Oral tradition and the making of the poems

The single most important thing to understand about Homeric poetry is that its language was built for performance without a text. That is not a romantic claim about bards and firelight; it is a technical observation about how the verse is put together, and it was established by close statistical study of the diction itself.

The formulaic system

Milman Parry, in work published in the late 1920s and early 1930s, showed that Homeric diction operates as an economical system. For a given character, in a given grammatical case, at a given position in the hexameter line, there is characteristically one fixed noun-epithet phrase and no competing alternative. The system has enormous extension — it covers a great many metrical situations — and remarkable thrift: it very rarely provides two ways of saying the same thing in the same metrical shape.

That combination is the signature of a traditional medium rather than an individual style. A writer composing at leisure accumulates alternatives, because alternatives are expressive. A singer composing in performance needs exactly one phrase that fits, and needs it instantly.

This explains a feature of the poems that puzzles new readers: the epithets often do not fit the moment. Odysseus is polymētis, “of many counsels”, when he is not counselling anything. The sea is “wine-dark” when nobody is looking at its colour. Achilles is “swift-footed” while sitting down. The epithet is not a comment on the scene; it is a metrical component doing structural work.

Parry, Lord and the comparative evidence

Parry then tested the model against living practice. With Albert Lord he recorded South Slavic epic singers in the 1930s, observing how illiterate performers composed long narrative poems in performance using an analogous system of formulae and typical scenes. Lord set the results out in The Singer of Tales (1960), which remains the standard statement of the position.

The comparative argument is powerful and has limits worth stating. The South Slavic tradition is not the Greek one, and no recorded South Slavic song approaches the scale or the architectural control of the Iliad. What the fieldwork established is that oral composition-in-performance of long narrative verse is possible and that it produces formulaic diction. What it did not establish, and what remains disputed, is how a poem of the Homeric epics’ scale and design came out of such a tradition.

Typical scenes and larger structures

Below the level of the phrase, the tradition supplies whole recurring patterns: arming, feasting, sacrifice, the arrival of a guest, the launching of a ship. The Odyssey is particularly rich in these, and its central moral concern — hospitality, xenia — is repeatedly examined by varying one standard scene. A stranger arrives; the sequence of greeting, seating, washing, feeding and only then questioning either runs correctly or is broken. Polyphemus breaks it by eating his guests. The suitors break it by consuming their host’s substance. The Phaeacians perform it perfectly. The poem thinks by variation on a pattern the audience already knows.

An artificial dialect

Nobody ever spoke Homeric Greek. The language of the epics is a Kunstsprache, an artificial poetic amalgam, principally Ionic but carrying Aeolic forms and some older material, with alternative forms of the same word available for different metrical shapes. Its layers are of different ages, and the mixture is itself evidence of a long tradition.

One detail is a small masterpiece of philological reasoning. The transmitted text does not write the letter digamma (a w-sound, ϝ), which had dropped out of Ionic Greek. But in a great many lines the metre only works if a word originally began with it. The sound is invisible in the text and audible in the rhythm. That is direct evidence that the verses were shaped at a period earlier than the text that transmits them.

Rhapsodes and performance

The performers of the poems in the classical period were called rhapsōdoi — plausibly “song-stitchers”. By that date they were reciters of a more or less fixed text rather than composers in performance, and they competed at festivals. Plato’s Ion is a portrait of one of them: a professional Homer-reciter who can hold an audience of twenty thousand and cannot explain what he knows.

Ancient tradition reports a rule at the Panathenaia in Athens requiring the Homeric poems to be recited in their proper order, one performer taking up where the last stopped. The rule is reported in later sources rather than contemporary ones. It is generally taken as good evidence that by the sixth century BCE there was an authoritative sequence to keep to — which implies something reasonably close to a fixed text.

How the text became fixed — the disputed part

This is where certainty ends. The competing accounts include:

• Dictation. A monumental composer dictated the poems to a scribe at some point after the adoption of the Greek alphabet, producing a text at a stroke.
• Gradual crystallisation. The text became progressively fixed over centuries of performance and transmission, with no single moment of creation — the view associated with the evolutionary model of Gregory Nagy.
• A Peisistratean recension. The tradition that the Athenian tyrants had the poems collected and ordered in the sixth century BCE. Reported in antiquity; its historicity is disputed and it should not be repeated as fact.

These are not variants of one story. They imply materially different things about what kind of object the transmitted text is. The evidence does not currently decide between them, and this platform does not pretend otherwise. See The Homeric Question.

Why this matters for reading and for adaptation

An oral-traditional poem is not a script with a missing performance. Repetition in Homer is structural, not redundant; the epithets are load-bearing; the “digressions” are how the medium supplies context. Modern adaptations routinely strip all of this, because film has its own redundancies and does not need the poem’s. That is a legitimate transposition, not a betrayal — but it does mean that a film cannot preserve the specific quality that makes Homeric narration what it is, and criticism that pretends otherwise is asking cinema to be verse.

The one thing an adaptation can preserve is the framed, retrospective structure: the hero telling his own story, with all the unreliability that implies. Our comparison of the poem and the 2026 film treats that as the central structural question.