Anne Carson’s Greek lives where its words give out.
On the morning of her Nobel Prize, Anne Carson’s agent reported that she said “Yikes.” The Swedish Academy praised her “playful dialogue with the classical tradition,” and the chair of its committee spoke of “the philologist’s obsession with words.” From Sappho to Sophocles, what she has returned to in Greek is less its stories than its cries, its broken tongues and the holes in its papyri: the places where the language stops being words.
“Yikes.” That was the word, according to her agent, Nicole Aragi, as the New York Times reported it, when Anne Carson learned on Thursday that she had won the Nobel Prize in Literature. A second answer came in her own voice, in a telephone interview with the Nobel Prize: “I think you’re all insane.” The third came from the Swedish Academy. Its permanent secretary, Ingrid Carlberg, had reached her by telephone, and reported that Carson was “for the moment in Iceland, where she’s listening to the wind.”
Two of the three reached the world in someone else’s sentence. In the interview she said a good deal more: tea and crackers, a floor that needed sweeping, a phone she kept turning off. Asked why the old stories are relevant now, she said she was not interested in relevance, and turned the question round: “I’m interested in us being relevant to them.” Of all of it, only one thing had no words in it, and it was not anything she said. It was what she was listening to.
The Academy’s citation praised a body of work that proceeds “in playful dialogue with the classical tradition.” The Times called her best known for her translations of ancient Greek and for “remixing them in a modern, playful voice.” Both are true, and both make it easy to read the Greek as material: Geryon and Herakles, Electra and Cassandra and Sappho, taken down from the shelf and set going again. Autobiography of Red is a Greek story retold from its first page to its last, so that reading is not wrong. At the same press conference, the chair of the Nobel Committee, Anders Olsson, placed her work somewhere else: at the intersection of “the philologist’s obsession with words and the poet’s desire for new meaning.” That is nearer the mark. It does not say where, in the words, the obsession settles. What her work keeps returning to is less what the Greek says than the places where it stops saying anything: the cry that is a word only in the dictionary, the tongue that breaks in the middle of a line, the hole in the papyrus that she prints as a bracket. The play is real. It is a philologist’s play, conducted with the text rather than around it, and it is played at the edge of language.
I
The Best Language
She became interested in Greek at high school. “It just seemed to me the best language,” she told the Times Magazine in 2013. “Best” is a large word for a translator. A translator spends a working life carrying one language into another and learns, line by line, what each will and will not hold. One thing Greek holds that English mostly does not is a working grammar for the moment when speech fails.
Greek tragedy is full of cries: oimoi, otototoi, pheu, aiai, oi. They are written like words and the dictionaries list them, but they behave like sounds. Lars Nordgren’s Greek Interjections (2015) is a study of them. Its reviewer in the Bryn Mawr Classical Review, Coulter H. George, describes nearly seventeen hundred instances, from the plays and fragments of the three tragedians and Aristophanes, and an analysis in which a cry can head a short phrase of its own: a dative “me,” for the one the cry falls on; a nominative for the one who cries it; a genitive for what it is cried about. George wondered whether the cry is always the head, since those words can also stand without one. Either way, the grammar is there.
English has its cries too, and enjoys them: oh, ah, ow, argh, and, as of Thursday, yikes. It even has one with grammar in it, “woe is me,” but the grammar is fossilised: it survives as a fixed phrase that nobody now builds on. Greek built one of its cries of grief the same way: oimoi is oi and moi, “ah” and “me,” and the standard lexicon glosses it “woe’s me.” There the grammar is alive: the endings still work, and a playwright can set a cry at the head of a line and build the line out from it.
That is one practical meaning of “best” for a translator of tragedy. Greek has a grammar for the scream, and English mostly has an interruption. The obvious English for oimoi is “alas,” and “alas” is a sigh where the Greek has a scream. Once a cry has been turned into an English word, it is no longer a cry. It is a word about one.
Sophocles’ Electra poses this problem for long stretches of the play. Its heroine has lost her father and is waiting for a brother who does not come, and Sophocles gives her long passages of lament. A translator who turns every cry into “alas” produces a heroine who sighs. A translator who leaves every cry in Greek produces a page an English audience can’t follow. Somewhere between the two lies a sound an English-speaking audience can hear as a scream without mistaking it for a word. The title Carson gave an essay on the play in 1996 names the problem: “Screaming in Translation.”
II
Bones of Sound
That essay, “Screaming in Translation: The Elektra of Sophokles,” appeared in Francis Dunn’s collection Sophocles’ “Electra” in Performance (1996). In 2001 her translation of the play appeared from Oxford. In its foreword, as Justin Maxwell quoted it in Rain Taxi that summer, she described its laments as a language “like listening to an X-ray,” and wrote: “Electra’s cries are just bones of sound.”
The play turns on a false report. An old servant arrives at the palace with news that Orestes, the brother Electra has waited years to see return, is dead. Orestes has sent him, and the news is a lie. In Sophocles the message takes half a line, and the other half is a remark on its own brevity:
τέθνηκ’ Ὀρέστης· ἐν βραχεῖ ξυνθεὶς λέγω.
Orestes is dead. I put it in brief.
Electra answers in the next line, 674:
οἲ ’γὼ τάλαιν’, ὄλωλα τῇδ’ ἐν ἡμέρᾳ.
The first three words are a cry with grammar in it: an interjection, the pronoun for “I,” and an adjective, “wretched,” agreeing with her. Then comes a comma and a sentence: “I am destroyed on this day.” In Carson’s English the line reads: “OI ’GO TALAINA. / My death begins now.”
She keeps the cry. She doesn’t translate it. She spells it in our letters, leaves it standing in capitals, and then the line crosses into English. In the edited Greek the crossing is a comma. In Carson it is a full stop and a line break, and the sound gets a line to itself. Translated, the first half of the line would force a choice between the noise and the grammar, between “Ah!” and “wretched me.” Left as it is, it keeps both, and the reader has to carry them. An explained scream is no longer a scream.
Cassandra, in Carson’s Agamemnon of 2009, gets the same treatment and then something more. Before Aeschylus’ prophetess says anything that can be construed, she screams, and Carson prints the scream as sound. As Emily Wilson quoted it in the Nation, it reads: “OTOTOI POPOI DA! / Apollo! / O!pollo! / Woepollo! / O!” The Greek cry is kept whole. Then the god’s name is broken open, and the breaking is English: Woepollo. Wilson, whose objections come in a moment, notes that nothing in the lines being translated corresponds to that wordplay, and that it is nonetheless in keeping with Aeschylus’ method. A few lines later Aeschylus makes the pun himself. Cassandra calls the god by name and then by a participle that sounds like it, Apollon, Apollon … apollōn emos, “Apollo, Apollo … my destroyer.” “Woepollo” carries his pun into English a few lines early.
III
Her Own Interpreter
An objection to this method comes from someone who understood it well, and she made it with care. Emily Wilson, a classicist and a translator of Greek, reviewed An Oresteia for the Nation in April 2009, and she granted the method what it had earned. Carson’s use of Greek for Electra’s laments, she wrote, “could be extremely effective onstage.” She gave a reason, too: a grief spoken in a language foreign to its audience “emphasizes Electra’s strangeness, as well as the solipsism of her stance.” The untranslated cry puts a woman outside the language of everyone around her, which is exactly where Sophocles’ Electra stands. She found Carson’s approach “more supple” in the Orestes, which she called “the most powerful play in her trilogy.”
Her reservations were about the Agamemnon, and they came in two weights. The lighter one concerns Cassandra. Wilson did not fault the cries or the play on Apollo’s name. What she found “somewhat less successful, to my ear” was what came after them. In Carson’s version Cassandra glosses her own prophecy, and it “seems jarring,” Wilson wrote, to make her “her own interpreter.” The heavier verdict concerns the play as a whole. The Agamemnon, she wrote, “does too much of the work for us” and “doesn’t leave alone enough” of what Carson had called, in a recent essay, “the deliberately untranslatable.”
That second verdict carries the weight because it is made in Carson’s own terms. Wilson measures the play against Carson’s own phrase and finds too much translated anyway. The Cassandra scene is the clearest instance. The scream is kept, and then a professor steps out from behind it to say what it meant.
That has to be granted. Cassandra’s curse is that no one believes her. In Aeschylus she is also, for long stretches, not understood. The chorus’s first answer to her scream is to ask why she raised it over Apollo at all. A Cassandra who annotates herself has been made intelligible, and her unintelligibility is half of what the scene is made of.
What survives is narrower than the claim that Carson leaves the cry alone. The cry is where she sets herself one of her hardest tests, and by Wilson’s measure she does not always pass it. In the Electra she keeps the sound and lets the sentence follow. In the Agamemnon the sentence comes back and explains the sound. The difference between the two is the one the method turns on: whether a translator can leave a thing standing without standing beside it.
IV
The Question Mark
The cry is sound past the edge of words. Its silent twin is the gap, the place where the papyrus is torn and the words are simply not there. One of Carson’s clearest statements about the gap is not in an essay but in a poem of 2000, “Essay on What I Think About Most,” from Men in the Off Hours. The poem takes a fragment of the Spartan poet Alkman that begins in the middle of things, with a verb that has no subject, and prints the missing subject as a bracketed question mark. It hears out the philologists, who would say the subject was named in verses now lost, turns on philology itself, and declines to guess. The question mark stays where it is.
It is a poem by a professor of classics, accusing her own discipline. In 2002, in the introduction to If Not, Winter, her translation of the fragments of Sappho, she took up the gap in prose, and was candid about what she was doing. Where the papyrus is damaged she used brackets, and she said plainly that they were not a record. They were “an aesthetic gesture toward the papyrological event rather than an accurate record of it.” Then, more simply: “Brackets are exciting.” Brackets, she wrote, “imply a free space of imaginal adventure.”
The book is full of them. Its title comes from fragment 22, which, as Mendelsohn reproduces the page, opens with a column of single brackets, a word hanging from one or two of them (“work,” “face”), and then a line that stands clear: “if not, winter.” Among the fragments is the poem where the cry and the gap meet: Sappho’s fragment 31, in which the speaker watches the woman she loves laughing with a man and loses her speech. Carson’s third stanza begins “no: tongue breaks and thin / fire is racing under skin.” The breaking tongue is the cry happening inside the body. The poem comes down to us because the ancient critic we call Longinus quoted it in his treatise On the Sublime, and it breaks off where the quotation ends, in the middle of a sentence: “But all is to be dared, because even a person of poverty.” Carson gives that break no bracket. Words that survive only in someone else’s quotation, she explains, are “intentionally incomplete.” The cut is the quoter’s, and she leaves it as he made it.
The objection to this half of her work came early, from a reader who admired her. Reviewing If Not, Winter in the New York Review of Books in 2003, Daniel Mendelsohn wrote that her body of work was, “to my mind,” “the most distinguished, original, and successful adaptation and reconfiguration of classical models” of the past generation. He praised her fragment 31 for catching “the odd, stilted quality of the Greek.” His objection was to what he called the “sentimental ‘fragmentist’ approach,” the notion “that fragments are beautiful because they are incomplete.” Whatever the state of the papyri, he wrote, “we do know that she wrote whole poems, not fragments.” To make the gaps the point is to make Sappho “reflect our own preoccupations.” He objected, too, to what he called her “selective application of her principles of representation.” The line that gives the book its title, he pointed out, has brackets in the Greek, and she printed it without them. A name nearby, missing its first letters, she completed (rightly, he allowed) from elsewhere in the corpus. And where a fragment of two words says “you burn us,” she printed “you burn me.” She did it, he reports, to keep what she called its “fragile heat,” and he argued that it cost “what the words actually say.”
This has to be granted, and her own note grants part of it: not every gap is marked, since that “would render the page a blizzard of marks.” The gaps are philology’s. The brackets are hers, set where the scholars found the holes, and set selectively. The question mark in the Alkman poem is hers too, placed where the subject should be and left empty. Mendelsohn presses the debt further. Her introduction, he noted, dryly chides the ancient scholars who quoted Sappho only to make a point about grammar or vocabulary. He answers that this is how much of Sappho survived at all. Fragment 31 is among the survivors. Her delight in the gap depends on the discipline her poem accuses. She conducts the quarrel in that discipline’s own notation, and that is what it means to say her play is philology.
Each objection, taken on its own ground, holds, and their main charges pull in opposite directions. Wilson faults Carson for filling a gap she should have left open: the gloss after the cry. Mendelsohn faults her chiefly for the opposite, for making too much of what is missing, as if the shape of the loss were the poem. The two don’t line up perfectly, since Mendelsohn also catches her quietly filling gaps. But in the main, one says she explains too much, and the other that she makes too much of the silence. Carson works in the narrow strip between them, and two of her most serious critics stand at its edges.
V
Listening to the Wind
The question beneath all of this is simple to ask and has no answer yet: whether a cry can be carried from one language into another without being explained on the way. A cry translated becomes a word about a cry. A cry left standing becomes a foreign object in an English line, and the reader has to decide what to do with it. A gap filled is no longer a gap. A gap left open may come to be loved for itself. Carson’s work does not settle any of this. It is a long record of attempts, some of which, by her critics’ account, fail. The attempts are the work.
The two halves of her practice meet in punctuation. In the Electra she puts a full stop where the edited Greek has a comma, so the cry stands in a sentence of its own before English begins. In If Not, Winter she puts brackets where the papyrus gives out, so the absence stands on the page as a shape. Both marks say the same thing: something is here that is not being said in English. One holds a sound and the other holds a silence. Neither tells the reader what is inside it. Each objection falls on one of those two marks: one was explained by a gloss, the other asked to carry more than it holds.
On Thursday morning the Academy’s permanent secretary reached her by telephone, and was told that she was in Iceland and listening to the wind. Of everything in the day’s reports, the wind is the one thing that needed no translator. It reached the rest of the world the way much of Greek lyric has: in someone else’s sentence, with the sound itself left behind where it was made. Somewhere in the North Atlantic, a woman who once spelled Electra’s scream in our letters was listening to a sound with no words in it.
❧
Sources: Carson’s “Yikes,” via her agent, Nicole Aragi, her schooling and the Times Magazine line of 2013 from Alex Marshall, “Nobel Prize in Literature,” The New York Times, October 8, 2026. Carson’s own words from the Nobel Prize’s telephone interview; Ingrid Carlberg and Anders Olsson from the Swedish Academy’s press conference, both October 8, 2026, checked by ear. The Academy’s citation as announced. Lars Nordgren’s Greek Interjections (De Gruyter, 2015) as described by Coulter H. George in the Bryn Mawr Classical Review 2016.11.17; oimoi from Liddell, Scott and Jones. The Greek of Sophocles’ Electra and Aeschylus’ Agamemnon from the Perseus Digital Library; translations from the Greek are the editor’s. “Screaming in Translation: The Elektra of Sophokles,” in Francis M. Dunn, ed., Sophocles’ “Electra” in Performance (Metzler, 1996), pp. 5–11, per the University of Waterloo’s Carson bibliography. Carson, Sophocles: Electra (Oxford University Press, 2001), p. 76; the foreword as quoted by Justin Maxwell in Rain Taxi, Summer 2001. Emily Wilson, “Violent Grace,” The Nation, April 8, 2009, which is also the source for the cry in Carson’s An Oresteia (2009). “Essay on What I Think About Most,” from Men in the Off Hours (2000), is described, not quoted, from the text printed in The Boston Globe, September 21, 2011. Carson, If Not, Winter: Fragments of Sappho (Knopf, 2002), the introduction and fragment 31; fragments 22 and 38 as reproduced and reported by Daniel Mendelsohn, “In Search of Sappho,” The New York Review of Books, August 14, 2003.
Header and interior images generated with Gemini for this essay; neither depicts an actual place, object or person, and the header was extended at the foot and cleared of a generated signature. Drafted and edited with Claude Opus 5.5 (the configured model; the serving model on any turn may differ); copy reading by Gemini.