I'm working on a topic in engineering and submit a paper to a journal about 6 months ago (which I also posted on ArXiV). The paper was recently rejected. Since the original submission, I've made significant progress on the problem, and I've written that progress and additional developments, into another paper which I plan to submit to a conference at the end of the month. I had originally been planning to cite the earlier paper, but in fact a lot of what was in the original paper is made essentially obsolete by my new developments.

To try to summarize the situation: paper 1 does "X and Y", is on ArXiV, and was rejected by a journal. Paper 2 does "X++ and Z", and I'm planning to submit this to a conference soon. Content "Y" is not significant enough to be it's own paper, but contains some important information that needs to be referred to in order for "Z" to make sense.

Should I still try to publish the original paper somewhere? Should I combine that paper with the new developments and publish that (i.e., publish "X++, Y, and Z")? Or, should I just leave the original one standing on ArXiV and forget about getting it through peer review?

  • Why was it rejected?
    – Buffy
    Jan 8, 2021 at 22:36
  • 1
    The reviewers wanted to see some extra content about "W" (some simulations of sensitivity to certain inputs) which we had thought seemed out of scope, that some of the content was superfluous (which I agree with now), that the conclusions were not quite what the journal wants (this is pretty fair), and that we didn't make enough contribution over and above a paper we took inspiration from (which I think is inaccurate, our formulation is similar, but not identical, and our analysis and conclusions are much stronger).
    – RJTK
    Jan 8, 2021 at 22:46
  • 1
    Both options can be good. It really depends on the paper(s) and the preferences of yourself and your coauthors.
    – Anyon
    Jan 8, 2021 at 22:54
  • 3
    That is the path I'm leaning towards, it makes sense to put a lot of it together and definitely makes for a better paper. I think my main concern, which I didn't articulate in the original question, is if it's okay to leave the arxiv paper standing and have a lot of repetition from that version (e.g. the introduction and literature review will be very similar) in another published paper. i.e., is it bad to have "dangling" arxiv papers? It seems a bit unavoidable if I post to arxiv before getting it accepted, which is the point of arxiv in the first place right?
    – RJTK
    Jan 8, 2021 at 23:38
  • 3
    It's fine to leave dangling papers on arxiv. You should put a comment in the "Comments" metadata for both papers on arxiv. Jan 9, 2021 at 0:33

1 Answer 1


This question was asked a long time ago, and the OP got advice in the comments, but it remains on the unanswered queue. I will try to answer it.

First, the situation described here is similar to another academia.SE thread. It's also related to a MathOverflow thread I recently wrote an answer for.

My view is that, early in your career, every publication counts, so I advise anyone in a similar position to still try to get the first paper published. If the publication system was more efficient, or if you worked more slowly, you would not have subsumed your own paper. I don't see why you, the author, should be punished for working fast.

The literature is full of examples where papers in a series were published out of order. This happens when the first is rejected and resubmitted while the second is already under review, or sometimes even if the papers are accepted in order, they can be published out of order due to journal backlog times.

Unless I viewed the first paper as embarrassing, I'd probably do just what the OP originally suggested. Have the second cite the first, resubmit the first somewhere lower, and send the second to the conference. You could also have the first paper cite the second, to point out that the techniques developed have already found an application. I much prefer this over a dangling arXiv paper. You never know when some future researcher might need that "superfluous" content, and if it's an eternal preprint, they might be less confident to rely on it. If you see some quick ways to edit and improve the paper based on the referee report, then do that before resubmitting, but try to delay as little as possible. It's ok to have two different approaches to the same thing, and get both published. The new referees for the first paper will probably not even be aware that you wrote the second paper. Or, if you cite it, then they are aware of it in a positive way. In math, it would be very rare for a referee to reject a paper because the author had written a subsequent one generalizing the first. Having the arXiv timestamp helps you a lot in that regard.

You must log in to answer this question.

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