n.Newfeed

Literature · 8 min read

How Oral Poets Like Avdo Međedović Composed Epics in Performance

For thousands of years, we assumed Homer memorized the Iliad. One afternoon in 1930s Yugoslavia blew that assumption apart forever.

When a largely illiterate Balkan farmer sat down and performed a 12,000-line epic poem from memory — longer than the Odyssey — the Harvard scholar watching him assumed he was witnessing an extraordinary feat of memorization. He wasn't. He was watching something far stranger, and far more interesting, than that.

The Night Milman Parry Heard Something Impossible

In the early 1930s, the classicist Milman Parry traveled to Yugoslavia with his assistant Albert Lord on what seemed, on paper, like a narrow academic errand: he wanted to study living oral poetry traditions to better understand how ancient Greek epic might have worked. What he found instead was a discovery that would quietly detonate assumptions scholars had held for two millennia.

The singer was Avdo Međedović, a Muslim farmer from Montenegro with almost no formal education. Parry and Lord asked him to perform an epic he had heard only once before, sung by another guslar — the Slavic term for these oral bards — at a gathering some time earlier. Avdo obliged. He sang for hours.

When the performance was transcribed and compared to the version Avdo had heard, the poem wasn't merely reproduced. It was expanded. Richer in detail, more elaborate in its battle scenes, more emotionally developed in its character moments. Avdo had taken someone else's epic and, in real time, made it bigger and better — without once having read a word of it, without a script, without rehearsal.

This single afternoon in the Balkans forced a reckoning with a question classicists had been dodging for centuries: did Homer memorize the Iliad, or did he do something like what Avdo did? The answer Parry and Lord arrived at was uncomfortable, and it has never quite been absorbed into popular understanding of what Homer was.


The Formula Machine: How Singers Actually Build Epics

The first thing to understand is that oral poets are not operating like actors running lines. They are operating more like jazz musicians — fluent in a set of compositional tools that allow them to construct, on the fly, something that sounds both new and familiar.

Parry identified these tools and called them formulas: pre-built metric phrases that could be slotted into a line of verse the way a prefabricated beam slots into a building under construction. "Rosy-fingered Dawn." "Swift-footed Achilles." "The wine-dark sea." These phrases aren't poetic decoration — they are solutions to a real-time engineering problem.

Epic meter is brutally demanding. In the Greek tradition, every line had to conform to dactylic hexameter, a strict rhythmic pattern of long and short syllables. A singer couldn't pause mid-performance to think of an elegant way to describe the sunrise. He reached for the formula, it fit the meter perfectly, and the narrative momentum never broke. The formula was cognitive scaffolding.

Above formulas, there were what Parry and Lord called themes: larger narrative building blocks, the recurring stock scenes that appear across epic traditions worldwide. The arming of a warrior before battle. The council of elders deliberating in a hall. The hero descending into the underworld. These scenes were not plagiarism — they were genre grammar, the chord progressions that a skilled singer could elaborate, compress, or inflect differently depending on the moment, the audience, and the dramatic need.

What Avdo Međedović possessed was not a better memory than other singers. He possessed a larger and more sophisticated repertoire of formulas and themes, which he could deploy with greater fluency. His performance of someone else's song was like a master jazz pianist reinterpreting a standard: the skeleton was borrowed, the execution was entirely his own.


Every Performance Was a Different Poem

Parry and Lord conducted a revealing experiment. They asked Yugoslav singers to perform the same song twice in a row — sometimes with only minutes between the two performances. Then they transcribed both versions and compared them.

No two performances were word-for-word identical. Not one. The narratives followed the same arc, the same characters made the same choices, the same battles were won and lost — but the actual verbal texture was different every time. Lines were expanded or compressed. Descriptive passages shifted. Occasionally whole scenes were rearranged.

The singers, when confronted with these transcripts, were baffled by the apparent contradiction. They had sung the same song, they insisted. To their minds, there was no inconsistency. This is because in oral cultures, sameness is not a property of words — it is a property of meaning. The song was the same because it told the same story, carried the same emotional weight, honored the same heroes. The idea that identical wording was required for identity was, from their perspective, an alien and somewhat absurd standard.

This has staggering implications for Homer. If Yugoslav oral singers in living memory couldn't reproduce their own performances word-for-word, the notion of a single original Iliad — one perfect text that Homer dictated and history preserved — becomes extremely difficult to defend. What we almost certainly have, in our modern editions of Homer, is one performance: a specific, celebrated instantiation of a tradition that had been sung in hundreds of different forms across generations. Writing didn't capture Homer. It arrested him.


What Writing Did to Storytelling (And Why It Was Violent)

The moment an oral epic is written down, something irreversible happens. One performance gets elevated to the status of the text. Every other version — every variation sung across decades or centuries, each one a legitimate and living expression of the tradition — is retroactively reclassified. Some become "variants." Others become "corruptions." The frozen text becomes the origin, and everything that doesn't match it becomes a deviation from something that was never actually fixed in the first place.

Parry-Lord scholarship has a term for what literacy does to oral tradition, though scholars apply it carefully: it colonizes. Writing doesn't neutrally record oral culture; it replaces the logic of oral culture with a completely different logic, one that privileges fixity, attribution, and singular authoritative texts. The living ecosystem gets pinned to a board and labeled.

You can watch this happen across multiple traditions simultaneously, once you know what to look for. The standardization of the Hebrew Bible during the Second Temple period. The editorial crystallization of the Mahabharata — itself roughly ten times the length of the Iliad and Odyssey combined — during the Gupta period. The recording of Norse sagas in medieval Icelandic manuscripts, performed by Christian monks who almost certainly smoothed over pagan elements they found troubling. Each of these moments represents a negotiation between an oral tradition's necessary fluidity and writing's demand for a single, authoritative version. And in every case, the negotiation ended the same way: fluidity lost.

What we have been doing ever since is treating those negotiated, compromised, historically contingent frozen moments as pure origins — as if the Iliad that exists is the Iliad that always was.


The Singer, the Audience, and the Co-Creation Problem

One of the features of oral performance that gets most thoroughly erased by writing is the audience. In a hall where an epic is being performed, the crowd is not a passive receiver of the text. It is an active shaping force.

A skilled guslar read the room constantly. A battle scene that drew shouts and forward-leaning attention would be extended, elaborated, given more detail and duration. A long genealogical passage that produced shuffling and glazed expressions would be compressed or cut. If a local chieftain of particular importance was in the hall, a singer might weave in flattering references to his ancestors or invoke his name in a moment of heroic association. The performance was not delivered to the audience — it was continuously calibrated by them.

This makes the audience something that literary culture has no comfortable category for: an invisible co-author. Modern concepts of authorship, copyright, and textual integrity depend on the idea that a work originates in a single creative mind and flows outward to passive recipients. The oral epic collapses this model entirely. The "author" is a node in a network, not a solitary origin point.

The contemporary resonances are hard to ignore. Algorithmic content feeds that expand what you engage with and compress what you skip are, structurally, doing something ancient guslar did by instinct. Live-streamers on Twitch who shape their narratives in real time based on chat responses are working in a tradition far older than broadcasting. The collaborative fan-fiction ecosystems where stories accumulate contributions from thousands of anonymous participants, each building on and modifying what came before — these look less like novelties and more like digital oral culture, arriving not through nostalgia but through the structural logic of interactive media reasserting something very old.


The Deeper Disturbance: Fixity as the Aberration

Here is the claim that Parry-Lord scholarship implies but rarely states with full force: for most of human history, it was expected that stories would change. Variation was not error. It was health. An epic that couldn't adapt to new audiences, new political contexts, new emotional needs — that epic was already dying. The ability to change was what kept it alive.

Our modern obsession with original texts, director's cuts, and authorial intent — the whole apparatus of fidelity to a fixed creative object — is historically exotic. It is a product of roughly five centuries of print culture, and even within that span, it has only fully dominated for perhaps two. Most human storytelling across most of human history operated under the opposite assumption: that the version you were hearing was this telling, not the telling, and that the story's vitality lay precisely in its capacity to be retold differently.

What Avdo Međedović demonstrated on that afternoon in Montenegro was not just a remarkable personal feat. He was demonstrating the normal operation of a technology — oral performance — that humanity had used for tens of thousands of years before anyone thought to write anything down. The Iliad we read today is not the Iliad. It is an Iliad: one brilliant, frozen moment in what was almost certainly an endlessly generative, endlessly varying tradition.

The more disturbing question this raises isn't about Homer. It's about us: what other living, adaptive traditions have we killed by insisting on writing them down — and then mistaken the corpse for the origin?

Want the reader to remember this spot? Sign in — it is free.

Next How Zora Neale Hurston's Barracoon Reached Print in 2018

The app

Keep going, or take it with you.

Free accounts read twenty pieces a day. Pro is US$1.99 a month or US$11.99 a year for unlimited reading.