The Endless Edit
Say you write a poem, and are pretty pleased with it. Perhaps you tinker with it for a while, adding this, deleting that, till you're satisfied, or you leave it sit and go on to other things. Some time passes, maybe two or three months, and you come back to the poem; you still like it but see something else you think would be a small improvement and you make another change, and maybe a few weeks later another. And so it goes on for a year or more. Is there a time to stop this endless tinkering?
My first response to this oft-heard question is to quote Paul Valery: A poem is never finished, only abandoned. Perhaps you know the poem is done when you can make no more changes to it—or have become uninterested in doing so. Perhaps the poem is done when it has sat there untouched for a long enough time to have gathered a little dust. You pick it up, dust it off, can't see anything more you want to do with it. It's done.
But there are other factors involved. And there are reasonable limits to the revision process.
One important factor in deciding to stop tinkering with a poem is time—that is, your own movement through time. If you rediscover a poem you wrote years ago, you could revise it again. But over the years, you have changed: you are no longer the same person you were back then, and (hopefully) your writing has improved and changed, as well. This presents you with a choice between tinkering with the poem to bring it into your own present-time style; or to abandon the poem, leave it unchanged, and if the topic still intrigues you, write a new poem in your current voice or style.
When I read a poet's collected works, and their style and writing quality have not changed, it sets off alarm bells: there's something wrong. People are supposed to change over time, acquiring experience, knowledge, and (hopefully) wisdom. If their art remains entirely static, there must be something wrong; one has to wonder if they have been static, too.
Another reasonable limit to the amount of tinkering you want to do is that you might never stop. It might become obsessive, with nothing ever being declared done or finished, and new editions constantly being published of the same poems. I guess that, most of the time, I'd rather see a poet keep writing new poems, even if they're on the same (obsessive?) topics, then constantly rewriting the same material. Novelty and freshness can be a virtue.
The main thing I'd caution against in tinkering is the temptation to overdo it. There is a reasonable limit to how much tinkering you should let yourself do; it will vary from poet to poet, but you need to watch out for crossing that obsessive/compulsive line. Part of learning to become a better writer lies in training the awareness of when to stop: this is good enough, it's not going to get any better, so just stop revising it. At the point, the determination becomes: do I release this poem into the world or not, rather than, do I continue to revise it or not.
I've seen lots of decent second or third drafts of poems get killed by over-revision. All the life and breath goes out of the poems, even as they become so polished that in some circles they'd be lauded as examples of technical perfection and mastery.
If the poem's subject matter is very compelling to you as the writer, you might make several attempts at the poem, till you get closer to where the poem wants to be. I have several poems where I was dissatisfied with the first attempt, but rather than beating my head against the wall trying to stubbornly revise something that was fighting back, I put aside those second (or fifth) revisions and set out to write the poem all over again. I started from scratch each time, with the same triggering image or memory or vision or event in mind, setting out on a completely fresh attempt.
Sometimes it's best not to fight too hard, but to let it go and start over.
Again, just to be clear, this is not revision: it's a whole new poem that starts from the same vision. The abandoned earlier versions are not drafts, not actually revisions, but separate poems; as separate poems, they succeed or fail on their own merits, rather than as drafts of something unfinished.
You could, after you produce a version that satisfies you, go back to the previous attempts and pull out and use a remarkable turn of phrase, or line, or image, that seems to have some life in it, still, and incorporate it into the new poem. But don't overdo that, either; too much old stuff shoe-horned into the new poem will turn the poem into a contraption, which is yet one more way to kill the life and breath in it.
There are lots of times when continuing to face a problem head-on only makes things worse, and sometimes one needs to approach things from an oblique angle, from the side. We might think of it as the aesthetic geometry of triangulation: finding a new angle of attack, when the old angle just isn't producing results.