Yet, speaking very generally, I have found that I often do a poem more harm through addition than subtraction. When I revise by grafting many new lines to the original, adding ornament to statements and mist to once-clear scenes, I usually ruin the poem rather than improve it. Redaction, on the other hand, is often better to clarify, direct, and reveal the poem. This is because I tend to reiterate sentiments or images when I first draft a poem, stating and restating ideas or moves in an effort to express them more perfectly. This stuttering, I have found, is just part of my generative process. When, later, I return to these poems to revise, I begin by choosing the clearest iteration, the swiftest image in the bunch; the rest, I dump.
But my poems are frequently over-fleshed by more than stammering attempts. Often whole segments of the poem are empty or dead. These vacant passages in the poem fall into the following three rough categories.
The Dead Head
Many of my poems don't actually get underway until the third line or the third stanza. I usually identify these false starts by their vagueness, commonness, or writhing lyric. If, in those early lines, I refer to anything meteorological, seasonal, or address any of the common abstractions (Time, Love, Home, etc), I am almost certainly clearing my throat while already occupying the stage.
The Empty Belly
Sometimes, after a rousing start, a poem will begin to putter about thoughtlessly. Poems that are three pages long almost certainly do this at one point or another; they begin to include the reader in their infatuations, flitting about with confidence, if not beauty. In my own poems, these passages identify themselves by a stiffening or slipping of tone and the sudden abundance of irrelevant anecdotes and analogies. If I, for example, describe the lint in the clothes dryer as the hair-clogs of angel dogs, it's time to pack it in.
Poems of the Empty Belly test the social contract between reader and writer, where the reader asks again and again, Where are we going? and the poet replies only, We're almost there.
The Sleeping Foot
The exit of a poem is easy to miss, and often the poet goes rushing past it. Half of revision for me is the process of finding the true end of the poem, whether it comes in the fourth line or the fortieth. Conclusions are probably most difficult because of the pressure of both the poet's and reader's anticipation. A poem that has an Empty Belly may be forgiven if the conclusion is sufficiently arresting. But without conclusion, the poem is purposeless. And the exasperated reader feels that they have been waved over by the poet only to be told, "Never mind."
And yet, conclusions are the most exasperating part of the poem, requiring both purpose and deftness, directness and flirtation. If I have to write new lines during my process of revision, it is most often to correct the conclusion. Often, though, the conclusion of a poem exists in some imperfect state within the draft of the poem. We only have to, as Heaney's old saw goes, dig to find it.
If one of my drafts concludes with either a bundle of pretty words or a stack of polemics, I know I have run by the conclusion, usually buried some lines above.
Finally, while I suggest that poets consider writing with an eraser, I also would caution against, 1.) cutting without close-reading, and 2.) overwriting files. Don't remove too many organs. The difference between a surgeon and a butcher is whether they pull out the spleen or the heart.