Loading essays…
Craft essay · Revision
An editorial argument for treating revision as an act of attentive listening rather than correction.
By Poets of The Household4 min read
Mary Karr opens her teaching, in The Art of Memoir, with what she calls a "big message": Listen up. The instruction is for prose, but it transfers cleanly to poetry, and especially to revision. Many poets, on a first reading of their own draft, hear what they meant to write. Revision begins when they hear what is on the page.
The editor's bias here is toward revision as listening rather than revision as correction. Correction assumes the poem has a target and the draft is a near-miss. Listening assumes the poem may be saying something the writer did not plan to say, and that the draft is partly evidence of what the poem wants to be. The two postures lead to different edits.
The cleanest worked example in English is Elizabeth Bishop's villanelle "One Art," whose seventeen drafts survive in her papers at the Vassar College archive. The scholar Brett C. Millier reconstructed the sequence and the published facsimiles are widely reproduced; the discussion here follows that reconstruction.
Bishop knew from the first draft that she wanted a villanelle — the form was fixed. The first draft, headed "How to Lose Things" or "The Gift of Losing Things," is in prose-like blocks with rhymes scrawled in the margins. By the second draft the first refrain has settled into its final form: The art of losing isn't hard to master. But the second refrain — the rhyme for master — does not arrive. Disaster does not appear until the fifth draft. The parenthetical (Write it!) that breaks the speaker's composure in the final stanza does not appear until the eleventh.
A first reading of the published poem suggests a kind of cool mastery. A reading of the seventeen drafts in sequence suggests something different: a poet listening to a draft over and over until the poem told her where its real subject was. The cool surface is the late-arrival; the urgency under it is what the drafts kept disclosing. Many poets find this the more useful model — revision as a long act of attention, not a polishing pass.
Donald Hall, in essays collected over decades (Goatfoot, Milktongue, Twinbird, Poetry and Ambition), returned often to the idea that a draft has its own sounds the writer has not chosen — what he called the "milktongue" of vowel-music, the "goatfoot" of meter, the "twinbird" of resemblance. Hall's point is that a poet who reads the draft only for sense will miss most of what is on the page. The vowels are doing something. The line breaks are doing something. The grammar is leaning one way or another. The reviser's job, in his reading, is to notice what the draft is already doing and decide whether to amplify it, change it, or let it stand.
Take a draft from a month ago. Read it aloud three times: once for sense, once for vowel-music only (suppress meaning), once line-by-line for break-tension. Make three small edits, one from each reading. Sit a week. Keep one.
Oliver's chapters on revision and the writing life are the source for the permission-to-swerve reading the essay develops.
Reproduces the seventeen drafts of Bishop's villanelle as held at the Vassar archive; the source for the draft-by-draft reconstruction in this essay.
Karr's 'Listen up' framing transfers cleanly from memoir revision to poem revision; the chapter 'In Praise of Revision' is the closest sustained statement.
The collected Hall essays where 'goatfoot, milktongue, twinbird' is articulated as a way of listening to a draft's sound-life independent of its sense.
Mary Oliver, in A Poetry Handbook, makes a related point about revision in a different register: the act of revision is partly an act of permission. The first draft commits the poet to a subject; revision is where the poet sometimes discovers that the subject was a near-thing for the real subject, and that the poem is asking to swerve. Many poets find this kind of swerve hardest to grant on the second draft and easier to grant on the fifth — by which time the writer has heard the draft enough to know what it is doing without them.
The following are not rules. They are practices many editors will recognise; some will be useful and others will not, depending on the poem.
The framing has its limits. Some revision is correction — a wrong word, a wrong line break, a comma that breaks the music. Some revision is structural decision — cut the second stanza, move the ending to the beginning. Naming all revision as "listening" risks dressing up a craft act in a mystical register that does not fit every case. The editor's bias here is toward the language of listening for the kind of revision that asks the writer to change what the poem is about, and toward the language of decision for the kind that asks the writer to change a tool the poem is using.
A final note. Several editors report that the best revisions happen on a printed copy, with a pencil, away from the screen. The evidence for this is anecdotal but consistent. Many poets find that what the screen rewards (small word-level changes) is not what most drafts need (a willingness to hear the larger thing the poem is doing). Whether the medium of revision matters as much as those reports suggest is open, but worth testing.