Analysis · Book XXII and its afterlife

The Twelve

The homecoming ends with a mass killing, and then with something the poem spends fewer lines on than it spends on a meal: twelve enslaved women are made to clean up the bodies and are then hanged in the courtyard. This essay is about what the text says, what translators have done with it, and what Margaret Atwood built out of the gap.

01What the poem actually says

After the suitors are dead, Odysseus asks his old nurse Eurycleia which of the household women have been disloyal. Her answer, in Book XXII, is an audit: there are fifty women in the house, and “twelve in all have misbehaved” — meaning they have slept with the suitors. Odysseus orders them brought in, made to carry out the corpses and scrub the hall, and then killed with swords.

And then the poem does something worth stopping on. The order is changed, and it is not changed by Odysseus. Telemachus — the boy the first four books were spent making sympathetic — objects that he will not give these women a “clean death”, because of how they treated him and his mother, and hangs them from a ship's cable instead. The manner of death is a status judgement, made by the youngest man in the house, and the poem lets it stand.

Immediately afterwards the goatherd Melanthius is mutilated and killed. Read together, the two punishments are a grammar: the disloyal free-born suitors get spears in a fair-ish fight, the enslaved get something else, and the poem's own vocabulary of clean and unclean death does the sorting.

02Who these women are

The single most clarifying fact is one that most English readers miss because translations blur it: these are not employees. They are dmōai — enslaved women, war captives or house-born, who could not refuse a suitor any more than they could refuse a master. The poem does not raise the question of consent, and its silence is the thing every modern retelling picks at.

The house's loyal slaves are rewarded by name — the swineherd Eumaeus and the cowherd Philoetius — so the poem is perfectly capable of treating slaves as moral agents when they choose the winning side. The twelve are judged by the same standard applied to people who had no ability to choose.

Figure 1 · Who is killed, and on what grounds

The suitorsfree-born, could have leftkilled in open fightMelanthiusenslaved; actively armed themmutilated, then killedThe twelve womenenslaved; could not refusehanged — no “clean death”Eumaeus & Philoetiusenslaved; chose the winning siderewarded, named, freedthe same status, opposite outcomes — so the poem can see slaves as agents when it wants to
The bottom box is the argument. If enslavement excused nothing for the twelve, it should have earned nothing for the two — and the poem rewards the two lavishly.

03What translation does to them

This scene is the best available demonstration that translation is interpretation. The Greek identifies the women plainly — the vocabulary is “women” and “slaves”. English versions have gone in both directions, and the popular story about it is only half right.

VersionWhat it does with the women
Butler, 1900 (this site)calls them simply women — but quietly drops Homer's detail about how long their feet went on twitching
Murray, Loeb 1919renders the noun literally, keeping their enslaved status visible in English
Mid-century and late-century verse translationsseveral add contemptuous epithets the Greek does not supply — the abuse is the translator's
Wilson, 2017restores “slaves” where others softened it, but a reviewer in BMCR notes she also writes “girls” where the Greek says women — so this version, too, makes a choice

The honest conclusion is not that older translators were villains and newer ones are innocent. It is that there is no neutral English for this passage, and every translator's solution tells you what they thought the scene was about.

04Atwood's answer

The Penelopiad (2005), written for Canongate's myth series, takes the two silences — Penelope's and the maids' — and gives them voices. Penelope narrates from the underworld, long dead, dryly reviewing her own legend. Interleaved with her account, the twelve return as a chorus, and the formal joke is that they never get to speak in the same register twice: they arrive as a skipping rhyme, a sea shanty, a ballad, an idyll, a mock-heroic drama, an anthropology lecture, a courtroom trial, a lament, a love song, an envoi. Their forms are all borrowed and all popular — the modes available to people who do not own the literary tradition they are trapped inside.

Two things that structure does, which criticism has been arguing about ever since. It reopens the question of guilt: if the twelve could not refuse, then the hanging is not justice but tidying, and the epic's happy ending has a body count it never audits. And it refuses to let Penelope off — her account keeps circling what she did and did not know about the women in her own house. The most common scholarly objection is that even here the twelve remain a device rather than persons: they get a chorus, not a story. Whether that is a failure of the book or an exact description of their position is, deliberately, left open.

05Why the scene matters to the rest of the poem

It matters because the Odyssey is otherwise the great poem of hospitality — of how you treat the powerless stranger at your door. Its monsters are defined by eating guests; its moral centre is a swineherd who feeds a beggar; its hero spends the second half of the book disguised as exactly the sort of person the household can mistreat without consequence. And then it ends with twelve people who had no power at all being killed for what happened to them.

You do not have to resolve that to read the poem. But it is the place where the Odyssey's own ethics push back hardest against its ending, and it explains why so much modern rewriting starts at exactly this point in the text.

06Sources and copyright

  • Homer, Odyssey XXII, in Butler's 1900 public-domain translation, published here in full; A. T. Murray's 1919 Loeb (also public domain) consulted for comparison
  • Margaret Atwood, The Penelopiad (Canongate, 2005) — described and discussed only; no text from it is reproduced here, as it remains in copyright
  • Corinne Pache's review of the Wilson translation, Bryn Mawr Classical Review 2018.10.58, on the maids' vocabulary
  • Scholarship on slavery in the Homeric household, and on the critical reception of the Penelopiad

Research pass 2026-08-24: 21 claims confirmed against two or more sources, 12 single-source, 3 disputed. One frequently cited article could not be verified in any database and has been left out rather than cited on trust.

← All analysis essays