Homer's Tune and the Lie of Tradition

Humours of the Homerist 1

The failure of my kidneys and a debilitating anaemia have for now put paid to my recording of Homer. But the project will resume! In the mean time, here is an essay, aspirationally the first in an unplanned sequence—as inspiration serves …

Homer’s Tune and the Lie of Tradition

¹

Despite the simple fact that Homer was sung, or could have been or was meant to be sung, but we have somehow lost the melody—until now the most straightforward, uncontested truth about the historical transmission of ancient μουσική—an industry of Homeric poetics, breeding apprentices, has all the same been built in the modern age by focussing solely on metre, in a way that is completely oblivious to verbal melody. I do not mean merely missing it or forgetful of it, but wilfully blind to the role of melody in the word-music of Homer and its composition. The prejudices of the approach have proliferated into a schoolroom orthodoxy, taught, it would seem, as an esoteric and professional, ‘alternative’ fact.

This industry is funded on the myth of an ‘oral tradition’. ‘Oral’ in this use is highly restricted, however. No aspect of the sensual, rational or even mnemonic qualities of melody—a key aesthetic and cognitive component, one should have thought, in human and other oral expression—plays any part in the performance from the mouth which is thought to be relevant to the substance transmitted in the hypothesised tradition.

The orthodoxy’s founding principle—were it to be formulated and expressed—is that Greek verse is quantitative; we are introduced to the term ‘quantitative metre’. The expression is largely tautological. It comes into focus only by contrast with the term, ‘stress metre’. English poetry, for example, is characterised by contrasting emphases, alternating stressed amongst unstressed syllables, inculcating a pattern of expectation which the natural internal stress patterns in words can counterpoint as well as reinforce—so introducing the joys of syncopation and emphasis into the dynamic meaning of the words. Absolute duration becomes variable, dependent equally on the words’ phonetic nature and the metabolism of the performer. My late teacher Elliot Zuckerman used to like to cite Hamlet’s lines, by way of illustration:

Absent thee from felicity awhile

And in this harsh world draw thy breath in pain

To tell my story.

Consider the length of time it takes to say the Latinate ‘Absent thee from felicity awhile,’ in relation to the Anglo-Saxon monosyllabic gasp of the line following, ‘And in this harsh world draw thy breath in pain … ’ Both are ‘equally’ English iambic pentameter.

Note that in the English case, the metrical description of Shakespeare’s verse is not possible, is not even conceivable, without a focus on the accentual profile and behaviour (the stress pattern) of his peculiar English. The conceit in the concept of ‘quantitative metre’, applied to ancient languages, is that native prosody is at best a reinforcing adornment upon a structure that is adequately analysed in terms of purely quantitative metrical patterns. This is the habit with Latin poetry, despite the fact that Latin stress patterns have been transmitted to modern readers and reciters, and do indeed seem musically to reinforce Latin meters—almost all of which were demonstrably borrowed from Greek! Whereas in the case of the Greek originals, the melodic contours indicated by the natively developed system of accent marking are now completely ignored in metrical or more general poetic analysis. The pitch patterns of Greek have been preserved! But against all common sense, they have heretofore been deemed irrelevant to the patterns of downbeats (‘ictus’) encoded by the metrical feet. The critics have left the composers a tune without a beat and a beat without a tune.

Homer’s lines, by contrast with Shakespeare’s pentameters, are like equal segments of a song composed with a given time signature: all of his lines could be recited each in the same length of time. (This is not my prescription for effective performance, but all the same it is a demonstrable fact.) Whole number ratios govern the differences in duration of the different notes or syllables. This reflects an origin of the movements, in both Homeric and modern classical music, in dance form. Measurement by steps, performed by the feet (literally) and the body in space and time, lends itself to such rational rhythms, whereas speech stresses are, in a word, freer in their use of phonetic contrasts (pitch change, duration, volume) for emphasis.

What distinguishes Greek and other ‘quantitative’ languages is that there is a natural inherent difference in durations of certain vowels that is not merely imposed on the language, as in English song, by the metrical and rhythmic templates of the composition. There is no evidence, however, to claim that there were whole number ratios between syllabic durations in everyday speech, as for example 2:1 or 3:2. But ω and ου generally take longer to pronounce than ο, and η and ει than ε. There are long and short versions (referents for the signs of) each of α, ι, and υ. This allows for the development of conventions for deploying certain syllables in the necessarily rational quantitative time relations of a dance movement. Such conventions would allow these movements to be measured equally well by the conventionally short and long syllables as by the time relations themselves. In this way one understands why certain ancient languages did not have to develop time signatures and staffs to notate the rhythmic patterns of their poetry; the syllables, defined by rules of quantity, are capable of encoding them on their own.

In English song movements, on the other hand, syllables can be either stretched or squeezed to create rhythmic effects inside the musical line; their absolute timing therefore becomes refreshingly elastic, in notation and performance. Grandmaster Flash takes more time, objectively, to rap ‘Don’t push me ’cos I’m close to the edge’ (nine syllables in eight beats) than he does to render ‘It’s like a jungle, sometimes it makes me wonder how I keep from going under’ (twenty in five). Each sequence has its stresses evenly spaced, however, in an absolute sense determined by the backing track.

English song settings are quantitative in the same sense as texts of Greek μουσική. The danced hexameter is a backing track. I therefore ask the reader to imagine what it would be like to describe English songs—by the Beatles, say—without any reference to their melodies, or even to nuances between predicates and epithets in the lyrics. ‘All the lonely people,’ after all, are just ‘people’ at the end of the day. Would there be any insight gained into Beatles songs from parsing them into a monotonous sequence of substantives and verbs? Or rather, only a nightmarish and strangely ludicrous sort of melody-free falsification? Songs without tunes and filler for words? And yet this is what is proposed in the premises of an oralist approach to Homeric poetry and poetics, no matter how softened by remnant humanity is the socialised product of the modern Homeric critic. It is a fact, but one which yields no insight into poetics, that the Beatles’ melodic compositions are exemplars of quantitative metre. The very same can be said for the melodic efforts of Homer.

The use of the notion ‘formula’ is pernicious to description, of language generally but especially of poetry. It should not be allowed to inform the premises of new students who seek to advance their knowledge of Homer, because it can only prejudice and distort their encounter. I know of no professional critic who uses the term, however well meaning, who does not in some place or some context claim that Homer’s usage in a particular instance is ‘formulaic’, to imply that to that extent its literal meaning does not count—as though a solution to a problem had been given. It is like answering the question ‘why do I feel heavy?’ with ‘because of gravity’; as so often in the modern academy, there is no real answer there, only the parody of a scholastic one. The ‘formula’ is pernicious in Homeric criticism because it leads to dismissal rather than insight into usage, explaining away rather than explaining. Certain segments and whole lines and passages are glossed, under various unsettled criteria, as formulaic, which glossing serves as a sort of conversation-stopper.

One should have no problem with the description ‘formulaic’ if it was intended, as it ought to be, as a criticism of the poetic achievement. The poet is, after all, accused of using ‘filler’. Students and translators resort to such description, however, as a sort of ‘get out of jail’ card when other explanations fail or do not satisfy; it is always an excuse not to think about possible associations and interpretations. The wheat of Homer’s meaning is separated from the chaff, which is so much filler that only finds its way into Homer’s lines because of ‘metrical necessity’.

Pseudo-problems are created and then solved, with an apparent, but illusory, forensic insight. There are, for example, ‘equivalent formulas’ of the same metrical shape, which violate the ‘economy’ pillar of Milman Parry’s theory—one concept one formula—which is obviously a key to any theory of extemporising under pressure of performance. Such a problem is grist for the Homeric critic’s mill, and I do not doubt the insight gained by using the oral theory as foil, as it were, to prove intent in the use of, say, ἱπποδάμοιο instead of the metrically equivalent ἀνδροφόνοιο as an epithet of Hector.

²

But metrical equivalence is, strictly, a mathematical notion. There is no equivalence in the actually different sounds of language. A rose by any other name (ῥόδος) would not smell as sweet. There is in fact nothing equivalent, in phonic or melodic or semantic substance, between so-called equivalent formulas except their metrical pattern—which is a thing experienced by both singer and audience primarily through, and never apart from, their melody, rhythm and meaning. The best that can be said for the melodic similarity or even equivalence of phrases is that it is a thing as yet unstudied in Homer’s text. Perhaps that will change in response to my work. But as it is, the erudite response to Homer’s intentionality in singing ἱππόδαμος rather than ἀνδροφόνος, would seem to be a somewhat resigned ‘duh?!’

But the oral-formulaic theory presumes far more upon Homer than merely to describe the poetry. It claims to be a theory of composition. Formulas are posited as ‘metrical building blocks’. We are proposing to compose the Beatles’ songs—unforgettable words and melodies—out of syllabic quanta with no fixed melody, no accentual patterns to give them rhythm, and only a blunt force ideographic meaning. No ‘Beatles’ need be involved. Metrical building blocks are capable only of building larger quanta, not rhythmic wholes like hexameter lines, and certainly not melodies or music or songs or poetry. All they are is toneless patterns of duration! It is neither an irony nor an accident that those who propose formulaic metrical building blocks developed in an oral tradition, as the elements of aesthetic composition in Homeric musical poetry, show themselves, in the act, to be musically illiterate. This is in no sense a kind of unitarian polemic: the critique here rests on the protagonists’ own foundations, on an evident category misunderstanding of the nature of alleged ‘metrical building blocks’ in relation to the crafted melodic pattern rhythmed out through whole hexameter lines. It takes more than a brick to build a house, let alone the Great Pyramid.

The notion of ‘tradition’ is carted in, because formulas had to be developed and ‘handed down’ to be perfected over unrecorded eons, for the compositional theory to become thinkable and possible, so that without any basis in ancient testimony or in practical experience, compositional formulas become ‘traditional formulas’. In this very formulation they take on a monolithic and totemic existence, thought by some to extend even beyond Homeric Greek to the solemn metaphysics of an Indo-European phraseology, an inference that is wholly unwarranted by fact or witness. There are in fact certain frozen collocations and proper names in Homer, such as Ἀργεϊφόντης, that do bear within them a diachronic history that invites forensic philology. But all such examples are distorted into indistinguishable muck by the wanton flood tide of the alleged and ubiquitous ‘traditional formulas’ with which Homer has been obliged to compose his lines. All this foisted on us by people who could not be bothered to propose how to sing these ostensibly frozen phrases, as against those phrases they allow to be non-formulaic, to venture an opinion whether they sounded good at all.

Small wonder that the Homer industry is prone to frame its Homeric criticism as the ‘discovery’ of anonymous artistry in usage, despite its subject’s ostensibly primitive, oral and formulaic origin and composition. This is a kind of colonialist sophistry that finally serves neither Homer nor his students. Students should not be ‘set up’ with dummy problems. If they find Homer formulaic, perhaps they have not actually heard his music. That is very likely, given the emphases and skills of his current teachers and translators. But perhaps they should credit their judgement, and seek out poetry instead that does not so offend or otherwise confuse them.

My own suspicion is that the reality of generational transmission, and hence tradition, is not an oral thing, but a physical one, grounded in the physical habit of dance. The continuity of Greek ‘folk dance’ is a trans-millennial thing, a real transmission which continues today, despite vast differences in its present linguistic and melodic accompaniment. The muscle memory of rhythm seems far more impervious to time and use than are habits of language and melody. This is a transmission by the feet and through the feet; even here the concept of ‘metre’ is an abstraction from the actual felt reality that is a footfall.

Consider the statement of Paul Maas, to my mind slightly gleeful: ‘we have no means of reading, reciting or hearing Greek poetry as it actually sounded. It may be possible for us to form a mental notion of it; but such a notion is too shadowy to serve as a basis for the scientific investigation of the subject.’

³

I say gleeful because this asserted impossibility and shadowiness is precisely the cue for the mathematically inclined to wade in and declare the abstraction of Greek metre to be the summary, knowable nature of Greek poetry. In relation to this tack perhaps oral theory seemed clever and freeing, in foisting onto Homer its own shadowy invention of the ‘metrical formula’—albeit as part of a mathematically elegant (if descriptively false) theory of economy and extension. The ignoring of the seemingly irrelevant accent marks in texts seems to have led not to the realisation that all we could know, sadly, about the sound and performance of Greek poetry was its metre, but to the delusion that metre was all there was to know. But the deeper aesthetic, or pseudo-religious, sentiment involved, in preferring colourless marble and unrecoverable melody to the once vibrant reality, has been forever immortalised in Keats’ Ode to a Grecian Urn: ‘Heard melodies are sweet, but those unheard/Are sweeter …’

* * *

Τhe central fact of the modern Humanities, is that language changes. This is a most awkward fact. If schools themselves are about the transmission of traditions over generations, this is a fact that suggests traditions may be false.

‘Tradition’ is in fact a claim upon and about the present, not the past. It is an answer to an anthropologist’s question, in the same way as to a child’s: why do we do it this way? Because ‘tradition’. Like ‘gravity’. It is not contested in the academy that language changes: it is an accepted fact of history and nature. All the same this fact is an affront to the reason, in that it calls into question the possibility of transmission of any kind, without the danger of miscommunication—whether synchronical or diachronical. Conservatism in general, and an educator’s emphasis on the continuity of tradition, stems (in part) from a reactive fear for this danger. The Protagorean flux of language is like an a priori principle in the empirical Humanities, not itself amenable to direct causal explanation. It would seem to be fundamental that at some level, usage changes the thing used, although the proximate historical causes of various kinds of language change may be as varied, and idiosyncratic, as the world’s languages themselves.

In the face of this pervasive fact of flux, ‘tradition’ must be seen not as a window to the past reality, but first and foremost as an assertion in the present, laying claim to property in the past. The fact is change; the reactive falsehood, seeking comfort, is called tradition.

Homeric scholars enjoy speculating about an ‘oral tradition’. It gives them a level of free play without practical, historical or physiological constraints. No attempts are made either to compose or improvise Greek hexameters; there is no interest in experiments being done and theories tested. All data comes only from comparison with observable exemplars, with accompanying assertions of parallelism if not similarity.

The Greek Dark Age: Fact or Fiction?

This academic fantasy life is projected onto the blank canvases of a so-called ‘Dark Age’ spanning unrecorded centuries. It was philologists who, once upon a time, knew that Linear B could not possibly be Greek. This was because the unfired tablets baked in a conflagration were dated far too early, by coordination with a suspect Egyptian chronology for Mycenaean ware. These scholars evidently knew something about the likely or possible rates of change in languages. But when it was discovered that Linear B was in fact Greek, this did not lead the philologists to demand a revision of the standard chronology. The Dark Age of Greece that results from this chronology is an academic imposition on the empirical reality, where excavated Mycenaean finds are in fact immediately contiguous with, or even overlap, Archaic ones:

fragments of geometrical vases . . . have been found on various sites in Greece together with late examples of Mycenaean pottery.

⁴

Archaeologists embarrass not only themselves, but the whole academy, with theories about ‘heirlooms’ to explain the presence of pristine Mycenaean objects deposited in archaic levels, after a purported five hundred or more years of human life which left no trace of itself on Earth.

There is no Dark Age stratum in the ground. Here is Denys Page on the site at Hissarlik, thought by some to be the location of Ilium:

for the site is barren of deposits which might be referred to the period c. 1100-700 B.C. Not one sherd of proto-geometric pottery is known to have been found at Troy—not by Schliemann, or by Doerpfeld, or by Blegen himself. We are now in effect asking what happened at Troy during the Dark Ages of Greece, from the the 11th to the 8th century B.C.: and this is the answer which we must accept—that there is nothing at Troy to fill the huge lacuna. For 2000 years men had left traces of their living there; some chapters in the story were brief and obscure, but there was never yet a chapter left wholly blank. Now at last there is silence, profound and prolonged for 400 years.

⁵

But the Homer people have not hesitated to frolic in that desolate playground, with mysterious bards wandering around non-existent chieftains’ dining rooms to support them. Their world has left no signs of life, agriculture or defecation. It was a cosmic conflagration that happened to bake the unfired clay tablets that preserved us Linear B. (The already fired tablets were obliterated in the blaze: that entire archive has disappeared, and all of the Mycenaean culture and economy has had to be reconstructed from the meagre snapshot in virgin clay left in the warehouse on doomsday.) The demise of the Mycenaean civilisation was clearly sudden and catastrophic. The cause of the hellish atmospheric phenomenon no doubt resulted in the migrations of peoples, languages and coastlines.

The Dark Age of Greece (unlike a Linear B tablet baked in a conflagration) is a fantasy unrelated to anything that was ever under the sun: it is purely the by-product of a forced chronology. Such is a fit place for moderns to locate any number of invented oral traditions that might be asked to produce an Homer. The Dark Age of Greek chronology is in fact a prime example of something created merely and simply ‘for the sake of the metre.’

As we shall have reason to note, the Greek language was no exception to the Heraclitean flux: changes continued inexorably among its various versions, from pre-Mycenaean times to the present. There is good evidence to suggest that by the Roman era, the Greek accent was already developing into a monosyllabic stress feature, which it remains today, on the syllable that used to contain the highest pitch point in a word. So on what basis does one suppose that the accent system bequeathed us in manuscripts that date from an era when the spoken accent had changed, conveys any data useful for the pronunciation of earlier classical texts, let alone the Homeric ones?

My Own Theory: How Homeric Pitch Dynamics Worked

The so-called ‘Byzantine’ rules for classical Greek accentuation, explicitly rooted in the practice of Hellenistic Alexandrian scholars and expressed in the surviving manuscripts, have yielded in W. S. Allen’s comparison with Vedic the notion that the Greek accent was in fact a ‘contonation’. In the Greek (and Latin) case this contonation is recessive under rule. The written marks in the manuscripts show the place within a vowel where the voice rose in pitch, and this is also the syllable where in later Greek the word came to be stressed. But in the classical practice there was also an automatic down-glide which followed the rise. In fact in situations where this down-glide is unable to occur within the word boundary—as for example in oxytone-final words not followed by a pause—the grave sign indicated that the rise itself was suppressed. And most important of all: when this down-glide occurred over a long or heavy syllable, it was more dynamically prominent than the preceding rise. This combination of the unmarked down-glide in pitch with long syllabic quantity turns out to be critical in the reinforcement of metrical ictus—and it goes unmarked in the accentual system. It is a ‘post-acute barytone’.

The only place where the contonation was fully signed was on long vowels bearing the circumflex, which graphically represents the rise and the fall in pitch. But in most cases the contonation was disyllabic; the syllable of rising pitch was marked acute, and the down-glide-bearing syllable was not marked. Hence there are at least two clues that a particular manuscript reflects the performance of an original recessive disyllabic contonation, rather than the later monosyllabic stress: 1) the use of the grave sign exclusively to signify the suppression of an acute, where the automatic following down-glide in pitch could not occur within a word or word-like boundary; and 2) the meaningful distinction between circumflex and acute when they are placed on long vowels. The circumflex means the voice rises in pitch within the first mora of the long vowel, then falls over the rest of it; the acute means that the voice rises within the second mora of a long vowel and falls on the following unmarked syllable. Under the description given by Allen for the ancient system, no more than one vowel mora may follow the down-glide in pitch; hence a circumflex can only recede to a long penult when the ultima is short.

Consider King Lear’s ‘Never, never, never, never, never!’ It is the anti-iambic line, the anti-line, because each and every word in it denies the ictus. The jolting of the dislocated rhythm conspires with situation and meaning to force upon the actor a transcendent histrionic moment, as Lear faces Cordelia’s death. We know this only because we know that ‘never’ has to be stressed on the penult, never the ultima. We take it for granted that we must know how to speak the language, how and where to stress the syllables, before we can begin to take in and enjoy Shakespeare’s poetry. And yet the oral theory of Homeric poetics, and the entire oral tradition it postulates, completely ignores the accent marks transmitted in our best texts of Homer.

There is no controversy: Homer’s Greek was a language with a pitch accent. No expert disagrees with this! Every last word came with a built-in pitch pattern, and hence every line and sentence comes in a melodic shape. The written accentual system indicates that this pattern was organised at the level of the vowel mora, smaller than the syllable. My new theory has shown when the syllable with the marked rising pitch was accentually prominent, or when instead the following down-glide in pitch was prominent so as to reinforce or syncopate the underlying beat. The theory describes syllables marked grave as denoting the suppression of pitch rise, indicating a state of tension rather than neutrality. That is, oxytone final syllables can always be ‘released’ in their melodic expression, by the positioning of pauses or the deployment of word-extending enclitics. Hence it is reasonable to conclude, as well as common sense to assume, that considerable attention was paid to the arrangement, positioning, release and suppression of pitch accents and clausal melodies in the composition of Greek poetry and prose. (This attention is not mirrored in the modern study of such poetry and prose.)

In some papyri of Homer we appear to find an unsystematic, somewhat zealous use of the grave, whose instructional intention has therefore to be speculated. We also find uses of the circumflex that violate the strictures on recession; both circumflex and acute seem to come to mean ‘accented’ simply, as though the monosyllabic stress had taken over without distinction between them. Neither of these things occur in the East Roman manuscripts. In the face of the very real change in practice around the Greek-speaking people who produced them, I argue that we have in these manuscripts a case of real transmission, perhaps reliant on continuing anachronistic traditions in rhapsodic accentual practice, but in themselves, purely graphic—written and scholarly. Now that is a rare brooch! The system of accentual marking of syllables, themselves organised metrically by conventions of length, allows the music of words from a different time, some of them now obscure, to be brought to life in the present. It is an achievement like that of a classical score. Forget the oral fantasies, this one-time Alexandrian graphic innovation, preserved in some manuscripts of the Eastern Rome, comes to constitute the real tradition that preserves for us the music of Homer.

As I post my recordings, I display a Greek text of Homer for each performed segment of the Odyssey, where the syllables determined to be prominent by the new theory of the Greek accent are set in boldface. The down-glide is generally prominent when it occurs on a long syllable. This determination is not univocal, however. There are times when the theory allows for a choice of emphasis between adjacent syllables; the ictus of the metre especially, and external pressure often put by the hexameter on syllabic quantities, influences the judgement. But judgement in such matters has always been key to the transmission, by historical individuals, of the Homeric text now extant. I presume to assume a place in this transmission. I make the case that the Roman-era manuscripts represent a proposition about the Homeric text, one perhaps not properly recognised as such by modern editors and commentators; and my added graffiti will treat this proposition as a lemma, an assumption, in its own demonstration: a concrete visual guide to the reality of Homer’s word music for those handicapped by literacy.

The Homeric compositions are miracles of oratorio, a stylised melodic declaration that achieves a sublimity of realisation in narrative and lyrical expression, a sublimity of sheer representation that is immediate, despite the fact that the medium of its art language is no longer alive. In the legacy of modern classical music, we find ourselves awash in such miracles. No quantity of well-meaning liner notes filled with detailed historical biography has ever resolved, or even circumscribed, the miracle of musical composition. Just what was Mozart’s change of diet in Vienna that caused a change of favoured key? How proximate were Miles Davis’ most pointedly modal solos to injections of heroin? Modern Academia offers us Homeric poetry as a traditional product of an illiterate Dark Age, which left no archaeological, geological or literary strata. Parry’s theory utterly ignores the melodic substance recorded in our best manuscripts, which is the substance that metre exists to serve. It is a complete travesty, a betrayal of the classicism that means to bring us forward the genius of Greek authors and composers.

None of these original inheritors of Homer’s music saw fit to explain, or explain away, the irreducible genius of his craft, except perhaps with the rather ambivalent suggestion that the singer was blind. Whether the composer is from Chios or Vienna, the music is the sound and the thing. It is disgraceful that Greek authors’ modern champions and defenders do not even attempt to sing the song of Homer, to try to register the experience of repeated melodic phrases, before filling introductory volumes of evidence-free verbiage in the way of wide-eyed students. Homer was the maverick who saw and seized the possibility, how the chanted bardic catalogue could turn into an epic one-man show. Once upon a time, her North Pole was in the Bear. Homer’s poetry comes to the ancient Greeks, and to us, from a vanished age: but thanks to the luminosity of her craft, not at all a dark one.

© A. P. David 2022, Templeton, California.

1. Draft of a paper to be published in Dramaturgias.

2. see Richard Sacks, The Traditional Phrase in Homer, Leiden: E. J. Brill, 1987, 220ff.; discussion in A. P. David, The Dance of the Muses: Choral Theory and Ancient Greek Poetics, Oxford: Oxford University Press, 2006, 169-71.

3. Paul Maas, Greek Metre, Oxford: Oxford University Press, 1962, 3-4.

4. E. A. Gardner, Ancient Athens, London: Macmillan, 1902, 157-8.

5. D. Page, “The Historical Sack of Troy”, Antiquity XXIII: 1959, 31.