The arXiv version of my article has two citations. Now that my paper is published in a journal, I want to know if these citations will be transferred to the journal version? Should I withdraw my article from arXiv?

  • 12
    Don't withdraw your article from the Arxiv. It is meant to be an archive. Withdrawing something is generally taken to indicate the original was too wrong to be corrected (the other major case is a paper that has been subsumed by another).
    – Jessica B
    Commented Aug 17, 2015 at 10:16
  • 6
    Note you can't withdraw from arxiv once the article is public. You can place a notification on the article with the word "withdrawn" in it, but the article is still there for everyone to read and cite however they want.
    – user4512
    Commented Aug 17, 2015 at 13:52

2 Answers 2


No, the citations will not change just because the paper is published in a journal (however, if the citations are themselves in preprints, then contacting the authors to make them aware of the publication would make sense, and would probably make them either update the citations in the preprint or at least change them for the final publication).

However, what you can do is update the arXiv paper to add a mention of where the paper has been published.

  • Even if you mention this to Arxiv, google scholar won't give correct bibtex info. I am going to avoid Arxiv in the future. Your university should offer a local preprint service, those have worked fine for me. It seems like Arxiv just wants to self-promote to me. Contacting authors of each article that cites you does not seem feasible to me.
    – dorien
    Commented Jul 11, 2017 at 4:39
  • @dorien I don't see how a local preprint service would be better. Nor do I have any idea what you mean by arXiv self-promoting. Nor for that matter what bibtex info you mean. Commented Jul 11, 2017 at 4:44

I'm guessing that your concern is whether the citations from the two versions will be combined in a citation database. The answer is that it depends on the citation database, and whether it considers the possibility of something like arXiv preprints, which cannot be taken for granted. Google Scholar certainly does handle this case: it will detect similarity between the two versions and attempt to merge them (or you can do it manually if it doesn't detect it on its own). I don't know whether other databases such as ISI or PubMed do, however.

  • 7
    ISI doesn't consider arXiv versions at all.
    – Arno
    Commented Aug 17, 2015 at 10:59
  • 6
    Similarly, MathSciNet does not include anything from the arXiv; papers are indexed once the are published. Commented Aug 17, 2015 at 11:37
  • @OswaldVeblen: this is not quite true anymore, MathSciNet will eventually link arXiv references to the pubished version using "cf. MRXXXXX" (or attempt to, anyway). However I am not sure whether such links are counted in their citation counts; I believe they are not.
    – Tom Church
    Commented Aug 17, 2015 at 22:31
  • 2
    @OswaldVeblen, and that unfortunate omission in MathSciNet is why by now it is primarily useful for historical, rather than contemporary, searches and such. Commented Aug 17, 2015 at 22:31
  • 1
    Just puzzling over this same issue. It seems that as of early 2019 MathSciNet does count a cite to the arXiv version of my paper as it if was to the paper. They seem to be doing this via the MRXXXXX reference as Tom Church indicated. Commented Feb 3, 2019 at 23:02

You must log in to answer this question.

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