In answering this question, I wondered:

Many fields are now beginning to embrace preprint servers such as arXiv. Typically, authors upload a manuscript to the preprint server at the same time as they submit it to a journal. On the whole, this is seen as acceptable, and not as 'dual publication'.

On the other hand, submission of the same manuscript to two 'proper' journals simultaneously is (in all fields I know) regarded as serious misconduct.

Where, precisely, is the boundary between 'acceptable' and 'unacceptable' conduct here? What distinguishes a 'preprint server' from a 'journal'? At first glance, one might say 'peer review'. However, journals come in many forms, including 'open review' (where the submitted manuscript is available for anyone to read and comment on immediately), and 'predatory' (where any submission can be accepted without review if you pay the fee). Similarly, many recognised preprint servers exercise some basic quality control on submissions.

  • I guess the main principle is that there shouldn't be two archival versions of the same publication. Whether or not a preprint was subjected to basic quality control doesn't make any difference - it's not archival either way. Things that undergo open review but are not eventually accepted are also clearly not archival. And indeed papers in workshops that don't publish proceedings are generally non-archival. Publications in predatory journals sadly count as archival, as I understand it, so you wouldn't then be able to publish the same thing elsewhere. Commented Oct 15, 2020 at 10:34
  • 1
    @StuartGolodetz Doesn't this just translate the question into "what distinguishes archival"? And I don't know any good definition of that except "I know it when I see it". It probably also doesn't help that arXiv by name and description is an archive (of preprints). Personally if we are using semantics, I'd rather work with intent here; a preprint is intended as something before (pre-) the final publication (print).
    – mlk
    Commented Oct 15, 2020 at 11:26
  • @mlk Yes, on some level. Although I don't think the definition of archival is that unclear to be honest, at least in the field I work in (computer vision) - basically, if it's a journal paper, a conference paper, or a workshop paper that appears in published proceedings, that counts, whereas if it's a preprint, a rejected submission or a workshop paper that doesn't appear in published proceedings, it doesn't. The waters are muddied slightly by journals allowing you to publish extensions of conference papers with e.g. 30% extra material, but the rules are usually explicitly stated in that case. Commented Oct 15, 2020 at 21:11
  • To be honest, most of the time I've found it's fairly clear what the situation is, and on the rare occasions when it's not, a simple email to an organiser of the conference/journal to which you're submitting will usually clarify things. For example, we once had a paper that explicitly got accepted as a non-archival poster (not appearing in the proceedings), with an email saying that we could later submit it to another venue. Instead, we resubmitted it to the same venue a year later, but decided to explicitly clarify with them that that was equally ok (it was). Commented Oct 15, 2020 at 21:16
  • And yes, before anybody says re. my comment above, it's ill-defined what "30% extra material" means, but in practice, it means "whatever the journal editor is willing to accept as being 30% extra material" :) Some things get smoothed over by introducing a human into the loop, in practice. Commented Oct 15, 2020 at 21:18

2 Answers 2


Although there is conceivably that "archival" issue, involving the potential question of "which is the authoritative version?", I don't think that is the genuine point here.

So far as I can observe, the contemporary issues here are two: first, about "getting credit" twice for "one piece of work", and, second, about making referees (working for free) waste their time and energy.

In mathematics, in the U.S., decades ago, while it was nice to have more papers rather than fewer, it seemed to me that the real gauge of a person was not paper-count at all, but the aggregate of their work/ideas. The idea of worry about "self-plagiarism" did not make much sense then, and it still does disturb me, insofar as it seems to me reasonable to write/speak about one's ideas "endlessly", without having to promise to never repeat oneself. For that matter, some things are worth repeating, after all.

The not-wasting-referees'-time objection is more legitimate, to my mind.

In particular, then, in math, in the U.S., putting things on arXiv scores no publication-status-points (in math, currently), for many purposes. So it cannot possibly lead to double-counting in any score-keeping game. And there's no refereeing, so no waste of peoples' time.

Similarly, if one does a pay-to-publish publication, whoever's time it may or may not waste, it doesn't score any points status-wise, so there's no double-counting. (Unless one imagines that one's CV-readers don't know the differences among journals/venues.)

Back to the "archival" issue: conceivably traditional journals will maintain their documents longer than others...? Unclear.


Your question is based on a misconception. There is no such thing as an unacceptable use of a preprint server. You can upload anything you like there (assuming it is allowed by moderators), but no one will consider it as a form of publication, either dual or non-dual.

And yes, the issue is precisely the lack of peer review. And if there are journals that also don’t do peer review or do peer review that’s at a very low standard, then publishing in those journals also won’t get you very far in an academic career.

You must log in to answer this question.

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