Marking a multiple choice quiz, should you penalise wrong answers? It's a ploy that deters guessers - it keeps them honest, stops them bluffing or taking too many chances.
If a stanza of a poem doesn't do anything, do you just skip over it or do you penalize it? Perhaps it's there for other readers, not you. But suppose it affects your enjoyment of the rest of the poem? Maybe it's crass, sexist, derivative, etc. Would you ignore it then? Suppose instead of a bad or pointless stanza it's a questionable line, or word, or line-break?
Obviously, it depends. But on what? The proportion of the whole that's affected; the nature of the flaw. But also it depends on the type of work being read, the reader, and why they're reading. Some poets (Selima Hill?) tend to write uneven pieces. Others (Heaney?) don't. The perceived unevenness may be because the work is multi-style or polyphonic, the reader not equally at home with the various styles/voices. Some people (especially if they're judging) will judge a poem by its worst line. Others won't mind panning for gold, seeking rare and beautiful wonders. My impression is that
- Non-narrative, discontinuous poems are more likely to be read with a pick'n'mix approach
- If a line of a poem is much better than the rest, it's sometimes said to "be worth the admission fee alone", the rest of the poem excused
I don't see why readers should be more lenient with non-narrative pieces, and though I can understand why a good goal might justify going to an otherwise ordinary game, a poem's not a live event - it can be edited.
The reader's strategy can in time affect the poet's writing. If readers are going to ignore the bits they don't get, the poet's more likely to add more stanzas or more obscurity - after all, there's nothing to lose. I think the current fashion amongst frequent poetry readers is more indulgent than it was a decade or so ago. Perhaps the increasingly competent and voluminous output of creative writing students makes readers crave for something "a bit different".
Some poems contain a 2-D constellation of quotable fragments. What should fill the gaps between them? Dead wood or padding? Options include
- Nothing - compacted fragment and nothing else
- White space (to "let the images breath"; to provide space for fields to be generated between the poles)
- Absorbent text that attempts to magnify/refract the effect of the powerful fragments
- Text that attempts to provide continuity between the fragments (which may either amplify the fragments or mask their effect)
- Text that's part of a different thread.
Sometimes when a reader and I disagree over a poem, I've asked them about certain phrases only to find that the reader's ignored them rather than try (and fail) to incorporate them into their interpretation. I don't see any problem with a poem containing self-contradictions, but if the reader edits those contradictions away, something's surely wrong somewhere.
Sometimes a reader (for personal/professional reasons) wants to like the poetry. Obscure poetry gives such readers more scope to do this - nothing needs to be explained and anything in it might be vital to the poem as a whole, so nothing can safely be deleted. Once some time has been invested in a text, it's rather hard to dismiss it, though some readers may give up, using variations of the "there are only so many hours in the day" argument
- "If the poet's not going to bother editing their work, I don't have time to do the editing for them"
- "I don't trust the poet. If there are so many phrases that make little sense to me, maybe the effect of the other words is delusionary, luck, a mirage. Maybe the poet's trying to bluff me"
I'm not against "hopeful monsters" as such though I'm less tolerant of padding (in the form of words or space) than many other readers are. Surplus words and line-breaks can make me wary of the other words and line-breaks that at first sight seemed effective. I begin to lose trust in the poet. Before long, the implicit reader-poet contract isn't worth the paper it's written on, so I start reading another book instead.