The Singer of Tales

Milman Parry, 1928: a Sorbonne doctorate arguing Homer’s epics were built from an oral tradition’s toolkit, not an author’s pen. 1933: he goes to Bosnia to check — and finds living singers who never tell the same song the same way twice. Albert Lord carried the proof for twenty-five years. It became The Singer of Tales.

In 1928, at the Sorbonne, a young American classicist named Milman Parry defended a doctorate under the linguist Antoine Meillet. The dissertation — written and published in French — argued that Homer’s style runs on fixed metrical formulas: recurring word-groups that sit in the same slots of the hexameter line, over and over, too systematically to be one writer’s habit. The implication was large: the Iliad and the Odyssey looked less like books composed by an author and more like the residue of an oral tradition. It was a strong argument built entirely on internal evidence. What Parry did next is why this page exists: he went to check.

Across two trips between 1933 and 1935, Parry went to Yugoslavia with his assistant Albert Lord and a local guide and interpreter, Nikola Vujnović. They concentrated on Bosnia, and the choice was deliberate: literacy there was lowest, so the oral tradition was considered purest — least contaminated by print. Over those two field seasons they recorded thousands of hours of living epic poetry, performed by illiterate singers called guslari, who accompany themselves on the gusle, a one-stringed fiddle, and can carry a song for hours at a stretch.

Never the same song twice

The finding fits in one sentence: ask the same singer for the same song twice, and you will never get the same wording twice. This is not a failure of memory — it is a different technology altogether. The guslari were not reciting a fixed text, because there was no fixed text. They were composing, live, from a toolkit the tradition had handed them: formulas — word-groups pre-fitted to the slots of the meter — and themes — repeatable content blocks like the arming of the hero, the feast, the launching of the ship. Master the toolkit and you can improvise an epic in real time, the way a jazz musician improvises over changes: the structure inherited, the performance yours, every telling new.

Turn back to Homer with that in hand, and the text itself testifies. Of the roughly 27,803 lines in the Iliad and Odyssey combined, about 9,200 are repetitions — from short recurring word-groups to entire type-scenes rerun nearly whole. One line in three repeats. In a written literary work that would be a flaw. In a formula-built oral poem it is the fingerprint of the method.

The twenty-five-year carry

Parry never got to write the theory up. On December 3, 1935, in Los Angeles, a gun discharged as he unpacked clothing from his luggage; he died at thirty-three. The fact needs no dressing. What it left behind was a young field assistant holding the recordings and an unwritten book. Albert Lord carried it — returning to the field, by the publisher’s own account of the work, into the 1950s — and published the synthesis in 1960 as The Singer of Tales, from Harvard University Press. The book extended the argument past Homer to Beowulf, the Chanson de Roland, and the Byzantine Digenis Akritas, and it is why the field speaks of the Parry–Lord theory. Since 1960, “Homer” has had to be reckoned with less as an author in the modern sense than as the name we give a tradition — or, on the theory’s own logic, perhaps the last and greatest of many singers, the one whose telling got written down and stuck.

The date nobody has

So when was it written down? The honest answer is the uncomfortable one: nobody knows, and the field has no consensus. The “8th century BCE” you will meet in every popular account is, by the encyclopedia’s own admission, “more often stated than argued for.” What actually anchors that date is the alphabet: the Greeks adapted Phoenician letters in the 8th century BCE, which makes it the earliest the poem could have been written down — an outer bound, not evidence that it was. Serious positions run much later; Robert Lamberton places the epics’ fixation as late as the mid-5th century BCE, astride the spread of general literacy. The defensible picture is not a date but a river: Bronze Age memory fossilized in the verse, Iron Age performance, an Archaic-period writing-down somewhere along the way, and centuries of rhapsodes still reciting — and lightly reshaping — the poem before the text froze.

Four hundred years of re-performance

And the song never stopped being performed — it changed instruments. George Chapman finished the first complete English Homer: his Odyssey came out in installments in 1614–15, and the whole works — Iliad and Odyssey bound together — in 1616, the Odyssey in rhyming couplets of iambic pentameter where his Iliad had used fourteeners. Chapman held the field until Alexander Pope’s version, completed in 1726. Two centuries after Chapman, in 1816, a twenty-year-old John Keats — who read no Greek — opened Chapman’s Homer and wrote “On First Looking into Chapman’s Homer,” a sonnet about exactly the vertigo this page is about: an old song landing like a new world. In 1708 Anne Dacier translated the Odyssey into French prose — the first woman documented to have translated the poem at all. And in 1996 Robert Fagles’s Penguin translation, with an introduction by Bernard Knox, became the modern touchstone, rendering Odysseus — the polytropos, literally the “many-turned” man — as the “man of twists and turns,” where Richmond Lattimore had gone with “man of many ways.”

Then, on November 7, 2017, Emily Wilson published hers with Norton — the first complete translation of the Odyssey by a woman into English (the qualifier matters; Dacier was there in French three centuries earlier). Wilson matched the Greek line for line — 12,110 lines — in plain, speakable iambic pentameter, and let one word do the arguing:

Tell me about a complicated man. / Muse, tell me how he wandered and was lost / when he had wrecked the holy town of Troy. — Emily Wilson, The Odyssey, Book I (Norton, 2017) · as published in The Paris Review (opens in new tab)

“Complicated,” for polytropos. Not tactical, not physical — psychological. It is the most-cited choice in every review of her translation because it declares the whole project: each generation re-performs the song for its own room, in its own register, from the same inherited units. Chapman’s couplets, Dacier’s prose, Fagles’s muscle, Wilson’s plainness — the guslari would recognize the move.

That is this week’s foundation, and it is also a working model. The Bosnian singers proved a thing can be traditional and never identical twice — built from inherited, tested units, composed fresh at every telling, no two performances the same and none of them false. A living knowledge tradition works the same way: cited to the bones, and never twice the same.