Recently we have written a research paper and posted a preprint on the Arxiv to obtain some non-official peer review and to increase the quality of our paper. Till now we have not submitted this paper to any conference or journal (it is unpublished), but we are interested in sending it to a conference.

While testing our paper for plagiarism using free sites, we obtained results showing 83% plagiarism from the version we posted on the Arxiv. In other words, the software is saying that we plagiarised our own paper. We also tried pulling down our paper from the Arxiv to avoid this plagiarism, but it’s not possible. Does that mean our paper will get rejected by all conferences? What should we do?

  • 13
    I agree with Chris, but let me just say that in my opinion your arXiv posting is actually a helpful thing, because it makes clear to people where the material was taken from
    – Yemon Choi
    Apr 20, 2016 at 19:26
  • 7
    I can't imagine that anyone who runs a plagarism checker and gets this result won't also see what's being flagged up and realise it's from the same people. Assuming they don't prohibit papers having been published in arXiv before (some do...) they'll be used to this sort of thing happening. Apr 20, 2016 at 20:06
  • 1
    Presuming that you cite your preprint on Arxiv.
    – Bill Barth
    Apr 20, 2016 at 20:31
  • 1
    Take it easy. Plagiarism software is spectacularly bad in detecting plagiarism in maths papers. It flags totally innocent things like "Let $G$ be a graph." as plagiarism... I used to work at a university in Asia that implemented mandatory plagiarism checks on PhD theses, and my last PhD student there had problems with these checks, despite having original research results in the text... Apr 20, 2016 at 21:22
  • 7
    Quoting the arXiv paper is the best way to go — What? There is absolutely no reason to quote an earlier version of the same paper! If you want to safeguard against this sort of thing, the proper defense is the cite the earlier version(s) — arXiv preprint, technical report, PhD thesis, or whatever — preferably in a footnote on the title page.
    – JeffE
    Mar 13, 2017 at 23:25

3 Answers 3


But while testing plagiarism using free sites we found that results were showing 83% plagiarism from our paper only which is posted on Arxiv.

Two questions:

  1. Why did you test your paper for plagiarism in the first place? I have submitted about 50 papers and have never done this...partly because I never thought of it and partly because I know haven't committed plagiarism because I wrote the papers! Were you in doubt whether you plagiarized or not? I don't get it.

  2. If I understood you correctly, you seem concerned that a plagiarism detector found your paper on the Arxiv. But that's super weird – a paper cannot plagiarize itself, and you are not the first or ten-thousandth person to post their preprint on the Arxiv. Why would that be a problem?

Based on what you said, it is surprising that you only got 83% plagiarism. I guess the reason that it's not 100% is that you did make some nontrivial modifications in your paper (and neglected to mention that in your question). OK, so...I recommend that you post the updated version on the Arxiv. Then the plagiarism detectors will either figure out what's going on or report that you have plagiarized 100% from a preprint with the same authors and title...a puzzle that even the most harried conference editor should be able to sort out.

Finally, since the question was originally posted on Math Overflow, let me say that in mathematics it would look absurd for a paper to cite its own Arxiv preprint. If that is actually a halfway standard convention in other fields, I would be very interested to see examples.

  • 5
    It's a bit odd to put questions in the answer section. Mar 14, 2017 at 4:27
  • 5
    it is surprising that you only got 83% plagiarism – it may also be due to the equations. The plagiarism checker probably considers the Arxiv’s PDF, whose equations are different than their TeX source (at least to the plagiarism checker) or a PDF of the same paper generated with another style file.
    – Wrzlprmft
    Mar 14, 2017 at 10:12

The situation as I understand it:

  • You wrote a manuscript A and uploaded it to the arXiv. You do not intend to publish manuscript A in a journal or conference.

  • You later wrote a manuscript B. You intend to submit B to a conference. This manuscript B shows 83% overlap with manuscript A.

Thus, you are worried to have problems with publishing manuscript B.

While both are your work there is a potential risk somebody may think you are committing self-plagiarism. (As if you would you try to publish both A and B in a journal or conference there could be an actual problem.) To avoid this perception you should make the status of A clear.

You can do this via:

  • Adding a footnote to the front page of A saying that it is a draft/technical report/etc. and not intended for publication. You can also include this remark as "comment" in the meta data on arXiv.

  • Mentioning the draft A in your paper B, like: "A preliminary version of these results with more detailed proofs is available on arXiv." or whatever fits you context.

Doing either of those should suffice, but doing both might add some extra safety.

(Assuming there is no policy that results must not be circulated beforehand for the conference where you submit. But in mathematics this should virtually never be the case.)

  • 4
    My reading of the original question is that OP wrote manuscript A, posted it to the arXiv, and then some time later wants to submit a revision A' of the same paper to a conference.
    – JeffE
    Mar 13, 2017 at 23:27
  • 1
    @JeffE The answer seems to coincide with your comment? Just let B = A'
    – user2768
    Feb 20, 2019 at 12:08

What you should do:

  1. Not worry about it.
  2. When you upload your paper to a conference/journal, include the arXive paper in your citations.
  3. Write a cover letter to the conference chair / editor stating that this paper is based work that you have posted as a pre-print on arXive, and state if/what is additionally in your submission
  • 10
    No, citing the arXiv version would be quite absurd. A paper should not cite itself. Mar 13, 2017 at 19:44
  • 1
    Citing itself? I assume that the paper being submitted for peer review is not identical to the paper on arXive, but contains the same general idea. If it is the same manuscript, then of course skip #2 and go to #3.
    – kabZX
    Mar 13, 2017 at 21:09
  • 4
    The main point of arXiv is to put the paper there first and then submit that exact paper (modulo some formatting) to journals/conferences (and then to update the paper to reflect reviewer comments so that it matches the published paper). Mar 14, 2017 at 8:10

You must log in to answer this question.

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