The scenario: a question has been asked "does A=B?". An answer is given that is wrong. And the questioner has accepted it. My questions:

  • This is math so there is (for many and this particular question) a definite right or wrong answer. The difficulty could be a misuse of definitions, or a typo in calculation that goes astray, or any number of things. But how do you communicate this to the answerer in a non-confrontational way.

    The 'rational' way would be to say "No, that is incorrect, here is the correct answer.". But that is inflammatory even in private, and stackexchange is pretty public. But to say 'Here is an alternate answer', or "Here is a better way to word it, by removing your 'not'", is either annoyingly postmodern or aggressively passive-aggressive.

  • it seems strange, in some instances, that the questioner, who is often asking out of (self-proclaimed ignorance) is the one expected to judge correctness of the answer. An approved answer tends to stop voting on other answers, even if the approved answer is wrong.

Do you have any thoughts/suggestions on either of these points?

  • 7
    $\begingroup$ Well, I once gave a wrong answer that was accepted, and another user pointed out that it was wrong in a way that was straightforward yet non-confrontational (or I found it so). If you want to see the exchange, it's here: math.stackexchange.com/questions/7130/… $\endgroup$ – Mike Spivey Mar 15 '11 at 16:16
  • $\begingroup$ that sounds good (another reason for 'relaxed' wording is that sometimes one's correction is wrong, in detail or overall). $\endgroup$ – Mitch Mar 15 '11 at 16:32
  • 1
    $\begingroup$ @Mitch: Which question is it? $\endgroup$ – Bill Dubuque Mar 15 '11 at 18:02
  • 2
    $\begingroup$ Commenting on the user's answer whose answer has been accepted the point where he did wrong or where you get a counterexample would help... $\endgroup$ – user8250 Mar 15 '11 at 18:28
  • $\begingroup$ @Bill: 1 - You are involved (that I know of), 2 - I realize the difficulty of not having a concrete example, but since this is (partly) a question about emotion, I think being distant is better, 3 - it may turn out that I myself am in fact the factually wrong one, and then it's a matter of the emotion of my own personal shame...which I naturally want to avoid :) $\endgroup$ – Mitch Mar 15 '11 at 18:38
  • 6
    $\begingroup$ @Mitch: I have no idea which question you refer to. It's not too infrequent that wrong, or nonoptimal answers are accepted. Generally the best one can do is to give a correct answer and politely explain any subtleties that may exist that lie at the basis of the error. Although, theoretically, we should always be able to decide if a proof is correct or not, this often is clouded in many subjective details in practice, since one never writes out fully formal proofs. So it's always a judgment call "reading between the lines". Not to mention purely subjective pedagogical judgments. $\endgroup$ – Bill Dubuque Mar 15 '11 at 18:57

I don't see what's inflammatory about "this is wrong, here is the correct answer." If you are right and the accepted answer is wrong it is in everyone's interest to know this. If you want to be extra diplomatic you might say something like "I think you have overlooked the possibility that X..."

Questioners should ideally wait before accepting an answer to see how the voting and other feedback turns out, and also to elicit additional answers. In addition, there are plenty of ways to tell if an answer to a question you can't solve is right; for example, if you missed some fundamental insight that is clearly explained.

Keep in mind that this software was basically designed to answer questions like "how do I get JavaScript to do Y?" and in that situation it's easy to verify a correct answer (either the code works or it doesn't). Unfortunately mathematics is not (always) this simple.

  • 9
    $\begingroup$ It's truly unfortunate that many folks accept answers so quickly. Who knows how many insightful answers we've missed due to this. Is there any way we can change this, perhaps by disallowing early acception? $\endgroup$ – Bill Dubuque Mar 15 '11 at 21:08
  • 3
    $\begingroup$ We don't have the capability to change any of those kinds of details. If you want to suggest something like that take it up with meta.SO. $\endgroup$ – Qiaochu Yuan Mar 15 '11 at 21:15
  • 16
    $\begingroup$ I agree with this and would go so far as to say that it in mathematics, it is generally expected that you will point out others' mistakes when you find them. It is possible to phrase it a bit more politely, e.g.: "I believe / If I am not mistaken, this answer is not correct...." Note that this connotes a polite amount of lack of certitude on the part of the writer but taken literally it is consistent with utter confidence in one's answer. E.g., I believe that 2+2 = 4, and if I am not mistaken there are infinitely many prime numbers. $\endgroup$ – Pete L. Clark Mar 15 '11 at 23:36
  • 2
    $\begingroup$ Your penultimate sentence is very astute, Pete! $\endgroup$ – Georges Elencwajg Mar 25 '11 at 20:19
  • $\begingroup$ Knowingly that we have to wait for several minutes before accepting answer. $\endgroup$ – Adola Feb 20 '17 at 13:03

If you can afford the time and effort to do it, I would recommend posting your own answer without commenting on anyone else's post. It might be a good idea to justify your answer more extensively than you would normally consider necessary. Let the community judge who is right and who is wrong.


You must log in to answer this question.

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