I have my whole PhD thesis split into two papers to be submitted to the same journal at the same time, and each of them needs to be cited by the other (circular citations). I can publish the first paper to arXiv and get the DOI and cite it in the second paper. But I also want the first paper to cite the second paper. But in order to cite the second paper in the first paper, the second paper needs to be published first, which then becomes the first paper and the same problem arises. I am thinking of this possibility of publishing the first paper in arXiv and get its DOI and then cite it in the second paper and upload the second paper to arXiv and then get its DOI and edit the first paper by adding the DOI of the second paper. But is it fine to do so? I am wondering if the journals would accept each paper citing the other in this way. Also, are there any other better ways having each paper citing the other?

  • It's not normal for papers to cite each other. Citations are for work that a paper builds on. Which of your papers builds on the other?
    – Bryan Krause
    Commented Sep 5, 2022 at 3:31
  • In my case both the papers in some sense depend on each other. They apparently can't be published as a single paper because it's a massive work. Commented Sep 5, 2022 at 3:37
  • 3
    What field is this? There are at least some where such circular references are fine, assuming the split into two papers is appropriate.
    – Anyon
    Commented Sep 5, 2022 at 4:06
  • It's in the field of robotics, and I am planning to send it in an Elsevier journal. My supervisor told me that it is possible to make circular citation, but I am wondering if this three-way method (publishing the first in preprint and then citing it in the second and then editing the first to include the citation to the second) is acceptable way of doing it. Commented Sep 5, 2022 at 4:13
  • 6
    @BryanKrause: In my world, circular references are perfectly commonplace, since papers discuss related work and researchers discuss their work in progress with each other, frequently leading to situations where two or more papers provide context for each other. In the most extreme case, two papers report on independent work leading to the same result, so they should of course cite each other. Commented Sep 5, 2022 at 16:23

3 Answers 3


Ask the journal. If the journal has an editor, ask the editor. You are not trying to circumvent the review process. You are not asking about your specific paper at this stage - this is a structural question about submissions.

Remember that the papers you submit will certainly not be published in their present form. There are always edits, refinements and so on. So it should be OK, in the form you are submitting in now, to say “Jacob (2022b) “My second paper”, submitted to…”.

  • If your papers are rejected, then you don’t need to worry about the format of the references they contain.
  • If your papers are accepted, the editor will tell you what to do about the references.
  • If one is accepted and the other is rejected, you and the editor between you can work out what to do about the references from the accepted one to the rejected one. Perhaps they can be omitted altogether.

Read the journal. Has it published interdependent papers before? If so, how did the references work? If not - you need to come up, at submission time, with a good argument as to why they should use this format for the first time.

  • Yes, the delay in publication due to review/revision works in your favor here. There will be time to update "temporary" citation statements.
    – Buffy
    Commented Sep 5, 2022 at 13:45

"But in order to cite the second paper in the first paper, the second paper needs to be published first"

In my opinion, that is simply incorrect. To cite a paper it should not be first published or peer reviewed. If you have it on arXiv you can simply cite the arXiv version. But even without it, you can cite it as an "unpublished manuscript".

If you are concerned about reviewers who will doubt the correctness of the results since they have not yet been peer reviewed, then you must first publish the first paper, without basing your results on paper 2 (but only citing it as a related work), and then only publish paper 2.

  • Thank you for your answer, and yes unpublished papers can be cited but that makes the citation untraceable. The intention in having circular citations in my papers is to ensure the reviewers of one paper can get to go through the other paper and if published the readers can get to trace the other (companion) paper, and when I said in order to cite the second paper in the first paper the second paper needs to be published, i was saying it in the context of my thesis. Commented Sep 6, 2022 at 4:42
  • 1
    Thanks, but your papers are on Arxiv, so I don't see how they're going to be "untraceable".
    – Dilworth
    Commented Sep 6, 2022 at 19:00

I agree with Martin Kochanski's answer, but wanted to add some thoughts. Now, I don't know the conventions in robotics, but I've been part of physics papers where we did exactly what you describe, i.e. updated arXiv preprints to have circular references before journal submission. Our intention was more to benefit preprint readers than the journal(s), however. The journal(s) can and should be notified of the related manuscript at time of submission. Cover letters are very useful for this. Note that this is true whether you include circular references in the submitted files or not, and whether you post preprint versions at all or not. It is also often possible to upload the related manuscript as supplementary information for reviewers, which is especially useful if the publication venue uses double blind referencing, you haven't posted preprints, you submit to different journals, or you have reason to expect the two papers are likely to be reviewed by different people.

  • Thank you for your answer. And yes i would be mentioning it in the cover letter. I was thinking of figuring out in advance whether the journal allows it or not, so that I could plan the submission accordingly rather than waiting for the first decision which may take some weeks. Commented Sep 6, 2022 at 4:43

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