Fact: When people ask questions on this site, often the question itself contains grammatical, stylistic, and other language errors.

Premise: The absence of pointing out the correction of all language mistakes gives a learner the illusion that the question is correct grammatically, structurally, stylistically etc. In particular, I find this confusing.

Question 1: Shouldn't people point out the non-focal, secondary language problems of a question within their answers to the question?

Question 2: What can I do to help me better understand what is acceptable and what is unacceptable language within the posted questions?

  • 1
    This question has been discussed here (see this question and this), but not since the site's very early days. In the event our practice has deviated substantially from the consensus expressed then, and I think it is well worth reopening. Commented Jun 14, 2014 at 18:57
  • I rolled back to the original version from choster's edit. I normally wouldn't do this, but I see a possibility that this question might be being interpreted incorrectly, and I'm just going to ask that any correction-to-the-question wait until after clarification. See my other comment asking for clarification. Commented Jun 15, 2014 at 0:55
  • Rushn - Can you please clarify your question? There are three possible ways of correcting non-focal language issues in a question: (1) edit the question directly, (2) provide a comment, (3) point out the correction in the answer. Can you confirm that your question is primarily directed at #3? In other words, you did not specifically ask why questions themselves are not proofed/edited/corrected more. Now you might like #1 or #2, but in this case for this question on meta.ell.se, you're focusing on why people don't put additional helpful comments in answers, right? Commented Jun 15, 2014 at 1:20
  • 5
    This gives a learner an illusion the question is grammatically, structurally, stylistically, etc. correct. There's a logic error there, in my mind. This is a learner's site; most questions are asked by people who are NOT fluent English speakers, and perhaps have only rudimentary English skills. No one should be looking at questions from that demographic and assuming the questions are syntactically and grammatically correct.
    – J.R. Mod
    Commented Jun 15, 2014 at 2:28
  • This gives a learner an illusion the question is grammatically, structurally, stylistically, etc. correct. In addition to J.R.'s comment, I doubt the basic validity of this premise. Note the OP is fairly new to this site. The scenario in which this would occur is some combination of someone not Internet savvy, new to this site, mistaking this site for an English blog or tutorial, and a Question that's borderline correct English. Similarly, language learner posts to bulletin boards contain responses that primarily answer the main question and only occasionally address other issues. Commented Jun 15, 2014 at 6:40
  • @CoolHandLouis "why people don't put additional helpful comments in answers" is correct. By not providing such comments you confuse the other learners who read a question,see " outrageous errors"as Hellion put it, and then wonder,what is up with that. This is what comes to my mind: I think a question is badly mangled (in my opinion),but no one points to it; may be the question is fine and it just me,who thinks the question is weird. So I start to question everything I had already learned.
    – Rushn
    Commented Jun 15, 2014 at 14:08
  • @!Rushn, I have edited your question based on my request for clarification and your response. Part of the change was to help focus it on making it more personally helpful to you while retaining the general question. I think this will be more productive for you. Please review my edit now to make sure it's precisely in line with your concerns. You can make further edits as needed using the "edit" button under the question. You can review my changes by clicking on "edited X time ago" at the end of the question, above my name, which is to the left of your name. Thanks - Commented Jun 15, 2014 at 16:10
  • 1
    @Rushn It may be worth your time looking through this meta post, specifically noting that we insert a space after commas. Precisely the situation that we're discussing - editing without highlighting issues - has occurred. However, see below for reasons on why we don't always edit every post for every grammatical error.
    – jimsug
    Commented Jun 15, 2014 at 22:17

6 Answers 6


I think there may be some fundamental big-picture misunderstandings here about what Stack Exchange is, what it isn't, and how it works. To some extent this is unavoidable.

My best advice for new users is to just enjoy participating in the Question and Answer side of things. It can take several months of active question asking/answering and then some participation on meta to get a feel for why things are the way they are.

Comment about Premise: I think you must be talking about a very limited set of questions which stick on the edge of good English. While it's plausible that some people who are less Internet savvy might be confused by an almost-correctly worded question, I don't think they would jump to conclude that everything about the question is perfect English. If this were a student handbook of questions and answers, one would surely think that the questions should be well-formed. Also, blogs and articles that are posed in question-answer format will undoubtedly have nearly-perfectly formed questions that the author can then succinctly answer.

But anyone who has encountered even a langauge learner message-board knows that a conversation between people tends to focus on some primary intent and other language may be less than stellar. And it doesn't take long for new users to recognize that Question Askers don't have command of the English language. They have to focus on one thing at a time.

Semi-rhetorical question: Shouldn't people point out the other language problems of a question within their answers to the question?

The answer to this is, at least in part, to provide some info about the way stackexchange works. There are three ways we can provide assistance to correct ancillary, non-focal language issues in a question: (1) edit the question directly, (2) provide a comment, (3) point out the correction in the answer.

1. Questions can be directly edited.

Many people -- especially those new to stackexchange -- may not realize that questions are often directly edited and corrected.1

There are functional/logistical reasons that questions are not edited more-than-they-are. These corrections are typically only noticed by the original question author and higher-level members who are interested in quality control.1 Also, people do not get points for editing questions even though such edits can take a significant amount of effort and skill. This is by design for Stack Exchange. There are many reasons for exercising restraint in directly editing questions. And this can leave some questions with less-than-ideal language. The other answers in this thread have addressed this well.

1. To some extent, when one sees problems within questions, they may be overestimating a laissez faire attitude since they don't directly see the edits made to the questions.

2. Comments can provide suggestions to improve answers.

In some cases, a comment can assist in the improvement of a question. If the change is made and the comment becomes irrelevant then it should be deleted. (It may be deleted by a moderator as well.) In other cases, it's easier to simply edit the question itself if the problem is simple.

3. The Answers themselves can point out ancillary problems with questions.

This is your point. A problem with this is a level of diminishing returns. Long answers are not read. Answers that are not read don't get points. It's often a quick and to-the-point answer that gets points.

In other cases, even very simiple questions can have complex answers and take hours or days of research to answer well. The person providing the answer has a choice: address everything in a monolithic post or address a single issue very well. It's just a lot more fun to answer a single question well than it is to correct every little mistake. And we know, as Answer Givers, that those secondary mistakes will often be cleaned up by a direct edit to the question, making any corrections we suggested orphaned and no longer relevant.

4. What can I do to help me better understand what is acceptable and what is unacceptable language within the posted questions?

Focus on the primary question only and answers given. If something within a question doesn't make sense to you, you have the followig options: - Post a comment asking for clarification. - Take that portion of the question, and ask it again as another question! If you want, you can reword it slightly. Also, if it makes sense, you can put a cross-reference link back to the original question if that would help. - Don't try to make sense of everything. You may be trying too hard. Read good English (like BBC articles) and ask questions here regarding issues you encounter. Don't try to learn language from questions posted by language learners. You'll run into too much noise, as you have discovered, and that can become overwhelming.


I am not at all unsympathetic to the opinion you express; as our SE Commissar Robert Cartaino put it before the site even entered public Beta, ‘What better way is there to learn English than to have someone correct your errors, especially in the supportive context of an "English Language Learners" site?’

There are, however, practical, methodological and in some cases personal considerations which argue against heavy editing of the sort you suggest. I feel sure that these have all fed our evident reluctance to perform much revision, so it may help shape the discussion to lay them out.

  1. On the practical side there is the sheer volume of work this would entail: a thorough revision can be quite as time-consuming as a satisfactory answer. We’re currently getting about 20 questions per day and producing about 35 answers. If, say, half of the questions submitted require extensive revision (I think that is an underestimate), you are talking about increasing the community’s output by the equivalent of about ten answers per day: nearly 30%.

    Now, most of the answering/editing work here is done by a very small cadre: at any given time, two-thirds to three-quarters of the answers come from about fifty people. I think you have to assume that those fifty people are already doing as much as their circumstances permit; for them to undertake extensive editing would almost inevitably cut into their answering time. But as it happens, the number of answers we produce—1.8 answers per question—is already the one metric in which this site falls significantly short of SE’s target. I think many of us would dislike seeing us fall further down that scale.

  2. On the methodological side there are some strong arguments against editing.

    • The unedited text is an index to the asker’s linguistic and English-language sophistication. I think most of us who answer here are careful to address questions at the level of learning the asker evinces—perhaps pitching our answers a little above that, so the learner has to stretch a little, but not so far above it that the learner gives up in despair. Heavy editing may eliminate important clues to what sort of answer is called for.

    • In many cases the question is so execrably expressed that readers have to guess at what meaning is intended, and inevitably many of those guesses are wrong. At least twice a week I find myself disagreeing with another answerer about what is actually being asked. That makes me very unwilling to change the question, for fear that I will misrepresent what the questioner wants. But if I leave the question untouched, the next reader has an opportunity to see what I did not and provide the answer called for.

    • Tacit correction in effect is the proofreading against which we have set our faces. We reject proofreading at bottom because it is the very least valuable form of instruction we can provide a learner. It violates a principle which I think is very generally held here: we give no answer and make no correction without an explanation which is not only valid for the current case but is to some extent generalizable to future cases (and we are careful to define how far that generalization can and cannot be extended). But where can you put the valued explanations called for by an edit?

  3. On the personal side ... I for one spend a great deal of my professional life rewriting other people’s bad English, and I have no desire to make a busman’s holiday of ELL! —That is probably not an aversion I share with many of my colleagues here. But most of us are, in one way or another, professional writers; and I think we do share a profound distaste for silencing or distorting the voices of others. Edit unto others as you would have them edit unto you.

    Me, I love the ‘bad English’ I read here. It is wholly unlike the self-complacent, pedestrian, corrupted bad English I read in my daily work: it is filled with effort and struggle and a desire to become better. It is filled with the drama of learning—a drama that gives me to understand much deeper than I ever have before the complexity and beauty of the language I serve. I have no desire to wrest these writers’ styles into conformity with mine.

I look forward to reading others’ thoughts on this subject. It is an important question and deserves close attention.

  • The question 25009 gave rise to this question of mine. I had to read it 5 times to understand. 2 people responded, but no one seemed to be concerned with its formatting. In my opinion,the person who asked that question should have been at least advised to format it properly. The advice in turn would have prompted the learner to figure out what was wrong with his question,promoting further learning.
    – Rushn
    Commented Jun 14, 2014 at 21:51
  • Or the question could have been rejected with a request to format it further to make it understandable to an average person.
    – Rushn
    Commented Jun 14, 2014 at 21:59
  • 4
    +1 for "The unedited text is an index to the asker’s linguistic and English-language sophistication." There are times I have been tempted to edit an O.P.'s question, but have decided not to, thinking, "If this question gets improved, no one will understand why such a 'fluent' speaker is asking such a fundamental question."
    – J.R. Mod
    Commented Jun 15, 2014 at 2:24
  • I think the biggest issue, when it comes to editing, is that we're not mind-readers. If all the evidence points to something, then proceed with caution, edit, then ask for confirmation.
    – jimsug
    Commented Jun 15, 2014 at 4:47
  • @jimsug Yes - ask, then edit to the answer. Commented Jun 15, 2014 at 10:34


I'll always edit formatting. But this isn't really an issue.

For an example, see Relative pronouns (where vs. which).



I'll try not to correct language in the body of the question. One of the reasons is straightforward and simple, as StoneyB (again) has pointed out: I usually have no way of knowing what the asker's intention is.

Furthermore I've rejected or improved several edits that try to change the phrasing of a question to make it "better", when it was completely comprehensible without the edits - these are usually a judgement on the style of a question, rather than grammatical corrections.

Grammatical corrections are fine, and style corrections are probably acceptable if they're made as part of an answer that answers the question (read: not an answer that explains the edit) and explained, or explained in a comment. Otherwise, all the user sees is that their post was edited, and has no idea why.


The idea behind graduated moderation privileges, on SE sites, I believe, is to distribute the workload. The reason moderators exist is to catch the exceptions, not to do all the work.

In addition to the 4 pro-tem moderators, there are 41 users with access to the 10k 2k moderator tools. I'm not sure how the tools they are afforded overlap, perhaps someone can enlighten me.

While I'm not sure whether the question:answer ratio is, as StoneyB suggests a symptom of our workload capacity.

It's certainly possible that it is an indication of the amount of work that we're able to take on, but I'm leaning toward the idea that questions are having answers accepted too quickly - for some reason, some users are less willing to answer questions that have accepted answers. I'm not privy to these kinds of statistics, and I'm not sure whether anyone else is, but I'd be interested to see whether questions receive fewer views from registered users after they have a question accepted.

I don't think that, if all 6,755 users were active and a quarter of them edited just one question a day, all ~7,647 of our questions would be fixed within a working week. Then it would just be up to the current core of fifty or so people to fix questions as they come in.

This isn't going to happen. I'd wager that less than one-sixth of our users have been active in the last four months, and that less than 300 of them have been active in the last month. At that rate, with each person editing one question a day, assuming there's no overlap, it'd take a little over a month, at the current rate, including new questions. A lot more work.

However, it is indisputable that attempting to enforce edits on every question would at the very best, be ignored by a sizable number of users, and at worst, would reduce output.

Explicit proofreading

The community, as a whole, has decided against explicit proofreading, I've raised the issue about implied proofreading, if such a classifier can be used. There are many examples of users correcting content in their answers, so it's clear that we correct it.

This, hopefully, provides a couple of benefits:

  • The asker is notified of the correction (in the form of a notification for the new answer);
  • The original text of the question gets indexed by search engines, and this may inadvertently direct search traffic to the question. While they are usually quite good at correcting input language, if it slips through, we attract traffic in this way. The corrected language in the answer hopefully attracts searches for the answer content.

The second is a very minor benefit, if it can be considered one at all.

Proofreading the question really only benefits the asker (which is the immediate goal), but is likely to be too localised to assist any future visitors.

Another possibility is that allowing proofreading will attract homework questions with little to no effort to solve it independently - such questions are usually closed in any case.


I feel obliged to respond to StoneyB's piece on the personal preferences toward correcting answers.

In the not-too-distant past, I would have leapt at the opportunity to correct someone else's English. However, these days, I have much more of a laissez-faire attitude.

I think that correction is welcome when it's asked for, and grammar policing when it isn't. While it's arguable that every person asking a question on ELL is asking for a correction of their post, the explicit question is my focus, and as long as the body of the question is comprehensible, I'll let it be. If there's something glaringly erroneous, as mentioned, I'll mention it in the post.

I spend most of my time dealing with language, and it's real, raw language with slips and "mistakes" and non-standard utterances. By the time I jump onto ELL to unwind (and that's the word with the meaning closest to what I'm after, so it's an irony-free label), I can deal with pretty much anything.

Also, I'm definitely going to edit questions to Australian English grammaticality, and I'm sure the majority of users are going to come from, or aspire to, either BrE or AmE.

(On a slightly unrelated topic, computers seem to think I speak New Zealandish.)

The short version:

  • Feel free to edit for formatting
  • Beware of editing for style and playing "guess what's in your head" with the asker
  • If you have the time and can guess what they're asking, edit away
  • Don't feel obliged to proofread
  • Computers think I speak New Zealandish
  • Your unrelated topic was great fun. They think I'm most likely a native speaker of (1) Norwegian (2) English (3) Swedish; if English, (1) Welsh (2) Australian or (3) Irish. If I could have remembered my AAVE I'd have really bollixed them! -but I just counted those as ungrammatical since I don't know the grammar. Commented Jun 14, 2014 at 22:36

Fact: Many questions contain errors and shortcomings.

Can't argue with that.

Premise: Some users may mistakenly assume uncorrected faults are actually valid usages.

That would be a foolish assumption. There's a limit to how much ELL can be expected to help fools.

Question 1: Should answers address peripheral shortcomings in question text?

Definitely not! Answers should be as clear and short as possible, and stick to the point.

Question 2: How can OP better understand what constitutes "unacceptable" language in question text?

I think this is the least important issue here, but it's the most complex...

  • Blatant errors should be corrected. The degree of agreement on this point is clearly shown here.
  • Clumsy (but comprehensible) phrasing should be retained (as a useful indicator of competence).
  • Incomprehensible text should be addressed by closevotes/comments requesting clarification
  • If you're still unsure what to do, either do nothing (probably best), or flag for moderator attention.

I would just mention that I know I'm not 100% consistent in applying these principles myself. And at the risk of straying from the point, I note that @Rushn (the OP here) has twice commented on apparent problems with question #25009: time or times (singular or plural). I see nothing about that question that "needs" to be changed. Personally I'd have suggested waiting time variance, but since one of the existing answerers says he's a statistician, I'm content to leave well alone there.


I agree completely with StoneyB's points, and mention only that when I see such a question I do try to edit it to eliminate the most outrageous errors, when it's clear that (a) I'm not changing the meaning, and (b) the errors aren't part of what's driving their question.

A couple of typos and some bad formatting, absolutely I will fix those with no hesitation. A badly mangled question whose intent can barely be discerned? I'll fix some punctuation and formatting to try and make it more intelligible, but as Stoney said, I don't want to remove cues or clues that others may pick up on better than I did. Errors from Russian speakers are often different than those of Italians or Chinese, for example, and may give an answerer a better basis to relate the answer given to the questioner's language of origin.


If we wanted to make sure all questions had exemplary grammar, then questions could not be posted until they were reviewed, edited, and approved. Such an approve-first process goes against the Stack Exchange model.

The solution? Remember who writes these questions, and do NOT assume that the questions are exemplary English worth emulating.

  • But we also learn how to properly ask a question. If the order of words in a question is out of whack, parts of speech are missing,that besides confusing us,sends us a message that it is OK to express yourself this way. The question #25009 on this site really threw me off.
    – Rushn
    Commented Jun 15, 2014 at 14:25

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .