After I put a paper on arXiv proving X, I found out someone had previously proved X already. What next?

Question

I recently put a paper on arXiv proving X. Shortly after, I received an email from L letting me know that my result was proven quite a while ago by G. I was not aware of the paper by G and there is no reference to it made in my arXiv upload.

This happened for two reasons, primarily it was my carelessness, which I have no excuse for. The other reason is that there was a more recent paper by B published after the paper by G that proves a special case of X that also does not cite G as the first person to prove it. I had seen the paper by B and not the paper by G, so I thought X was still unproven outside of B's special case.

At this point, I do not know what to do. I have already responded to L thanking them for letting me know that my result was already proven, and I intend to update my paper when I get the chance with a new abstract, introduction, conclusion, and citations that acknowledge the paper by G.

However, outside of those two things, I do not know what to do, so here are some questions:

1. After I update the paper, should I email L again letting them know that I resolved the issue?
2. Should I send an email to G apologizing for the lack of citation?
3. Should I let someone (either G or L) know that the paper by B also misses their paper?
4. Should I let B know about the paper by G?
5. Is there anything else I should do?

Answer 1

I second @Buffy that in principle there is nothing else to do, as long as you update the preprint. I'd add that depending on the circumstances, the necessary update may turn out to be quite minimal - for example, if I was notified that my result appears in, say, a letter from Euler to Bernoulli, then I would probably add just a line to the abstract, not rewrite the whole thing, and also add a thank-you to the person who notified me (I would probably ask them first). Of course, if the paper of G is more recent and accessible then Euler, or if your proof falls along the same lines as G's, or if the two proofs are actually completely different and it is of interest to compare them - then the rewrites would be more substantial. All this would also affect the publishability of your work, so I would encourage you to ask a colleague with experience in the field to advise if/where you should submit your paper.

I would also encourage you to consider the opportunities that present themselves in your situation. You came in contact with L, who presumably is also a researcher in this field, you are now aware of the paper of G and can explore the work based on it, and possibly contact B to share that knowledge. While it is very much dependent on the importance/usefulness/applicability of methods and results in your or other fields, and the people involved, this is an opportunity to consider inviting B, G, and L to a collaboration. While "I rediscovered your result" is probably not a strong incentive for G, if it is actually "I rediscovered this result, because it turns out it is very interesting in this other field", then the incentive to have a discussion is stronger, and ditto for B (although let me stress that if you reach out to B, make sure that the message cannot be misconstrued as an accusation that they didn't do their due diligence reviewing the literature). Who knows what you can discover together, instead of working on the same topic in isolation. Again, it may be helpful to ask a colleague familiar with both the field and your result for advice here.

Finally, let me say that rediscovering known (or better said: published) results happens to all of us at some stage. I personally was once in your current position, and also was either L or G in various other situations, and I am also just now starting a collaboration that emerged from exactly this kind of 'unintentional networking'.

TL;DR: correcting the preprint and thanking L is all that is generally expected in this situation. Consider using your new information about the prior work to approach L, B, or G for a possible collaboration.

Answer 2

There doesn't seem to be anything else to do. Don't beat yourself up over missing a paper. There is a lot out there and it can be hard to see it all.

However, your paper may not be publishable, given that the result has already appeared. You can possibly do some extension(s) that would be publishable, but old results aren't very "novel", to publishers.

In some fields, like math, a different proof of an existing result might be worth publishing, provided that it provides some new techniques or insights into a general class of problems.

Answer 3

This kind of thing happens reasonably often, and can occur due to reasonable oversights of the available literature on a topic. Sometimes a writer misses a relevant paper, and in the worst case, that paper renders their own work of minimal or no value. Now that you are aware of the existence of this other paper, I recommend you undertake the following process:

• Review and revise your existing paper to incorporate this new literature. At a minimum, this would entail telling your reader that the result was previously proven in G. You will need to make a decision on whether it is still worthwhile to include your own proof (is it a different approach?), whether to include the proof in G, and so on. Upload the revised version to arXiv to replace the present version.

• Once you have revised and uploaded your paper, you will need to decide whether it has any residual value given the existence of this previous paper. If you decide it doesn't have any residual value, you can withdraw your paper from arXiv. You can then treat it as a "working paper" on your own system that you might be able to come back to later if you have any further insight that would add value to the paper.

• In regard to the decision of whether to withdraw your paper, the main issue is to decide whether "it's ready". As pointed out in the comments below, rediscovery of an idea can be useful in science, even if it leads to some redundancy in papers. Nevertheless, you might decide that the paper can later be extended in another way or refocused to add some novel results. In the latter case you might decide to withdraw it for now and keep it as a working paper for later work.