If there is an answer to a question that I have asked on MathOverflow or MathStackExchange which helps me to prove a result appropriate for a published paper, what should I do:

(1) Cite the MO or MSE question as a bibliography item.

(2) Mention the username of the one who answered my question in the acknowledgements or where my question is cited.

(3) Email the one who answered my question, and ask him/her what he/she prefers, including adding him/her as a second author of the paper; the problem is that it is not always possible to find his/her email address, especially when the username is not the exact name of the user (as in my case..).

I guess that an answer to my current question depends on how significant the answer on MO or MSE is for the proof of the result in my paper; well, the answer in itself is quite 'trivial' either because a graduate student can understand it or because it relies on a known result in another field of mathematics (after being familiar with the known result, my question becomes easy).

Thank you very much.

  • 3
    $\begingroup$ Some related posts on Academia Stack Exchange (I am pretty sure more can be found): How to acknowledge a MathOverflow user? and How to acknowledge a MathOverflow user? $\endgroup$ Commented Jul 29, 2018 at 9:04
  • 1
    $\begingroup$ Thank you for the nice link. Although it seems to me plausible to just cite their answer as a bibliography item, I will contact them offering co-authorship, as I promised to Francois Ziegler. $\endgroup$
    – user237522
    Commented Jul 29, 2018 at 9:20
  • 12
    $\begingroup$ @MortezaAzad Actually I think the right place is Academia StackExchange rather than MathOverflow Meta (because it's more about academic ethic than about usage of MathOverflow). Of course, the question of where the question belongs, that belongs in MathOverflow Meta. 😊 $\endgroup$
    – Gro-Tsen
    Commented Jul 29, 2018 at 9:44
  • 10
    $\begingroup$ Just to add to my comment below, I am sometimes mentioned in the "Acknowledgments" for helping the paper in this or that way, and that is fine with me. I think this is the standard way if we ask a question from a colleague (in person, by email, at MO, etc.). Offering collaborationship should be reserved to cases where the contribution is really substantial or decisive. $\endgroup$
    – GH from MO
    Commented Jul 29, 2018 at 10:38
  • 1
    $\begingroup$ If the proof relies on a MathOverflow Post, it might even be good to repeat the whole thing- of course with giving full credit. Theoretically MathOverflow questions can get deleted and so a proof could become incomplete. $\endgroup$ Commented Jul 29, 2018 at 14:37
  • 1
    $\begingroup$ A flip side of the question is, what should you do if a proof from your MO answer was used almost word for word in a paper written by the question-asker (this happened to a colleague). $\endgroup$
    – user123627
    Commented Aug 7, 2018 at 17:02

4 Answers 4


When I was a graduate student, this situation actually happened to me in the inverse direction, when I asked a question on MathOverflow which one of the answerers (a MathOverflow celebrity) found interesting and deep enough to be expanded to a paper. So let me add some details of the procedure that my co-authors and I followed in that case as an example. Hopefully, it might be helpful. (You can find some other similar stories of joint research papers stemmed from MathOverflow on this Meta Post).

  • My original question received two really nice partial answers. The answers were using techniques that (up to appropriate modifications) could be applied to some other cases and could potentially provide a full characterization of the subject of the original post.

  • As I had no contact information (email/website address) attached to my MathOverflow profile at the time, one of the answerers (who eventually became the corresponding author of the joint paper) left a comment in my post asking me to email him.

  • I reached out to that colleague and he brought up the idea of writing a joint paper with him and the other answerer of the post which I cheerfully accepted. He also provided a primary draft of the paper (containing a brief introduction, the answers and some additional results of him) which we circulated amongst us until a nearly final version was prepared.

  • In the paper, the corresponding MathOverflow question is properly cited and it is also mentioned that the original asker of that question is one of the authors. Later, a link to the paper was added to the MathOverflow post as well.

As a final remark, I believe that the proper citing (i.e. The link to the post + The contributor's real name) of any idea that you see on MathOverflow and use in your work is ethically essential no matter if it is substantial or merely a source of inspiration.

It is also highly recommended to reach out to the askers/answerers in order to inform them that you are going to use their contribution to MathOverflow in your work as well as asking about their real name if they are posting under the pseudonym here on this site.

It also might be considered inappropriate if you don't offer a co-authorship to them in your ongoing project. As part of the academic etiquette, it is always good to give enough credit to your colleagues. You may leave it to them to decide to collaborate on the paper or not. Note that most of the good researchers actually make the right decision in such cases depending on the degree of essentiality of their contribution to your project.

  • 3
    $\begingroup$ Thank you very much for your answer! Before asking the above question, I intended to cite (+acknowledge) the answer to my question, but was not sure about co-authorship. Now I see that it is better to contact the answerer and let he/she decide if they think they should be co-authors. $\endgroup$
    – user237522
    Commented Jul 29, 2018 at 10:57
  • $\begingroup$ @user237522 You are welcome! MathOverflow could be such an enjoyable academic forum if we, the users, commit ourselves to observe some basic contribution principles and try to always be friendly and helpful! Good luck with your project my friend! :-) $\endgroup$ Commented Jul 29, 2018 at 11:03
  • $\begingroup$ I agree with what you say about MathOverflow. Thank you very much for your kind words! $\endgroup$
    – user237522
    Commented Jul 29, 2018 at 11:15
  • $\begingroup$ @MortezaAzad, probably it's what you meant, but I would even strengthen "could be" to "is"! $\endgroup$
    – LSpice
    Commented Aug 3, 2018 at 17:58

On my opinion, you may begin with option (3), and if the person who answered the question opts out of co-authorship, then do (1), and (optional) (2).

I already have some experience with this (people cite my MO answers as regular references, and I cite people who answered my questions). Journals never objected these references. Personally, I consider contributions to MO as regular publications, and never agree to be a co-author on the sole basis that I answered an MO question. On the other hand, this can lead to a further collaboration.



I believe you should always offer co-authorship; it’s up to the other guy to decline if s/he deems the contribution too small to justify it. Anything looks trivial once proved — that is our curse — and even e.g. connecting a question to “a known result in another field” can be a crucial contribution.

(If it’s so trivial that you can decide it’s negligible, then you shouldn’t have asked it here :-)

If they seem hard to identify, you can always address them here using the @ mechanism.

  • $\begingroup$ Thanks for the answer. I do not think that their answer is negligible, and I intended (at least) to add it as a bibliography item. Anyway, I will contact them via @ and let them decide about co-authorship. $\endgroup$
    – user237522
    Commented Jul 29, 2018 at 8:59
  • 21
    $\begingroup$ I don't suppose you offer co-authorship every time you use someone else's published result (whether in a peer-reviewed journal or in a preprint, e.g., on the arXiv): why should it be different for a MathOverflow answer? I don't see any reason to treat MO posts as different from any form of publication: acknowledge and cite properly what they did, yes, ask how to best do that certainly, but offer co-authorship only when the idea is a central part and more than a tool in the paper being written. $\endgroup$
    – Gro-Tsen
    Commented Jul 29, 2018 at 9:42
  • 3
    $\begingroup$ @Gro-Tsen I agree that this is not MO or internet-specific at all. The key point is: because they asked (so links about acknowledging existing old posts don’t apply). In my experience, once you ask there is an implied contract that anything stemming from the resulting conversation is potential joint work. If your paper proves Fermat and the other guy just taught you Schur’s lemma, trust that s/he will decline (or suggest splitting off the joint part into its own paper). $\endgroup$ Commented Jul 29, 2018 at 9:57
  • 9
    $\begingroup$ @Gro-Tsen: There is a difference between finding the answer in published material (including MO), and asking a specific question and getting an answer to it. $\endgroup$
    – GH from MO
    Commented Jul 29, 2018 at 10:03
  • $\begingroup$ Thank you very much all of you! In my case, the answer is a tool in the paper being written; however, since they helped me in finding that tool, I will ask them to decide whether they want co-authorship or not. $\endgroup$
    – user237522
    Commented Jul 29, 2018 at 10:15
  • 4
    $\begingroup$ OK, I missed the fact that OP was talking about an answer to their question while researching that particular topic, and not just some answer they found on MathOverflow. In this case, I agree that this is a priori reason for joint publication. $\endgroup$
    – Gro-Tsen
    Commented Jul 29, 2018 at 11:40
  • 7
    $\begingroup$ @FrancoisZiegler I agree with your comment that this is not MO specific, but I don't believe that you should always offer co-authorship when you ask someone a question and they can answer it. It depends on how it fits into your paper---it could be something very minor---and even suggesting writing part as a joint paper (or writing an appendix) is not always appropriate in my mind. $\endgroup$
    – Kimball
    Commented Jul 29, 2018 at 11:59
  • $\begingroup$ @Gro-Tsen, I apologize for the misunderstanding; perhaps I should have emphasized that it was an answer to my question. Anyway, it is interesting and helpful to read the comments. $\endgroup$
    – user237522
    Commented Jul 29, 2018 at 12:28
  • 3
    $\begingroup$ @Kimball Maybe I was too “one size fits all”, and should have prefaced my answer with “when in doubt”. But then, the OP was in doubt. Most often the other guy will decline, saying “anyone could have done it” (mathematicians are good at that), but better err on the side of letting them be the judge. Unless it’s so obvious you dare not even suggest they’d like credit for it (i.e., you’re not in doubt). $\endgroup$ Commented Jul 29, 2018 at 13:15
  • 4
    $\begingroup$ @user237522 May I suggest that you unaccept my answer? Now that it’s both community wiki and maybe not universally agreed upon. (And, BTW, I do waive coauthorship if you end up using this advice ;-) $\endgroup$ Commented Jul 29, 2018 at 13:34

As someone who is not a professional mathematician, I offer a different perspective.

I think (1), (2), and (3) are important, and all should be considered. However I would suggest a different approach to suggesting coauthorship (3). There is (in my view) a risk similar to resume padding when (3) is practiced too much. In order for MathOverflow to assume a proper place in the assessment of an individual's influence on the literature, I think it is important to keep a certain distinction between the idea of authoring journal papers (or even ArXived postings) and the idea of authored MathOverflow posts. The journaled papers suggest an idea (whether true or not) of thorough and ongoing scholarship, while MathOverflow postings are often quick (and sometimes thorough, but often not) responses to solve a specific issue.

Some of us find it easier to do this quick work, and accept credit accordingly. This is not to say coauthorship shouldn't be offered nor that it should not result from such a collaboration. It is to say that the degree of the contribution should be weighted accordingly. More specifically, if the coauthor is willing to help with the writing and additional research of the resulting article, then it makes sense and preserves (to me) the status and meaning implied by coauthorship of a refereed article. I see the danger of this meaning being altered in a bad way or diluted if (as I read the other answers) it is used as an automatic option, and the person who made the answer is considered as having done much of the research and scholarship, when in fact they haven't.

In short, consider and consult before doing (3), and do it with care and mindfulness of the profession. Of course (1) and (2) should always be followed.

Gerhard "Can Amateurs Still Receive Payment?" Paseman, 2018.07.29.

  • 1
    $\begingroup$ Thank you for your answer. I agree that one has to be careful. I hope/think that most mathematicians will be fair enough to decline co-authorship, when it is not required. (BTW, since I do not have enough experience in publishing papers, I will start with option (3), in order to be fair as much as I can). $\endgroup$
    – user237522
    Commented Jul 29, 2018 at 19:13

You must log in to answer this question.

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