8. Sounding Like Homer: The Proof …1
I am ready to make the case for the authenticity of a performance of Homer that follows the instructions encoded in texts derived from East Roman manuscripts. These instructions separate not only words but lines: Homeric hexameters are musical wholes shaped by the pulse points of a dance movement based around the uniquely Greek isochronous dactyl. These incantatory compositions, Homeric lines, cannot be derived from imagined Indo-European cola, or any other metrical segments supposed to be grounded in lingual rhythm. We shall begin by answering questions we raised initially; to remind you: why is it that a text of Homer might need ‘correction’, in the form of accent marks, for a Greek speaker to learn how to perform the verse? If he already knows the words in their prosody, what would be the point? What are the sources of possible danger whereby, left to themselves, students who knew how to read Greek might all the same fall into a ‘bad habit’ when reciting Greek poetry from a written text?
In the Venetus A lemmas, as in the papyri, accents were not used all over the place, as Nagy points out; only at selected places in texts written in scriptio continua. Clearly, the students whom these Alexandrian texts served, mostly knew where the words were accented; as surely as they knew how to pick out, register and parse what we call ‘words’ as their eyes scanned the continuous stream of letters. In the presentation of these manuscripts, and ever since, Greek words were essentially prosodic units. But there is evidence, still present in the East Roman texts, that on occasion the registration of unity stretched over what we would call phrases, perhaps in the way that English ‘catch phrases’, containing opacities frozen in tradition (‘batten the hatches’, ‘rock ’n’ roll’, ‘good bye’ or ‘what’s up?’ ‘wassup?’ and ‘’sup?’) or concocted for advertising or comedy (‘shut the fuck up!’ ‘what the fuck?!’ and ‘’the fuck?!’, ‘STFU’, ‘WTF’), stop being parsed as separate, extractable words and register always together. We sometimes observe a suspension of normal sandhi rules internal to these phrases in Homeric Greek, and also a secondary accentual contonation, as is standard with the suffixes called ‘enclitics’, but extending also perhaps rearward from the phrase’s terminal accent. We recall that the guslars thought in terms of ‘sound groups, not individual words’ which did not always coincide; and that they only understood what Parry and Lord meant by ‘word’ when they saw them separated in a written record of their performance.2 I submit that there is nothing self-evident about the existence or substance of Greek words except as prosodic units or groupings.
There is no call to invent the ‘pressure’ of a metre in order to generate such unitary phrases in usage. But when words or phrases are arranged so as to exploit their inherent rhythms and melodies inside larger structures like epic lines—that is, in poetry—factors that affect or even alter natural prosody may come into play. And naturally it is in the recording of poetry where one finds the most need for direction, and hence the most use of accent marks in early texts.
How were native speakers likely to trip up? To begin with, a portion of Homer’s diction had become obscure in the historical period. The first scholia from the classical era served as glossaries for this lost vocabulary, although classical teachers did not limit themselves to guesses about diction; allegorical exegesis also characterises the pre-Alexandrian era. One presumes that forgotten or unfamiliar words were also obscure as to their accent; the same goes for compound or cultic names and idiomatic phrases, whose meaning was from the earliest times the subject of speculation. But the focus on prosodic error, or simply on prosodic recording, seems only to have come to a head in Alexandria, resulting in the deployment of Aristophanes’ now familiar signs. One may speculate, but only speculate, that the change taking place in the contemporary Greek language toward a stress prosody at the former high pitch location in words, may have played a role in motivating in Alexandria the scholarly recording, and so preserving, of the accentual patterns that characterised the poetry that was by now already considered classical.
The main and generic source of performance error, however, comes from the ictus of poetic feet. Modern students often learn a technique of scansion, as a way to drill on metres, where they punch every long downbeat (or thesis). The result is monotonous but also weirdly hypnotic. The dancer, after all, kept to this repeated beat without complaint, like a hypnotised waltzer. Ancient students also would have had an instinct, native to human rhythm, for stressing alternate beats; Greek was as iambic in that way as English. Even within the accentual contonation, the theory has it that one mora or syllable was more prominent than the other; in some cases it was only the surrounding environment that determined which one. Errors would have the effect of neutering any syncopation or ‘special effect’ intended by a composer, for example, in favour of a performed regularity. Hence knowledgeable correctors would have been as keen to prohibit as to prescribe; that is why so many graves are deployed in early papyri at the many places where a student may be tempted to accent to keep in rhythm. This is one ‘bad habit’ that the negative marking (grave) seeks to forestall. Only when it was decided to display words as prosodic units, separating them graphically, could the accent system become both universal and economical in its prescriptions and prohibitions.
In the period when the recessive contonation was operant, it would seem that gradations of emphasis between syllables would have been at their most nuanced and subtle, in relation to the monosyllabic accentual patterns of later Greek. To begin with, syllables outside those involved in the contonation retained their strong vowel values and full quantities, a situation we find only in some song settings of languages with stress accents, like English. Then there was the suppressed onset of the contonation when it could not be completed within the word; it is not certain whether some rise in pitch was still registered on acutes turned grave, and if there was a ‘knock-on’ effect on the initial syllables of words following. Finally, within the word’s accentual contonation itself there was also a contrast between prominence/non-prominence, reflected in placement with respect to metrical ictus, which depended on the interaction of the pitch elements with quantity: when the down-glide fell on a long or closed syllable, it predominated over the marked rise; otherwise the rise itself was prominent.
It seems very likely that there was a stage, reflected in some papyri, when the accent became monosyllabic and indicated by high pitch. The circumflex and acute came to mean the same thing: ‘accented’. The significance of the circumflex and the disyllabic contonation was lost. In that circumstance it is inevitable that native speakers would misinterpret and so distort the musical instruction intended by the accent marks, which usually did not mark the prominent long down-glides in any case. There may have been ‘correctors’ who still remembered how the contonation actually worked; the usual suspects also would have included local correctors who only had to pretend that they knew, to earn their pay from the Greekless. There is a tragedy here that descriptivists might respond to cathartically: the very economy of method in the deployment of signs by Aristophanes of Byzantium, preserving the contours of the original contonation by marking only the location of its onset of rising pitch, led to misunderstanding and distortion of the message when the very notion of a disyllabic, melodic and quantitative implementation became foreign to the speakers’ habits of monosyllabic vocal stress.
In the meantime of course compositional prosody itself was changing; almost certainly there were some ‘correctors’ from Nagy’s ‘pre-Byzantine’ world who wished to reflect the prosodic or musical styles of their own day—to set the old texts to a ‘New Music’. The citharodes may already have been doing this long since to hexameters with the newfangled, seven-stringed or greater instrument (κίθαρα, λύρα) in different modal scales. Here is West:
Classical writers distinguish rhapsodes from citharodes. The latter sang the poetry of Homer and others to melodies of their own, accompanying themselves on the cithara, and they looked back to Terpander as the famous exponent of this art. Homer [by contrast] was thought of as a rhapsode …3
This Terpander is credited with increasing the number of strings in archaic times from the four of the Homeric phorminx to a lyre of seven or eight, filling out the modal octave. West discusses what he calls the ‘dual-tradition hypothesis’, as between performers of hexameters with and without an accompanying musical instrument (rhapsodes versus citharodes), but avers that ‘[t]he dual-tradition hypothesis is supported by no ancient testimony, and it lacks intrinsic plausibility.’
I conclude that Homeric ‘singing’ was truly singing, in that it was based on definite notes and intervals, but that it was at the same time a stylized form of speech, the rise and the fall of the voice being governed by the melodic accent of the words.4
It is therefore not excluded that the Bankes Papyrus, and other provincial papyri displaying idiosyncratic accentuation outside the Alexandrian standard, may in fact represent citharodes’ occasional librettos.
I do not have it in me to attempt to approximate the melodic innovation of a citharode. There is something far more concrete, tethered and doable about attempting such musical patterns as the word accents themselves deliver. The dactylic beat is a real pulse that shapes the voice. You may imagine on your own how a rhapsode might wield his stick in the service of truth.
The circumflex and especially the post-acute barytone were no longer in later periods the heavy accentual phenomena around which poetry and choreia were composed. And of course at some point the high-pitch accent took on the qualities of stress, where the vowel grade and quantity of unstressed syllables were weakened. But the East Roman manuscripts derived from Alexandria manifestly preserved the reality and the effects of the classical contonation; there is on this level every reason to believe that the Venetus A is written in the same language as Homer spoke, and preserves his reinforcing prosody as well as it preserves his dactylic feet. I have argued that Latin has never lost this contonation; it is preserved in Latin’s received stress pattern. The recessive rule for Latin is that the contonation must begin (that is, the pitch-rise must occur), where possible, on the second mora before the ultima. Because in Latin (unlike Greek) the rule is indifferent as to the quantity of the ultima, the circumflex is more prevalent in Latin than in Greek. Every long Latin penult is a circumflex! One hears it still on Italian penults. The classical Greek, Latin and Vedic/Sanskrit ‘contonation’ must best be understood to have an Indo-European provenance, but its development into multiple recessive phenotypes seems to me to remain an historical problem as yet unsolved.5
Some part of the manuscript evolution may have involved the transition, after Alexander and in Rome, to the uses and needs of non-native speakers. But the school of thought or scholarly tradition that led to the superb Venetus A manuscript proposes a text based on the best authorities at the University of Alexandria, who included geometers, with their demonstrations and their lemmas, as well as grammarians. (This is clearly not true of some of the local papyri which contain no apparati.) The final result of the process can be understood on its own terms as an inevitable development in the representation of a language where word boundaries and tonal prosody were mutually definitive empirical phenomena. Even native students could use such guidance.
Let us take stock of what can be said about the historical authority of these fully developed East Roman manuscripts and their system of prosodic marking. Nagy could not be more wrong when he infers that what he calls ‘Byzantine’ writing practices are ‘newer’ and therefore ‘less accurate’ than what came before. The opposite is true: these practices finally express what had been latent in the representation of what was a word-level accent, by the separation of words and by focusing on terminal intonation. Lower case letters also have proved very useful things for many purposes. In distinguishing between lower and upper cases in writing, these manuscripts can be said to have initiated a new kind of visual registration of language, including new kinds of distinction and emphasis sometimes not possible in unwritten speech. They can even be credited with a usage influenced by geometry that has been sadly lost in the modern scholarly habit: where we have ‘footnotes’—with no connection to dancing feet, except perhaps their odour—these manuscripts had lemmas supported by argument. Sometimes ‘newer’ actually means ‘innovative’ and ‘better fit for purpose’: this East Roman era is, with great good fortune, one of those times.
And in the face of what Nagy himself calls a ‘sea change’ in the accentual practices of living contemporary Greek, the East Roman manuscripts like the Venetus A show an extraordinary scholarly diligence and economy of method in transmitting the original word-level prosody, despite the contemporary reality and the anciently lost representation of key phonemes like the ubiquitous -w- of the digamma. They preserve the quantities of Homer’s original words: the dactylic hexameter and its ictus have made sure of this. It is known that the position of the rising pitch in a word did not change over the course of Greek’s history—unless it once presented, say, prior to the antepenult—and persists today as the location of the stressed syllable. Therefore in the vast majority of cases, there is no error in the marking of these accents by the Alexandrian grammarians.
Consider this analysis of West’s, citing Wackernagel:
The Alexandrian scholars and the grammatical tradition that derived from them attached importance to the study of Homeric accentuation, and record a number of particular accentuations that cannot have been established either from the living Greek language or from theory and analogy, but must have been preserved by a continuous tradition of oral performance from early times: such accentuation as ἀλεωρή, γάρ αὐτον, δηιοτής, θαμειαί, ταρφειαί, καυστειρῆς, ἀγυιαί, Τρωιούς, ὅθί σφισι, and others.6 That rhapsodes continued to perform Homer in Hellenistic times, and indeed much later, is known from agonistic inscriptions … It is not the case that Alexandrian scholars kept their noses in books and ignored performing artists of their times … Homeric scholars were naturally familiar with the sound of rhapsodes’ voices. These rhapsodes performed Homer in such a way that the word accents were audible, and they were taken to have at least some authority in the matter. Rightly so, seeing that their accentuation had peculiar features which appear genuinely ancient. How ancient? As ancient as the times when the words concerned were in use in the living language. That implies a delivery in something not too far removed from speech tones at least as early as Homer, if not earlier.7
That the scholar-critics used to frequent the theatre is evident not only on the level of the language they recorded, but on that of the mimesis they witnessed. Consider the scholia to Iliad II.360 recorded in Venetus A; when Nestor ‘turns away’ mid-speech from addressing the Achaeans to speak directly to Agamemnon, the scholia refer to τὸ σχῆμα ἀποστροφή, ‘the figure apostrophe’.8 The frame-scholium explains that Nestor literally turns his speech away from one to the other (ἀποστρέφει δὲ τὸν λόγον ἀπὸ τῶν Ἑλλήνων πρὸς τὸν Ἀγαμέμνονα). Of course neither addressee is really there at all. There is only the rhapsode, and very likely the audience implicated by him—the scholiast gives this away when he calls the Achaeans ‘the Hellenes’. Homer never calls these people ‘Hellenes’! The scholiast is describing how the rhapsode turns from addressing his Hellenic audience, who on this side of the fourth wall have been sitting in for the Achaeans, and addresses Agamemnon directly.
The use of ‘schema’ suggests that there was a stylised gesture involved when the rhapsode turned away from the theatre of Greeks to conjure them an invisible king. It comes with the territory for a solo thespian, with no guitar to hide behind, that he must populate the imagination, and that he can summon appearances from the dead, the invisible, or from thin air; these story-teller’s wonders are nigh impossible with the machinery of staged drama. The rhapsode has power over what is not seen, and can layer the forces and agents in his kind of narrative without the constraints of other tellers’ media.9 At times he is like a medium, being possessed in turn by these different agents, human and divine, and even the recovery of ‘himself’ and his narrator’s voice can thus become a dramatic moment.
Apostrophe has been treated as a literary figure, a narrative technique, even as a quirk born of metrical necessity in the Homeric context. But ‘schema’ and ‘schemata’ refer in the context of the Athenian stage to the repertoire of gestures used by choruses of dancers, as well as solo dancers (Herodotus 6.129) and actors. It is surely the most obvious gesture for a solo actor doing a one-man-show playing many parts, to turn to his imagined interlocutor and address him or her; the rendering of dialogue by one performer requires this, apart from the vividness of the effect, which can be made to surpass what can be achieved by two actors. This ‘apostrophe’ happens not only between the rhapsode’s audience and a protagonist, but sometimes in the opposite direction—at least this is how I imagine the turn to the familiar character whom he addresses in the second person, as in the celebrated cases of Patroclus and Eumaeus: here I think the ‘you’ is addressed to the audience itself, to whom these endearing characters are felt specially to belong. But in the present instance the rhapsode is already embodying Nestor, and it is as Nestor that he turns directly to Agamemnon. This move, already in character, is also described as ‘apostrophe’ by the scholiasts. And of course the rhapsode will at times embody Patroclus and Eumaeus as well, and speak as them.
Really apostrophe, ‘turning away’, is on a spectrum with the regular announcement between speeches that the narrator is now ‘exchanging his place’ (ἀμειβόμενος), usually translated as though the character himself is ‘replying’. Apostrophe is, however, as marked and vivid a ‘breaking the fourth wall’, in relation to the rendering of the usual exchange between speeches, as the critics (apart from the metrical dogmatists) have made it out to be.
There is, of course, a difference of Alexandrian scholarly opinion on occasion (στενάχων vs. στεναχῶν). But West’s inference from Wackernagel’s data that the preservation of particular anomalous accentuations implies a living tradition of rhapsodic performance, which preserved such pronunciation in the face of a changing synchronic picture—a tradition with which the grammarians were very likely familiar in performance—is richly to the purpose. It once again bespeaks an almost cultish conservatism, which begins not so much with the thespian or concert-planning instincts of rhapsodic performers, but originally with the pre-recension demands from their audience in Solon’s Athens.
It is difficult to overstate the significance of this demand by an audience, from the performers, for their Homer in its proper order. Consider the scale of its implications. This was a generation that knew its poems in their sequence, composed as written literature or otherwise, but expressed an anxiety about hearing any change in the order, as though they knew they were dealing with a record as fragile as memory, which had to be repeated as it was for fear of its loss or distortion, including the words whose meaning was forgotten and those whose accentuations were foreign to their own use. Recently an anxiety was reported in the news about a priest who had misspoken the lines repeated in Catholic baptism, that twenty years’ worth of baptisms had been invalidated. (The priest had been saying ‘we’ instead of ‘I baptise’; the royal ‘we’ was not considered as a possible mitigation; yet English-speaking Roman Catholics in my lifetime have started saying ‘We believe’ instead of translating Credo as they used to do.) There is perhaps a comparable anxiety expressed both popularly and among a professional elite about the ‘restoration’ of works of visual art. Nothing must be added! Only various accumulated pollutants must be removed. And yet there are some who find that the restored Sistine Chapel beams at them like a billboard. Anxiety breeds anxiety as preservation of Michelangelo takes on a present colouration and the actual artwork a metaphysical reality.
Nowhere is this anxiety more fraught about the loss of what is held preciously in the memory, because it belongs to a world that can never otherwise return, than in the restoration of ruins from antiquity in Greece and elsewhere. No one dares restore the colours to the pediments in their Hindu vitality, at the risk of losing our psychic connection; instead everything must be preserved in each generation exactly as it now is, or was at the first northern encounter; or else protected in museums with the fragmentary ruins reproduced on the sites in facscimile. Hence we continue to dwell in the limbo of the Pentelic mausoleum, some of us imagining the coloured life of the Parthenon and its polis, but living in an eternal present of the off-white marble ruin. Nothing must be changed: roped off and professionally maintained, the living ruin ensures a psychic connection to the historical one. This professionalism is a metaphor for the Alexandrian scholars—and perhaps the late rhapsodes as well—who preserved the alien sounds in performing and writing Homer in manuscripts, as though freezing Meleager’s brand against consumption in the embers. Ion the rhapsode cast a spell, himself suspended like an iron ring in the magnet’s energy (Plato, Ion 533d ff.); but everyone knows that a spell no longer works unless one gets the exact original words in the exact original order.
All that is required to unlock these conserved texts musically is the knowledge that the marked accent once signalled the beginning of a contonation that was usually disyllabic, where the second element was dynamically prominent when long, at last to make sense of the various patterns of reinforcement of the ictus of Greek metres. Hence these manuscripts preserve the original prosodic map of the Homeric composition. The quality of some syllables surely changed: digamma disappeared, along with intervocalic sigma, and metathesis happened—though in a way that did not alter the prosodic reinforcement of the underlying rhythm. Hence a student who practices the contonation, and becomes familiar with its interplay with the hexameter, can feel with genuine confidence that he is within the Homeric sound-world. If the Attic spelling of Homer still sounded like Homer, so can we. That is the achievement of the methodical and historically conscious approach to prosody in the late Roman manuscripts.
It is time for students of classical Greek to put down their other work and learn how the disyllabic contonation works in the performing of texts, in prose rhetoric and in all the kinds of poetry. I find that this is not intuitive work for people raised on classical paradigms. To the best of my knowledge, none of the modern languages in which classical texts are studied bring native experience to bear in implementing a disyllabic, polyvalent accentuation. But it is essential that this training be done—for all the Greek poets—but for Homer’s sake to begin with. I suspect it is only the rhythmic, prosodic, musical life of his words in the mouth and on the breath that will finally cure students of Homer of the notion that there is any rhythmic, prosodic, or musical actuality to the idea of a metrical formula. To look at his score without hearing it, Mozart also is full of repeating formulas. They are a feast for the tone-deaf and for fans of statistics and Lego™.
A metrical formula is an abstraction stripped of syncopation, melody, or meaning. It is therefore impossible, in the nature of things, that it be a unit of poetic or musical composition. The emperor has had no clothes for a long, long time in the domain of Homeric criticism—for longer than my lifetime. Yet I can imagine a generation of students who will become practiced in the Homeric music. I was never the boy in class who was any good at poetry. But I imagine a generation of critics to come who will start to feel with some confidence, and be able to express it to others through the melodic movement denoted in the Homeric manuscripts, when it is that Homer sings, and when it is that Homer nods.
It is indeed a moment of arrival to be able to criticise Homer’s music. The Indo-European Homer was beyond the realm of such criticism. But on two authoritative and wise observations in French, the failed colometric derivations of the hexameter die a death. Meillet observes that the dactyl is ‘une innovation du Grec.’10 He is referring to the time-equality of its arsis and thesis: the dactyl is an isometric foot born to be stepped in dance. It is not a foot that represents a stylisation of any language. It is therefore critical to note that the later Aeolic cola used by some to derive the hexameter itself—whether the glyconic or its cadence movement called the pherecratean—must contain a dactyl.11 Hence it is illegitimate to seek an Indo-European comparate for them, like the Vedic gayatri, that does not: Greek metre, and Greek folk dance, uniquely showcase this isochronic step.
Pierre Chantraine observes that ‘il apparait que le rhythme naturel de la langue grecque s’adaptait mal à la métrique rigide de l’hexamètre dactylique.’12 Homer’s hexameter lines show vowel and syllable lengthening and shortening (metri gratia) as well as spondaic substitution and even morphological adaptations;13 Aeloic lyric cola, and lyric dactyls generally, do not. It is therefore impossible, logically and factually, for lyric segments that show no sign of the maladaptation Chantraine describes, to join together or expand so as to form such a licensed rollicking as we find amply displayed in Homer’s hexameters. Greek choral lyrics are language-driven dance; Homer’s poetry, on the other hand, is dance-driven language.
Homer the Greek spoke an Indo-European language, all right. But his musical composition shows a movement through its syllables that was uniquely, and forever, a pas de Grecques. The rhythm of these syllables, and the syncopated melody they carry, survive in the Venetus A via Constantinople and Venice. They will sound again—resounding!
Originally from A. P. David, ‘Singing Homer’s Spell: The Disyllabic Contonation and the Proposition Made by East Roman Manuscripts’, Dramaturgias: Revista do Laboratório de Dramaturgia da Universidade de Brasília, 19:7 (2022), 805-66; since revised.
see A. B. Lord, The Singer of Tales, Cambridge, Mass.: Harvard University Press, 1960, 25, 99-123; quoted above from Steve Reece, ‘Some Homeric Etymologies in the Light of Oral-Formulaic Theory’, The Classical World 93:2, 1999, 189-90.
M. L. West, ‘The Singing of Homer and the Modes of Early Greek Music’, 113-14, emphasis and bracketed comment added.
Ibid., 115.
David, ‘The Classical Interlude: A Law of Recessive Tonal Accent for “Classical”-Era Indo-European’, Dramaturgias: Revista do Laboratório de Dramaturgia, 4:2 (2017), 229-37.
citing J. Wackernagel, Kleine Schriften, 880-1, 1102-7, 1154-78.
West, 114.
Lara Pagani, ‘The Iliad “Textscholien” in the Venetus A’, in Marco Ercoles, Lara Pagani, Filippomaria Pontani and Giuseppe Ucciardello, eds., Approaches to Greek Poetry, Berlin: De Gruyter, 2018, 91
For an excellent account, for example, of the rendering of the figure of Eris through the telling of the aristeia of Diomedes, see Katherine Kretler, One Man Show: Poetics and Presence in the Iliad and Odyssey, Washington DC: Center for Hellenic Studies, 2020, 59 ff.
Antoine Meillet, Les Origines indo-européennes des mètres grecs, Paris, 1923, 57.
For a discussion of the relation between dactyl and cretic in lyric cola, see David, The Dance of the Muses: Choral Theory and Ancient Greek Poetics, Oxford: Oxford University Press, 2006, 236-45.
Pierre Chantraine, Grammaire Homérique, Paris: Librairie C. Klincksieck, 1942, 94.
David, The Dance of the Muses, 158-9.