Epic Singers and Oral Tradition

  Use the following persistent identifier: http://nrs.harvard.edu/urn-3:hul.ebook:CHS_LordA.Epic_Singers_and_Oral_Tradition.1991.

5. Homer as an Oral-Traditional Poet*

Some misconceptions have arisen about the “oral theory” and about the quality of the Serbo-Croatian oral-traditional epic and its possible relevance to an understanding of the Homeric poems. They are the subject of this paper. Much of its burden is to demonstrate to the Homerist how the superb singer in the South Slavic tradition can make clear the relationship of the gifted individual to the other singers in the tradition. By observing what he does, we can learn the ways in which his songs are superior to those of his fellows, because his thought is on a higher level than theirs, even as Homer’s surpassed the songs of lesser bards for the same reason.

The theory assumes no such thing. There is no body of formulas to which all singers are restricted. Singers remember phrases they have heard and lines they have used, but they do not “memorize” set formulas.

On the other hand, there is a peculiar depth to some oral-traditional epithets—for certain of them have their own histories—which stems from a combination of metrical thrift and special meaning. Take “swift-footed Achilles” for example.

If one looks at the usage a little differently, one might say that Homer had three ways of saying “swift-footed Achilles”; πόδας ὠκὺς Ἀχιλλεύς, ὠκὺς Ἀχιλλεύς, and ποδάρκης δῖος Ἀχιλλεύς (swift-footed, godlike Achilles). These noun-epithet combinations around the idea of “swift-footed” may be fraught with meaning from ancient traditional stories of childhood deeds, as Gregory Nagy has pointed out in The Best of the Achaeans, on the basis of Pindar’s “Nemean 3.” [3] In other words, viewed in this way, ὠκὺς Ἀχιλλεύς is not an alternative for δῖος, although metrically equivalent to it, but for πόδας ὠκὺς Ἀχιλλεύς and πόδάρκης δῖος Ἀχιλλεύς, its equivalent in sense. I might point out that M. W. M. Pope failed to see this equivalence, because he was tied too closely to metrics. The fact that there are three ways of saying “swift-footed Achilles” argues that there is, or was at one time, a value to the meaning of the epithet significant enough to {73|74} warrant the invention over time of three different ways of expressing it metrically. Metrics are important. Vivante’s book on epithets in Homer proceeds as if Homer’s works might just as well have been written in prose. [4] But there are other values as well as metrics, as, for example, the kind of meaning indicated by Gregory Nagy and developed further by Joseph Nagy in his book on the childhood deeds of the Irish hero Finn MacCumail. [5]

Even though there are South Slavic singers whose style may seem mechanical (and there probably were ancient Greek singers of that caliber {74|75} also) a look at Avdo Međedović’s formulaic style easily disproves that statement. There is nothing mechanical in his composition of lines.

The repeated couplets do not have to be gnomic in content, but may express any oft-repeated idea that can be expressed in two lines. Indeed the couplet is frequently expanded by a line or two. The main thing is that there be a more or less stable block of lines that is frequently repeated and plays an important role in oral-traditional composition.

There are instances of repeated couplets in the Homeric poems as in Serbo-Croatian epic. Take, for example, Iliad 8.66-67 and Iliad 11.84-85:

ὄϕρα μὲν ἠὼς ἦν καὶ ἀέξετο ἱερὸν ἦμαρ,
τόϕρα μάλ’ ἀμϕοτέρων βέλε’ ἥπτετο, πῖπτε δὲ λαός.

So long as it was early morning and the sacred daylight
increasing, so long the thrown weapons of both took hold
and men dropped under them.

[Lattimore translation]

Or compare Odyssey 19.600:

19.600          ὣς εἰποῦσ’ ἀνέβαιν’ ὑπερώϊα σιγαλόεντα,
19.601          οὐκ οἴη, ἅμα τῇ γε καὶ ἀμφίπολοι κίον ἄλλαι.

So she spoke, and went back up to her shining chamber,
not alone, since others, her women, went to attend her.


Such lines and couplets form a special group of repeated lines. Just as the formulas come to the singer’s lips by conditioned reflex without conscious remembering, it may be correct to say that the multiformed couplets too are unconsciously remembered whole; they are certainly not memorized.


The use of the term “improvisation” in referring to the method of composition of the South Slavic oral-traditional epic has caused some misunderstandings. The Oxford English Dictionary defines improvising as making up a song or words “on the spur of the moment,” “extempore.” It quotes Tobias George Smollett in Travels Through France and Italy, who mentions “improvisatori, the greatest curiosities, unique within Italy, individuals who have the surprising talent of reciting verses extempore on any subject you propose.” The OED in its discussion also uses the word “unpremeditated” and defines improvisation as “the production or execution of anything off hand.” This is a far cry from the technique of composition of oral-traditional epic.

My own preferred term for that type of composition is “composition by formula and theme.” “Composition in performance” or possibly “recomposition in performance” are satisfactory terms as long as one does not equate them with improvisation, which, to my mind, means to make up a new nontraditional song from predominantly nontraditional elements.

There is considerable confusion on the difference between improvisation and the creation of new songs in an oral tradition of singing narrative songs. It is a complex problem, but this much may be said at present. New songs in a living tradition of epic are forged from traditional formulas and themes and deal with traditional subjects. A new {76|77} song in this genre has new names, but almost everything else in it has appeared before in the tradition in one form or another. The improvising of shorter topical songs “on any subject” “on the spur of the moment” is a very different matter, especially since the subjects may be, and usually are, nontraditional, thus requiring a new vocabulary.

The outsider misunderstands, when he is told that singers can make up songs “on any subject,” and asks for subjects outside the tradition. The result is a curiosity that proves nothing except that the singer normally does not compose a song extempore about “any subject,” but only about certain kinds of subjects, for which he has the materials in his experience.


In The Singer of Tales I described how a singer learns to sing by listening to singers and gradually absorbing—learning, if you will, but not memorizing—the often used phrases, lines, or even couplets and small blocks, which he hears. I believe that this is what we do when we learn to speak a language. Although some of the phrases may be comparatively stable, because they express the often used thoughts of the poetry, they are not irrevocably fixed. There are several ways of saying the same thing. They emerge from usage and are not imposed by the {77|78} “tradition” on the singer. There are phrases that singers use because they help in saying what they want to say and in forming lines. They do not have to use them, yet they feel no compunction about doing so. They do not avoid them; many of them are necessary for making lines, but there are other ways and sometimes singers in the natural way of singing use those other ways.

Some singers do not vary the common phrases (formulas) much, others do. Some singers use many well-known expressions, others use fewer. Neologisms are not unknown. In short, singers in practice may vary a great deal in their use of the common phrases according to their individual ability and sense of “style,” if I may use that term. Each singer in a tradition is an individual and his style is discernible as his. The “tradition” is not a rigid monolith outside the singer, but as dynamic as the singers who operate in it, who form and constantly reform it. Homer’s originality is within the tradition, not in the sense of being within limits but in the sense of being within potentials that can be realized by superior individuals. The tradition includes individuality. If one can understand the processes one can see the potentialities being expressed.

It is as important, if not more so, to understand what is meant by “tradition” as it is to appreciate the flexibility and complexity of individual idiosyncrasies within it. A tradition is easy to define but not easy to know. It is the sum of all its parts, by which I mean that it consists of all the singers of epic, good and bad, and all the performances, also good or bad, of all the songs, likewise of variable quality, in the course of the life of a culture. There is cohesiveness within a tradition, because singers learn from other singers both the songs and the technique of making them. A golden thread of family relationships runs through the tradition vertically and horizontally. Some elements in a tradition are fleeting and others long-lasting. We generally think of the latter as marking a tradition, and correctly so, but the transitory elements should be recognized as well, because they represent the momentary response of the individual poet to an idea. Here belong the neologisms and the hapax legomena. Avdo depicts one of the chief men of the northern Border, Tale of Orašac (who is sometimes a crude character on the surface, but a very complex one), as addressing the commander of the armies of the Border as “an old ox” (bivolice stari). [14] Tale asks him “Have you recently fallen {78|79} ill?” (Jesi li se freško obolijo?). The question is a little like saying, “What’s wrong with you?” Freško (freshly!) is clearly a neologism and quite possibly also a hapax legomenon. Avdo likes such words, and frequently enlivens his songs with “colloquialisms,” or borrowings. Another example is kolega, “colleague,” in the phrase, referring to one of the nobles in the assembly, moj dobri kolega, “my good colleague!” Although such usages may not have permanency, inasmuch as they are created by a traditional poet operating within a tradition, they are part of the fabric of the tradition.

One should not suppose that a traditional singer can quickly become an Ezra Pound or a T. S. Eliot, although he may already be a Homer. If he is good enough, if he too is a genius, within the bounds of his tradition he can be a Homer. But he can never be a Pound or an Eliot. There is nothing that Homer, or any traditional singer, does that requires writing, contrary to what has frequently been said. On the other hand, Pound and Eliot are inconceivable without writing. They choose their words, they construct a “fixed” text, although they may later shift the position of some words or substitute others. The very self-conscious way in which that is done argues an attitude toward a text that is not part of the mentality, or “mind-set,” of oral-traditional poets. The non-traditional poet uses a variety of references, often very individualistic and, except to a given few, obscure. The singer of oral-traditional epic cannot do all that for several reasons, and the act of writing itself is probably the least important of them. What Pound and Eliot—and many others, of course—did was the result of a long period of historical development of literature from many times and places. Theirs is a style that belongs primarily to short, for the most part non-narrative poetic forms, but it is not limited to them. The style is the result of a long period during which there was a seeking for new and different forms and effects, not merely an individual style within a traditional framework, but a style so totally different that no one else could have created it. This type of poetics is not that of oral-traditional poetry. It is ludicrous to say that there is no difference between oral-traditional poetics and written literary nontraditional poetics! Homer should not be treated as if he were writing like Pound or Eliot.

In a review of a new edition of W. B. Yeats’s poems Seamus Heaney wrote (New York Times Book Review, March 18, 1984, p. 1): “Stonework, facing and planing the blocks, the deliberate exercise of strength, the exclusive intent of the craftsman absorbed in this task—it all fits with what we know of Yeats as an artist in verse. He would work long at shaping a poem, handling and testing a line until it pressed down with the greatest semantic density and conducted the right musical strain from the lines before and after. The evidence is in the manuscripts, but it is also there in the perfect feel of the individual stanzas.” The oral-traditional singer of epic, composing in performance, even in a dictated performance with a scribe, does not, indeed cannot, compose in this way.

The singers of oral-traditional epic in Yugoslavia say that they think about songs between singings in “public.” They may even practice them to themselves or with a small, intimate group. Such rumination—”practicing” may imply too conscious an activity—should be differentiated from the early period of learning; although it is akin to it, it is less intensive. The process of learning is always going on. Every singing is involved in it to some degree, because the singer always learns in conjunction {80|81} with a song. When the singer is thinking about a song, he is thinking about its story; and when he sings to himself he is thinking about the telling of it in song. To us the song may be inseparable from the text, but to the singer a text emerges from the song at each performance even when the singer sings only to himself.

As he thinks about a song, remembering it, singing it to himself, if he is an Avdo Međedović, he will find himself embellishing his catalogues or envisioning his assemblies and realizing anew his descriptions of heroes and horses. These themes are not composed “on the spur of the moment.” Although not textually fixed, they have a kind of stability. It is not merely by chance, for example, that the two scenes of hospitality on the hero’s journey to Buda in Avdo’s “Smailagić Meho” are differentiated so carefully from one another. In one the sons of the family aid their father, in the other it is the daughters. The conversations are different, too, one comparing the older and the younger generations—a favorite subject of Avdo’s—the other disclosing the possible untrustworthiness of the vizier in Buda. Both conversations are significant in the song, presenting ideas and information that serve the overall purposes of the song. They are the result of Avdo’s thinking of the story and of its telling. They were not “improvised” extempore.

Version A

Kud gođ skita, za Aliju pita. Wherever he wandered, he asked for Alija.
Kazaše ga u gradu Kajniđu. He said he was in the city of Kajniđa.
Kad tatarin pod Kajniđu dođe, When the messenger came to Kajniđa,
Pa eto ga uz čaršiju prođe. He passed along the main street.
Pa prilazi novom bazdrđanu, He approached the new shopkeeper, [81]
Te upita za Alino dvore. And he asked for Alija’s house.
Bazdrđan mu dvore ukazao. The shopkeeper pointed out the house.
Kad tatarin na kapiju dođe, When the messenger came to the gate.
Pa zadrma halkom na vratima. He beat with the knocker on the door.
Zveknu halka, a jeknu kapija. The knocker rang, the gate resounded.
Doma nema Đerđelez Alije, Đerđelez Alija was not at home,
Samo stara Alijina majka. Only Alija’s old mother.

Salih Ugljanin actually put those words in those positions, yet he did so subconsciously, following the traditional techniques of his inherited poetic language. The patterns of assonance, alliteration, and line structure, the balancing of words by analogy and antithesis are all part and parcel of the traditional singer’s language. In this he is unlike the non-traditional poet who, after considerable thought, consciously chooses the exact words he or she wishes and places them precisely where he or she wants them for particular effect, for nontraditional reasons of which the poet is perfectly aware. Yeats’s poetics is his own. Salih’s poetic art is not uniquely his; it is inherited and shared by others, valued by his listeners as the proper expression of ideas held in common.

Parry collected three versions of Salih’s “Song of Bagdad.” The passage just given (version A) is from the earliest of them, dictated to Nikola Vujnović on July 23, 1934. A second version was sung for the records on July 24, 1934, and a third sung on November 22, 1934. [17] The corresponding passages in the later two versions are as follows:

Version B Version C
November 22, 1934 July 24, 1934
[lines 131-142] [lines 113-128]
Kud skitaše, za Aljiju pita, Kud goj skita za Aljiju pita;
Za gaziju Đerđeljez Aljiju.  
Kazaše ga u gradu Kajniđi. Kazaše ga u gradu Kajniđu,
  U Kajniđi gradu carevome. [115] [82]
Tera tatar u gradu Kajniđi.  
A kad dođe, do u Kajniž siđe, [135] Kad tatarin pod Kajniđu dođe,
Ej! traži kulu Đerđeljez Aljije.  
  Pa otide novijem sokakom
  Kroz čaršiju od grada Kajniđe,
  Pa upita mlada bozdrđana,
  Bozdrđana upita tatarin: [120]
  “Kamo dvore Đerđeljez Alije?”
Kazaše mu kulu i kapiju. Bazdrđan mu dvore ukazao.
Tatar brže tera na kapiju, Ćera tatar novijem sokakom.
  Kad Aljiji na kapiju dođe,
Pa zatrupa halkom na vratima. Pa zatrupa halkom na vratima. [125]
Jeknu halka a goljema vrata. [140] Jeknu halka a goljema vrata.
Nemaše tu Đerđeljez Aljije, Tu nemaše Đerđeljez Aljije,
Samo stara do penđera majka. Samo stara na odaji majka.

The text in the column on the left was sung the day after Salih had dictated the same song in the first excerpt quoted earlier (version A) and that in the column on the right was sung four months later. The fine pairing and balancing of verbs in the dictated version are not found in either of the sung texts. In dictating Salih got into a kind of rhythm, which manifests itself in this passage by pairing and balancing. The tendency to alliteration, assonance, rhyme, and chiasmus, however, is strong in all three “performances.” Such acoustical patterns and arrangements are clearly organizing forces in the structure of groups, or blocks, of lines. To a certain degree, within blocks of lines, they have mnemonic value, but it is clear from these three examples that the singer has not memorized a fixed text. Only one line is common to all three texts, “Kazaše ga u gradu Kajniđu,” “He said that he was in the city of Kajniđa,” but note that in the July 24 version the name of the city is in the dative, “Kajniđi.”The first two or three lines form a block. The first is built on “za Aliju pita,” “he asked for Alija.” “Kud god(j) skita,” “wherever he wandered,” in versions A and C results in internal rhyme. Version B, curiously, breaks that pattern. I would suggest that Salih had in mind, or, {83|84} better, was possibly influenced by “kazaše,” “he said,” in line 3 of version B, which is in line 2 of the other texts, with the potential of creating the following two lines:

Kud skitaše, za Al’ju pitaše,
Kazaše ga u gradu Kajniđu.

All that is speculation, of course, yet version B is notable for its second line, “za gaziju Đerđelez Aliju,” “for the hero Đerđelez Alija,” which repeats the internal rhyme of the putative first line and typically elaborates on the content of that line. The structure is not unlike the appositives of Anglo-Saxon poetry. The third line of version B and the second of versions A and C round out the block with the chiasmus of alliterations k-g-g-k. Just as the second line of version B elaborates on the preceding line in that version, so the third line of version C elaborates on its second line, and in so doing it makes a chiastic arrangement of consonants between the two lines:

Kazaše ga u gradu Kajniđu,
U Kajniđu gradu carevome.
K           g     g           K
           K           g

The B version is the only one of the three to have a line indicating the journey of the sultan’s messenger from Stambol to Kajniđa, “tera tatar u gradu Kajniđi,” “the messenger rode to the city of Kajniđa. ” The clause “tera tatar,” with its attractive alliteration, is, however, a constant in the passage in the sung texts. Later in version B the messenger goes to Alija’s gate, “Tatar brže tera na kapiju,” “The messenger quickly rode to the gate,” and in version C he goes along the street, “ćera tatar novijem sokakom,” “the messenger rode along the new street.”


Whatever sense of textuality there may be in the oral formulaic type of composition is to be found in the blocks of lines like those just discussed and the themes that are made up of such blocks. One expects a degree of verbal correspondence among instances of the same theme by the same singer in one or more songs, and it is the subject matter itself and the blocks of lines used to express it that cause that expectation.

A theme is a repeated passage and it is useful in composition because it {84|85} has a group of lines basic to it which by modification and addition of other lines may be adapted to a particular context.

It is clear from these parallel passages of description of battle from two different songs by the same singer that (1) he has not memorized the theme, that (2) he repeats blocks of two, three, or four lines more or less verbatim, that (3) he sometimes alters the order of groups of lines, that (4) he has a shorter and a longer form of some groups (as in the first lines of the theme), that (5) there are some groups close in meaning but different in wording (indicated by the bracketed but joined lines), and {88|89} that (6) he has adapted the theme somewhat to the particular battle (indicated by the asterisked lines). This type of composition is very characteristic of oral-traditional formulaic style.

The armings of Agamemnon, Patroclus, and Achilles begin with the same three lines as that of Paris. The first epithet in these lines is καλάς, modifying κνημῖδας in the preceding line, and I would like to suggest that one of the factors in its choice was that it alliterates with the “k’s” of that line. The only other epithet in these lines immediately follows καλάς. It is άργυρέοισιν, modifying ἐπισφυρίοις, and I should like to suggest in this case also that the choice was in part influenced by the fact that it assonates with the last word in its line, ἀραρυίας, even though it does not modify that word.

The arming of Paris continues with an idea peculiar to that hero’s corselet (line 3.333):

οἷο κασιγνήτοιο Λυκάονος ἥρμοσε δ’ αὐιτῷ.

it belonged to his brother Lykaon, but it fitted him also.

This picks up the “k” alliteration again. After the description of the greaves, the donning of the armor is resumed in lines 3.334-338:

ἀμφὶ δ’ ἄρ’ ὤμοισιν βάλετο ξίφος ἀργυρόηλον
χάλκεον, αὐτὰρ ἔπειτα σάκος μέγα τε στιβαρόν τε·
κρατὶ· δ’ ἐπ’ ἰφθίμῳ κυνέην εὔτυκτον ἔθηκεν
ἵππουριν· δεινὸν δὲ λόφος καθύπερθεν ἔνευεν·
εἵλετο δ’ ἄλκιμον ἔγχος, ὅ οἱ παλάμηφιν ἀρήρει.

Across his shoulders he slung the sword with the nails of silver,
a bronze sword, and above it the great shield, huge and heavy.
Over his powerful head he set the well-fashioned helmet
with the horse-hair crest, and the plumes nodded terribly above it.
He took up a strong-shafted spear that fitted his hand’s grip.

The same lines are used in the arming of Patroclus, and there too they are preceded by a line peculiar to that hero, referring to his breastplate, or corselet:

16.134          ποικίλον ἀστερόεντα ποδώκεος Αἰακίδαο.

elaborate, and starry, of swift-footed Aiakides.

The last line of the run, which mentions the spears, is changed from one spear taken up by Paris to two taken up by Patroclus. {90|91}

3.338          εἵλετο δ’ ἄλκιμον ἔγχος, ὅ οἱ παλάμηφιν ἀρήρει.

He took up a powerful spear that fitted his hand’s grip.

16.139          εἵλετο δ’ ἄλκιμα δοῦρε, τά oἱ παλάμηφιν ἀρήρει.

He took up two powerful spears that fitted his hand’s grip.

The arming of Paris ends with that line, but that of Patroclus continues with what he did not take, Achilles’ Pelian ash spear. In other words, the basic lines in each case have been adapted to the hero of the moment, Paris or Patroclus.

But the first three lines we considered were used to introduce the arming of Agamemnon and of Achilles as well. In the case of Agamemnon, lines 11.20-28 describe the special corselet Agamemnon put on:

τόν ποτέ οἱ Κινύρης δῶκε ξεινήϊον εἶναι.
πεύθετο γὰρ Κύπρονδε μέγα κλέος, οὕνεκ’ Ἀχαιοὶ
ἐς Τροίην νήεσσιν ἀναπλεύσεσθαι ἔμελλον·
τοὔνεκά oἱ τὸν δῶκε χαριζόμενος βασιλῆϊ.
τοῦ δ’ ἤτοι δέκα οἶμοι ἔσαν μέλανος κυάνοιο,
δώδεκα δὲ χρυσοῖο καὶ εἴκοσι κασσωιτέροιο·
κυάνεοι δὲ δράκοντες ὀρωρέχατο προτὶ δειρὴν
τρεῖς ἑκάτερθ’, ἴρισσιν ἐοικότες, ἅς τε Κρονίων
ἐν νέφεϊ στήριξε, τέρας μερόπων ἀνθρώπων.

that Kinyras had given him once, to be a guest present,
for the great fame and rumour of war had carried to Kypros
how the Achaians were to sail against Troy in their vessels.
Therefore he gave the king as a gift of grace this corselet.
Now there were ten circles of deep cobalt upon it,
and twelve of gold and twenty of tin. And toward the opening
at the throat there were rearing up three serpents of cobalt
on either side, like rainbows, which the son of Kronos
has marked upon the clouds, to be a portent to mortals.

After this special passage the lines in the two other armings (those of Paris and Patroclus) reappear in that of Agamemnon slightly changed in 11.29-31:

ἀμφὶ δ’ ἄρ’ ὤμοισιν βάλετο ξίφος· ἐν δέ οί ἧλοι
χρύσειοι πάμφαινον, ἀτὰρ περὶ κουλεὸν ἦεν
ἀργύρεον, χρυσέοισιν ἀορτήρεσσιν ἀρηρός. {91|92}

Across his shoulders he slung the sword, and the nails upon it
were golden and glittered, and closing about it the scabbard
was silver, and gold was upon the swordstraps that held it.

Before the first line has ended, the sword’s description (29-31) has begun. That description is followed by that of the shield, a special and ornate passage that is unparalleled in the other passages (11.32-40).

ἄν δ’ ἕλετ’ ἀμφιβρότην πολυδαίδαλον ἀσπίδα θοῦριν,
καλήν, ἣν πέρι μὲν κύκλοι δέκα χάλκεοι ἦσαν,
ἐν δέ οἱ ὀμφαλοὶ ἦσαν ἐείκοσι κασσιτέροιο
λευκοί, ἐν δὲ μέσοισιν ἔην μέλανος κυάνοιο.
τῇ δ’ ἐπὶ μὲν Γοργὼ Βλοσυρῶπις ἐστεφὰνωτο
δεινὸν δερκομένη, περὶ δὲ Δεῖμός τε Φόβος τε.
τῆς δ’ ἐξ ἀργύρεος τελαμὼν ἦν· αὐτὰρ ἐπ’ αὐτοῦ
κυάνεος ἐλέλικτο δράκων, κεφαλαὶ δέ οἱ ἦσαν
τρεῖς ἀμφιστρεφέες, ἑνὸς αὐχένος ἐκπεφυυῖαι.

And he took up the man-enclosing elaborate stark shield,
a thing of splendour. There were ten circles of bronze upon it,
and set about it were twenty knobs of tin, pale-shining,
and in the very centre another knob of dark cobalt.
And circled in the midst of all was the blank-eyed face of the Gorgon
with her stare of horror, and Fear was inscribed upon it, and Terror.
The strap of the shield had silver upon it, and there also on it
was coiled a cobalt snake, and there were three heads upon him
twisted to look backward and grown from a single neck, all three.

The basic lines then reappear for another brief spell, also somewhat modified in lines 11.41-45:

κρατὶ δ’ ἐπ’ ἀμφίφαλον κυνέην θέτο τετραφάληρον
ἵππουριν δεινὸν δὲ λόφος καθύπερθεν ἔνευεν.
εἵλετο δ’ ἄλκιμα δοῦρε δύω, κεκορυθμένα χαλκῷ
ὀξέα τῆλε δὲ χαλκὸς ἀπ’ αὐτόφιν οὐρανὸν εἴσω

Upon his head he set the helmet, two-horned, four-sheeted,
with the horse-hair crest, and the plumes nodded terribly above it.
Then he caught up two strong spears edged with sharp bronze
and the brazen heads flashed from him deep into heaven. {92|93}

In the last passage I have underlined the changes from the “basic lines” of the theme, although one should also note that line 43 has the ἄλκιμα δοῦρε found in the arming of Patroclus in 16.139. The arming of Agamemnon, which ends with line 44 in the passage just quoted, or, more strictly, after the first words of line 45, is a combination of modified basic lines and special passages pertaining to it alone. The preceding examples illustrate admirably the typical thematic structure and composition in the Iliad, which are in method of composition so very similar to that in the South Slavic epic.

Let me comment very briefly on the second point first. In the Moslem Slavic Balkans the singing of epic was fostered in the courts of the pashas and beys, courts and nobles as rich if not richer and more luxurious than the courts and nobles of medieval Europe. The personages of the epics themselves were nobility of various ranks, witness the assembly in “Smailagić Meho” quoted earlier. These are men “dobroga soja i odžaka,” “of good seed and family,” as the poetry puts it. {93|94}

One might add that in contrast to the lack of information available about the circumstances of composition of the medieval poetry we have an abundance of evidence of the exact way in which the Serbo-Croatian songs in the Parry Collection were recorded. We know that they were collected under rigidly meticulous field conditions, with statements by the singers themselves on the phonograph discs about their lives, what songs they knew, and how and from whom they learned them.

In addition to their concern about a Yugoslav peasant epic, some Homerists have felt that the South Slavic oral-traditional epic is so inferior in quality that it has no relevance to the Homeric poems and should be ignored. As a mild and genuine example of that attitude, and with all due respect to a great scholar, I refer to an article published by Albin Lesky in 1954. In it he wrote:

We must now have reached the kernel of our problem: the question of the orality or literacy of Homeric composition can be decided only from stylistic analysis … Are the structure and execution of the Homeric narrative such that one can think of them as having come into being in pure orality? Here what Wolfgang Schadewaldt has showed us about the {94|95} Iliad has special weight. He not only had made the magnificent plan of the poem impressively visible, but he has demonstrated in detail the techniques of foreshadowing, retarding, and combining, which have made the Homeric poem from a plain traditional type of narrative into a great work of art … Let us be reminded at least of two paradigmatic cases: the decision of Zeus, which the nodding of his head in Book 1 announces as effective, is revealed in Books 8, 11, and 15, bit by bit, through the prophecy of the god. The death of Achilles, which will not be told in the Iliad, but nevertheless is included in it with such artistry, emerges ever more and more strikingly before our mind’s eye through a series of speeches beginning with the words of Thetis (18.96) and reaching to the words of the dying Hector (22.358). All that these two examples represent is unthinkable in an orally conceived epic. From time to time we feel sure that in many cases an oral composition exhibits a remarkable technique of “conjointure,” but when we read the summary of the frequently mentioned epic of Avdo Međedović in Bowra’s book [1952], the gap between it and Homeric art leaves a decisive impression. One can scarcely expect that the publication of the complete text will change any of that. [25]

I am acutely aware of the very valuable role played by Lesky in respect to Parry’s work in Germany, for which I am enormously grateful. Hence my first reaction to this quotation was astonishment that such a reputable scholar as Albin Lesky would make final judgment— and it is clearly a final judgment—on a 13,000-line song from the two-page summary that I gave Bowra in 1949! Could one make a proper judgment of the Iliad from a two-page summary? Would one have noticed from such a summary, intended simply to sketch the narrative, that the plan of Zeus was mentioned with increasing involvement in Books 1, 8, 11, and 15? Would the successive foreshadowings of the death of Achilles have emerged from such a summary? I think not. Something must have fogged Lesky’s vision. He seems to have been ready to condemn a work merely on report, without having read it!

Moreover, what is so amazing about the references in Books 8, 11, and 15 to the plan of Zeus initiated in Book 1 of the Iliad that it would have required writing? Are oral-traditional singers supposed to forget one of the main threads of their song, if other action intervenes? How absurd! An oral-traditional singer knows his song. Like any storyteller he knows what he has said and he knows what he is going to say. One does not need writing to tell a story well and artfully. {95|96}

Between Books 1 and 8 of the Iliad Homer turns his attention away from the will of Zeus and that eminence’s plan. Most explanations attempt to demonstrate that Homer’s plan is far superior to that of Zeus. Would an oral-traditional singer have been able to stay away for thousands of lines from one of the main subjects of his song, expressed forcefully and very elaborately near the beginning of his song, and revert to it eventually? I believe that something similar happens in Avdo Međedović’s long song, “The Wedding of Smailagić Meho.” [26] In it, one of the significant elements brought into the story by Meho’s betrothed, Fatima, is the exile of her father to Bagdad by the traitorous vizier in Buda. For thousands of lines no mention is made of this and one begins to wonder whether Avdo will ever resolve Zajim Alibeg’s complication in the story, or whether he has simply forgotten it. Eventually, however, after several oblique references to it, Avdo returns to the question of Fatima’s father and brings him back from exile. The story of the exiled Zajimbeg was not in Avdo’s source. This is Avdo’s addition to the version he heard, and he does not forget it, in spite of much intervening action. But Zajim’s return is important to Avdo’s concept of this epic. The fate of Zajim, whose family and succession have been threatened and virtually destroyed by the vizier, is compared with the fate of Meho’s father Smail, old and impotent, the future of whose line was also threatened by the same enemy. Both houses are involved, and Avdo tells fully his meaningful tale of succession and loyalty. Fatima’s family is a counterpoint to Meho’s, as Agamemnon’s return in the Odyssey is a counterpoint to the return of Odysseus. Avdo has kept complete control of his narrative, even as we feel that Homer has of his. Had Lesky read the South Slavic song with the same perspicacity with which he read the Iliad, I am sure that he would have noted this.


In “The Wedding of Smailagić Meho” Avdo’s skill in building suspense and in genuine use of irony is worthy of attention, because that song too concerns the loyalty of Bosnians and the treachery of the vizier in Buda, influenced in this case not by the “seven kings” but by the Christian enemy, General Peter, and traitors around the sultan. When Smailaga gives Polonius-like instructions to his son Meho before the young man’s departure for Buda, he stresses that the vizier is their true friend and that Meho and Osman must act accordingly. Avdo and the audience know the irony of this and Avdo plays it to the hilt. When Meho and Osman return from Buda and report to Smail that the vizier is a traitor, Smail even threatens to kill his son for such a traitorous statement. The audience is on edge as to what will happen, even though they know that eventually the old man will be persuaded of the truth.


Most of the characteristics assigned by Lesky to the world of literacy were forged in the world of orality. Homer’s skill did not come from writing, although writing provided the opportunity for the poet of the Iliad and the Odyssey to compose those poems in the form we have them for someone to write them down. Had he not the skill before the writing, had the tradition not afforded him not only the “epic material” but also the conceptual and stylistic potential to create those masterpieces, the occasion would have availed him nothing. So much for writing, either from dictation—which still seems to me to be the only normal way in Homer’s time—or by the singer-poet himself. In the 1930s the writing of a scribe to the dictation of the singer provided an opportunity to Avdo Međedović for a 12,311-line version of the song of “Smailagić Meho.” But he did not need anybody’s writing to produce the 13,326-line version of “Osmanbeg Delibegović i Pavičević Luka” which he sang for the phonograph records shortly afterward. No writing was involved in it. If Avdo seized the chance offered by a stable audience, with or without a scribe, to tell his song to his fullest, then Homer too could have done the same. In fact they both did!

It would be appropriate to see what Avdo could do without writing when he was singing at 18-20 decasyllabic lines a minute, not singing a memorized text, but one whose lines he composed as he went along. Analysis of that song from its full text, rather than a judgment from a summary, will eventually demonstrate its qualities. David Bynum’s account of the circumstances of the singing of it and the dictating of “The Wedding of Vlahinjić Alija” is instructive:

Avdo Međedović began to dictate the Ženidba Vlahinjić Alije on the twelfth of July, 1935, the same day on which he finished dictating The Wedding of Smailagić Meho. Having already that day dictated the final 255 verses of the previous composition, he composed also the first 1,290 lines of the new story. The dictation continued through verse 2,979 on the thirteenth of July. The next day, 14 July, Parry intervened to defer the dictating work then in progress and have Avdo start the Wedding of Ali Vlahinjić all over again from the beginning, singing it this time for the recording microphone. That day and the next, 15 July, sufficed to complete the singing of the epic, and on 16 July the dictation of the same {98|99} story, interrupted two days earlier, was resumed and carried forward a few hundred lines to verse 3,292.

Parry then again interrupted the dictation to have Avdo sing the rather different epic of Osmanbeg Delibegović i Pavičević Luka. This continued without interruption for four days, 17-20 July, and yielded 7,132 verses of text on the phonograph records. Except for a single day of rest (2 July), Avdo had by this time been dictating and singing between one and two thousand epic verses every day for twenty-three days, and Parry gave him a well-earned vacation on 21-23 July. When Avdo returned to composing poetry on 24 July, he resumed the dictation of the Wedding of Vlahinjić Alija, which he had left off in a slightly less than half-finished state a week ago. On 25 July he finally completed that tale, having been given every opportunity that distraction, fatigue, or the lapse of time could contribute to induce him to forget or to change the elements of the Vlahinjić Alija story, or to confuse it with other epics.

Bynum notes that the same process of interruption and distraction occurred when Avdo sang Osmanbeg Delibegović i Pavičević Luka. Three other epics (Ženidba Vlahinjić Alije, Gavran Harambaša i Sirdar Mujo, and Robovanje Tala u Ozimu) intervened, along with “three further days of rest (27-29 July).” Moreover, “many hours of conversations before the microphone about still other epics and the events in Avdo’s own life (Parry Texts 12436, 12443, and 12445) were introduced between the day (20 July) when Avdo left off singing Osmanbeg Delibegović i Pavičević Luka at verse 7,132 and his completion of the epic for the phonograph records on 1-3 August.” As Bynum points out,

The ability of the great oral-traditional singer to produce an intricate and artistic structure extending over many lines is illustrated dramatically in the songs of Avdo Međedović. The first episode of “Osmanbeg Delibegović” contains an example of ring-composition. It goes like this:

(1) Osmanbeg, an old man of great dignity and importance, rises early and his wife prepares strong coffee for him, and gives him his cape and pipe. (Agrli, “strong, bitter,” appears only in this passage in the song.)

(2) In sadness he goes with his steward to the ramparts of the city.

(3) He settles there where the cannon are and looks over the plain of Osek, whence a rider appears, who is described at length as he approaches.

(4a) Osmanbeg inquires who he is, apologizing for breaking custom.

(4b) Silić Jusuf identifies himself. He has learned from his widowed mother that he has a maternal uncle, Osmanbeg. He asks why Osmanbeg has never visited his sister.

(4c) Osmanbeg replies that he was so saddened by his son-in-law Mahmut Pasha’s death that he had retired and not gone anywhere. He did not know of the birth of his nephew.

(3) Osmanbeg orders the cannon to be fired to summon the lords of Osek.

(2) With Jusuf he returns to his house rejoicing.

(1) The lords arrive. There is feasting, and congratulations are offered to Osmanbeg. The lords spend the night at Osmanbeg’s house at his insistence.

The center of the ring is the meeting of Osmanbeg and Jusuf, and especially the discovery of family relationship, uncle and nephew, that will form one of the most significant elements in the song. Continuing the story, we find a duplication of pattern with marked differences amid similarities.

The lords of Osek have spent the night with Osmanbeg. The next morning they arise early—as did he in the opening scene of the song—and the sun rises over the mountains. They have their morning drinks. Osmanbeg goes to the window and looks out over the plain. A rider appears, a Christian general, who is described as fully as Silić Jusuf was, actually even more fully. Osmanbeg calls his lords to look at the wondrous sight and gives orders that the new arrival be brought to him, but that the general should leave his weapons at the door. He speculates on why a Christian general has sought him out. {100|101}

The contrast between the two scenes is dramatic. The tired, sad, discouraged Osmanbeg at the beginning of the first episode, alone except for wife and servant, drinking his strong and bitter coffee, is contrasted with the ebullient Osmanbeg of the second episode, surrounded by the lords of Osek, drinking their morning drinks—no stark and bitter coffee here.

Instead of the rider speaking from his horse to Osman on the ramparts, we have a formal and boisterous meeting with the newcomer, a representative of the enemy camp, magnificent, but perhaps deadly.

As we go further into the song, the rings multiply and spread, and there is an intensification of dramatic contrast with the first episode which carries the story forward.


In Avdo’s song of “Smailagić Meho” there are two levels, or interwoven strands, the manner of which is typical also of Homer. As touched upon earlier, there is a contrast between the situation of two families, both of which are loyal to the sovereign. One of these, of which the hero Meho is the scion and fond hope for the future, is favored by the ruling power, in direct contact with it, in a region of the state that is a model of loyalty. The whole point of the story is the contrast with the fate of the other loyal family, whose seat is outside of that region, and whose patriarch is in exile. He and his daughter, Meho’s betrothed, are separated by traitors. Their family will die out. In one case we have a family solid and assured of a future great destiny—in the other, a split family destined to extinction. The balance and contrast are clear. What good is loyalty, when the ruler is surrounded by traitors? Where is justice to be found and how is it to be attained? The hero is the intermediary, because he momentarily flirted with the idea of treachery, which gives him the right to be the avenging hand of justice against the traitor.

Another level in the song is the personal and mythic one of the initiatory hero. This is ancient, but in Avdo’s story two or even three stages of growth are combined: the proving of the hero’s worth as a man in battle, his gaining of a bride, and eventually the assumption of his father’s position of leadership.

It seems that the processes of oral-traditional narrative patterns overlap in time and in the course of generations, layers of story coalesce. The result is a fabric of great complexity and richness, when the weaver is sensitive to the nuances of his inherited material and the full possibilities of its poetic art. {101|102}

Up to this point I have discussed Homer’s traditional art on the level of formula and theme. I wish to conclude with the role of tradition in the larger view of Homer’s songs. In the last years of my teaching I found myself lecturing on the Iliad and then the Odyssey to a class that had already read Genesis 1-39, the Enuma Elish, the Gilgamesh epic, and Hesiod’s Theogony. On the assumption that narratives of this sort were current in tradition in ancient Greece and the Near East and were likely to have been known in one form or another to the Greek traditional bards and to Homer among them, I began to view Homer’s poems from the standpoint of these, for the most part traditional, stories. On the well-established principle that in oral-traditional narrative significant elements are repeated under a variety of forms, or multiforms, I related Achilles and Sarpedon to other semidivine heroes such as Heracles and Gilgamesh and numerous others. It seemed to me that the section of the Theogony that catalogues the heroes who came after Zeus was relevant to the Iliad, which gives honored reference to Heracles. These semi-divine creatures had problems because of their semihumanity, and through them mythic thought, expressing itself in anthropomorphic imagery, struggled with the dilemma of mortal immortals, and that of their progenitors who found it difficult to accept their children’s mortality. One of the oral-traditional ways to understand Achilles and Thetis is through the multiforms of Zeus and Sarpedon or Ares and his son Askalaphos, to say nothing of their counterparts in Aeneas and Aphrodite. Tradition and the traditional singers understood the dilemmas in the traditional stories of gods and mortals which we call myths and which came into being to help bridge the inconsistencies in human life, and Homer set them forth with all the poignancy and pathos and sense of tragic inevitability at his command. The traditional narratives taught the singers that it all started when—to quote Genesis 6—”the sons of god mingled with the daughters of men. ” The special breed of semidivine heroes explained the divine part of human beings, but paradoxically it also showed how mortality touched even those glorious ones. That surely is one of the main threads of the Iliad.

Another key to an understanding of the gods in Homer may be found in the War of the Gods in the Theogony and elsewhere (for example, in the Enuma Elish), a generational war, a war for power. The Theogony echoes traditional myths that relate the rise to power of Zeus. It has occurred to me that the Iliad reflects the struggle by Zeus to maintain his power, his sovereignty, with some gods on his side, others precariously {102|103} opposing him, or wishing they could, and squabbling among themselves. Much more than Troy or an Achaean victory or even the wrath of Achilles was at stake, although the relationship of Thetis to Zeus, and the favor he owed her, were part of the struggle of the gods. The divine conflict was nicely interwoven with the theme of mortality of the semi-divines, possibly by Homer himself, but probably by the generations before him. Book 5, Diomedes’ book, prepares us for the full diapason of Book 21, when even the rivers join in the battle and Achilles is nearly swamped by gigantic forces, a magnificent figure striking out against a cosmos enraged at his rage and excess in heroically challenging his fate. The cosmic forces, even as they themselves strove together, were disrupted by the hero. The gods had allowed Diomedes for a brief spell to discern on the battlefield who were gods and who men, but Achilles took on the cosmos itself in his anguish after the death of his substitute Patroclus, a pawn in the plan of Zeus, and in his fight for glory in a short life. Traditional narratives hold in their many folds, in numerous forms, deep meanings on Life and Death. Homer learned these as a traditional singer, and with the mastery of genius he retold them in traditional form in the Iliad and the Odyssey. {103|104}


[ back ] * The original form of this paper was read at a Homer symposium at the University of Pennsylvania, March 22-23, 1984. It has not been published before.

[ back ] 1. Austin, 1975, 17-18.

[ back ] 2. Pope, 1963, 12.

[ back ] 3. Nagy, G., 1979, 325-326.

[ back ] 4. Vivante, 1982.

[ back ] 5. Nagy, J., 1985.

[ back ] 6. Hoekstra, 1964; Hainsworth, 1968; Janko, 1981.

[ back ] 7. Lord, A., 1960, 43-44.

[ back ] 8. Lord, A., 1975.

[ back ] 9. Ramo Babić, “Ženidba Smailagić Meha,” “The Wedding of Smailagić Meho,” Lord Text no. 8, lines 191-192, housed with the Milman Parry Collection, Widener Library, Harvard University, Cambridge, Mass.

[ back ] 10. Goold, 1977, 31.

[ back ] 11. Lord, A., 1960, 272-275.

[ back ] 12. Opland, 1983.

[ back ] 13. Menendez Pidal, 1967, 201.

[ back ] 14. At Iliad 1.159, Achilles calls his commander in chief, Agamemnon, κυνώπα, “dog-eyed,” “having the looks of a dog.”

[ back ] 15. My thanks to Jonathan F. McKeage for the quotation from Eliot and for pointing me to Eliot’s notes on it. The Waste Land and Other Poems (New York: Harcourt, Brace & World, 1934); the text of the poem is on pages 29-46; the lines quoted, the final lines of the poem, are on page 46; the notes to the poem are on pages 47-54.

[ back ] 16. Parry, M., 1953, no. 2, lines 113-128. See also Lord, A., 1960, 54-57.

[ back ] 17. Parry, M., 1954, no. 1, lines 131-142, and no. 3, lines 108-128.

[ back ] 18. Ibid., nos. 4 and 12.

[ back ] 19. Goold, 1977.

[ back ] 20. See Arend, 1933, 92 ff., J. I. Armstrong’s splendid article (1958) on the arming scenes, and Bernard Fenik’s treatment of them (Fenik, 1968, 73-74, 78-79, 191).

[ back ] 21. The translations of the Iliad used in this paper are by Richmond Lattimore. See Lattimore, 1951. The Greek text used is that of Thomas W. Allen in the Oxford Classical Texts (Homeri opera).

[ back ] 22. Parry, A., 1971, 439-440.

[ back ] 23. Huxley, 1969, Appendix, “Some Irish Analogues,” 191-196.

[ back ] 24. I refer to the work of Daniel Melia at the University of California, Berkeley, Joseph Nagy and Patrick Ford at the University of California, Los Angeles, and John Collins at the University of Cork, Ireland.

[ back ] 25. Lesky, 1954, 7-8. Translation my own.

[ back ] 26. Međedović, 1974a and 1974b.

[ back ] 27. Parry, M., 1953 and 1954, nos. 1-3.

[ back ] 28. There is a synopsis of this song, Parry Text no. 677 in Parry, M., 1954, Synopsis IX, 285-288, with notes on page 428.

[ back ] 29. Lesky, 1954, 8.

[ back ] 30. Međedović, 1980, x-xi.