We have a mistake in a published paper and we will send a Corrigendum to the journal. Concerning arXiv we have two options:

  1. Replace v1 by v2, where v2 has no longer the mistake, and state what we did in the arXiv meta data, adding a footnote in the arXiv-version, telling that v1 was wrong.
  2. Or: Also keep the mistake of v1 on arXiv visible and upload the corrigendum to arXiv too. (But then, what to do with the original v1-arXiv-preprint? Add there, that the result is corrected in a further publication?)

and a question:

  1. Should the very same authors appear (even though one has not worked for the correction but also did not participated in the mistake)?

Is there a canonical way to proceed?


1 Answer 1


#1 is best: use arXiv's "replace" option to upload a new v2 with the same arXiv: id, which has the corrections incorporated into the paper. The arXiv preprint has its own life, independent from that of the published paper, and this makes the preprint as useful and convenient as possible - people following the arXiv id or URL for the original preprint will get the corrected version automatically.

In the comments field, I would put something like "incorporates published corrigendum". If the corrigendum has its own DOI, I would include it in the comment as well. (And while you're at it, make sure you've included the DOI of the paper itself in the appropriate arXiv field.)

Keep the authors field unchanged. People don't really need to know which authors actually wrote the mistake or the correction, so long as all of them accept responsibility and approve the correction. (Unless there are special circumstances - e.g. the correction is itself a major scientific advance, or the "mistake" was something like data fraud and the guilty party needs to be outed, in which cases you could consider adding a note within the paper itself.)

Note that the original v1 will continue to be available for those who click it explicitly, so anyone who really wants to see the history or compare the versions can do so.

  • 3
    If the corrigendum has authors not belonging to the original authors, then they should probably be added as authors to the integrated arXiv version.
    – Arno
    Oct 4, 2018 at 12:39
  • 2
    If following this option, it's also good to explicitly point out in v2 where changes have been made (maybe in a footnote), to help minimize confusion among readers who may find differences between the published and arXiv versions --- you can't count on readers to notice the comments field on arXiv, especially once they've downloaded the pdf. Oct 4, 2018 at 13:33

You must log in to answer this question.

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