiences in every corner of the Greek and Roman world, and the endurance of different degrees of orality despite the emergence of a book culture – are crucial points that must be taken into account in the study of the perception of orality in ancient literary critics' reflections on archaic epic poetry. We will return to this point later.

Theocritus' verses were born in a context where metaliterary reflections by poets themselves were usual; and it is noteworthy that thoughts on the oral dimension of archaic epic were part of such considerations. However, reflections of this sort had far more ancient origins. The primacy of poets in such matters was already recognized by Aristotle who, in a passage from book III of the Rhetoric affirms:

Arist., Rhet. 3, 1404a

ἤρξαντο μὲν οὖν κινῆσαι τὸ πρῶτον, ὥσπερ πέφυκεν, οἱ ποιηταί· τὰ γὰρ ὀνόματα μιμήματα ἐστίν, ὑπῆρξεν δὲ καὶ ἡ φωνὴ πάντων μιμητικώτατον τῶν μορίων ἡμῖν· διὸ καὶ αἱ τέχναι συνέστησαν ἥ τε ῥαψῳδία καὶἡὑποκριτικὴ καὶ ἄλλαι γε.

The poets, as was natural, were the first to give an impulse to style; for words are imitations, and the voice also, which of all our parts is best adapted for imitation, was ready to hand; thus the arts of the rhapsodists, actors, and others, were fashioned. (transl. J. H. Freese)

According to the Aristotelian perspective, poets – as was natural – were the first to give an impulse to the study of style; and since words are a kind of imitation, they could make use of the voice, the most imitative human faculty of all, in order to create the arts, i.e. rhapsody, the art of speaking, the actor's art, and others. It is clear that the imitative nature of both words and the voice implies the

This 'oral' dimension of the reading practices is highlighted by Cavallo 1998, who states that "la maniera più abituale di leggere era comunque, a qualsiasi livello e con qualsiasi funzione, quella a voce alta […]. La lettura poteva essere diretta o anche fatta da un lettore interposto tra il libro e chi lo ascoltava, uditorio o individuo; nel caso di certi componimenti poetici, più voci lettrici si alternavano, secondo la struttura del testo. Queste pratiche spiegano anche l'interazione assai stretta tra scrittura letteraria e lettura" (47). On the development of this practice beyond the Late Antiquity, in the complex cultural system of Bysanzium see Cavallo 2007, 87–106. Cf. Dion. Thrax, Ars grammatica 2 Lallot. On the role of the reading aloud in the grammarian and teaching practices see Del Corso 2005, 21–30, and Cavallo 2007, 21–28.

absolute centrality of the oral dimension of the literary communication system. Therefore, all literary activity primarily depends on the use of the oral mode of communication, with its expressive potential in terms of nuances of sound, musical modulations, and the ability to arrange syntagms, formulas and expressions to create new harmonies. Only in a second phase does this fundamental use of the voice provide the necessary starting point for the creation of the various literary genres, which correspond to the Aristotelian arts (αἱ τέχναι).

This idea is further developed and defined in the Poetics, starting at the very beginning of the treatise. Here, after having defined poetry, in all its possible forms and articulations, as the manifestation of an essentially mimetic activity²¹ which is performed by variously combining the use of rhythm (including dance),²² language and music, Aristotle deals with epic poetry. He states:

#### Arist., Poet. 1447a-b

ἅπασαι μὲν ποιοῦνται τὴν μίμησιν ἐν ῥυθμῷ καὶ λόγῳ καὶ ἁρμονίᾳ, τούτοις δ' ἢ χωρὶς ἢ μεμιγμένοις· […] ἡ δὲ ἐποποιία μόνον τοῖς λόγοις ψιλοῖς ἢ τοῖς μέτροις καὶ τούτοις μιγνῦσα μετ' ἀλλήλων εἴθ' ἑνί τινι γένει χρωμένη τῶν μέτρων, ‹ἡ ἀνώνυμος› τυγχάνουσα²³ μέχρι τοῦ νῦν.

All the poetic arts mentioned produce mimesis in rhythm, language, and melody, whether separately or in combinations. […] while epic poetry uses either plain language or metrical forms (whether combinations of these, or some one class of metres), that poetry which remains so far unnamed. (transl. S. Halliwell, with modifications)

Arist., Poet. 1447a 14–16 ἐποποιία δὴ καὶ ἡ τῆς τραγῳδίας ποίησις ἔτι δὲ κωμῳδία καὶ ἡ διθυραμβοποιητικὴ καὶ τῆς αὐλητικῆς ἡ πλείστη καὶ κιθαριστικῆς πᾶσαι τυγχάνουσιν οὖσαι μιμήσεις τὸ σύνολον, "now, epic and tragic poetry, as well as comedy, dithyramb, and most music for aulos and lyre, are all, taken as a whole, kinds of mimesis" (transl. S. Halliwell).

See Gallavotti 2003, pp. 123–124.

It is worth to note that this is another problematic passage of the text, sinceA(cod. Parisinus 1741, 10th–11th century) and the Latin translation of William of Moerbeke present only τυγχάνουσα (existens) and various interpreters have suggested different solutions for the text, i.e. Bernays conjectured ‹ἀνώνυμος› τυγχάνουσα on the basis of the arabic version of the treatise (quae est sine appelatione), whereas Lobel suggested ἀνώνυμοι τυγχάνουσιν οὖσαι, a possibility which is accepted also by Kassell 1965, even if he does not judged useful the participle, so that he preferred the syntagm ἀνώνυμοι τυγχάνουσι; finally a different conjecture was proposed by Suckow: τυγχάνει οὖσα, accepted also by Tarán 2012. For an overview on the textual problems of this passage see Tarán 2012, 230–231. For the interpretation of the expression ἡ ἀνώνυμος, suggested by Gallavotti 2003, as a reference to the iambic, tetrameter, elegiac and epodic poetry, which implies only a way of the mimesis as the epic poetry, even if it hasn't got neither a defined subject-matter nor a mode of the mimesis, see Gallavotti 2003, 125.

As is well known, this passage presents many textual problems, since the word ἐποποιία has not been transmitted unanimously.²⁴ For the interpretation of this textual section, then, I will adopt the perspective suggested by Carlo Gallavotti, who considers ἐποποιία a genuine reading. According to this point of view, Aristotle indicates the constitutive elements of epic poetry and its mimetic character, which consists in using only words and metres,²⁵ without any dance or music. Though the expression λόγος ψιλός is frequently adopted to indicate prose in opposition to poetry,²⁶ in this context – as Carlo Gallavotti has clearly demonstrated²⁷ – it indicates plain and simple words, which together with the hexameter form the backbone of epic poetry. Indeed, this idea of the centrality of the spoken word in the epic literary genre, like the importance assigned to sound and rhythm, is not exclusive to the Aristotelian reflection, but occurs – in almost the same terms – in Plato as well. In the dialogue between Socrates and Phaedrus, in a passage devoted to demonstrating the superiority of orality over writing in philosophical practice, Socrates deals with different literary genres, including epic poetry:

#### Plat., Phaedr. 278b–d

ΣΩ. οὐκοῦν ἤδη πεπαίσθω μετρίως ἡμῖν τὰ περὶ λόγων· καὶ σύ τε ἐλθὼν φράζε Λυσίᾳ ὅτι νὼ καταβάντε ἐς τὸ Νυμφῶν νᾶμά τε καὶ μουσεῖον ἠκούσαμεν λόγων, (c) οἳ ἐπέστελλον λέγειν Λυσίᾳ τε καὶ εἴ τις ἄλλος συντίθησι λόγους, καὶ Ὁμήρῳ καὶ εἴ τις ἄλλος αὖ ποίησιν ψιλὴν ἢ ἐν ᾠδῇ συντέθηκε, τρίτον δὲ Σόλωνι καὶ ὅστις ἐν πολιτικοῖς λόγοις νόμους ὀνομάζων συγγράμματα ἔγραψεν· εἰ μὲν εἰδὼς ᾗ τὸ ἀληθὲς ἔχει συνέθηκε ταῦτα, καὶ ἔχων βοηθεῖν, εἰς ἔλεγχον ἰὼν περὶ ὧν ἔγραψε, καὶ λέγων αὐτὸς δυνατὸς τὰ γεγραμμένα φαῦλα ἀποδεῖξαι, οὔ τι τῶνδε ἐπωνυμίαν ἔχοντα δεῖ (d) λέγεσθαι τὸν τοιοῦτον, ἀλλ' ἐφ' οἷς ἐσπούδακεν ἐκείνων.

SOCRATES. We have amused ourselves with talk about words long enough. Go and tell Lysias that you and I came down to the fountain and sacred place of the nymphs, and heard words which they told us to repeat to Lysias and anyone else who composed speeches, and to Homer or any other who has composed poetry with or without musical accompaniment, and third to Solon and whoever has written political compositions which he calls laws: – If

See Gallavotti 2003, app. ad loc.: ἐποποιία is transmitted by A (cod. Parisinus 1741, 10th–11th century), but it is not translated in the arabic version of the treatise, on the basis of a previous Syriac text, by Abū Bišr (10th century), and, then, it was deleted by Ueberweg. The term is inserted in the text with brackets by the majority of editors, as, for example, Christ 1882, Hardy 1932, Kassell 1965, and more recently Tarán 2012, who, however, adfirms in the comment that the word ἐποποιία "must be excised because it yields no reasonable sense" (Tarán 2012, 226).

For such interpretation of this debated passage of the Poetics see Gallavotti 2003, 124–125. Cf. e. g. Plat., Menex. 239c; Arist., Rhet. 1404b 14, 33.

See Gallavotti 2003, 124–125, who specifies that "è solo il contestoadare l'esatto valore di questa terminologia che non è univoca, e non è ancora rigidamente fissata dall'uso scolastico" (125).

he has composed his writings with knowledge of the truth, and is able to support them by discussion of that which he has written, and has the power to show by his own speech that the written words are of little worth, such a man ought not to derive his title from such writings, but from the serious pursuit which underlies them. (transl. H. N. Fowler)

We are almost at the end of the dialogue, and of its famous polemic against writing and book culture.²⁸ Socrates arranges a verbal game of mirrors to emphasize the triumph of the spoken word over writing: with his characteristic irony, he asks Phaedrus to tell Lysias that a voice told them, in the sacred place of the nymphs,²⁹ to repeat to all orators, to Homer and to the other poets who composed their verses with or without musical accompaniment, that, if they are looking for the absolute truth and are able to defend their written work by their own speech, they can be named philosophers or lovers of wisdom. All other philosophical implications aside, it is clear that, even in the struggle which Plato wages against the writing system, epic poetry is perceived as a literary genre based on the spoken word, on the mere spoken word – not as much as rhetoric, but in opposition to lyric poetry, where the spoken word is mixed with music. So, in the wider Platonic frame of a philosophical conception which exalts oral communication in opposition to the system of writing, epic poetry continues to be considered a 'spoken word-centered' literary genre.

In later reflections, when Plato's philosophical perspective will be mostly overlooked or neglected, it will be usual to focus on Homeric epic as wordbased poetry, whose standard form of dissemination is oral delivery, as the 'sound' of the words plays a substantial role in terms of the final rhetorical effect, even if obviously it is different from other 'musical' genres. The pre-eminence of such an aspect as the main and distinctive characteristic of the genre is evident – to mention only one example – in the essay On Literary Composition by Dionysius of Halicarnassus. In the context of the analysis of composition, which is considered more important than the selection of words,³⁰ Dionysius describes in detail the distinction between music and speech (the μουσική-λογική opposition) and the different approaches that the various genres have with these tools of the communicative system.³¹ Even if this polarity is normally interpreted

In the huge bibliography about the complex dialectic between orality and writing culture in the Phaedrus see, at least, the overviews in Trabattoni 2005, 86–101, and Cerri 2007, 99–106, 109–119. On this passage of the dialogue see also Hunter, in this volume.

This is the environment in which the dialogue is set: cf. Plat., Phaedr. 230b, 241e, 263d. Cf. Dion. Hal., de comp. verb. 1–2.

For an analysis of the relationship between words and musical aspects in the literary communication from the perspective of Dionysius of Halicarnassus see Gentili 1990, in particular 7–10, 20–21. A deep survey of the stylistic functions of the reflections elaborated by Dionysius

as a reference to the opposition between poetry and prose, Dionysius further specifies that there are certain genres which use the voice's tune to create particular effects upon the ear:

Dion. Hal., de comp. verb. 11

δεδειγμένης τῆς διαφορᾶς ᾗ διαφέρει μουσικὴ λογικῆς λοιπὸν ἂν εἴη κἀκεῖνα λέγειν, ὅτι τὸ μὲν τῆς φωνῆς μέλος, λέγω δὲ οὐ τῆς ᾠδικῆς ἀλλὰ τῆς ψιλῆς, ἐὰν ἡδέως διατιθῇ τὴν ἀκοήν, εὐμελὲς λέγοιτ' ἄν, ἀλλ' οὐκ ἐμμελές […].

Now that the difference between music and speech has been shown, a few remaining points may be made. If the intonation of the voice – not in song but in ordinary conversation – has a pleasant effect upon the ear, it will be called "song-like" rather than "singing".

This analysis of composition and its 'musical' aspect, then, implies a subtle distinction between different kinds of literary genres characterized by the use of the voice; indeed, Dionysius is not referring here just to prose, in opposition to poetry, but also to a specific kind of poetic discourse, in which the tune of the voice – a voice not singing but speaking as in an ordinary conversation – can have a pleasant effect on the listener, being full of melody yet not tuneful, i.e. metrically sung. The emphasis, then, is on the differences between epos and lyric poetry, especially in the perspective of the effects these literary genres have on the audience through the specific use of a mixture of words or of words and music.

The evidence from Dionysius of Halicarnassus highlights how fundamental the oral dimension of epic poetry was in ancient literary critics' reflections: the centrality of this aspect is twofold, as it concerns both the prevalence of words and speaking in the frame of the hexametric arrangement of the narrative material, and a close attention to the effect that epic poetry should have on listeners. It is possible here to outline an evolution: the kind of orality which literary critics discuss is no longer the original one, namely a communicative system used both during the phase of the composition of a text and during that of its circulation; however, it is still a defining feature of the genre, which on account of its effect on the audience can be adapted to a world where the spoken word still plays a crucial role in communicative dynamics, albeit not exclusively within the confines of poetry.

This 'new' perspective on the oral dimension of epic poetry, based on an attention to the listener, is emphasized in many examples of ancient literary criticism, among which the case of the treatise On Sublimity stands out for its programmatic and constant aim both to present itself as a useful work for

of Halicarnassus about the phonic nature of the words and of their sequences in a text is in Rossi 2020b.

specific addressees, namely public figures,³² and to analyze the multifarious characteristics of the sublime, especially in relations to their effect on the listener.³³ This is evident from the very beginning of the treatise:

#### Subl. 1, 2, 4

ἐπεὶ δὲ ἐνεκελεύσω καὶ ἡμᾶς τι περὶ ὕψους πάντως εἰς σὴν ὑπομνηματίσασθαι χάριν, φέρε, εἴ τι δὴ δοκοῦμεν ἀνδράσι πολιτικοῖς τεθεωρηκέναι χρήσιμον ἐπισκεψώμεθα. […] οὐ γὰρ εἰς πειθὼ τοὺς ἀκροωμένους ἀλλ' εἰς ἔκστασιν ἄγει τὰ ὑπερφυᾶ· πάντη δέ γε σὺν ἐκπλήξει τοῦ πιθανοῦ καὶ τοῦ πρὸς χάριν ἀεὶ κρατεῖ τὸ θαυμάσιον, εἴγε τὸ μὲν πιθανὸν ὡς τὰ πολλὰ ἐφ' ἡμῖν, ταῦτα δὲ δυναστείαν καὶ βίαν ἄμαχον προσφέροντα παντὸς ἐπάνω τοῦ ἀκροωμένου καθίσταται.

You have urged me to set down a few notes on sublimity for your own use. Let us then consider whether there is anything in my observations which may be thought useful to public men. […] For grandeur produces ecstacy rather than persuasion in the hearer; and the combination of wonder and astonishment always proves superior to the merely persuasive and pleasant. This is because persuasion is on the whole something we can control, whereas amazement and wonder exert invincible power and force and get the better of every hearer. (transl. D. A. Russell)

 The interest of the author for the paideia of the addressee of the treatese, Postumius Florus Terentianus, is clear in many passages of the work, in which he is continuosly named with affection and friendship (Subl. 1, 1; 1, 2; 1, 3; 1, 4; 4, 3; 9, 6; 9, 10; 12, 4; 15; 26, 2; 29, 2; 39, 1; 44, 1). On the profile of the addressee see Russell 1964, 59, Mazzucchi 2010, 131–132. The attention for the public figures, the πολιτικοὶ ἄνδρες, as main addressees of the cultural and literary message of the treatese is underlined by Russell 1964, 60 –61, and Mazzucchi 2010, 132. The latter one considers the expression as a possible reference to the specific social group of the magistrates, on the basis of Strab. 1, 1, 22, who, at the beginning of his Geography, states that the work is conceived for a statesman and a general reader, i.e. the public at large, adding also a profile of the education of the statesman, based on a round of courses usual both for freemen and students of philosophy. The other evidence taken into account by Mazzucchi is Dion. Hal., de Dem. 15, who, not so differently from Strabo, describes the training of a statesman and of the men engaged in the commercial affairs as an education focused on a round of courses, which allows them to acquaire a refined diction. It is worth to note, however, that in these passages the expression πολιτικοὶ ἄνδρες tends to assume a broader meaning than the specific definition of magistrates, since it alludes to many figures of the political context of a polis, i.e. orators, magistrates, statesman. Forasimilar interpretation of the expression see Halliwell 2021, 128–129, with further examples. Moreover, in the case of Strabo's Geography, the broad and various spectrum of political and, in general, public figures of the polis, who are alluded in the expression, correspondes to an ideal reader of the work, who should have all the necessary and enciclopedic knowledge to understand the multifarious perspectives of the geographical discourse. On this aspect see Dueck 2000, 186–187, with further bibliography.

 For an analysis of the influence which the sublime, as presented in the Pseudo-Longinus' treatese, has on the hearers, especially thanks to the peculiar intersubjectivity between authors and audience, is provided by Halliwell 2012, 328–367.

In the first chapter of the work the focus on the oral component of the literary message, in every literary genre, is designed to highlight the listener's perspective, as well as to achieve the pragmatic goal of providing a rhetorical education useful for the addressee's (possible) political career and, more generally, for many aspects of civic life. It is evident, then, that, starting from these two cornerstones, the oral dimension of literary communication serves as an essential parameter for the kind of literary criticism and rhetorical analysis developed by the author of the treatise.³⁴

Moreover, in the treatise the absolute centrality of this 'new' orality, i.e. an oral dimension of the literary message grasped in relation to the listener's perception, is closely linked with the sublime, as it emerges in the chapter dealing with the marks of true sublimity:

Subl. 7, 2–3

φύσει γάρ πως ὑπὸ τἀληθοῦς ὕψους ἐπαίρεταί τε ἡμῶν ἡ ψυχὴ καὶ γαῦρόν τι ἀνάστημα λαμβάνουσα πληροῦται χαρᾶς καὶ μεγαλαυχίας, ὡς αὐτὴ γεννήσασα ὅπερ ἤκουσεν. 3. ὅταν οὖν ὑπ' ἀνδρὸς ἔμφρονος καὶ ἐμπείρου λόγων πολλάκις ἀκουόμενόν τι πρὸς μεγαλοφροσύνην τὴν ψυχὴν μὴ συνδιατιθῇ μηδ' ἐγκαταλείπῃ τῇ διανοίᾳ πλεῖον τοῦ λεγομένου τὸ ἀναθεωρούμενον, πίπτῃ δέ, ἂν αὐτὸ συνεχὲς ἐπισκοπῇς, εἰς ἀπαύξησιν, οὐκ ἂν ἔτ' ἀληθὲς ὕψος εἴη μέχρι μόνης τῆς ἀκοῆς σῳζόμενον. τοῦτο γὰρ τῷ ὄντι μέγα, οὗ πολλὴ μὲν ἡ ἀναθεώρησις, δύσκολος δὲ μᾶλλον δ' ἀδύνατος ἡ κατεξανάστασις, ἰσχυρὰ δὲ ἡ μνήμη καὶ δυσεξάλειπτος.

It is our nature to be elevated and exalted by true sublimity. Filled with joy and pride, we come to believe we have created what we have only heard. When a man of sense and literary experience hears something many times over, and it fails to dispose his mind to greatness or to leave him with more to reflect upon than was contained in the mere words, but comes instead to seem valueless on repeated inspection, this is not true sublimity: it endures only forthe moment of hearing. Real sublimity contains much food for reflection, is difficult or rather impossible to resist, and makes a strong and ineffaceable impression on the memory. (transl. D. A. Russell)

In this passage the oral/aural dimension of the literary message and the audience's action of listening are not only essential for the performative aspects of the text and for its publication and further diffusion, but also appear to be crucial for the reception of the text, and for the first steps in the process of interpreting the

For another, strong case of the centrality of the orality in the communicative system of the polis, in relationship to a specific public in the context of the funeral speeches, and for the powerful effects of this genre of discourses on the audience see the introductive frame of the dialogue in Plat., Menex. 234a-235c. See Robinson 2018, 176–181, with further bibliography. For an overview on the position of the Menexenus in the tradition of the Athenian funeral speech see Pappas/Zelcer 2015, 58–76.

text – as well as for any further critical analysis of it. In other words, understanding the sublime in a literary text is, at first, an 'aural experience', which must be consolidated via reflection and then through one's memory, but only if it is a real kind of sublimity and not a false one. Donald A. Russell has convincingly defined this aspect of the treatise as attention to the "psychology of the audience".³⁵ This perspective seems to be based precisely on a reconsideration of the oral dimension of literary communication, focused partly on the performative delivery of the text – with the pragmatic aims it serves in contemporary political life – and partly on its various effects on the public.

This continuous attention to the listener thus becomes an indispensable critical tool to analyze the main characteristics of a literary text of the Classical tradition, including the Homeric poems. Just to give an example, in the famous chapter 10 of the treatise, dedicated to the selection and arrangement of the material as one of the possible sources of the sublime, the author affirms:

Subl. 10, 1, 5

φέρε νῦν, εἴ τι καὶ ἕτερον ἔχοιμεν ὑψηλοὺς ποιεῖν τοὺς λόγους δυνάμενον, ἐπισκεψώμεθα. οὐκοῦν ἐπειδὴ πᾶσι τοῖς πράγμασι φύσει συνεδρεύει τινὰ μόρια ταῖς ὕλαις συνυπάρχοντα, ἐξ ἀνάγκης γένοιτ' ἂν ἡμῖν ὕψους αἴτιον τὸ τῶν ἐμφερομένων ἐκλέγειν ἀεὶ τὰ καιριώτατα καὶ ταῦτα τῇ πρὸς ἄλληλα ἐπισυνθέσει καθάπερ ἕν τι σῶμα ποιεῖν δύνασθαι· ὃ μὲν γὰρ τῇ ἐκλογῇ τὸν ἀκροατὴν τῶν λημμάτων, ὃ δὲ τῇ πυκνώσει τῶν ἐκλελεγμένων προσάγεται. […]

ὁ δὲ Ὅμηρος πῶς; ἓν γὰρ ἀπὸ πολλῶν λεγέσθω (Il. 15, 624–628) ἐν δ' ἔπεσ', ὡς ὅτε κῦμα θοῇ ἐν νηὶ πέσῃσι λάβρον ὑπαὶ νεφέων ἀνεμοτρεφές, ἡ δέ τε πᾶσα ἄχνῃ ὑπεκρύφθη, ἀνέμοιο δὲ δεινὸς ἀήτης ἱστίῳ ἐμβρέμεται, τρομέουσι δέ τε φρένα ναῦται 5 δειδιότες· τυτθὸν γὰρ ὑπὲκ θανάτοιο φέρονται.

Now have we any other means of making our writing sublime? Every topic naturally includes certain elements which are inherent in its raw material. It follows that sublimity will be achieved if we consistenly select the most important of these inherent features and learn to organize them as a unity by combining one with another. The first of these procedures attracts the listener by the selection of details, the second by the compression of those selected.

[…]

Now compare it with Homer (I select one example out of many):

he fell upon them as upon a swift falls a wave,

huge, wind-reared by the clouds. The ship

is curtained in foam, a hideous blast of wind

See Russell 1964, 8; see also Mazzucchi 1992, 157. For other cases of a special attention of the literary critics on the impact which the rhetorical effects could have on the audience cf. Arist., Rhet. 3, 11, 6; Quint. 8, 2, 21; Demetr., de eloc. 222.

#### 158 Laura Lulli

roars in the sail. The sailors shudder in terror: they being carried away from under death, but only just.

(transl. D. A. Russell, with modifications)

In this passage, on the one hand, the selection of the words and, on the other, the condensation and arrangement of the chosen elements are considered aspects of the text capable of attracting the listener's attention. In light of this, different literary examples are analyzed in the chapter and, after the well-known quotation of Sappho's ode (fr. 31 V.), the author mentions five verses from Iliad book XV, in which the effect of Hector's arrival on the battlefield is compared to a disastrous shipwreck caused by a terrible storm.

Sensitivity to the oral dimension of the text, grasped through the listener's perception of it, has now fully become a critical tool to analyze the cultural heritage of the literary tradition and to provide ever new and effective resources for the rhetorical education of the future elites. This is clear from many passages of the essay, which are dedicated to figures of speech, and particularly to phantasia, i.e. any thought capable of generating words and speeches (Subl. 15, 1–2, 3), and to the imaginary second person, i.e. a kind of variation of person (Subl. 26).

Subl. 15, 1–2, 3

ὄγκου καὶ μεγαληγορίας καὶ ἀγῶνος ἐπὶ τούτοις, ὦ νεανία, καὶ αἱ φαντασίαι παρασκευαστικώταται· οὕτω γοῦν <ἡμεῖς>, εἰδωλοποιίας <δ'> αὐτὰς ἔνιοι λέγουσι καλεῖται μὲν γὰρ κοινῶς φαντασία πᾶν τὸ ὁπωσοῦν ἐννόημα γεννητικὸν λόγου παριστάμενον· ἤδη δ' ἐπὶ τούτων κεκράτηκε τοὔνομα ὅταν ἃ λέγεις ὑπ' ἐνθουσιασμοῦ καὶ πάθους βλέπειν δοκῇς καὶ ὑπ' ὄψιν τιθῇς τοῖς ἀκούουσιν. 2. ὡς δ' ἕτερόν τι ἡ ῥητορικὴ φαντασία βούλεται καὶ ἕτερον ἡ παρὰ ποιηταῖς οὐκ ἂν λάθοι σε, οὐδ' ὅτι τῆς μὲν ἐν ποιήσει τέλος ἐστὶν ἔκπληξις, τῆς δ' ἐν λόγοις ἐνάργεια, ἀμφότεραι δ' ὅμως τό τε <παθητικὸν> ἐπιζητοῦσι καὶ τὸ συγκεκινημένον. […] 3. ἔστι μὲν οὖν φιλοπονώτατος ὁ Εὐριπίδης δύο ταυτὶ πάθη, μανίας τε καὶ ἔρωτας, ἐκτραγῳδῆσαι, κἀν τούτοις ὡς οὐκ οἶδ' εἴ τισιν ἑτέροις ἐπιτυχέστατος, οὐ μὴν ἀλλὰ καὶ ταῖς ἄλλαις ἐπιτίθεσθαι φαντασίαις οὐκ ἄτολμος. ἥκιστά γέ τοι μεγαλοφυὴς ὢν ὅμως τὴν αὐτὸς αὑτοῦ φύσιν ἐν πολλοῖς γενέσθαι τραγικὴν προσηνάγκασε, καὶ παρ' ἕκαστα ἐπὶ τῶν μεγεθῶν, ὡς ὁ ποιητής (Il. 20, 170–171),

οὐρῇ δὲ πλευράς τε καὶ ἰσχίον ἀμφοτέρωθεν μαστίεται, ἑὲ δ' αὐτὸν ἐποτρύνει μαχέσασθαι.

Another thing which is very productive of grandeur, magnificence and urgency, my young friend, is visualization (phantasia). I use this word for what some people call image-production. The term phantasia is used generally for anything which in any way suggests a thought productive of speech; but the word has also come into fashion for the situation in which enthusiasm and emotion make the speaker see what he is saying and bring it virtually before his audience. It will not escape you that rhetorical visualization has a different intention from that of the poets: in poetry the aim is astonishment, in oratory it is clarity. Both, however, seek emotion and excitement. […] Now Euripides devotes most pains to producing a tragic effect with two emotions, madness and love. In these he is supremely successful. At the same time, he does not lack the courage to attempt other types of visualization. Though not formed by nature for grandeur, he often forces himself to be tragic. When the moment for greatness comes, he (in Homer's words)

whips flank and buttocks with his tail

and drives himself to fight (transl. D. A. Russell)

Subl. 26

ἐναγώνιος δ' ὁμοίως καὶ ἡ τῶν προσώπων ἀντιμετάθεσις, καὶ πολλάκις ἐν μέσοις τοῖς κινδύνοις ποιοῦσα τὸν ἀκροατὴν δοκεῖν στρέφεσθαι·

φαίης κ' ἀκμῆτας καὶ ἀτειρέας …

ἄντεσθ' ἐν πολέμῳ· ὣς ἐσσυμένως ἐμάχοντο. (Il. 15, 697–698)

καὶὁἌρατος·

μὴ κείνῳ ἐνὶ μηνὶ περικλύζοιο θαλάσσῃ. (Arat., Phaen. 287)

2. ὧδέ που καὶὁἩρόδοτος· "ἀπὸ δὲ Ἐλεφαντίνης πόλεως ἄνω πλεύσεαι, καὶ ἔπειτα ἀφίξῃ ἐς πεδίον λεῖον· διεξελθὼν δὲ τοῦτο τὸ χωρίον αὖθις εἰς ἕτερον πλοῖον ἐμβὰς πλεύσεαι δύ' ἡμέρας, ἔπειτα ἥξεις ἐς πόλιν μεγάλην, ᾗ ὄνομα Μερόη." (Hdt. 2, 29, 2, 3, 6) ὁρᾷς, ὦ ἑταῖρε, ὡς παραλαβών σου τὴν ψυχὴν διὰ τῶν τόπων ἄγει τὴν ἀκοὴν ὄψιν ποιῶν; πάντα δὲ τὰ τοιαῦτα πρὸς αὐτὰ ἀπερειδόμενα τὰ πρόσωπα ἐπ' αὐτῶν ἵστησι τὸν ἀκροατὴν τῶν ἐνεργουμένων. καὶ ὅταν ὡς οὐ πρὸς ἅπαντας, ἀλλ' ὡς πρὸς μόνον τινὰ λαλῇς,

Τυδείδην δ' οὐκ ἂν γνοίης ποτέροισι μετείη (Il. 5, 85)

ἐμπαθέστερόν τε αὐτὸν ἅμα καὶ προσεκτικώτερον καὶ ἀγῶνος ἔμπλεων ἀποτελέσεις, ταῖς εἰς ἑαυτὸν προσφωνήσεσιν ἐξεγειρόμενον.

Urgency may also be conveyed by the replacement of one grammatical person by another. It often gives the hearer the sense of being in the midst of the danger himself.

you would say they were tireless, never wearied in war,

so eagerly they fought. (Homer)

may you never be drenched in the sea in that month! (Aratus).

"You will sail upstream from Elephantine, and then you will come to a smooth plain. After crossing this, you will embark on another boat and sail for two days. Then you will come to a great city called Meroe" (Herodotus). Do you see, my friend, how he grips your mind and takes it on tour through all these places, making hearing as good as seeing? All such forms of expression, being directed to an actual person, bring the hearer into the presence of real events. Moreover, if you speak as though to an individual and not to a large company, you will affect him more and make him more attentive and excited, because the personal address stimulates:

you could not tell with whom Tydides stood. (transl. D. A. Russell)

Regarding phantasia, the centrality of the effect this figure can have on the listener is reaffirmed from the beginning of the chapter 15, with an explanation about the differences between the characteristics and aim of rhetorical visualization in poetry and in oratory. This subject-matter is illustrated by many examples from Euripides, whose phantasia often attains the sublime not naturally, but thanks to the rules of the tragic literary genre. Moreover, to elucidate this aspect, the author quotes Il. 20, 170 –171, which describes Achilles' fury in battle. Also the variation of person, in chapter 26, is analyzed as a tool which is useful to give the listener the sense of being in the middle of a dangerous situation, and it is illustrated with examples from historiography (Herodotus)³⁶ and epic poetry (Homer, Aratus). In this case, orality is conceived of from the perspective of the rhetorical ability – found in many different literary genres – to capture the listener's attention, to the point of even bringing a sort of 'synaesthetic' experience about that involves both hearing, i.e. the first sense used for the perception and enjoyment ofaliterary work, and sight, i.e. the sense subsequently activated for the mental representation of what one has heard and for the consequent activation of the imaginative and reflective faculties.³⁷

Many other examples could be mentioned to try to elucidate, if only by approximation, the definition and/or re-definition of orality in Greek literary criticism. Anyway, looking at these passages, it seems clear that ancient literary criticism was fully aware of the power of oral communication, which had become not only the 'medium' for literary works, together with the instrument of writing, but also a parameter for the analysis and evaluation of the past literary tradition. This process allowed orality to find application even on another, pragmatic level. Indeed, the oral dimension of the literary experience became crucial for the acquisition of rhetorical skills by the future members of the elite as an essential part of their higher education.

We can, therefore, identify traces of the original orality of the Homeric poems, their endurance, re-adaptation and transformation within the complex development of the communication system in Greco-Roman culture; and we can surely observe that, in order to acquire a high position in ancient society, it was necessary to continue to listen to Homer and to the sound of his verses – to Achilles' cries and Odysseus' tales.

For the freedom of citation in the case of the Herodotean passage, which is both atticized and abbreviated, see Russell 1964, 144. A similar use of the citation by the author of the treatise is evident also for Homeric (cf. 9, 8) and Platonic (cf. 32, 5) passages.

According to Russell 1964, 144, this element would be common in the Greek culture where a crucial value was attributed to the sight, especially in the cognitive and critical processes; cf. e.g. Heracl. fr.B101a D.–K.; Hdt. 1, 8.

### References


Guarducci, M. (1929), "Poeti vaganti e conferenzieri nell'età ellenistica", in: Memorie della Classe di Scienze morali e storiche dell'Accademia dei Lincei VI 2, 629–665.

	- and Contexts, Berlin New York 2011, 27–40.

Mazzucchi, C. M. (2010<sup>2</sup> ), Dionisio Longino. Del Sublime, Milano (1992<sup>1</sup> ).


Russell, D. A. (1964), 'Longinus'. On the Sublime, Oxford.

Sbardella, L. (2004a), "Teocrito pindarico. Il κέρδος, la fama e la poesia omerica in Nemea 7. 17–31 e nell'Idillio XVI", in: Seminari Romani di Cultura Greca 7. 1, 2004, 64–83.

Sbardella, L. (2004b), "Χῖος ἀοιδός: l'Omero di Teocrito", in: R. Pretagostini/E. Dettori (eds.), La cultura ellenistica: l'opera letteraria e l'esegesi antica. Atti del Convegno COFIN 2001, Università di Roma "Tor Vergata", 22–24 settembre 2003, Rome, 81–94.

Schironi, F. (2018), The Best of the Grammarians: Aristarchus of Samothrace on the Iliad, Ann Arbor.

Schmidt, M. (2002), "The Homer of the Scholia: What Is Explained to the Reader?", in: F. Montanari (ed.), with the collab. of P. Ascheri, Omero tremila anni dopo. Atti del Congresso di Genova, 6–8 luglio 2000, Rome, 159–183.

Tarán, L. (2012), Greek and Latin, Edition of the Greek Text, in: Aristotle Poetics. Editio maior of the Greek text with Historical Introductions and Philological Commentaries, by L. Tarán (Greek and Latin, Edition of the Greek text), D. Gutas (Arabic and Syriac), Leiden – Boston.

West, M. L. (2001), Studies in the Text and Transmission of the Iliad, München – Leipzig.

Trabattoni, F. (2005), La verità nascosta: oralità e scrittura in Platone e nella Grecia classica, Rome.

Zanker, G. (1987), Realism in Alexandrian Poetry: a Literature and its Audience, London.

Ziegler, K. (1988), L'epos ellenistico. Un capitolo dimenticato della poesia greca, appendix Ennio poeta epico ellenistico by F. De Martino, transl. G. Aquaro, premises M. Fantuzzi, Bari (= K. Ziegler, Das hellenistische Epos: ein vergessenes Kapitel griechischer Dichtung, mit einem Anhang "Ennius als hellenistischer Epiker", Leipzig – Berlin 1966<sup>2</sup> , 1934<sup>1</sup> ).

### Richard Hunter Homer and 'the Elegists': an Ancient Construction of Difference

Abstract: This paper considers an account in book XIV of Athenaeus' Deipnosophistai of how Homer differed from other archaic poets in terms of musicality. It is argued that this construction, which owes much to Plato, has close links to other ancient accounts of literary and cultural change over time and foreshadows in various respects the modern concern with distinctions between 'oral' and 'written' poetry.

Keywords: Aristoxenos; Athenaeus; Dio Chrysostom; Hesiod; Homer; Longinus.

A repeated modern complaint about the difficulty of understanding the poetic landscape of archaic Greece is the apparent failure of the ancients to theorise 'didactic poetry' until very late in antiquity. How were poetic differences constructed, staged and received in the early periods? In the absence of ancient discussion, we are forced to fall back on our own attempts to schematise,¹ and on the reception of early poetry in the Hellenistic and Imperial periods. The matter is pressing for the subject of this volume, if only because of the fact (and I think it is a fact) that the ancients were on the whole neither very bothered as to whether Homer was an oral poet (on any model of orality) nor would they have very readily understood what is at stake for us in a 'sistema comunicativo orale'. They did, however, care very much about Homer and Hesiod and about differences between them and between them and other poets, both other hexameter poets and those whom we call 'the elegists'. If we want to understand the early Greek poetic landscape, then we can hardly ignore the various efforts of later Greeks, both scholarly and less so, to make distinctions within the poetic corpus available to them.

Aristotle is obviously a key figure here, though how precisely his influence percolated through into later criticism offers, of course, a very thorny nest of issues. Stress is usually given to the emphasis in the Poetics upon mimēsis as a fundamental criterion of poetry, although in antiquity that left the door open to notoriously problematic cases such as Empedocles. In this paper I consider just one later attempt to draw lines of division between Homer and other early poets, but one whose premises still affect, I believe, the way we conceive of

For ways into such modern attempts see Hunter 2014, chapter 2, Sider 2014.

OpenAccess. © 2022 Richard Hunter, published by De Gruyter. This work is licensed under the Creative Commons Attribution 4.0 International License. https://doi.org/10.1515/9783110751963-010

the archaic poetic landscape. This text is a well known passage in book XIV of Athenaeus' Deipnosophistai, a book largely devoted to a history of Greek music; the sources of this account will not be my primary concern here.

In the course of book XIV the deipnosophist Masurius delivers a long comparison of the contemporary state of music with that which prevailed in the past:²

#### Athen. 14, 631e-3c

(631e) τὸ δὲ παλαιὸν ἐτηρεῖτο περὶ τὴν μουσικὴν τὸ καλὸν καὶ πάντ᾿ εἶχε κατὰ τὴν τέχνην τὸν οἰκεῖον αὑτοῖς κόσμον· διόπερ ἦσαν ἴδιοι καθ᾿ ἑκάστην ἁρμονίαν αὐλοὶ καὶ ἑκάστοις αὐλητῶν ὑπῆρχον αὐλοὶ ἑκάστῃ ἁρμονίᾳ πρόσφοροι ἐν τοῖς ἀγῶσι. Πρόνομος δ᾿ ὁ Θηβαῖος πρῶτος ηὔλησεν ἀπὸ τῶν αὐτῶν <πάσας> τὰς ἁρμονίας. νῦν δὲ εἰκῇ καὶ ἀλόγως ἅπτονται τῆς μουσικῆς. καὶ πάλαι μὲν τὸ παρὰ τοῖς ὄχλοις εὐδοκιμεῖν σημεῖον (631f) ἦν κακοτεχνίας· ὅθεν καὶ ᾿Aσωπόδωρος ὁ Φλιάσιος κροταλιζομένου ποτέ τινος τῶν αὐλητῶν διατρίβων αὐτὸς ἔτι ἐν τῷ ὑποσκηνίῳ, "τί τοῦτ᾿;" εἶπεν· "δῆλον ὅτι μέγα κακὸν γέγονεν," ὡς οὐκ ἂν ἄλλως ἐν τοῖς πολλοῖς εὐδοκιμήσαντος. οἶδα δέ τινας τοῦθ᾿ ἱστορήσαντας ὡς A᾿ ντιγενείδου εἰπόντος. καίτοι οἱ καθ᾿ ἡμᾶς γε τέλος ποιοῦνται τῆς τέχνης τὴν παρὰ τοῖς θεάτροις (632a) εὐημερίαν. διόπερ ᾿Aριστόξενος ἐν τοῖς Συμμίκτοις Συμποτικοῖς, ὅμοιον, φησί, ποιοῦμεν Ποσειδωνιάταις τοῖς ἐν τῷ Τυρσηνικῷ κόλπῳ κατοικοῦσιν. οἷς συνέβη τὰ μὲν ἐξ ἀρχῆς Ἕλλησιν οὖσιν ἐκβεβαρβαρῶσθαι Τυρρηνοῖς ἢ Ῥωμαίοις γεγονόσι, καὶ τήν τε φωνὴν μεταβεβληκέναι τά τε λοιπὰ τῶν ἐπιτηδευμάτων, ἄγειν δὲ μίαν τινὰ αὐτοὺς τῶν ἑορτῶν τῶν Ἑλληνικῶν ἔτι καὶ νῦν, ἐν ᾗ συνιόντες ἀναμιμνήσκονται τῶν ἀρχαίων ἐκείνων ὀνομάτων τε καὶ νομίμων καὶ ἀπολοφυράμενοι (632b) πρὸς ἀλλήλους καὶ ἀποδακρύσαντες ἀπέρχονται. οὕτω δὴ οὖν, φησί, καὶ ἡμεῖς, ἐπειδὴ καὶ τὰ θέατρα ἐκβεβαρβάρωται καὶ εἰς μεγάλην διαφθορὰν προελήλυθεν ἡ πάνδημος αὕτη μουσική, καθ᾿ αὑτοὺς γενόμενοι ὀλίγοι ἀναμιμνησκόμεθα οἵα ἦν ἡ μουσική. ταῦτα μὲν ὁ ᾿Aριστόξενος· κἀμοὶ δὲ διὰ τοῦτο φαίνεται φιλοσοφητέον εἶναι περὶ μουσικῆς. καὶ γὰρ Πυθαγόρας ὁ Σάμιος τηλικαύτην δόξαν ἔχων ἐπὶ φιλοσοφίᾳ καταφανής ἐστιν ἐκ πολλῶν οὐ παρέργως ἁψάμενος (632c) μουσικῆς· ὅς γε καὶ τὴν τοῦ παντὸς οὐσίαν διὰ μουσικῆς ἀποφαίνει συγκειμένην. τὸ δ᾿ ὅλον ἔοικεν ἡ παλαιὰ τῶν Ἑλλήνων σοφία τῇ μουσικῇ μάλιστ᾿ εἶναι δεδομένη. καὶ διὰ τοῦτο τῶν μὲν θεῶν A᾿ πόλλωνα, τῶν δὲ ἡμιθέων Ὀρφέα μουσικώτατον καὶ σοφώτατον ἔκρινον, καὶ πάντας τοὺς χρωμένους τῇ τέχνῃ ταύτῃ σοφιστὰς ἀπεκάλουν, ὥσπερ καὶ Αἰσχύλος ἐποίησε· 'εἴτ᾿ οὖν σοφιστὴς † καλὰ † παραπαίων χέλυν'. (632d) ὅτι δὲ πρὸς τὴν μουσικὴν οἰκειότατα διέκειντο οἱ ἀρχαῖοι δῆλον καὶ ἐξ Ὁμήρου· ὃς διὰ τὸ μεμελοποιηκέναι πᾶσαν ἑαυτοῦ τὴν ποίησιν ἀφροντιστὶ πολλοὺς ἀκεφάλους ποιεῖ στίχους καὶ λαγαρούς, ἔτι δὲ μειούρους. Ξενοφάνης δὲ καὶ Σόλων καὶ Θέογνις καὶ Φωκυλίδης, ἔτι δὲ Περίανδρος ὁ Κορίνθιος ἐλεγειοποιὸς καὶ τῶν λοιπῶν οἱ μὴ προσάγοντες πρὸς τὰ ποιήματα μελῳδίαν ἐκπονοῦσι τοὺς στίχους τοῖς ἀριθμοῖς καὶ τῇ τάξει τῶν μέτρων καὶ σκοποῦσιν ὅπως αὐτῶν μηθεὶς <μήτε> ἀκέφαλος ἔσται μήτε λαγαρὸς μήτε μείουρος. (632e) ἀκέφαλοι δέ εἰσιν οἱ ἐπὶ τῆς ἀρχῆς τὴν χωλότητα ἔχοντες


I pass over various textual and interpretative problems which do not affect the argument of this paper.

λαγαροὶ δὲ οἱ ἐν μέσῳ, οἷον·

αἶψα δ᾿ ἄρ᾿ Αἰνείαν φίλον υἱὸν ᾿Aγχίσαο [φίλον υἱὸν Meineke: υἱὸν φίλον] τῶν αὖθ᾿ ἡγείσθην ᾿Aσκληπιοῦ δύο παῖδε (Il. 2, 731)

μείουροι δ᾿ εἰσὶν οἱ ἐπὶ τῆς ἐκβολῆς, οἷον·

Τρῶες δ᾿ ἐρρίγησαν, ὅπως ἴδον αἰόλον ὄφιν (Il. 12, 208) καλὴ Κασσιέπεια θεοῖς δέμας ἐοικυῖα τοῦ φέρον ἐμπλήσας ἀσκὸν μέγαν, ἐν δὲ καὶ ἤια (Od. 9, 212)

διετήρησαν δὲ μάλιστα τῶν Ἑλλήνων Λακεδαιμόνιοι τὴν μουσικήν, πλείστῃ αὐτῇ χρώμενοι, καὶ συχνοὶ παρ᾿ αὐτοῖς ἐγένοντο μελῶν ποιηταί. τηροῦσιν δὲ καὶ νῦν τὰς ἀρχαίας ᾠδὰς ἐπιμελῶς (633a) πολυμαθεῖς τε εἰς ταύτας εἰσὶ καὶ ἀκριβεῖς […] (633b) συνέβαινε δὲ τὸ μὲν παλαιὸν φιλομουσεῖν τοὺς Ἕλληνας· μετὰ δὲ ταῦτα γενομένης ἀταξίας καὶ γηρασάντων σχεδὸν ἁπάντων τῶν ἀρχαίων νομίμων ἥ τε προαίρεσις αὕτη κατελύθη καὶ τρόποι μουσικῆς φαῦλοι κατεδείχθησαν, οἷς ἕκαστος τῶν χρωμένων ἀντὶ μὲν πρᾳότητος (633c) περιεποιεῖτο μαλακίαν, ἀντὶ δὲ σωφροσύνης ἀκολασίαν καὶ ἄνεσιν. ἔσται δ᾿ ἴσως τοῦτο <ἔτι> μᾶλλον καὶ ἐπὶ πλέον προαχθήσεται, ἐὰν μή τις ἀγάγῃ πάλιν εἰς τοὐμφανὲς τὴν πάτριον μουσικήν·

Efforts were made in ancient times to keep music beautiful, and all its technical elements served to maintain its proper organization; this is why every scale had a specific set of pipes, and every pipe-player owned pipes suited to each scale used in the competitions. Pronomus of Thebes was the first person to play all the scales using a single set of pipes. Nowadays, on the other hand, people approach music in a random, careless manner. In the past, moreover, being a popular favourite was regarded as evidence of bad technique; as a consequence, when a pipe-player got applause once while Asopodorus of Phlius (SH 224) was still killing time backstage, he said: "What was that? Apparently something terrible happened!", as if that were the only way the crowd could have given the other man a favourable reception. But I am aware that some authorities report that it was Antigeneidas who said this. The fact is that our contemporaries regard a successful public performance as the ultimate aim of their craft. This is why Aristoxenos says in his Sympotic Miscellany (fr. 124 Wehrli): 'We act like the inhabitants of the Posidonia located on the Tyrrhenian Gulf. What happened to them is that they were originally Greeks but have turned into barbarians and become Etruscans or Romans, and their language has changed, along with all their other practices. They continue today to celebrate only one Greek festival, in which they get together and imitate their ancient way of speaking and behaving; after they wail about them with one another and cry their hearts out, they go back home. We are actually in the same situation, he says; for our theatres have been barbarized, and popular music itself has been utterly degraded, and only a few of us recall privately what music was once like'. Thus Aristoxenos; but it seems to me as well on this account that music deserves to be a subject of philosophical inquiry. There is in fact considerable evidence that Pythagoras of Samos, who has such a great reputation in philosophy, took a more than passing interest in music; indeed, he insists that music holds the fabric of the entire universe together. And by and large ancient Greek wisdom (sophia) appears to have been closely connected with music. This is why they regarded Apollo and Orpheus as the most musical and the wisest of the gods and demigods, respectively, and why they referred to everyone who practised this art as a sophistês ("wise man, intellectual"), as Aeschylus (fr. 314) wrote: 'So then, a sophistês † beautiful † striking a false note on a tortoise-shell lyre'. That the ancients were intimately familiar with music is also apparent from Homer, who in the course of composing as melos all his poetry produces without thought numerous headless and hollow, as well as tapering lines. Xenophanes, Solon, Theognis, and Phocylides, as well as the elegiac poet Periander of Corinth and the others who do not set their poems to music, carefully construct their lines as regards the number and arrangement of the metrical units, and see to it that none of them will be headless, hollow, or tapering. Headless lines are those that limp at the beginning: ' when in fact they came to the ships and the Hellespont. (Il. 23, 2) 'a strap made of a slaughtered bull's hide was stretched over it'. Hollow lines are those (that limp) in the middle, for example: 'straightaway, then, Aeneas the beloved son of Anchises'; 'The two sons of Asclepius, again, were their leaders'. (Il. 2, 731) Tapering lines are those (that limp) at the end, for example: 'The Trojans shuddered when they saw the glistening serpent'. (Il. 12, 208) 'Lovely Cassiepeia, like to the gods in appearance'. 'I filled a large goatskin sack with this and brought it; provisions were inside'. (Od. 9, 212). The Spartans were the Greeks who preserved their music most faithfully, inasmuch as they were deeply immersed in it, and they produced large numbers of lyric poets. Even today they carefully preserve their ancient songs and have a great deal of precise information about them […].

In ancient times, the fact was that the Greeks enjoyed music. But afterward, when the situation became chaotic and almost all the old customs grew antiquated, this tendency was abandoned and low-class musical styles emerged; everyone who composed in them opted for effeminacy rather than gentleness, and for loose licentiousness rather than self-control. This state of affairs will perhaps continue to develop and grow even more pronounced, unless the traditional style of music is brought back to general attention. [trans. S. D. Olson, adapted]

The basic claims are: "in the old days" (τὸ παλαιόν) the proper κόσμος of music was preserved, whereas, these days, people (οἱ καθ᾽ ἡμᾶς) approach much music εἰκῃ καὶ ἀλόγως; what mattered in the past was τέχνη, whereas now it is popular acclaim for performances of a μουσική which is πάνδημος; so important was music in the past that Homer "set all his poetry to music",³ literally "made all his poetry melos", a fact that accounts for the many 'faulty' verses in the poems, whereas non-musical poets (I shall come back to their identity) do not make such 'errors'. It is contemporary Sparta which best preserves traces of and information about the music of the past; the current degenerate state of music is the result of the abandonment of the ancient customs as they "grew old", leading to an ἀταξία characterised by μαλακία, ἀκολασία and ἄνεσις, and this will just get worse unless the πάτριος μουσική is given back its place in the limelight.

This is the natural sense of μεμελοποιηκέναι (so, e. g., Gulick's translation, West 1981, 113); Olson's translation is mistaken here.

If all of this sounds very familiar, it should: behind this jeremiad and behind Aristoxenos lies of course Plato's famous account in book III of the Laws of the decline 'over time' of poetry and music from a properly ordered world in which different genres were kept separate and music was in the hands of experts, not the ignorant mob (there was no rowdy applause), to an ἄμουσος παρανομία and a chaotic θεατροκρατία in which the only criterion of success was giving the mob what it wanted (Leg. 3, 700a7–1b3). Echoes of the Platonic account in any subsequent 'learned' narrative of musical and poetic history are hardly unexpected, but it might still surprise (I think), however familiar we are with the habits of thought in the Second Sophistic, that a writer of the late second-early 3rd century AD should essentially adopt the Platonic schema to explain the current state of affairs, some five hundred and fifty years after Plato. What kind of chronology does Athenaeus have in mind? What are we to infer from the elision of time between Plato and Athenaeus? Whatever the answer, there are important (if familiar) lessons here about the writing of cultural history in antiquity. The explicit citation of Aristoxenos (fr. 124 W<sup>2</sup> ), who is presumably also echoing Plato,⁴ is a very standard way of bolstering one's argument through the citation of important authorities, but it is perhaps too easy to forget that Aristoxenos too was writing in the middle-late 4th century BC, though he is apparently adduced to explain a contemporary (i.e. c. AD 200) situation. As Paola Ceccarelli puts it, "Aristoxenos is […] appropriated by Masurius as giving evidence for the Deipnosophists' own times".⁵ We might, with some possible justification, plead that Masurius has in mind a broad distinction between 'classical' (τὸ παλαιόν) and 'post-classical᾽ (οἱ καθ᾽ ἡμᾶς) times, with the change happening at some very vaguely defined time in what we call the 4th century BC, so that Aristoxenos and the deipnosophists are in fact grouped in the same period; such a structure would certainly fit a prominent attitude towards the past in the Second Sophistic.⁶ Nevertheless, the strangeness of such an attitude, if that is indeed what we are dealing with, deserves regularly to be defamiliarised. In antiquity, literary history was also cultural and political history.⁷

Masurius apparently claims that Homer's 'faulty' verses (though he does not use such a description) are the result of the fact that he set all his poetry to

So rightly, e. g., Barker 1984, 291 n. 159; see also D'Alessio 2020, 65– 67.

Ceccarelli 2000, 282. Ceccarelli offers an important study of the 'fluidity of time' and the interplay between different temporal and enunciative levels in book XIV of the Deipnosophistai. One might compare chapter 44 of Longinus, On the Sublime where some modern critics have understood the complaint about "our time" (κατὰ τὸν ἡμέτερον αἰῶνα) as referring to everything after the Classical period of Greece; the matter is, of course, hotly disputed.

See Hunter 2009, 14–15.

music, whereas "Xenophanes, Solon, Theognis, and Phocylides, and Periander of Corinth, the elegiac poet, and the rest" did not, or, literally, "did not apply μελῳδία to their poems". It is presumably important that Masurius/Athenaeus does not explicitly state a chronological relationship between these poets and Homer, although I think it would not be overbold to suggest that the unspoken assumption is that they came after the epic poet; the fact that they did not employ μελῳδία for their poems may in fact be a small-scale example of the thesis about the decline of music over time which is at the centre of this section of the text. The relationship between Homer and some other archaic poetry (and I will return to the identity of the poets) is here brought into a larger view of cultural history which we find time and again in the so-called Second Sophistic: this is, to put it at its simplest, a view that 'in the old days' the world was a musical and poetic one, but now it is (quite literally) a more prosaic place, where very few people actually understand music properly.⁸ Masurius' view that Homer "made all his poetry melos" is in fact anything but a mainstream ancient view; it is, however, of a piece with, and presumably is fashioned to reflect, a widespread general view of the history of Greek poetry and music. The idea that Homer 'sang' and sang to the music of an instrument would not surprise anyone,⁹ but the distinction which Masurius makes between Homer and other early poets opens up a very different set of questions.

Before pursuing the matter further, I should perhaps stress that Masurius seems here to be making a different claim than the familiar ancient linkage between Homer and μέλος, a claim which is rooted in, and takes part of its inspiration from, ancient performance culture. In Athen. 14, 620c, for example, the narrator reports Chamaileon (fr. 28 Wehrli = 30 Martano–Matelli–Mirhady) as stating in his On Stesichorus that "not only the poems of Homer, but also those of Hesiod and Archilochus, and moreover (ἔτι δέ) those of Mimnermus and Phocylides were sung" (μελῳδηθῆναι).¹⁰ What Chamaileon meant by this

The most striking and innovative version of that cultural history is placed in the mouth of Theon in Plutarch's On the Oracles at Delphi; I discuss this text in Hunter forthcoming b.

There isahelpful collection of material in Koller 1956. The standard ancient critical dichotomy of poetry is 'unaccompanied verse' and 'sung verse' or lyric, cf., e. g., Arist., Poet. 1449b29–30; the former may be ποίησις ψιλή (e.g. Plat., Phaedr. 278c2) or even just μέτρον (e. g. Dion. Hal., de comp. verb. 11, 3 (= II 37, 20 U-R), whereas the latter is usually μέλος or ᾠδή. Homer naturally falls into the former category. On this aspect see also Lulli, in this volume, 150–154.

Olson's translation "recited" neither renders the natural sense of μελῳδεῖν, though that verb, admittedly, covers a wider range of forms of delivery than our "sing", nor acknowledges that most of the citations in this passage of Athenaeus use the rather different ῥαψῳδεῖν.

may be debated, but the context in Athenaeus is plainly concerned with public performances subsequent to the poets themselves. So too, in a famous passage of [Plutarch], De musica (1132c), Heraclides Ponticus (fr. 157 Wehrli = 109 Schütrumpf) is reported as having stated that Terpander "set to music (μέλη περιτιθέντα) his own ἔπη and those of Homer and sung them in the contests". It seems to be something like this scenario that Chamaileon is being used to support in 620c, whatever the original context of that claim. What is not apparently at stake in 620c is anything like what we would call 'generic difference' between Homer and other poets; what is important there, rather, is the history of Greek performance culture, in which, no doubt, all manner of 'mixed' styles could (and did) flourish.¹¹

To return to 631d, and the distinction between Homer and some other early poets. There is another pattern here which seems to point to a chronological structuring within the passage. Homer's concern with music led him to create many 'faulty' verses ἀφροντιστί (the adverb happens only to appear here and once in Origen); this is perhaps not so much "carelessly" (Olson), as "without thought" (Gulick) or rather "without being bothered", i.e. Homer had more important fish to fry than getting the hexameter correct and regular. What are indeed 'faults' criticised (and defended) in the metrics of oracular verses by Plutarch's characters in the On the Oracles at Delphi¹² are in Homer, according to Masurius, signs that this is not what was most important. The idea must be that the music of the phorminx and the 'singing' of the verse, which Martin West sought to reconstruct in a famous paper, not only 'covered up', as we would say, the metrical irregularities but also explained them: Homeric irregularities in the hexameter both prove musical performance and the fact of musical performance justifies (as it were) the irregularities.

Our attempts to get behind Masurius' argument here are not helped by his choice of examples. Four Homeric verses, including those which are standard examples of metrical 'faults' in such a grammatical context, one further verse confected from two Homeric verses (a familiar phenomenon in later critical and grammatical texts), and two otherwise unidentified verses, which (I believe) appear nowhere else in our remains of Greek literature; one of them would, however, fit easily enough into the Iliad, and the other is often treated (as by Martin West) as a one-off variant of a transmitted Iliadic verse, namely 8, 305 (Teucer trying to shoot Hector):

The proposer of an honorific decree in 2nd-century AD Ephesos styles himself ποιητής […] μελοποιὸς καὶ ῥαψωιδός (IEph. 22), and such lists are at least suggestive for what happened in performance, as well as for the performers themselves.

Cf., e. g., 397c-d, interestingly citing these same three 'faults' in the hexameter.

#### καλὴ Καστιάνειρα δέμας ἐϊκυῖα θεῇσι

Why, however, someone should wish to make Kassiepeia, a name best known as the wife of Kepheus and mother of Andromeda, the name of Gorgythion's mother remains a mystery on this standard explanation of the verse which Masurius cites. Of the 'non-Homeric' verses (though that may be a tendentious way to classify them), one is transmitted in Athenaeus without any obvious metrical fault and so one has to be introduced by emendation, and the other (as West on Il. 8, 305 notes) perhaps requires us to read εἰοικυῖα, with a metrical lengthening attested by a minority of witnesses to Iliad 18, 418 but accepted there by modern editors.¹³ Thus, Masurius certainly offers a very singular collection of examples to support a very singular argument.

The other poets named in Athenaeus "labour" (ἐκπονοῦσι) over their verses and make sure (I paraphrase) that no grammarian can accuse them of metrical failings.¹⁴ If this narrative also sounds familiar, then (again) it should, for it seems very close to the famous chapter of 'Longinus' On the Sublime (33) which contrasts sublime poets (Homer, Archilochus, Pindar) who do "make mistakes" with poets such as Apollonius and Eratosthenes who do not; there is here a different spin put upon Homer's 'mistakes', but the structural analogy of the two passages seems very clear. In place of Athenaeus' ἀφροντιστί, 'Longinus' refers to παροράματα δι' ἀμέλειαν (Subl. 33, 4) and, guided by Athenaeus, we should seriously consider including metrical slips in the παροράματα to which 'Longinus' refers. The Longinian chapter has been called a "manifesto directed against what we may call the Callimachean ideal"¹⁵ and, despite the presence of Ion and Bacchylides among the flawless B-list, it does indeed seem to be one of the earliest passages of Greek criticism to acknowledge Hellenistic poetry as a separate entity. ἐκπονοῦσι in Athenaeus evokes the same ideal of poetic labor associated with the self-conscious, 'flawless' poetry of the Hellenistic and Roman periods.¹⁶ Masurius' distinction places the non-Homeric poets whom he names in a new world of non-musical ἀκρίβεια, a world which else-

See Chantraine, 1953, I 129, 424, noting (129) that one could read νεήνισσιν FεFικυῖαι, though the error would be hard to explain.

Kaibel proposed to insert e. g. <τελείους> after μέτρων, so that the sense would be "constructed their verses <as complete> with regard to the number and arrangement of the metrical feet […]"; this seems, on balance, unnecessary, if the prefix in ἐκπονοῦσι is given its full weight. Russell 1989, 308. For further discussion and bibliography on On the sublime 33 see Hunter 2011, 230 –234.

Cf. Theocr., Id. 7, 51 with Hunter's note. For further links between gnomic-didactic poetry and Hellenistic poetry see Hunter forthcoming a.

where may be associated with written, rather than performed, poetry;¹⁷ their literary distance from Homer is, roughly speaking, that which separates Bacchylides from Pindar in 'Longinus'. One particularly intriguing possibility is that Athenaeus' distinction between verse set to music and other verse points forward in certain respects towards our own modern concern with 'oral' vs 'written' poetry. It is of course important not to confuse the dynamics of composition with the dynamics of performance, but the ancients knew, no less than we do, when what seemed to be clear poetic facts needed explanation.

Why is this particular group of poets chosen to be distinguished from Homer? Martin West described them as "the elegists",¹⁸ and he seemed to think that this notice in Athenaeus derived from someone who assumed, without much real knowledge and almost certainly erroneously, that all gnomic poets wrote elegiacs.¹⁹ There may be an important element of truth in the latter claim (and it is likely enough, I think, that no one of Athenaeus' day had read any verse by Periander of Corinth), but the description of Periander as ἐλεγειοποιός does not necessarily suggest that it is elegiac metre which most binds these poets together, despite the fact that most critics of Athenaeus' day would indeed have accepted that elegy was (relatively speaking) non-musical. Rather, this list seems to offer some kind of mixture between traditional Sages and gnomic/didactic poets, who are both set in opposition to Homer. If so, this passage of Athenaeus would be a further example of a contrast which was an important theme in the writing of poetic history in the Second Sophistic. Dio's young Alexander, for instance, puts this contrast at the centre of his display to his father of his scholastic knowledge and of the reasons why he will give attention to no poet other than Homer:

#### Dio Chrys. 2, 4–5

τὰ μὲν οὖν ἄλλα ποιήματα ἔγωγε ἡγοῦμαι τὰ μὲν συμποτικὰ αὐτῶν, τὰ δὲ ἐρωτικά, τὰ δὲ ἐγκώμια ἀθλητῶν τε καὶ ἵππων νικώντων, τὰ δ᾿ ἐπὶ τοῖς τεθνεῶσι θρήνους, τὰ δὲ γέλωτος ἕνεκεν ἢ λοιδορίας πεποιημένα, ὥσπερ τὰ τῶν κωμῳδοδιδασκάλων καὶ τὰ τοῦ Παρίου ποιητοῦ. ἴσως δέ τινα αὐτῶν καὶ δημοτικὰ λέγοιτ᾿ ἄν, συμβουλεύοντα καὶ παραινοῦντα τοῖς πολλοῖς καὶ ἰδιώταις, καθάπερ οἶμαι τὰ Φωκυλίδου καὶ Θεόγνιδος.

Of other poems, I consider some to be sympotic, some erotic, some encomia of athletes and victorious horses, some to be dirges for the dead, and some to be composed for the sake of laughter and ridicule, like the poems of the comic poets and the Parian poet [Archilochus].

See Hunter 2003.

West 1981, 113.

See also the entries for Bias, Periander and Phocylides in West 1992.

Perhaps some of them might also be called popular, since they give advice and direction to the mass of private citizens, as, for example, the poetry of Phocylides and Theognis.²⁰

Dio here perhaps wants us to feel something of a specifically Aristotelian, or even just a peripatetic, flavour, for it was that poet and that school which, more than any other, were associated with catalogues and distinctions between types,²¹ and Aristotle had of course been Alexander's teacher. Somewhat later in the same treatise, Alexander turns his attention to how Homeric poetry itself fills all the roles traditionally ascribed to gnomic and didactic poetry:

Dio Chrys. 2, 44

δεῖ δὲ τοῦ ποιητοῦ τὰ μὲν ὡς συμβουλεύοντος καὶ παραινοῦντος ἀποδέχεσθαι, τὰ δὲ ὡς ἐξηγουμένου μόνον, πολλὰ δὲ ὡς ὀνειδίζοντος καὶ καταγελῶντoς.

Some parts of the poet we must understand to be giving advice and direction, some to be merely narrating, and many offering reproof and mockery.

Homer, a poet διδάσκων καὶ παραινῶν (2, 48), is himself the only epic, gnomic and didactic poet whom kings require. There is much that could be said here about Dio's Alexander's debt both to Plato and to the grammatical tradition best preserved for us in the Homeric scholia,²² but the structure of Alexander's analysis helps, I think, to confirm what is of primary importance in Masurius' contrast in Athenaeus. This imperial contrast between Homer and gnomic/didactic poetry may suggest that the latter is constructed as belonging to a period in which poetry was already losing a central place in Greek culture and becoming, rather, a separate literary activity in the hands of specialists. If there is anything to this suggestion, it is at least worth asking also whether the new prominence of

Hesiod is a very loud absence here, but that is explained by the particular attention then paid to him in the following chapters of Dio 2.

Alexander's use of δημοτικά applied to song is hard to parallel, as indeed his slight apology for the term suggests; I wonder whether we are to understand that this is his attempt at defining a genre a là Hellenistic scholarship, adjective in –ικός and all; one might have expected Theognis and Phocylides to be classed as something like παραινετικά. What marks Alexander's classificatory attempt, however, and marks it as typical of him, is that it foregrounds the implied audience ('ordinary people', rather than kings), whereas classificatory labels, as 2, 4 shows, are normally drawn from the subject or the occasion of the poem. Perhaps the closest analogue to Alexander's classification is Proclus' proem to Hesiod's Works and Days (p. 1 Pertusi), which marks the purpose of the poem as παιδευτικός, but notes that it concerns "the management of the household and the non-political life (ἡ ἀπράγμων ζωή)" and our "private life" (ἴδιος βίος). Alexander's choice to cite Lycurgus as a leader who "learned from Homer" and was indeed an ἐπαινέτης Ὁμήρου (2, 44–45) is another direct response to Plato's challenge to Homer in Resp. 10, cf. 599d8, 606e1, Hunter 2018, 31–32.

'didactic poetry' in the Hellenistic and imperial periods – the poetry of such as Nicander, whom Plutarch groups with Empedocles, Parmenides and Theognis (Mor. 16c),²³ Oppian and the astronomical and geographical poets – offered some of the stimulus to this scholarly reconstruction of the distant poetic past. However that may be, what we can say is that part at least of the ancient critical tradition set archaic didactic and gnomic poetry against Homer in a way which markedly resembles one view (most familiar to us from chapter 33 of On the sublime) of the development of Greek poetry set over a much longer time span; or rather, perhaps, a long development over time was retrojected back to explain patterns of difference conceived as internal to the archaic period. This might seem a strange thing to do, but if we try to imagine what it was like, even as early as the 3rd century BC, let alone in the 2nd and 3rd century AD, to seek to write the history of archaic poetry, it will begin to seem much less strange.

As a kind of necessary footnote, the absence of Hesiod from Masurius' account deserves notice. Modern scholars have looked very hard for metrically 'deficient' hexameters in Hesiod, of the kind to which Masurius points in Homer, but the search has yielded at best uncertain results.²⁴ On which side of Masurius' border would the poet from Ascra fall? How 'musical' were Hesiodic verses thought to be, and does an answer to that question have anything to do with the status which Hesiod subsequently enjoyed in the Hellenistic and Imperial periods? The fact that, in the Classical period, rhapsodes would perform Hesiod (as well as Homer), even if Plato's Ion refused to join in, does not provide the answer we are looking for, as Masurius' doctrine, whatever its sources, is clearly, as we have seen, an intellectual and historical construction which has its roots, at least partially, in Homeric versification. It might be tempting, however, to think that Hesiod is not mentioned precisely because he does not fit properly into that construction; here, as elsewhere, Hesiod resists categorisation.

### References

Barker, A. (1984), Greek Musical Writings, vol. I, Cambridge. Ceccarelli, P. (2000), "Dance and Desserts. An Analysis of Book Fourteen", in: D. Braund/J. Wilkins (eds.), Athenaeus and His World, Exeter, 272–291.

Chantraine, P. (1953), Grammaire homérique, Paris.

See Hunter/Russell 2011, 88.

See West 1966, 91–93.


# Sonja Zeman Paradoxes of 'Orality': A Comparison between Homeric Oral Poetry and the Heroic and Courtly Epics in Middle High German

The processes of composition, memorization and performance in oral poetry turn out to be more complex than was once supposed. Finnegan 1977, 86

Abstract: On the premise that cross-linguistic comparisons are a suitable tool for investigating the mechanisms of the 'oral' communicative system within the linguistic structure of historical texts, the paper undertakes a comparison between the Homeric tradition and Middle High German (MHG) heroic and courtly epics. Based on a revision of traditional accounts of 'orality' in older stages of language and an analysis of different grammatical features which are commonly seen as 'oral', this paper shows that the main difference between the investigated texts is not primarily shaped by medial aspects of 'orality', but rather by the pragmatic presupposition of the poet with respect to narrative truth. This will allow for a more fine-grained view of both the communicative triangle SPEAKER – ADDRESSEE – WORLD as a historically variable constellation as well as of the relationship between the different dimensions of orality and their diachronic mechanisms.

Keywords: Oral poetry; Middle High German; Homer; deixis; narrativity.

### 1 The "Vortex" of Orality

The 'oral' communicative systems of older stages of language have retained their fascination until today. Decades after the seminal work by Parry and Lord, we have gained a more detailed view of the Homeric epics and their 'formulaic poetry' (e. g. Bakker 1997b, Bozzone 2010, Bozzone 2021). In addition, many other characteristics of the Homeric epics have been investigated within the realm of orality. Although this has led to valuable insights, it has also been criticized that orality had to serve as an explanation for too many features that are not directly linked to the written vs. spoken distinction:

Despiteasignificant amount of revisionist work, the concept of orality remains something of a vortex into which a range of only partly related issues have been sucked: authorial originality/communal property; impromptu composition/meditated composition; authorial and audience alienation/immediacy. The relevance of orality to these issues is not in dispute; the problem is that they do not vary along specifically oral/literate axes.¹

Furthermore, it has been shown that a dichotomous difference between written and spoken language is not adequate to capture the various aspects of oral composition, performance, and transmission within older stages of language, and that the concept of orality rather comprises several different aspects.² However, which of these different aspects of orality have which effects on the linguistic structure has remained an open question. In order to win a more fine-grained view of the interplay between the different aspects of orality in historical stages of language, and, as such, of the general 'mechanisms' of the oral communicative system, I will zoom out of the archaic epics by investigating orality in Middle High German (MHG). Based on a revision of previous accounts of orality and a distinction between different aspects of orality, I will compare the Homeric epics with two different types of MHG epics and show that such a comparison allows for investigating the 'vortex' in a more systematic way.

The paper is structured as follows: Section 2 starts from the observation that many 'oral' linguistic features have been evaluated inconsistently – or in some cases even in contradictory fashion – in the literature. This calls for a distinction between different dimensions of orality which is based on a revision of approaches on historical orality in the Classical, Romance and German scholarly tradition. The benefit of this distinction is illustrated in Section 3, which offers a comparison between several 'oral' features in the MHG courtly and heroic epics. Since the two epics are situated in the same period of time and the same cultural context but reflect different medial constellations, their analysis allows for insights with respect to the effects of the different dimensions of orality on the linguistic structure. In this respect, my analysis shows that the main difference between the two poems lies within the attitude of the narrator towards the told story. In Section 4, this result is discussed in comparison with the Homeric epics, which leads to hypotheses with respect to the relationship between the different dimensions of orality and the oral predisposition within a specific cultural context. In sum, it is argued in Section 5 that the distinction between different dimensions of orality allows for dissolving the paradoxical descriptions of oral features, and, as such, a clearer view of the 'vortex'.

Hall 2008, 279.

See e.g. Koch/Oesterreicher 1985; Bakker 1998, Bakker 2005; Zeman 2016a, Zeman 2016b.

So although 'oral' features of Homeric epics will be the starting point, the paper does not claim to say anything new about the Iliad or the Odyssey but is concerned with the concept of orality in its historical dimension per se, the general mechanisms of oral communicative systems.

### 2 Paradoxes of 'Orality'

If we have a look into Hall's 'vortex' of orality, we find many different linguistic features of epic poems that have been classified as 'oral' such as formula, metrical patterns, repetitions, orthographic alternation, anacolutha, paratactic structure, 'illogical' narrative chronology, tense alternations, interjections, modal particles, deictics referring to the 'here and now', epithets, a general 'fluidity' of the text, etc. However, almost every feature of this (not comprehensive) list has been controversially discussed with respect to how oral it actually is. One famous example in this respect is ring composition, i.e. the representation of thematic elements in a pattern like A-B-C-B-A, which has commonly been seen as a "fundamental structural device of Homeric narrative".³ On the one hand, ring composition has been described as a pattern that reflects "the activity of performance and composition itself"⁴ and "could well have been evolved by oral poets".⁵ As linked to the medial conditions within oral performance, ring composition would thus be an oral feature par excellence. This assumption has also been supported by Person, who argues that ring composition can be understood as an expansion of common practices found in everyday oral conversation.⁶ Minchin, on the other hand, has argued that the A-B-C-B-A pattern refers "primarily to the pragmatics of storytelling",⁷ and is a feature which is only indirectly linked to the oral predisposition of Homer. Inasimilar way, Douglas has argued that ring composition is a cognitive pattern "hardwired in the brain"⁸ and, as such, found in many narratives all over the world, but that it is not specifically characteristic for oral composition. In addition, ring composition has also been characterized as a literate principle of elaboration that cannot be deduced from the principles of oral composition.⁹ According to Whitmann, for example, ring


Nimis 1998, 65.

composition might have been a mnemonic device in its original function, but constitutes an artistic architectonic principle in the Iliad.¹⁰ As this short overview shows, one single feature can thus be evaluated as an oral compositional device and as a pattern of artistic elaboration at the same time.

The main problem of such inconsistent evaluations leads back to the fact that the investigation of orality in older stages of language is based on a paradox itself: we are looking for oral residues of language in written texts. So what does 'oral' mean? In order to cope with this methodological problem, it is common to distinguish between 'medial' vs. 'conceptual'¹¹ or 'cognitive'¹² orality. Medial orality refers to the technical dimension, i.e. the fact that words are spoken, and is dichotomously opposed to written language. 'Cognitive' or 'conceptual' orality, on the other hand, is a gradual concept that refers to the fact thata(written or spoken) utterance can more or less reflect an 'oral style', regardless of its actual medial realization. Historical orality, i.e. orality that is only preserved in written texts of historical stages of language, can therefore not be anything else than conceptual/cognitive orality by definition.Which properties are characteristic of such an 'oral style', however, has remained an open question.¹³ Koch and Oesterreicher themselves have defined conceptual orality as a mixed bag ("bunte Mischung"),¹⁴ i.e. as a space in which components of language of proximity and language of distance combine and constitute particular linguistic constellations.¹⁵ In order to gain a more nuanced view of conceptual orality, it is thus necessary to have a closer look at these different constellations. In other words: we first have to examine what is inside the bag and disentangle the different aspects of orality and, second, specify their relationship to each other.

This step is crucial since the heterogeneous 'oral' features, such as formula, metrical patterns, repetitions etc. are obviously not oral in the same way. Whereas features like syntactic breaks and metrical patterns are supposed to reflect the fact that the poems have been composed 'on line' simultaneously to their reception, and thus are conditioned by cognitive parsing restrictions in spoken language, deictics and interjections are not directly linked to orality in a technical sense; rather they are linguistic devices that create an impression of

Whitmann 1958, 98.

Koch/Oesterreicher 1985.

Fleischman 1990a, Fleischman 1990b.

See Hennig 2009 and Zeman 2016a, Zeman 2016b for discussion.

Koch/Oesterreicher 2007, 351.

"Nun können wir das konzeptionelle Kontinuum definieren als den Raum, in dem nähe- und distanzsprachliche Komponenten im Rahmen der einzelnen Parameter sich mischen und damit bestimmte Äußerungsformen konstituieren." (Koch/Oesterreicher 1985, 21).

proximity between the poet, his audience, and the story world. Such features are commonly referred to in terms of 'vividness', 'immediacy', and 'enargeia' and, as such, linked to concepts like 'language of proximity' ("Nähesprache")¹⁶ and 'involvement'.¹⁷ Conceptual orality thus combines primary aspects of orality in the medial sense of the term, and secondary phenomena like stylistic markers of deictic immediacy, i.e. in Lord's words, aspects in and aspects for performance.¹⁸

The difference between the medial production of the poem and the conceptualization of proximity is crucial for every act of linguistic communication, whether it takes place today or has taken place 2800 years ago. What complicates the matter of historical orality is the common assumption that older stages of language are 'more oral' in general, since "[t]he oral mental habits of all languages that have not grammaticalized writing necessarily leave their mark on linguistic structure".¹⁹ The language of Classical and Medieval documents has therefore been claimed to be "essentially a spoken language".²⁰ It has, however, remained an open question to what an extent the style of texts of 'oral' cultures is 'more oral', and how cultural predispositions relating to the variable relationship between literacy and orality shape the 'oral style' of the poems, see Fig. 1.

It is thus the oral predisposition that is a variable of unknown type. Yet, this variable is crucial since investigations of 'orality' usually go beyond mere linguistic features and textual style. In the tradition of Ong's 1982 conception of 'primary orality', studies often do not speak of orality in order to describe "what happens when someone talks, but to label a period or a culture as different with respect to our own, literate culture".²¹ In investigations of the Homeric epics, for example, 'oral' often not only refers to the fact that oral poetry is composed during oral performance,²² but also that oral poetry is composed "in a manner evolved over many generations by singers of tales who did not know how to write".²³ Whereas oral transmission has been a crucial aspect for Homeric oral poetry, it is, however, certainly not a feature of every epic poem that has been evaluated as 'oral'. In this respect, Finnegan 1977 has objected that the characteristics of oral poetry identified by Parry and Lord cannot adequately describe the variety of oral story telling. In order to take into account cross-linguis-

Koch/Oesterreicher 1985; Ágel/Hennig 2006.

Chafe 1982.

Lord 1960, 13.

Fleischman 1990b, 22; see also Fitzmaurice/Taavitsainen 2007, 19, 22.

Fleischman 1990b, 24; similarly also Slings 1992, 100.

Bakker 1998, 33; see also Hall 2008.

Lord 1960, 5.

Lord 1960, 4.

Fig. 1: Dimensions of 'orality' (adapted from Zeman 2016a, 183).

tic diversity, Foley 2002 has proposed to differentiate between composition, performance and transmission of the poem. The evaluation of these three individual parameters allows for a more fine-grained classification of an instance of oral poetry, which can be described according to the 'oral' constellation of the parameters (see Tab. 1).

Tab. 1: Foley's 2002 differentiation of 'oral' constellations.²⁴


As seen in Tab. 1, the differentiation is in particular useful to describe the possible constellations of the "voices from the past", which can display different values with respect to composition, performance and reception. According to Foley and Ramey, both the Medieval and the Homeric epics belong to the "voices from the past", i.e. textual artefacts that "reach us only in writing, but various kinds of internal and external evidence argue that they derive from oral traditions."²⁵ In order to compare the different "voices from the past" category, it is thus necessary to determine the specific values with regard to the oral constellation, see also Finnegan:

There turn out to be different combinations of processes of composition, memorization and performance, with differing relationships between them according to cultural traditions, genres and individual poets. There are several ways – and not just one determined way suitable for 'the oral mind' – in which human beings can engage in the complex processes of poetic composition.²⁶

The distinction can furthermore account for the fact that diachronically, there is no abrupt transition from orality to literacy. As "voices from the past", both Homeric and medieval epic poems are not purely oral as they diverge in many ways and degrees from Ong's 1982 criteria of primary orality.²⁷ Both Homeric and Medieval epics have also been described as transitional, semi-oral or postoral texts. These are, however, problematic terms, as they presuppose a straight line of development from orality to literacy, whereas it is nowadays commonly accepted that 'oral' is a gradient property, allowing for different constellations with respect to the relationship of spoken and written language within a society. As such, orality is not "incompatible with writing",²⁸ and orality and literacy are not contradictory concepts. This holds in particular for the Medieval poems that are rooted both in oral tradition and the written word. As will be seen in the following, even within a rather short period of time, text genres can differ with respect to their oral constellations.

In addition, we have to take into account that oral epic poems are not instances of ordinary everyday storytelling but of 'special speech'²⁹ that displays a high degree of elaboration. As such, oral poetry is characterized by the interplay between oral traces within the text and artistic devices like metre and formulae. It

Foley/Ramey 2012, 85.

Finnegan 1977, 86.

See Stanley 1993, 274 for the Iliad, Haferland 2019 for the MHG Nibelungenlied.

Bakker 1998, 35.

Bakker 1998, 38.

is thus obvious that there is no straight development from transcription to elaboration, either.

In sum, historical orality cuts across the oral/literal axis by comprising different pragmatic features on different linguistic levels. In the following discussion, 'orality' is therefore used as a general term under which heterogeneous phenomena that can be attributed to different dimensions of orality are discussed. What is at stake is the question whether there are systematic dependencies between specific linguistic features and aspects of orality. In order to approach this question, a comparison between MHG heroic and the courtly epic poems seems promising, since both types of poems belong to the same temporal and cultural context but display different oral values with respect to the parameters of composition, performance and transmission.

### 3 'Oral' features in Middle High German and Homeric Epic Poems

For the MHG epic poems, it is commonly accepted that orality played "a crucial role in shaping the grammar (in the linguist's sense) of medieval vernaculars and, consequently, the linguistic structure of our texts".³⁰ It is, however, also evident that the MHG epics are rooted both in the oral tradition and the written word. This can be seen in the fact that many authors of epic poems emphasize their book-based erudition and their knowledge of the written sources of the narrated story. MHG epic poems could be both presented orally and read as texts, but it is assumed that free oral presentation of the poems was rather the rule than the exception.³¹ MHG epics are thus not situated in a state of primary orality in the sense of Ong, but in a 'third kind of medial condition'.³²

There are, furthermore, crucial differences between two different genres of epic poems, i.e. the heroic and the courtly epics.³³ Heroic epics like the Ortnit refer to Germanic epic cycles that were traditionally known. Their authors are not mentioned within the poem and remain anonymous. The story starts immediately or after only a short prologue. The courtly epics, on the other hand, tell stories that originate from French sources and are tied to courtly knighthood. They can frequently be ascribed to a specific author, who often mentions himself

Fleischman 1990, 22.

Haferland 2004, 463.

Däumer 2013.

Philipowski 2007, Haferland 2019.

in a prologue that precedes the actual story and often asserts its correctness. These genre differences³⁴ have been seen in connection with different medial conditions. It is assumed that heroic epics have been transmitted without aid of the written word before their transcription as texts. The courtly epics, in contrast, were probably composed and transmitted as written texts. Indications for the latter are frequent assertions by the author that the 'truth' of the narrated texts is ensured by written sources. The fundamental difference is thus seen in the fact that MHG heroic epics are the result of oral transmission, whereas courtly epics are composed, performed and transmitted with aid of the written word (see Tab. 2).³⁵


Tab. 2: The medial constellation of MHG heroic and courtly epics.

In order to investigate whether and how these two different medial constellations leave different traces within the epic poems, I will compare two canonical examples of heroic and courtly epics (both dated around 1200 AD), i.e. Nibelungenlied B and Tristan.³⁶ The Nibelungenlied is the oldest large-scale epic of MHG. Its status as an oral epos is controversial since it has been seen both as a 'book epos', i.e. an epos that is composed in written form but intended to be read and performed,³⁷ and as an epos based on oral transmission. Arguments for the first view are the fact that it does not display 'Stabreimdichtung' which is characteristic for early Germanic heroic epics. Furthermore, it also integrates schemata and topoi from the literary

See Philipowski 2007, 49–57 with further references.

See Philipowski 2007, 44: "Doch der grundlegende Unterschied zwischen höfischer Epik und Heldenepik ist, dass letztere das Resultat eines Tradierungsprozesses ist, der sich unserer Kenntnis nach m ü n d l i c h vollzogen hat, während die höfische Epik aller Wahrscheinlichkeit nach s c h r i f t g e s tütz t entsteht, schriftgestützt vorgetragen und schriftlich überliefert[…] wird." (emphasis in original).

The following editions have been used: Reichert 2020 and Marold 2016.

See Heusler 1956, Müller 2012, and Heinzle 2015.

tradition of the courtly epos, which is unusual for heroic epics.³⁸ There are also important differences when compared to the Homeric epics: There is no formula-thesaurus,³⁹ but there are stereotypical patterns for the representation of thematic scenes like battles, festivals, arrivals and departures.⁴⁰ Moreover, it has been questioned whether the large-scale design and the consistency of the text would have presupposed composition by aid of the written word. On the other hand, its metrical-stanzaic form, the emphasis on visual and spatial representations, as well as incongruencies in the text, have been taken as an argument that the production of the text was based on memory⁴¹ and that the text was "without doubt"⁴² composed for recitation in an oral performance. In sum, the Nibelungenlied is thus based on a mixture of literary principles and codification through memory.⁴³ Tristan is a classic courtly epic. We can assume that it was composed with the aid of writing and that its content was transmitted by written sources, but that it was probably performed orally.

In the following section, I will examine whether and how these differences with respect to the medial predisposition are reflected in the linguistic structure of the poems. In order to do so, I will apply a method designed by Ágel and Hennig that is based on Koch and Oesterreicher's differentiation between 'language of proximity' and 'language of distance' and has already been successfully applied to Early New High and New High German texts.⁴⁴ The benefit of this method is that it breaks down orality in different linguistic parameters, i.e. "time" (i.e. phenomena of serialization such as left-/right dislocation; paratactic structures etc.), "role" (i.e., signs of interaction between the discourse participants like vocatives and imperatives), "situation" (i.e. spatial, temporal and personal deixis), "code" (the interplay between verbal and non-verbal means, emotional evocations and interjections) and "medium" (i.e. traces of spoken language within the texts, such as phonic words which neglect graphical word boundaries, e.g. shouldya instead of should you). All these parameters are derived from universal-pragmatic conditions of the 'language of proximity' and comprise features that are supposed to be oral in general.

Haferland 2019, 39.

See in detail Miedema 2011, 38–44 who shows that the epithets in the Nibelungenlied are neither fixed patterns nor used very frequently.

Haferland 2019, 58.

Haferland, 2019, 55.

Müller 2012, 315.

Haferland 2019, 60.

Ágel/Hennig 2006.

The method has been developed to provide a tool that allows for situating a historical text within the continuum between the two poles of Koch and Oesterreicher's 'language of proximity' and 'language of distance' and to compare different texts with respect to their 'value of proximity', but this will not be the aim in the following.⁴⁵ Instead, the method is used in order to compare the different parameters of 'orality' and investigate the differences between the heroic and the courtly epics and their relationship with respect to their medial predisposition so as to gain a more fine-grained view of the interplay between the different linguistic features.

The analysis is based on an extract of 4000 words for each text. The text extracts have been randomly selected, but attention has been paid that both text excerpts involve the same amount of 'dialogical' (i.e. direct speech, comments of the narrator) and 'narrative' passages. In compliance with the method of Ágel and Hennig, each occurrence of an oral feature is counted as one point. Next to micro-structural features within the different parameters, also macro-structural characteristics like sentence length and the relationship between main and subordinate clauses have been analyzed. The results are summarized in Tab. 3.⁴⁶

See Zeman 2016b for discussion.

The analysis does not contain all features that are taken into account by Ágel/Hennig 2006. The frequency of 'phonic' words, for example, has not been included in Table 3 since most of the instances are conventionalized word contaminations (e.g. mirz ~ mir ez, "me it") so that it can be assumed that they are not direct reflections of spoken language. Furthermore, the low frequency of these contaminations does not allow for any conclusions. Also, 'Ausklammerungen', i.e. the positioning of sentential constituents outside the verbal bracket have not been included in Tab. 3. Since MHG does not have the same fixed word order as Present High German and often displays 'partial sentential frames', they would require are more detailed analysis with respect to their status as an oral feature in MHG. For the present analysis, I therefore restrict the analysis for the time parameter to the clearer cases of 'left-dislocation'.


Tab. 3: Comparison between Nibelungenlied B and Tristan with respect to oral features.

If we look at the sum of microstructural features in total, we see that the frequency of oral features within Tristan is higher than within the Nibelungenlied. The supposedly more oral character of the Nibelungenlied is thus not reflected within the frequency of oral features. But only if we look at the parameters individually, the differences between the two epic genres become more apparent. First, Tab. 3 shows a significant difference between narrated and dialogic passages, which concerns in particular the parameter of Code, Role and Situation. This result is expectable as it can be assumed that the amount of e.g. speaker-oriented deictics is higher in direct speech and comments of the narrator than in narrative passages. It is, however, remarkable that interjections and emotives are in particular characteristic for the dialogic passages within the courtly epics. Without the code parameter, the amount of oral features within the Nibelungenlied and Tristan are more or less comparable. The time parameter is not sensible to the narrative – dialogic distinction in the same way.With respect to the macro-structure, there is no indication for a difference between narrated and dialogic passages either, but a tendency that the amount of subordinated sentences is lower within the Nibelungenlied and that the average length of a sentence is shorter.

In the following, some exemplary features are discussed in detail in order to examine how the statistical data correlates with functional differences between the poems.

### 3.1 Time Parameter

The time parameter comprises phenomena that are dependent on the time during the production of discourse. For oral communication, it is assumed that the simultaneity between composition, production and reception leads to a more 'aggregative' and 'additive' (in contrast to 'integrative' and 'subordinated') discourse structure. A striking example is left-dislocation,⁴⁷ i.e. the positioning of a sentential constituent outside the sentence like in My aunt, she used to sing folk songs, which has been seen as characteristic for oral narratives in general.⁴⁸ Dislocation patterns can frequently be found in MHG epic poems. In (1), the subject is referred to by a NP and a subsequent pronoun, so that the subject is presented as a separate intonation unit.

. Vier hvndert swertdegene, | di solden tragn chleit | mit samt Sivride. " knight's attendants | THEY were about to wear knightly clothes | together with Siegfried." (Nibelungenlied B , –)

Such patterns have also been described as a common feature of the Homeric epics⁴⁹ and as "a clear manifestation of their oral syntax",⁵⁰ see 2., where the "pendant nominative"⁵¹ (Patroklos dé) presents the subject as an isolated referent within its own intonation unit.

. Πάτροκλος δ' | εἷος μὲν ᾿Aχαιοί τε Τρῶές τε τείχεος ἀμφεμάχοντο | "And Patroclus, | as long as the Achaeans and Trojans were fighting around the wall, |"

θοάων ἔκτοθι νηῶν, | τόφρ' ὅ γ' | ἐνὶ κλισίῃ ἀγαπήνορος Εὐρυπύλοιο | far from the swift ships, | all the while HE, | in the tent of pleasant Eurypylus, | (Il. , –; example from de Kreij , )

Such dislocation patterns have also been seen as an indication that for oral syntax, not the sentence but the intonation unit constitutes the basic element.⁵² The

'Left-dislocation' is, of course, a literal term since it presupposes a typeface from left to right. Since it is a common term, I keep it nevertheless.

Chafe 1994, 67–68.

See e.g. Bakker 1997b, chapter 5; Bonifazi/Elmer 2011; de Kreij 2016; Ready 2019.

De Jong 2012, 122; quoted in de Kreij 2016, 152.

"Pendant nominative" is the term of traditional grammar; see Bakker who argues that the phrase is better described as the theme of the upcoming description (Bakker 1997b, 102). Chafe 1994; see Bakker 1998, 39 for the Homeric epics.

frequent use of left-dislocation in the Homeric epics has therefore been taken as an argument that the Homeric text is "characterized by a thoroughly oral conception, and so very far removed from our conception of a written text".⁵³

It is interesting to see that within the Nibelungenlied, left-dislocation tends to be particularly frequent in narrative passages. This supports the hypothesis that left-dislocation serves the discourse function of "framing" or "priming" the discourse referents, which marks the protagonist as the center of the following lines of discourse.⁵⁴ As such, it can be expected that left-dislocations are characteristic for narrative passages. Once conventionalized, the pattern can become a metanarrative signal for scene shifts.⁵⁵ As such, left-dislocation is not just oral, but also part of the narrative syntax. The data suggests that the courtly epics adopt this narrative strategy also for dialogical passages. More empirical analyses would be necessary to examine this in detail. Yet, the data supports the observation made above that it is crucial to take into account the difference between narrative and dialogical passages.

### 3.2 Situation, Code and Role Parameter

As seen in Tab. 2, the parameters of situation and role display comparable values, and, as shown in the following, share a similar function in referring to the communicative frame between the narrator and the audience, i.e. to the 'here and now' of discourse (see 3.–4.).


seht, dâ hôrte er unde vernam/in gazzen unde in strâzen / look, there he heard and got to know/in alleys and in streets

von klage al solch gelâzen,/daz ez in muote starke, due to laments such a behavior that it troubled him very much.' (Tristan –)

Bakker 1998, 41.

Bakker 1997b, 86–111.

De Kreij 2016, 164.

In 3.–4., the narrator addresses the audience directly by an exclamative (hei waz "alas what") and an imperative of a verb of perception (seht "look"). Both examples presuppose the 'here and now' of the telling moment as a shared communicative situation between poet and audience. In addition, they evoke the impression that the narrating and perceiving act and the narrated events are happening simultaneously before the narrator's and the audience's mental eyes. As such, these linguistic features are not oral in the medial sense of the term, but are instances of 'language of proximity', 'immediacy' and 'involvement'.

While both the heroic and the courtly epics are characterized by linguistic means that refer to the communicative situation, there are differences with respect to the established relationship between the poet, the audience and the story told. For the Nibelungenlied, exclamative utterances like in 3. are quite common and are usually insertions by the narrator (i.e. 35 of 41 instances of the pattern 'hey + exclamative'), whereas in the courtly epic Tristan, similar exclamations with â, ach ("alas") (e.g. â welh ein castêl! ("alas what a castle!"); Tristan 3159) and ôwê ("woe!") are rarely used by the narrator but are common for characters' speech. This suggests that the narrator plays a different role within the two epics. This is supported by the finding that evaluative comments about the events within the story world are more frequent within the courtly epics. In addition, metanarrative comments that reflect the production and representation of the story are characteristic only for the courtly epics:

. wie gevâhe ich nû mîn sprechen an,/daz ich den werden houbetman 'How do I now begin my speaking/that I prepare the dignified protagonist /

Tristanden sô bereite ze sîner swertleite,/daz man ez gerne verneme?" Tristan in such a way for his knightly accolade/that one would like to hear it willingly?' [Tristan –]

In 5., the narrator is not narrating the events, but reflecting about the further representation of the story. This serves two different functions. First, 5. evokes the impression that the epic is originating from a dynamic on-line production process, in which the story evolves while speaking. On the other hand, the comment also serves as a retardation of the knightly accolade of the protagonist, which constitutes an important peak in the story. Before this important scene is finally described, an extensive metanarrative excursus about the poetic problems of ekphrasis is inserted (4595–5000). Such metatextual reflections are uncommon for the heroic epics.

A linguistic feature which plays an important role in such metatextual comments is the present tense, a feature which has cross-linguistically been seen as one of the most prominent expressions of a 'vivid style'⁵⁶ and 'oral narration'.⁵⁷ Traditionally, the so-called 'Historical Present' (HP) has been described as a rhetorical device that dramatizes the story "by making the audience feel as if they were present at the time of the experience, witnessing events as they occurred".⁵⁸ However, the HP in its narrow sense – i.e. a present tense which is used (i) in narrative contexts (ii) in alternation with past tenses and (iii) refers to events 'on-plot' – is not documented in every oral tradition and does occur neither in the Homeric epics⁵⁹ nor in the older stages of German.⁶⁰ In MHG, the present tense does not referentially denote the sequential steps of an action on the plotline, but is restricted to the level of discourse, as seen in 5.

As the 'author's present',⁶¹ the present tense establishes a shared communicative frame that includes two different relations: the immediacy between the poet and the audience, as well as the immediacy between the communication partners and the representation of the narrated events. With respect to the question of how oral the present tense is, it becomes thus obvious that it is not so much a feature of orality in the technical sense of medial orality, but of performed narrativity. The present tense functions as a meta-linguistic device that "establishes the discourse as something other than narration".⁶²

Similar observations have been made on the "immediate present" in the Iliad. According to Bakker, the present tense is one of the most important devices in order to indicate an immediate mode in Chafe's sense, and, as such, a direct interaction between the poet and the audience.⁶³ But why then do the Homeric and the MHG narrators not make use of the Historical present? With respect to the Homeric narrator, Bakker 2005 suggests that "the Greek epic tradition aims at something other than mere vividness or pretended immediacy":

Recreating the past, reviving the crucial events of the epic world as models for the present may be the concern of any tradition of epic poetry, but the Homeric tradition appears to go one step beyond such an unreflective immediacy. […] as I will argue, the implicit poetics of

 Koch/Oesterreicher 2011, 74–75; similarly also Fleischman 1986, 203; Wolfson 1982; Fludernik 1991, Fludernik 1992.

Fleischman 1990, 306.

Koch/Oesterreicher 2011, 170.

Fleischman 1990a, 376 n. 22.

Bakker 2005, 96.

Herchenbach 1911; Boezinger 1912; see for an overview Zeman 2013.

Hempel 1966.

Bakker 2005, 94.

the Homeric tradition reveal that the "true" poetic version of the epic events is better than the real thing: besides the urge to create the presence and nearness of the epic events, Homeric epic, I will suggest, is also concerned with distance […].⁶⁴

Bakker argues that the oral characteristics of the Homeric epics are not only features of immediacy between the communication partners but concern the whole communicative situation, including the relationship between the poet and the audience as well as their relationship towards the 'narrative truth', i.e. the "acknowledged correspondence between a statement and a state of affairs referred to".⁶⁵ One of his arguments is the use of a construction that Bakker discusses as symptomatic for Homeric discourse: μέλλω + inf. μέλλω has been classified as a (semi-)auxiliary that denotes a present intention or arrangement for the realization of a future state of affairs.⁶⁶ In its literal meaning, μέλλειν denotes the subject's mental state of thinking about doing something while it is not determined whether the intended action actually will take place in the course of the story or not. There is, however, another use where the realization of the event in the future is indicated as certain, while the intention of the subject is not at issue or even contradicted as in 6.:


"but in fact he was never to return from the ships and to bring his report back to Hector"⁶⁷ (Il. 10, 336)

In 6., μέλλω + inf. indicates what will happen in the future course of events against the will of the protagonist, since it is known from context (and the common knowledge of the audience) that it is not Dolon's intention not to return. Uses like 6. have been termed as 'destiny in the past' or 'future of fate'. According to Bakker, the central feature of this construction is not its temporal meaning, but the divergence between two different states of consciousness:⁶⁸ the nar-

Bakker 2005, 96.

Bakker 2005, 92.

Wakker 2007, 169. See on the semantics of μέλλω also Markopoulos 2008.

Translation by de Jong 2007, 25.

Bakker 2005, 100.

rator knows what is going to happen, while the protagonist is unaware. A very similar effect is also induced in MHG by the past form of the modal verb suln ("shall")+inf. (see 7.).

. ['they (i.e. Tristan and Isolde) were afraid of what happened later, namely that which later on deprived them of joy and brought them into distress]: daz was daz, daz diu schoene Îsôt/dem manne werden solte,/

That was that that the beautiful Isolde was to become the lover of the man dem sî niht werden wolte. of whom she did not want to become [a lover].' [Tristan –]

sollte + inf. and ἔμελλεν + inf. display a similar semantic pattern. Both suln ("shall, be obliged to") and μέλλειν ("intend to")+inf. are constructions that involve a projective meaning, a past marker and presuppose an external modal force.⁶⁹ As such, they inherently unite present plan, future realization and retrospection. The projectivity of the denoted event refers to a time interval that is posterior to a reference point that is already past as seen from the perspective of the narrator. Linked with that, the narrator is talking about the events to come, i.e. the construction does not refer to the representational, but to the presentational level of discourse in the sense of Kroon.⁷⁰ This perspective 'from outside' is reinforced by the fact that the future realization of the event lies outside the control of the focalized character.⁷¹ In both 6. and 7., the events will happen against the will of the protagonists: Dolon intends to come back to the Trojans, and Isolde does not want to commit adultery, but neither of them can change their destiny.

What can these observations tell us with respect to the oral character of the epic poems? Bakker is certainly right to state that the construction is more than "immediacy" in terms of enargeia. What is at issue is not so much the visual reviving of the past in the narrative present but the simultaneous activation of two different reference frames at the same time, i.e. the discourse world of the speaker and the story world as well as the correspondence between the denoted event and the fact of its future realization which can only be foreseen by the narrator. As such, the 'future of fate' is restricted to narrative passages. Interestingly,

See in detail Zeman 2019.

Kroon 1998, 207.

The focalized character is the protagonist on the story level who constitutes the reference point for narrative perspectivization. It could be defined as the most salient mental subject in the story. As such, it does not necessarily coincide with the syntactic subject.

the 'future of fate' reading of MHG sollte + inf. is used in courtly epics, but not in the Nibelungenlied. In the heroic epos, prolepses are prototypically indicated by the past tense and the adverbial sît/sint/sider ("later on"), a pattern which is documented 100 times in 9504 verses (see 8.–9.).


sît/sint/sider ("later on") are not explicit markers of the future, but can also be used to indicate the next step in the successive progression of events. The preference for this pattern is in line with the assumption that the heroic epics are primarily focused on the sequence of events within the story world, whereas the courtly epics are characterized by a strong tendency to evaluate and comment on the narrated events, metanarrative reflection on the representation of the story and the narrator's dialogues, and dialogues with allegoric instances ("oh, it is you, Lady Aventiure, how is the dear hero doing?"; Wolfram von Eschenbach, Parzival, IX,7). This might suggest that the MHG epic poems – more similar to the prose narrator in Ancient Greek than to Homer – rather fictionalize an oral performance situation than directly reflect it.

More cross-linguistic investigations would be necessary in order to draw such conclusions, but there are several indications that the degree of fictionalization is an important indicator for oral poetry. Foley argued that the inventive and ironic use of traditional language in Homer supports the impression that the Homeric text is located "at some remove from its roots in oral tradition".⁷² The 'roots' of oral performance have been characterized by the fact that there is no difference between the author and the teller of the story, since author and narrator are the same person. The invention of a textual, fictional voice as "the substitute of the absent author's actual voice"⁷³ could thus be seen as a more general development linked to the fictionalization of the text.

Foley 1993, 278.

Bakker 1998, 32.

### 4 'Orality' and Narrative Truth

In sum, this analysis has shown that different parameters of orality show different effects on the linguistic structure of heroic and courtly epics. Whereas leftdislocation and length of sentence are subject to variation, the situation parameter is rather stable in comparison between the two epic genres. It differs, however, with respect to narrative vs. dialogic passages. This distribution supports the relevance of the distinction between the medial and deictic dimension of orality since it can be assumed that an increased sentence length of the courtly epics is facilitated by the fact that the courtly epics were composed and performed by use of the written word. In contrast, features of the situation parameter like interjections, personal pronouns and imperatives are not 'oral' in the medial sense of the term, but rather concern the relationship between the communicative partners. As such, speaker- and hearer-oriented deictics in general can be assumed to be universal features within passages of direct speech. The traces of deictic elements within the narrative passages, on the other hand, can be seen as reflections of the particular communicative relationship between the poet and the audience.

Based on the qualitative analysis, this general result can be specified in two respects. First, the increased amount of interjections and emotive expressions in the dialogical passages indicates that direct speeches in the courtly epics are more affective than in the heroic epics, as it has been also suggested in literary studies. In the heroic epics, direct speech is often represented as an action which has consequences in the development of the story, whereas it gives little insight into the inner life of the protagonists.⁷⁴ For the courtly epics, in contrast, direct speech is often used to represent thoughts and emotions of the protagonists as a motivation for further actions. This increased focus on the representation of the inner world by focalization techniques has been seen as one of the most important developments in courtly narration.⁷⁵

Furthermore, both genres of MHG epic poems are characterized by deictic means that invoke the impression of 'immediacy' of the reported events. It is thus not only the proximity between the poet and the audience in a (conceptualized) shared reference frame of discourse which is at stake, but also their relation to the narrative events, i.e. the whole communicative triangle SPEAKER – AD-DRESSEE – WORLD. This is important to note since the difference between the two epic genres refers particularly to the relationship between the speaker and

Philipowski 2007, 68.

Hübner 2003, 86.

the represented narrative events, as has been shown within the qualitative analysis of grammatical features like the present tense, solte and ἔμελλεν + inf. Whereas both the heroic and the courtly epics display traces of the communicative process and the relationship between the speaker and the addressee, it is characteristic only for the courtly epics that the narrator comments both on the behaviour of the protagonists, the events within the story world and the representation of the story. Such metatextual evaluations as well as allegories and personifications can be seen as an indication of a distanced, self-reflective relation between the poet and his representation of the story. In literary studies, this increased distance with respect to the content of the story as well as its representation is linked to a stronger focus on the discourse level and the beginning fictionalization of the 'narrator': whereas within the heroic epics, author and narrator are indistinguishable of each other, the narrator becomes more and more a fictional voice within the courtly epics. As a result, more than one level of communication has to be distinguished: The 'real' situation where the performance of the poem takes place, and the displaced narrative communication between narrator and the (implied) audience within the text.

Metatextuality appears thus to be the most important difference between the two epic poems. This focus on 'reflective thinking' has often been interpreted as a consequence of written narrations.⁷⁶ At first sight, it seems natural to assume that written narration facilitates elaboration and reflective thinking. But this relationship is not straightforward. As Finnegan has shown, reflective thinking is not uncommon in oral poetry either, as many poems rely on long processes of preparation.⁷⁷ The observations thus do not add up to a straight line of development from online to reflective thinking. Rather, it seems to be a gradual scale that does not parallel with the oral – written distinction completely, but is fundamental with respect to the differences within the various genres.⁷⁸

This fact is also important when comparing the MHG epics to the Homeric tradition. As seen above, the Iliad and the MHG epics share oral features like the segmentation into intonation units and deictics of proximity that simulate a simultaneity between composition and performance. The conceptualization of the performance situation as a shared communicative space between the poet and the audience is thus basically the same in both Homer and the MHG epics and might constitute a universal feature of oral poetry. There are, however, differences with respect to how the communicative triangle is conceptualized

Butzer 1995, 161; Philipowski 2007, 52 n. 44; Hall 2008, 285.

Finnegan 1977, 80–84.

See also Hall 2008, 285: "the distinction between online and reflective thinking provides a mechanism more fundamental than literacy for explaining differences in human behavior."

within the poems. These differences refer in particular to the narrative structure (e.g. the emotionalization of the dialogues in the courtly epics) and the metatextual stance of the narrator. In both respects, the Homeric epics are more similar to the MHG heroic epics than the courtly epics. This is in line with how both traditions are characterized within the literature, see Tab. 4.⁷⁹


Tab. 4: The medial constellation of Homeric and MHG courtly epics.

As Tab. 4 shows, the differences cannot directly be traced back to online processing or the communicative situation between the singer and the audience but refer to the attitude of the narrator to the story world. The link to orality is thus only an indirect one. Rather, the difference concerns the 'narrative truth', i.e. the way the interaction of the relationship between communication partners and their relation towards the relationship between the told and its validation as 'true' is conceptualized. This has been seen in close connection to the epistemo-

As shown above, the oral constellations are more complex than summarized simplistically in Tab. 2. It can nevertheless serve as an outline of arguments that are commonly discussed with respect to orality in Homeric and MHG oral poetry.

logical presupposition of what is conceptualized as 'reality' within a speech community and is, as such, linked to the transmission of the content of the stories: whereas the content of the heroic epics is given by tradition and conserved by iterative re-actualization in performance, the reconstruction of the 'right version' of the story within the courtly epics by the authors relies on literate erudition. This is obvious in how the narrators emphasize their elaborate literature research in order to present the 'right' version: cf. e.g. the narrator's prologue in Tristan (131–162), which ensures that he will tell the story rehte ("rightly", 134) by referring to the 'right' literary source, namely the "well read" Thômas von Britanje.⁸⁰

. Ich weiz wol, ir ist vil gewesen, "I know well, there are many people die von Tristande hânt gelesen; who have read about Tristan; und ist ir doch nicht vil gewesen, however, there aren't many people die von ihm rehte haben gelesen. who have read about Tristan rightly […] aber als ich gesprochen hân, but as I said, daz sî niht rehte haben gelesen, [the reason] that they haven't read rightly, daz ist, als ich iu sage, gewesen: this has been, as I tell you: sine sprâchen in der rihte niht, they did not tell in the right way als Thômas von Britanje giht, as Thomas from Britain does, der âventiure meister was who was the master of âventiure und an britûnschen buochen las and read in Breton books […] Als der von Tristande seit, The way he tells about Tristan, die rihte und die wârheit the right way and the truth begunde ich sêre suochen I began searching in beider hande buochen in books both in walschen und latînen Romance and Latin und begunde mich des pînen, and started to take pains

daz ich in sîner rihte in that I in his correctness rihte diese tihte." rectify these epic facts."

The truth of the poem is thus guaranteed by the literacy of the poet and the knowledge stored in books. For the MHG poet of the courtly epics, truth lies within the external world which can be rectified by historical studies. This is different in the Homeric tradition where the source of storytelling is seen in the poet's mental act of the present performance. According to Bakker, truth in Homer is an emergent concept that arises within the production of discourse. Since "the

Similarly also the prolog of Der arme Heinrich by Hartmann von Aue (6–11) and the afterword in Wolfram von Eschenbach's Parzival (827).

'true' poetic version of the epic events is better than the real thing",⁸¹ it is the present moment of performance that constitutes the "moment of truth"⁸² where the story is 're-created' based on the shared knowledge of tradition. In this sense, the source of narrative truth in Homer has been seen in the "acts of remembering and forgetting",⁸³ "judged on the basis of its tradition"⁸⁴ and, as such, as an active mental process. Unlike the Homeric epics that do not refer to an "objective reality independent of the narrator (the epic singer), for the epic past exists only as perception, both in the memory of the singer and the imagination of his audience",⁸⁵ the MHG poems conceptualize memory as a passive thesaurus⁸⁶ where facts about the external world are 'stored'. While the reactivation and reassurance of a speech community's collective memory as shared knowledge can be seen a crucial factor in all oral traditions, the presupposition for this oral storage of cultural information can thus be essentially different.

### 5 Conclusion: More Paradoxes

The observations above lead us to the following conclusions. First, this paper has shown that 'orality' research combines (at least) three different aspects, i.e. 1. 'oral' residues as traces of 'online composition', 2. the communicative constellation between the poet, the audience and the story world, and 3. the oral predisposition within a specific cultural context. These different dimensions of orality have to be kept apart in order to gain a more precise view of the relationship between the general structures of oral communication and the cross-linguistic differences in language use since parameters 1. and 2. operate differently in different genres as well as in their diachronic development and seem to be linked differently to the oral predisposition (3.).

1. 'Oral' residues as traces of 'online composition'

With respect to the time parameter, we can assume a slow tendency towards increasing word length and decreasing dislocation structures, while the occurrences of deictic means remain relatively constant. This result is consistent with a study on the oral features


Bakker 1997a, 17; Bakker 2005, 96.

Bakker 2005, 113.

Bäuml 1997, 39.

Bäuml 1997, 42.

in Early New High German (1350–1650), which shows that the time parameter is more subject to change than the situation parameter.⁸⁷

2. The communicative constellation between the poet, the audience and the story world Whereas the performance situation is basically the same for both the Homeric and the MHG epics, one important difference in the linguistic structure of oral poetry has been seen in the relationship between the narrator and the story world, as shown with respect to the metatextual features and the fictionalization of the narrative voice. In contrast to the Homeric tradition, courtly epics are characterized by a distance from their narrative source. It is not the memory of the poet but the external evidence which is responsible for the reconstruction of the story. As such, this difference is only indirectly linked to orality and rather concerns the validation of 'narrative truth'.

In sum, the discussion has shown that oral poetry in historical texts is based on more than one paradox:


All these paradoxes leave their traces within the linguistic structure, making it hard to evaluate any given linguistic feature along a dichotomous contrast 'written' vs. 'oral'. In this respect, the evaluation of the ring composition (in order to close this ring) seems not contradictory anymore, since elaboration and oral composition are not opposed to each other but instances of different aspects of orality. Differentiating between these aspects of orality can thus lead to a more consistent and maybe more conciliating description of the characteristics of oral poetry.

The comparison between the MHG heroic and courtly epics has thus shown that such analyses are suited in order to investigate the interplay between the different factors of orality and, by doing so, allow for a more fine-grained view of the mechanisms of oral tradition – thus granting us a more systematic view into the vortex.

See Hennig 2009.

See Mellmann 2014 and Ready 2019, 29 according to whom epics are stories of 'it is said' or 'they say'. As such, they are naturally displaced from their communicative source but re‐actualized within the particular performance.

Finnegan 1977, 109; emphasis by S. Z.

### References


Fleischman, S. (1990b), "The Discourse of the Medieval", in: Speculum 65, 19–37.


Lord, A. B. (1960), The Singer of Tale, Cambridge, MA – London.

Markopoulos, T. (2008), The Future in Greek. From Ancient to Medieval, Oxford.

Marold, K. (2016), Gottfried von Straßburg. Tristan, Berlin - New York.

Mellmann, K. (2014), "Gibt es einen epischen Modus? Käte Hamburgers Logik der Dichtung evolutionspsychologisch gelesen", in: E. Hárs/M. Horváth/E. Szabó (eds.), Universalien? Über die Natur der Literatur, Trier, 109–130.

Miedema, N. R. (2011), Einführung in das 'Nibelungenlied', Darmstadt.

Minchin, E. (2001), Homer and the Resources of Memory: Some Applications of Cognitive Theory to the Iliad and the Odyssey, Oxford.

Müller, J.-D. (2012), "Medieval German Literature: Literacy, Orality and Semi-Orality", in: K. Reichl (ed.), Medieval Oral Literature, Berlin - New York, 295–334.

Nimis, S. A. (1998), "Ring-Composition and Linearity in Homer", in: A. McKay (ed.), Signs of Orality. The Oral Tradition and its Influence in the Greek and Roman World, Leiden, 65–78.

Person, R. F. (2016), "From Grammar in Everyday Conversation to Special Grammar in Oral Traditions: A Case Study of Ring Composition", in: M. Antović/C. Pagán Cánovas (eds.), Oral Poetics and Cognitive Science, Berlin – New York, 30–51.

Philipowski, K. (2007), "Strophisches und stichisches Sprechen. Medientheoretische Überlegungen zur Figurenrede in höfischer- und Heldenepik", in: N. Miedema/F. Hundsnurscher (ed.), Formen und Funktionen der Figurenrede in der mittelhochdeutschen Großepik, Berlin – New York, 43–71.

Ready, J. L. (2019), Orality, Textuality, and the Homeric Epics. An Interdisciplinary Study of Oral Texts, Dictated Texts, and Wild Texts, Oxford.

Reichert, H. (2020), Nibelungenlied B = Der Nibelunge Nôt, manuscript St. Gallen (~1250 AD) (https://www.hs-augsburg.de/~harsch/germanica/Chronologie/12Jh/Nibelungen/ nib\_b\_00.html).

Sale, W. M. (1996), "Homer and Avdo: Investigating Orality through External Consistency", in: I. Worthington (ed.), Voice into Text: Orality and Literacy in Ancient Greece, Leiden, 21–42.

Slings, S. R. (1992), "Written and Spoken Language: An Exercise in the Pragmatics of the Greek Sentence", in: Classical Philology 87/2, 95–109.

Stanley, K. (1993), The Shield of Homer. Narrative Structure in the 'Iliad', Princeton.

Wakker, G. C. (2007), "Intentions and Future Realisations in Herodotus", in: R. Allan/M. Buijs (eds.), The Language of Literature: Linguistic Approaches to Classical Texts, Leiden, 168–187.

Whitman, C. H. (1958), Homer and the Heroic Tradition, Cambridge, MA.

Wolfson, N. (1982), CHP. The Conversational Historical Present in American English Narrative, Dordrecht.

Zeman, S. (2013), "Vergangenheit als Gegenwart? Zur Diachronie des Historischen Präsens", in: P. M. Vogel (ed.), Sprachwandel und seine Reflexe im Neuhochdeutschen, Berlin–New York, 236–256.

Zeman, S. (2016a), "Orality, Visualization, and the Historical Mind: The 'Visual Present' in (Semi)oral Epic Poems and its Implications for a Theory of Cognitive Oral Poetics", in: M. Antović/C. Pagán Cánovas (eds.), Oral Poetics and Cognitive Science, Berlin–New York, 168–195.


### Franco Montanari Epos and Orality: Conclusive Remarks and Open Questions

First of all, I wish to thank the organisers of this very interesting conference for giving me the honour of chairing the final roundtable. Given the broadness and richness of the topics under discussion, the title "Conclusive Remarks and Open Questions" sounds particularly demanding and complex. I feel like a soldier entrusted with a heavy flag, which he is struggling to hold: he still does his best to fulfil his duty and to bear the heavy burden – if only with trembling, unsteady hands – until the end of the parade (when he will possibly collapse…).

The conference programme has certainly been rich and remarkably varied, and the sessions have all met the expectations raised by the programme itself. Many of the speakers have taken part in the final roundtable in order to add further points and details to their papers, or to comment on other contributions. The speakers are free to take account of what has been said on this occasion in the final versions of their papers. Besides, it would be a bit much – and rather presumptuous on my part – to try and sum up the many different topics that have been discussed in the conference, and which have partly been taken up again in the final roundtable. So I will only mention a few topics that have struck me as being particularly interesting and relevant (and which have stuck in my head…).

I should start by saying that this conference has focused particularly on the fact that "literacy" and "orality" (or "oral poetry" or "oral theory") are not unambiguous concepts, which can be set in contrast sic et simpliciter – as is all too often the case – in a superficial way: "literacy vs. orality" does not mean much and does not correctly frame the issue of Homeric poetry's nature. Neither is it a monolithic, fixed concept; rather, they are both marked in a cultural sense by several factors. The occurrence of changes in the text, changes that we can reconstruct or even verify, alters the meaning and the features of Homer as "oral poetry", but certainly does not erase them. For example, Cassio¹ has highlighted a highly significant phenomenon which, from Hoekstra onwards, has been seen to reflect the linguistic stratification and internal history of the epic language. The fact that this phenomenon actually occurred and can safely be traced does not at all rule out that Homer should be regarded as "oral poetry", or – in other words (to quote a well-known definition by Luigi Enrico Rossi) – that Homer should be considered a "testimony of oral poetry". A very different

See Cassio in this volume, 41–50.

OpenAccess. © 2022 Franco Montanari, published by De Gruyter. This work is licensed under the Creative Commons Attribution 4.0 International License. https://doi.org/10.1515/9783110751963-012

textual-historical phenomenon which cannot be forgotten or overlooked is the various alternate versions of the Homeric text, as documented by the so-called "wild papyri". The variants which are in these papyri show numerous additional verses in the text of the Iliad and Odyssey that alter their composition, yet without in any way altering the overall structure of the poems, which would appear to have been quite consolidated by then. The phenomenon endures until the 1st century BC, when it disappears, which is to say that the consolidation of the numerus versuum, which had hitherto fluctuated quite a bit, has also rightly been connected to the influence of Aristarchus of Samothrace (and of his school) as regards this aspect of the poems and their textuality. We must bear in mind that only a small part of this phenomenon is known to us, owing to the complete randomness of the papyrus fragments' retrieval and publication: perhaps we should ask ourselves what its significance would be if, per chance, it were known to us in full – but this remains an idle question. However, I should stress once more that, despite its importance, this aspect of the text's history, just like the linguistic stratification and internal history of the epic language – in other words, the detection of changes in the form of the text over the course of its history (for it is well known that the Homeric poems' textual tradition is far from unambiguous) – cannot be seen to cast doubt on, or indeed rule out, the fact that Homer's poetry is in every respect "oral poetry" or a "testimony of oral poetry" (a definition which I find particularly compelling and useful).

But let us return to the most important and significant aspect of this conference, its 'core business': this seems to be clearly identified and summed up by the title, which speaks of the "mechanisms" of the oral communication system. For decades, after the popularisation of Milman Parry and Albert Lord's theory, an increasing number of discussions contrasted "orality" and "writing" as two globally and monolithically opposed concepts: Homer was born as an oral work or as a written work – basically, what we had were two parties, so to speak, divided over their idea of the genesis of the poems. Since then, however, studies in this field have made enormous progress and, as a consequence, broadened views and perspectives to a remarkable degree. This conference has certainly shown as much.

I have already recalled how "literacy" and "orality" should not be interpreted as unambiguous and monolithic concepts: not only can they be defined in different and nuanced ways, but each represents distinct and multifaceted phenomena. This has emerged in various papers. It has been emphasised by Egbert Bakker, for example, who – among other things – has explored the importance and cognitive function of the formulary system (a phenomenon typical of oral poetry and thus connected with epic's oral origins) even in the case of, and with respect to, written texts.² The possible coexistence of oral and written texts at a certain stage in the history of archaic Greek epic and the particularly broad variety of related phenomena is an aspect that has been discussed in various ways and which must be strongly emphasised, because – although by now it has become a fully accepted aspect of discussions on problems related to oral communication – it will provide plenty of food for thought once the proceeds of this conference are published. If my memory serves me right, the first – or at any rate one of the first – to talk about the coexistence of oral and written texts was Albrecht Dihle, in his all too often overlooked 1970 book Homer-Probleme. I would say that today this aspect – namely, the fact that oral and written texts, "literacy" and "orality", coexisted for a certain period – is taken for granted and belongs to the varied picture that is painted of the oral communication system and to the debate surrounding written vs. oral or oralderived literature.What remains open is the discussion concerning all the problems connected to this phenomenon in all of its aspects: for example, the features of performances, particularly in transitional phases (Lord introduced the concept of "transitional texts"), the phenomena connected to textless intertextuality, so-called interfomularity and/or intertraditionality, and everything which has to do with the debate on the discrepancy between repetition and recurrence (where repetition is intentional whereas recurrence is symptomatic) – along with the related issue of so-called "doublets". I would further point to the enduring relevance of the idea that the coexistence in Homeric diction of unquestionably formulaic expressions and what are certainly, or in all likelihood, non-formulaic ones bears witness to an objective and creative tension between the traditional formulaic system and a poet's individual system (a form of 'originality'): this is a perspective which current approaches to the overall topic of "literacy vs. orality", far from overshadowing it, brings into play and enriches it.

All in all, then, this conference has presented us with a gamut of topics and approaches to the varied issue of the oral communication system in relation to archaic Greek epic. Certainly, this short conclusive session has hardly touched upon all its aspects, but I trust that it has provided an idea of the overall picture and of its countless facets. I here wish to dwell on one further aspect discussed in the conference, an aspect which has struck me – no doubt because of my own limitations – as the most innovative and least common one. At any rate, it is the aspect that is least familiar to me. I am talking about the topic of the cognitive function and the role it plays in the context of the oral communication system in archaic Greece. I must admit that my knowledge in this field is very limited, and I hope it will not lead me to draw wrong conclusions. The cognitive function has to do with the

See Bakker in this volume, 19–39.

process of acquiring knowledge through thought, the senses, and experience. As scholars of Greek culture, this is bound to remind us first of all of Aristotle, with his strong interest in the inner workings of the mind and how they influence human experience through the formation of mental images. Aristotle was keen to base his studies on the cognitive function of the human mind upon collecting empirical evidence through in-depth observation and conscious experimentation. In our case, it seems to me that the problem primarily concerns the poet and the activation of his cognitive functions in order to make use of memory and his knowledge for the purposes of composition and performance. Studies based on the functioning of the human mind – essentially, studies in individual and social psychology – are brought into play in order to understand both the process of acquiring knowledge by the poet-aoidos and the process of retrieval and use of this knowledge during performances. Such issues (which I already mentioned above in relation to Bakker's paper) have chiefly been explored by Elizabeth Minchin (as regards its more general aspects).³ Moreover, Chiara Bozzone have brought cognitive linguistics into play, first of all to study "transitional" stages, insofar as these can be identified in the surviving Homeric texts.⁴

Finally, I wish to add a brief reference to a topic which has played only a marginal role in this conference, and which has only been mentioned by Giampiero Scarfoglio.⁵ I am referring to the development of Neo-Analysis, which – if I am not mistaken – is the most unexpected and least know branch of Homeric scholarship among our conference's participants. Hence, I would like to take the opportunity to explain why it is relevant and worthy of attention here.You can take this as my own personal contribution. Neo-Analysis developed in Homeric studies in the first half of the 20th century and reached its peak (and codification) with Wolfgang Kullmann's famous book Die Quellen der Ilias, published in 1960. This research strand fell within the typically German tradition of Quellenforschung and despite its name, which clearly recalled the old analysis, it was an explicitly unitarian theory (usually positing a single author for the Iliad and another for the Odyssey). For decades, Neo-Analysis stood in sharp and direct contrast to the theory of orality (probably in all of its different aspects), and it can also be seen to reflect the opposition between German Homeric studies and the Anglophone (or Anglo-American) tradition. According to Neo-Analysis, the sources of the Iliad (which has always been its centre focus) were only conceivable as written texts, which already placed not just the Homeric poems but also its more or

See Minchin in this volume, 1–18.

See Bozzone in this volume, 51–82.

See Scafoglio in this volume, 83–102.

less direct antecedents outside the oral phase of archaic Greek epic. However, things changed quite radically later, thanks to the new perspective promoted by Kullmann himself (who can probably be described as one of the greatest 20th-century scholars of the ancient world). The shift in question dates from the 1990s, when he abandoned the term Neo-Analysis in favour of Motivgeschichte. It is important to bear in mind that this kind of approach to and analysis of the Homeric poems only deals with content forms (Motive) and does not at all touch upon forms of expression (the issue of formularity and all stylistic aspects). Roughly speaking, with regard to internal allusions in the Homeric poems, Kullmann stated during those same years: "it seems that these allusions quote fixed texts […] This may, but need not, mean that these were written texts". The distinction between "fixed texts" and "written texts" was a striking innovation that opened up an entirely new path which was later developed in a particularly fruitful way: the transformation of Neo-Analysis into Motivgeschichte, and the possibility that the sources of the epic poems might be fixed yet unwritten texts, which is to say stable oral texts, or – if we prefer – texts that were more or less fixed yet still fully oral (and thus the object of performances). This constituted a substantial change. Among other things, it took into account many of the 'innovations' in the research on poetry and oral communication which I have outlined above, including the tension between poetic composition as a collective process and the individual poet's creative interventions, not to mention phenomena that can be traced back to so-called transitional stages or texts – to use the terminology introduced by Lord – between the oral phase and that in which the poems became fixed in writing. I have no time here to dwell on the development of the path that was thus opened, and which has led to the present-day situation, in which we can speak of a 'historic compromise' between oral theory and the now outdated "Neo-Analysis", which has turned into the study of a Motivgeschichte. This is something quite different, albeit based on many of the same historical assumptions as the old Neo-Analysis, starting from its focus on the contents of the mythical narrative of epic poetry, on their variants, internal allusions, the implications of mythical motifs, and so on (while also preserving the traditional avoidance of aspects related to forms of expression). Most importantly and decisively: this new approach no longer takes it for granted, or suggests it is necessary to assume, that the Quellen were written texts; rather, it assumes that they were fixed oral texts, which is to say poems belonging to a mature stage of oral poetry (a stage which scholars now regard as a necessary stepping stone in the historical development of oral poetry). This 'historic compromise', initially promoted by the champions of Neo-Analysis themselves, has practically given rise to a new research strand which moves beyond the previous, clear-cut, and irreducible contrast between Neo-Analysis and the studies on orality. This has developed independently based on the assumption that Motivgeschichte studies do not imply the existence and use of written texts, but rather that of fixed oral texts which can be conceived of as sources, despite the fact that they may still display a relative degree of stability and possible variations. The concept of "intertextuality" here finds extensive and significant application, which proves crucial for this research's new perspective. "What is now called 'Oral Neoanalysis' is in fact an Oral, Intertextual Neoanalysis"⁶ which does not ignore the dialectical relationship between the view of poetic composition as a collective process based on a consolidated tradition and the creative intervention of an individual poet, operating on the Motive of the epic narrative.

Although I cannot dwell on these topics any further here, I thought it might be interesting to add another piece to the already broad and varied puzzle presented by this conference. We have been dealing with a remarkable range of different, complex, and multifaceted mechanisms at play in the oral communication system of archaic Greek culture in relation to epic, and we have realised that this system presents a wide range of aspects that no scholar can hope to master on his or her own. This picture grows all the richer and more diverse as new advances are made and new paths are opened as the research progresses. It is easy to foresee that the volume stemming from this conference will long constitute a point of reference for this remarkably broad topic, which is so difficult to cover in full, even for specialists. We can expect this picture to be further enriched sooner or later and to offer new voices, new paths, and new perspectives: in this respect, in addition to having a picture of the situation – a status quaestionis, so to speak – what we have before us is, de facto, an open road.

### References

Dihle, A. (1970), Homer-Probleme, Opladen.

Kullmann, W. (1960), Die Quellen der Ilias (troischer Sagenkreis), Wiesbaden.

Montanari, F. (2012), "The Homeric Question Today", in: F. Montanari/A. Rengakos/C. Tsagalis (eds.), Homeric Contexts. Neoanalysis and the Interpretation of Oral Poetry, Berlin – Boston, 1–10.

Rengakos, A. (2020), "Neoanalysis and Oral Poetry: A Historic Compromise?", in: A. Rengakos/P. Finglass/B. Zimmermann (eds.), More than Homer Knew – Studies on Homer and His Ancient Commentators, Berlin – Boston, 37–49.

A basic overview may be found in Montanari 2012, Rengakos 2020. The quote is from Rengakos 2020, 46.

### Index Notable Things

Achilles 1, 4, 6, 11–16, 36, 55, 92, 94, 97, 133, 146, 159f. Adornment 23 f. Androktasia 122, 124–129, Archilochus 170, 172f. Aristarchus of Samothrace 208 Aristeia 122, 124, 127–129 Aristotle 60f., 84, 88, 145, 150–152, 165, 174, 210 Aristoxenos 165, 167, 169 Athenaeus 165f., 169, 171–174 Attention magnet 14 Audience 4, 6–12, 14–16, 21, 23f., 29, 34, 37, 52, 55, 63, 70, 74f., 77 f., 88–90, 92, 103f., 108–114, 118f., 126, 131, 133, 135f., 148f., 154–158, 174, 178, 181, 190–193, 196–198, 200f. Bacchylides 172f. Battle scene 119f., 124 Brit-Pop pronunciation 73 Catalogue 3, 119f., 124–126, 128f., 136, 174 Chamaileon 170f. Chunking, chunks 54–56, 59f., 67 Clytemnestra 86, 88f., 91, 93, 96–98 Cognitive function 208–210 Cognitive linguistics 210 Cognitive script 2 Collocations 56–59 Communication system IX, 118, 131, 149, 151, 160, 208f., 212 Composition 2–5, 19f., 60, 71, 90, 95, 117–119, 124, 129, 131, 135, 148, 152– 154, 173, 178–180, 182–186, 189, 197, 200f., 208, 210–212 Deictic 179–181, 188, 196f., 200 Dialect mixing 72, 76 Didactic poetry 165, 172, 174 f. Dihle, Albrecht 209

Dislocation 186–190, 196, 200 Doublets 209 Elaboration 179f., 183f., 197, 201 Elegy, archaic 173 Emotion and memory 14 Enactivist theory 6 Epic cycle 83–86, 88–90, 93, 95–99, 113, 184 Exclamative 191 Fictionalization 195, 197, 201 Fixed texts 21, 211 Formula 19f., 22, 31 f., 53–57, 90, 93, 179f., 186 Formularity 19f., 51–56, 62f., 68, 77f., 211 Formulaic epithets 92 Framing 190 Herodotus 48, 57, 117–137, 159f. Hesiod 32, 46, 68, 85, 135, 165, 170, 174f. Historical present 192 Homeric diction 52, 69, 71, 209 Imperative 47–49, 186, 188, 191, 196 Individual system 209 Interformularity 33 Intertextuality 33, 103f., 113, 209, 212 Intonation units 63f., 189, 197 Kullmann, Wolfgang 210f. Kunstsprache 51–53, 68–78 Landscape, description of 5–10 Language of proximity 180f., 186f. Literacy 19, 21, 181, 183, 197, 199, 207– 209 Literary criticism 145–147, 154, 156, 160 Lord, Albert Bates VIII, 1 f., 19, 53, 177, 181, 208f., 211

Mechanisms of composition 117, 131, 135 Memorization 22, 119, 183

OpenAccess. © 2022, published by De Gruyter. This work is licensed under the Creative Commons Attribution 4.0 International License. https://doi.org/10.1515/9783110751963-013

Memory – episodic 2 f. – narrowing 14 – spatial 3, 5 – visual 3, 5 – working 54f., 59 f., 63, 68, 77 f. Meta-language 37 Metatextuality 197 Meter 22, 51–53, 56, 60–68, 70f., 76–78, Metricality 60–62 Mimnermus 170 Motivgeschichte 211 f. Motif 110, 121, 124, 126, 129, 131, 136, 211 Motor resonance 7 f. Multitext 25 Narrativity 192 Narrator 29–31, 95, 107–109, 113, 133, 137, 170, 178, 187, 188, 190–192, 194f., 197–201 Neo-Analysis 98, 210f. – Oral Intertextual 212 numerus versuum 26, 208 Odysseus 24, 27, 35, 42f., 45, 49, 85–93, 96f., 103–113, 133f., 160 Oral-formulaic composition 20 Oral poetry 21, 31, 83, 90, 177, 181–183, 195, 197f., 201, 207f., 211 Oral texts 41, 78, 211 f. Oral theory 92, 98, 207, 211 Oral tradition 3–5, 22, 27, 53, 55, 71, 83, 85 f., 89–93, 97f., 103, 120, 132, 183f., 192, 195, 200f. Oral transmission 2, 26, 69, 181, 185 Orality VII–IX, 19, 21, 52f., 60, 77, 93, 117– 119, 131, 135 f., 146, 150, 152–154, 156, 160, 165, 177–187, 192, 196, 198, 200f. Oresteia-theme 86, 88–90, 92–94, 96, 98 Originality 178, 209 Papyri 26f., 33, 35 f., 41 f., 47f., 208 – 'Guéraud papyrus' (P. Sorb. inv. 2245) 49 Parry, Milman VIII, 1, 19, 27, 31 f., 52, 60, 70, 91 f., 177, 181, 208 Penelope 86, 97, 103, 107–109, 111 f., 133f.

Performance 2f., 20 f., 23–26, 32f., 52 f., 55, 60, 67, 69, 72–74, 76–78, 86, 90, 103f., 109, 114, 118f., 131, 133, 135, 148, 167, 170f., 173, 178f., 181–186, 195, 197, 199–201, 210 Periander 168, 170, 173 Phase theory 69 f., 76 Phocylides 168, 170, 173f. Pindar 68, 147, 172f. Plato 22 f., 26, 36, 89, 152f., 160, 169, 174 f. Plus-verses 27, 31, 33, 35 Plutarch 170f., 175 Prefabricated expressions 52f., 55 f. Prosodic regularization 63–68 Quantitative proof of orality 55–60 Quellen 210f. Randomness 31, 208 Recurring elements 121, 126–129 Repetition 3 f., 31–33, 36f., 54, 57, 60, 93, 119, 179f., 209 Rhapsode 21–23, 25, 33, 36 Ring composition 4, 118, 179, 201 Rossi, Luigi Enrico 207 Scribe 24, 29, 34, 36f. Shield of Achilles 6–9, 146 Simultaneity 189, 197 Solon 48, 152, 168, 170 Speech 3 f., 36, 62–64, 66, 68, 72, 86, 88, 90, 92–99, 118, 131–137, 152–154, 156, 158, 183, 187f., 191, 196, 199–201 Sportscaster 63, 66 Teiresias 35, 103–110, 112f. Telegonus 85, 103, 109–113 Telegony 103f., 106, 109–113 Terpander 171 Testimony of oral poetry 207f. Textual criticism VII, 19f., 26 Theognis 168, 170, 174f. Theory of mind 4 f., 11, 15 Thucydides 135–137 Traditional referentiality 32, 55, 77 Transitional phase 209

πρόφρων 42–46

Transitional texts 209 Transmission 19–21, 24, 26, 36, 109, 117, 178, 184, 199 Typical scene 45f., 119, 124–126, 131 Xenophanes 168, 170 Writing VII, IX, 2, 15, 19f., 24 f., 32, 88, 117f., 132, 152f., 157, 160, 169, 173, 181, 183, 186, 208, 211 Written texts 26, 48, 83, 135, 180, 185, \*ἔασσα / \*ἴασσα 45f. \*(ϝ)έκασσα 45f. (ϝ)εκών 44 εἴπατε 47, 49 εἶπα 48f. εἶπε 48 εἶπον and εἶπα 46f., 49 ἔνειμας 47f. ἔνειμε 47f. μέλλω + inf. 193f. πρόφρασσα 42–46

209–212

### Index Discussed Passages

Aristotle – Poet. – 1447a–b 151 – Rhet. – 3, 1404a 150 Athenaeus – Deipn. – 14, 631e–3c 166 Democritus – B21 D.–K. 37 Dio Chrysostom – 2, 4–5 173 Dionysius of Halicarnassus – de comp. verb. 11 154 Herodotus – 1, 82 130 – 7, 60–99 125 – 7, 208–224 127 – 7, 224–225 129 – 9, 22–23 129 – 9, 64 130 – 9, 67 130 Hesiod – Theog. 108 47 Homer – Il. – 1, 52–67 12 – 1, 365–392 11 – 1, 407–412 12 – 2, 37–40 30 – 2, 419–420 29 – 2, 484–770 124f. – 3, 302–305 27–30 – 4, 79–80 34 – 6, 466–471 10 – 9, 367–377 13

– 10, 290 43 – 10, 336 193 – 15, 377–378 29f. – 15, 390–392 189 – 18, 483–489 7 f. – 18, 490–496 8 – 21, 500 43 – Od. – 1, 28–43 86–96 – 1, 40–41 92f. – 3, 427 47 – 4, 538–541 35f. – 9, 363 36 – 9, 487–490 35 – 10, 386 43 – 10, 496–500 35f. – 11, 119–137 104–107 – 13, 391 43 – 21, 198 47 – 23, 281 ff. 104 [Longinus] – Subl. – 1, 2, 4 155f. – 7, 2–3 156f. – 10, 1, 5 157f. – 15, 1–2, 3 158f. – 26 159f. – 33 172 Nibelungenlied – B 3, 4 195 – B 4, 4 195 – B 28, 1–2 189 – B 212, 4 190 Photius – Bibl. 239, 17–19 84f. Plato – Ion – 530b5–c6 23 – 530d6–8 23

OpenAccess. © 2022, published by De Gruyter. This work is licensed under the Creative Commons Attribution 4.0 International License. https://doi.org/10.1515/9783110751963-014

– Leg. – 3, 700a7–1b3 169 – Phaedr. – 278b–d 152f.

#### Proclus

– Chrest. 301–302 88f.

Theocritus – Id. 16, 16–21 147f. Tristan – 131–162 199f. – 4589–4593 191 – 6022–6027 190f. – 12401–12406 194

Xenophon – Symp. 3, 5 20 f.