I am writing a paper that is based on previous work I have done that I have published as a pre-print. I am submitting it to a conference. However it is significantly improved and different. The submission process is double-blind do I have to cite the pre-print? The idea and contribution is mostly the same.

I am in Machine Learning / Deep Learning field of research.

  • 1
    Why not just update the preprint to be identical? Then it would be clear that there is no plagiarism. Jan 10, 2021 at 1:28
  • @AnonymousPhysicist excellent suggestion. Is it acceptable to add authors to the pre-print?
    – iordanis
    Jan 10, 2021 at 16:47
  • @AnonymousPhysicist I think this is the most fit answer. If you want to articulate it, in a post I am glad to accept it.
    – iordanis
    Jan 13, 2021 at 20:32
  • The contradictory notion of a "significantly improved and different" preprint is what gives rise to the diverging answers here. What is it? The manuscript submitted to the conference before peer review, after peer review, or some earlier draft? The ambiguity isn't your fault, the term is unclear.
    – henning
    Feb 9, 2021 at 10:01

4 Answers 4


Across most of the social sciences, the answer would be: No, the published manuscript should not cite its own pre-print.

The pre-print is rather considered a draft version that happens to have circulated prior to its acceptance at a journal, where it transforms into its final version.

This seems to be grounded in the paradox of the published preprint being regarded as a preliminary, non-refereed and thus fictitiously unpublished paper (e.g., a Google search for "preprints are unpublished" yields enough pages with confusing semantics of that kind).


A few days ago, I asked Springer about whether I need to cite my own paper's preprint, and I was told by an editor:

"Citation of preprint is not mandatory. But it is a good practice to do that. I suggest you to please cite the preprint. It usually helps in cross-linking of the paper".

I needed to figure out how to cite it, and found that when citing your own preprint which is going to be published, it is typically cited using the @unpublished entry if you are using BibTex.
As an example:

  author    = {AuthorName},
  title     = {The Paper Title},
  note      = {Preprint at \url{https://somerxiv.com/preprint/some/link/}},
  year      = {2023},
  doi       = {the/doi/url/if/available},
  publisher = {Name of the Rxiv},

Note the need to use \url so that no errors are thrown.

I cited it in the "Related Work" section of my paper, where I simply mentioned "The preprint of this paper is located at \cite{paperUniqueName}".

The topic of citing preprints in general is under some debate.


Yes, you should always cite your own related work just as you would that of any other person. To fail to do so leaves you open to a charge of self-plagiarism.

Worse, in a double blind review, the reviewers might claim you are plagiarizing if they know of the old work and can't connect it.

But I'm assuming that the new work extends the old and isn't just a submission of what is in the pre-print.

  • it is mostly a submission of the pre-print. Is this something allowed / acceptable?
    – iordanis
    Jan 9, 2021 at 22:59
  • 2
    This might be a question for the editor. "Mostly..." is hard to judge from a distance.
    – Buffy
    Jan 9, 2021 at 23:04
  • 5
    I've never seen a paper that cited its own preprint. Jan 10, 2021 at 1:29
  • 3
    I have, e.g., scholar.google.co.uk/…
    – user2768
    Feb 9, 2021 at 8:52
  • 2
    I have seen papers citing the preprint when there are significant differences between the two; e.g., "We claim that A=B; detailed computations can be found in the preprint/Arxiv version of this paper [12]". (Q: Why would one remove material from the preprint? A: Reviewers.) Feb 9, 2021 at 9:41

Yes. Otherwise, you could keep writing the same paper over and over again.

You must log in to answer this question.

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