On arXiv one can license a preprint on one of the following (see also a note on arXiv on licenses):

For cases when it is fine to submit a paper on arXiv.org perpetual (see also Does publishing a paper on arXiv prevent me from submitting it to a non-open access journal?), when I can safely use CC-BY (Creative Commons Attribution license)?

Ideally, I would use it all the time, as it is the most permissible (=open) license, which requires attribution. (However, I used it only once, it a case where I don't intend to send it to a journal; so, except for that case, when one should (or shouldn't) be afraid of using CC-BY on arXiv?)

EDIT: Here the emphasis is on arXiv perpetual vs CC, not on choosing between CC licenses.


2 Answers 2


Using a CC-BY license on arxiv.org will be fine whenever you don't need to retain exclusive rights. CC-BY is a non-exclusive license, and you are free to grant as many non-exclusive licenses as you wish. However, you won't be able to grant a exclusive license thereafter, which is what most commercial publishers are looking for. The typical "transfer of copyright" would essentially be an exclusive license in these cases.

If you intend to publish the paper submitted to arxiv.org with an open access publisher, it should generally work with a CC-BY license. Most open access publishers require only a non-exclusive license to publish a paper, so there would be no conflict here. It may however not be possible to go to a publisher which requires either an exclusive license or a copyright transfer afterwards.

Interestingly, a copyright transfer may be less restrictive than an exclusive license in this case. In principle, it should be possible to transfer copyright even if a license have already been granted to someone else (that's how a company would buy software from another company that has already been licensed). So, why not try the following with your next paper:

  1. Submit it to arxiv under a CC-BY license.
  2. Submit to a commercial publisher, and add a statement like this to the copyright transfer agreement:

"This agreement is with the understanding that a non-revocable, non-exclusive license under the terms CC-BY has been granted to a arxiv.org."

I wonder how they'd react...

  • 2
    I wouldn't try using a CC license and then submitting unless you are seriously willing to find another journal if it doesn't work (since it may very well not work). Also, even non-profit publishers often ask for terms in compatible with some or all CC licenses. Commented Apr 30, 2013 at 20:31
  • 8
    There is at least one case of ACM dropping a paper from a proceedings replacing it with a one-line note linking to the arXiv copy in this case. See: r6.ca/blog/20110930T012533Z.html Commented Apr 30, 2013 at 22:18
  • @MikaelVejdemo-Johansson Thanks for the link. I can understand that they like public domain even less than CC-BY...
    – silvado
    Commented May 1, 2013 at 18:34
  • A "transfer of copyright" is completely different from an exclusive license, and I am not aware of any reason why you cannot perform a typical academic copyright after having made the document available under an open licence. BTW, I have performed steps 1 and 2 on every paper I've submitted and I only have ever had trouble with the ACM because they have an on-line only copyright transfer form. Commented Sep 20, 2013 at 18:13
  • 1
    Note that "transfer of copyright" is not even a thing in some jurisdictions (e.g. in Germany). If German researchers sign something like this for US publishers, afaik, the passage is void.
    – Raphael
    Commented Jan 28, 2017 at 9:21

There's no way to know except by checking with the journals you have in mind. In mathematics, almost all journals will be OK with the default arXiv license, but most of them have copyright terms that would be incompatible with the CC licenses. Some may make an exception for you, but I would guess that many commercial journals would not (and probably some non-commercial ones, too - for example, the ACM is not friendly towards this sort of thing). Because this would often involve an exception to posted policies, I don't think there's any way to find out systematically besides asking individually. They may also be reluctant to make any general statements that could be interpreted as a precedent.

So in mathematics, I would be wary of the CC options unless you are sure you will submit only to the minority of journals that allow this. (And, of course, you should never choose any license without agreement from all coauthors.)

  • I never saw a statement explicitly allowing using a CC license (except for case when a journal publishes on such license) Commented Apr 30, 2013 at 18:37
  • I've never seen it explicitly mentioned, but some copyright forms for free journals are compatible with CC BY (e.g., combinatorics.org/ojs/index.php/eljc/about/…). Aside from these cases, which you can determine by reading the copyright form carefully, or journals that use CC licenses themselves, you would have to ask for an exception. Commented Apr 30, 2013 at 20:27
  • I'd be somewhat surprised at a math journal overruling the decisions of their editorial board over copyright concerns. But at any rate, it's the acceptance that matters, not the publication. I'd be happy to list a paper as accepted on my CV in this situation. Commented Apr 30, 2013 at 21:01
  • I don't think it would be looked at as overruling the editorial board. The acceptance is always conditional on dealing with the copyright issues, and if the author is unable to do so then the acceptance is moot (and might be officially rescinded, depending on the publisher's policies, which would complicate listing it on a CV - at the very least the wording would have to be very careful). Furthermore, I'd bet many editorial boards would be unsympathetic. Commented Apr 30, 2013 at 23:50
  • 2
    The principle isn't that authors must be allowed to use CC BY licenses (personally I don't, as the arXiv's default license seems to cover the rights that I want), but rather that acceptance decisions should be made for mathematical reasons and not corporate reasons. As far as quality of theorem is concerned, everything after acceptance is irrelevant. Commented May 1, 2013 at 4:51

You must log in to answer this question.

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