Is it fair to change a question after an answer has been provided?

Question

When this question was originally asked, the poster was wondering what to say about Keats's "To Autumn", as in his opinion there wasn't a whole lot to the poem. I spent about two hours putting together a response showing that in fact, the OP had missed the depth of the poem. After my answer was posted, a few commentors claimed that the original question wasn't well-formulated. A moderator then changed the question to something else entirely. The revised question as formulated by the moderator is much more trivial; it asks for a list of literary devices used in the poem.

That is not the question I answered, and my answer has nothing to do with the literary devices Keats used. I would not have taken the time to frame my response if that had been the question.

It's one thing to close a question as off-topic if the question doesn't meet the standards of the site. But to change the question completely and then leave it open is grossly unfair to those who have put in the time and effort to answer the question as originally asked. Frankly, I won't waste my time crafting answers any more if I can't reliably assume that the question I'm answering will, in fact, be the question I'm answering. If questions can be changed so completely by moderators after the fact, then no answer is safe.

Answer 1

Here's the guidelines as I see them.

Avoid changing a question's meaning

To begin with, you're absolutely correct - once a question has been written, it can be edited (Stack Exchange gives users a lot of authority, including proposing edits to other people's posts), but its meaning should not be changed. For precisely the reasons you state: people have invested time and effort into posts, and you can't change the meaning of the question without invalidating existing answers. A question undergoing substantial revision finds its answers ceasing to make sense.

So this is true whether you are a moderator, a random community member, or the person posting the question to begin with. Once you've asked the question, that's what the question is. You can clarify, but you shouldn't change -- if you need something new answered, the correct thing to do is open up an additional question.

The problem is questions that are ambiguous to begin with.

What is an ambiguous question?

Some questions are ambiguous -- they could be interpreted in several different ways, or nobody's sure how to interpret them at all.

Some typical examples:

• Questions that ask more than one thing. Otherwise you get people answering one question and not the other, and then trying to vote and compare between two answers to two different questions.
• Questions that are overly broad or unclear. If a question isn't really clear or specific enough to be answered, than by definition, attempts to answer it are likely to confuse issues even further. Answers can't match the question if you don't know what the question is.
• Questions that participants are misunderstanding and answering the wrong thing. Maybe the OP thought he was being clear, but people keep latching on to some side detail and understanding that as the real question. Again, we have answers that are trying to match the question, but missing the point.
• Questions that get answers that don't seem to match the question. Even if the question looks kosher, if the answers you actually get seem not to be addressing the same question, then you've got a problem. Whatever the reason, if the answers don't match the question, there's a problem here to be resolved -- maybe it's one particular poster; maybe it's a larger misunderstanding.

And here's the thing: All those problems we mentioned with questions being changed after they're written? With ambiguous questions, they're baked in at the start. Because if questions and answers are mismatched, then you've got people spending effort on answers that aren't helpful, and people posting and getting answers that don't address their question, and good content getting deleted because eventually the question does get focused (or closed).

So in these cases, "Don't change the question's meaning" is irrelevant. The question doesn't have one single meaning. The question needs to be fixed -- and that might mean invalidating some answers.

Don't answer ambiguous questions

This follows from what we've seen up until now. An ambiguous question is not aided by a good answer. An ambiguous question is aided by closing it, putting in on hold, commenting for clarification, and doing whatever it takes to remove the ambiguity.

If your answer has been invalidated, re-post with an appropriate question

There are still hiccups, and sometimes good content gets written that's mismatched with its question (or where the question gets closed).

The simple solution for these cases is: ask a good question, that is matched to your answer. And post your answer. Stack Exchange encourages self-answered Q&A. This is an awesome way to keep your work, and make the site better.

---

In conclusion, let me say a few words about the particular case at hand.

In this case, we had an ambiguous question. It could be read as per the title, "What can you say when there's not a lot to say?" -- a meta-question about writing literary criticism, using two particular poems as examples. Or it could be a very specific question -- not "How would one effectively comment on writing in this specific context (of not having anything to say about it)?", but rather "What comments can be made about this specific poem?". To be blunt, I read the first draft as clearly being the first of the two -- not asking specifically about "To Autumn" at all; merely using it as an example of a type, of poems that have beautiful imagery but little depth.

As I said, your analysis is great, and good reading. That does not guarantee that it is an answer to the question.

Please consider: If you had, hypothetically, misinterpreted OP's question (as could be assumed, if his clarification really had rendered your answer was irrelevant), what would that actually mean? It would mean telling OP he can't have his question answered, can't get an answer to what he actually asked, because you misinterpreted it and spent effort on a misguided answer! That's not any good either.

Instead, we rely on community moderation (that is, everybody experienced and active enough participates in moderating the site), on discussing issues as they come up, and hopefully on a healthy dose of good will.

In this case, the original question was poorly framed, but you interpreted it basically correctly, and the question can be edited in a way that preserves your answer perfectly. In another case, a different solution could have been found -- such as splitting off to a new question of your own. And in all cases, we're trying to discuss issues respectfully, looking for whatever solution will leave the question, the content, the site, in the best condition possible. In the long run, that takes precedence over any one post.

I hope this answers your question. From my experience, ambiguously-worded questions are a significant issue on "soft" Stack Exchange sites. So it goes :)

Answer 2

In general, any edit to a question should maintain the spirit of the question. If it changes the question radically, it is not a good edit.

In this case, though, the question itself is very broad; basically it's asking "give me enough information to write a commentary, for my studies."
The edit made the question specific to a single poem, to keep it on-topic.

I don't think your answer was in vain. But it's true that it doesn't answer the much narrower question which devices are used.

I see that in the mean time, you've edited the title. I think this new title is a good one, within the spirit of the original question and not narrowing it down only to which devices are used.

Don't get discouraged; during a private beta, everything is still very much in flux, and questions and answers can get changed radically. When Pets Stack Exchange was in private beta, I asked why I should not give water to a wounded bird. It was turned into a question why not to give water or food to a wounded bird, but that was not my intention; I understood not to give them food. I just wondered why not to give them water. It was closed and reopened, too.
(For fun, check the edit history - it now says that it was my bird. It wasn't; it was a bird that I found outside. But general animal care was deemed off-topic, so somebody made an edit saying it was "my" bird. A little white lie put into my mouth!)

So - expect these big changes during private beta. The community is still making up its mind on many things. Feel free to edit and comment (as you did), just make sure the edits are in line with the OP's intentions.

Answer 3

A few things:

1. Your answer to that question is easily one of the best answers we've had on the site so far. We should do whatever we can to keep it.

2. The question (as originally posted) was a horrible question. However, the edits that I believe you were complaining about didn't improve the question. The only people who would want a list of "literary devices" used in a specific poem are people who just don't understand literature that well. That edit should not have been made.

3. You did the correct thing by re-editing the question to get at it's original meaning: asking about how to interpret "To Autumn". That's how Stack Exchange should work: if someone makes a bad edit, the edit should be improved or rolled back.

4. I hope you stick around; this site desperately needs people with your expertise.

Answer 4

Thanks to @VixenPopuli, @Standback, and @Hamlet for their thoughtful responses. (Great usernames all, btw.) I regret the somewhat belligerent tone of my question. I will accept one of the other answers eventually. But I wanted to respond to some of the points those answers have raised, and, being verbose (har har), at a length rather longer than would fit into comments.

First, I'm aware that questions asked on SE sites are subject to revision by any user. I've been on Stack Overflow for some four years now and have revised questions, had my questions revised, and in some cases, received downvotes after the OP edited his/her question in a way that made my answer seem patently off-the-mark. So I'm not proposing that a question, once asked, should be pickled in amber.

That said, there's a difference between an edit that clarifies, and one that changes. In this specific case, both the question and the question title were edited, the former by the OP and the latter by a moderator. The edits drastically changed the purport of the question, and I agree with @Hamlet that those edits were for the worse. The original title was:

What to say when there's not a lot to say?

And the question elaborated:

"To Autumn" is the 'other' kind of poem in my opinion. It's been described as "nearly perfect, but doesn't have a lot to say". If I'm not wrong, its only aim is to describe/praise Autumn. How would one effectively comment on Keats' writing in such a context? The poem is bursting with imagery, but how do you comment on it?

This was the meat of the question as originally asked. It strikes me as being a perfectly valid question. The OP says s/he could be wrong, but thinks, based on someone's earlier commentary, that "To Autumn" is essentially a pretty but empty poem, and asks how one could write about it. Note that the question is not about pretty but empty poems in general; the question is specifically, How would one effectively comment on Keats' writing in such a context?

Well, one effectively comments on Keats's writing by showing that the attributed context and the earlier commentary is wrong. That is exactly what I did. Thereafter, @Standback commented to the OP:

That's also why you got an answer that doesn't actually address your question, but only your example - because the question itself isn't very clear. Can you explain what it is you're trying to do, and what you are looking for in an answer?

The OP is being explicitly asked to revise the question because a moderator thinks the provided answer doesn't address it. However, on what basis is @Standback saying the answer "doesn't actually address the question"? That is not for the moderator to decide. If the OP thinks it doesn't address it, the OP should be free to ignore the answer, downvote it, or clarify the question. But the moderator, who after all has the power to close the question if it seems off-topic or unclear, should not decide for the asker whether a provided answer is adequate.

Subsequently, the OP "clarified" that what s/he was really after was some help in writing a commentary on "To Autumn" for his/her homework. Two points here:

• The edit does have the merit of making clear to everybody's satisfaction what was, to my reading, perfectly clear from the statement of the original question: How would one effectively comment on Keats' writing in such a context? The edit underlines the fact that the question was never intended to be a general question about a certain kind of poem, but a specific question about "To Autumn".
• Nevertheless, if the request to compile a list of commentary for homework help had been in the original question, I would not have answered it. I'm not in the business of doing someone's homework for them.

To my mind, this edit makes the question unsuitable for the site. If the OP is "looking for commentary on devices in 'To Autumn'", then the proper place to look is a library catalog, or Google. A list of bibliographic references constitutes basic research, and my understanding is that users are expected to do basic research toward their question before asking it.

Based on this edit, @Rand al'Thor changed the question title to reflect the new ask. (He—I assume "Randal" is a he, apologies if misgendering—also deleted the literary-analysis tag, for reasons that remain unclear to me.) As @Hamlet says,

The only people who would want a list of "literary devices" used in a specific poem are people who just don't understand literature that well. That edit should not have been made.

I take this to mean the edit to both the title (from the moderator) and the question text (from the OP). By this point, the title and intent of the question had changed so much, the answer I had given basically had nothing to do with them. I would have answered neither any question with that title, nor any question with that intent.

The reason I'm going over this ground (readily accessible from the history of edits) is that I think we need a lighter touch. All the changes to the question resulted from moderator interventions to what was a perfectly legitimate question to begin with. I disagree, obviously, with the sentiment that the question was ambiguous to begin with. The title was poorly-worded, but the original question was well-formulated and specific enough to warrant a well-formulated, specific answer. If a moderator believes a question is unclear or off-topic, the moderator is free to vote to close it. But those who answer a question in depth and detail do so because they found the question clear and on-topic; that should count for something.

TL;DR: It is not for moderators to say what a question means, or whether a given answer addresses it. That's between the OP and the answerer. I understand that in a closed beta, things are in flux and a close eye needs to be kept on them. Nevertheless, I suggest we use a lighter touch in demanding that questions and answers conform to our ideas of what they should be.