I am writing a short paper in the area of combinatorics.

When the paper is complete, I would like to be able to submit it to arXiv.

The reasons that I would like to submit to arXiv are:

  1. To obtain a date and time stamp from a central authority so that I can prove the work is mine.
  2. To promote access to the paper under the Creative Commons Attribution-Noncommercial-ShareAlike license

Although I have created an account on arXiv, because I am not affiliated to any academic institution, it appears that it might be difficult to get the paper accepted by arXiv.

For example see,


The following information is also required for submission:

Institutional affiliation for the author(s) must be provided. Official report number(s) from the author(s) institution(s) must be provided.

Also the following blog article does not appear encouraging:

Does The Arxiv Blacklist Authors ?

Does anyone on mathoverflow have any advice on how to submit to arXiv (or a similar database) without an academic affiliation?

Thanks in advance for any advice.

  • 3
    $\begingroup$ I suggest that you also get one or two other people to check your work for both correctness and readability. $\endgroup$
    – Deane Yang
    Apr 12, 2010 at 12:27
  • 2
    $\begingroup$ One of my mentors,Dr.Melvyn Nathanson,is a HUGE advocate of publishing on Archive and he encourages everyone looking for a career in mathematics to begin publishing there.I could ask him about your question,Paul. $\endgroup$ Apr 12, 2010 at 20:24
  • 4
    $\begingroup$ @Deanne - that response is kind of off topic - wont't vote it down - but I won't vote it up either. Actually, once I get the time stamp sorted, I will probably just start a blog and try to build up a small following from zero. Thankfully one does not need an endorser to open an account with Wordpress. $\endgroup$ Apr 13, 2010 at 4:23
  • 2
    $\begingroup$ @Paul: So, what happened? $\endgroup$ Aug 27, 2010 at 13:19
  • 4
    $\begingroup$ @Paul: I agree with this criterion: put something on arXiv only if it is close to journal quality! $\endgroup$ Aug 28, 2010 at 12:51

6 Answers 6


I think you need to be endorsed first. See this link: http://arxiv.org/help/endorsement

  • 2
    $\begingroup$ Yes, that's right. I'm not sure what the best way of going about this is (I have received completely unsolicited requests for endorsement and have turned them down on the grounds that the rules say that I must personally know the author. In truth, if I were willing to vouch for the results of the manuscript to a reasonable degree, I would probably be willing to endorse it.) You could post a link to your paper on some math forum (I'm not sure MO is the right place for this; maybe sci.math.research?) and you might find someone to endorse it. $\endgroup$ Jan 25, 2010 at 4:06
  • $\begingroup$ @Douglas, thank you for your prompt answer. I knew about the requirement to be endorsed. The requirement to have academic affiliation seemed additional though. @Pete - thank you for the suggestion. I might resort to that - cross post to sci.math.research and MO with a summary of the paper and a link to a PDF, then see whether someone would be willing to endorse the paper for arXiv. But even then there is the requirement for academic affiliation. $\endgroup$ Jan 25, 2010 at 4:24
  • 3
    $\begingroup$ Yeah, I didn't need to be endorsed since I have an affiliation. I'm afraid, I can't endorse (not enough papers). you don't necessarily need to know the person, just check that their paper is not clap-trap: "You should know the person that you endorse or you should see the paper that the person intends to submit." $\endgroup$ Jan 25, 2010 at 4:53
  • 1
    $\begingroup$ @Paul: My guess is that you don't need academic affiliation if you have an endorsement. At least, like Douglas, I did have academic affiliation, and didn't need endorsement. Since endorsement seems to only be required without academic affiliation, it seems logical that you should not need to have both. $\endgroup$ Jan 25, 2010 at 6:33
  • 10
    $\begingroup$ Pete- from the arXiv help: "You should know the person that you endorse or you should see the paper that the person intends to submit." $\endgroup$
    – Ben Webster
    Jan 25, 2010 at 15:51

For your first goal (authoritative time stamp), check out this site. Although it may not be well known, it maintains a good audit trail and should be pretty hard to contest. It could also be useful for stamping work in progress, even if you are not ready to make it public.

  • $\begingroup$ Thanks +1 - would give you +10 if I could. That is a very useful looking service and might fit my needs very well. $\endgroup$ Apr 13, 2010 at 4:01

It seems to be possible to submit a paper without being its author. If I remember it right, on the second page of the submission form, you have a question "Are you an author of the paper?" and "No" is an acceptable answer. So, you may just find somebody who is a "registered author" (this is just a technical term describing a person's relationship to the arXiv, not his relationship to the paper being submitted) and ask him to do the submission for you (of course, he'll want to look at what he's going to submit first but the authorship will remain exclusively yours though his name will appear in the "submitted by" field).

I agree that this "academic affiliation" thing is somewhat annoying but it is, probably, some sort of a safeguard against completely random people posting some total junk on arXiv. I guess it can be overridden by the arXiv administrators on a case by case basis but my guess would also be that to do this, they should be first convinced that what you are going to post is a high quality stuff. So, before writing to them, it would be nice to have a few papers published somewhere and a few known people to certify that you aren't just one more crank, which creates some sort of catch 22 in your situation.

  • $\begingroup$ Thank you. That's a very considered answer. I agree a balance needs to be struck. Perhaps academic affiliation is no absolute guarantee of quality though. What I might end up doing is posting a summary of the paper here on MO with a link to a PDF somewhere. Then the reputation of MO can work its magic. If the consensus is that the paper <i>is</i> junk then I will just let it fade away. On the other hand, if the consensus is generally favorable, then probably someone will be willing to endorse and submit for me. $\endgroup$ Jan 25, 2010 at 6:41
  • 3
    $\begingroup$ @Paul, just make sure if you post on MO about your paper, you do so as a good question! We don't want questions of the form "What do you think of my paper", or any variations, thank you! :-) $\endgroup$ Jan 26, 2010 at 1:29
  • $\begingroup$ @Scott - OK understood - I will be sure to respect that. I was new to MO at that point and did not fully understand the ground rules. $\endgroup$ Apr 12, 2010 at 5:25
  • $\begingroup$ The academic affiliation appears to be optional: it may allow to bypass the need for endorsement and thus streamline the process. As for the authorship, I imagine that the intent was to allow researchers to delegate this kind of tedious submission to their subordinates. $\endgroup$ Aug 21, 2010 at 3:38

Official report number(s) from the author(s) institution(s) must be provided.

I have absolutely no idea what that phrase means. I have a few papers on the arXiv and have never knowingly provided that information. With the proviso that I'm not privy to the internal workings of the arXiv, I would recommend just trying to submit and seeing what it tells you to do - I deem it highly unlikely that you'll get a message saying "You have no affiliation, never darken our doors again!" but more likely "We notice that you have not provided an academic affiliation; therefore, in order for your article to be properly submitted, you need to do X, Y, and Z.".

There are, of course, other ways of getting timestamps and of making your work public. If you want to know whether or not it is of sufficient quality to be worth publishing, you should track down someone in the field who you could ask for an opinion.

(But if you do that, don't just send them the manuscript with a brief note saying "Please give me your opinion on the attached."! Write a specific letter to that person, preferably with a fair amount of flattery - don't overdo it - and ask a specific question. If you want to know "Has this been done before", you could ask that here, I think, but if you want to know "Is this decent quality work", then you should not ask it here.)

  • $\begingroup$ Andrew, thanks for the advice - much appreciated. $\endgroup$ Jan 25, 2010 at 11:03
  • 11
    $\begingroup$ I posted papers to the arXiv for the first time last month and just left the "Official report number(s) from the author(s) institution(s)" entry blank. Although at arxiv.org/help/submit it seems that it is required to enter some sort of report number, when you actually get there it seems to say to enter it only if it exists. Perhaps a better way to say it is that you should enter a official institution report number if and only if such a number exists. My guess is that they ask for this so that appropriate credit/sourcing/whatever is given to the institution. Maybe. $\endgroup$ Apr 11, 2010 at 11:20
  • $\begingroup$ @Philip - +1 for the report on your experince with the arXiv $\endgroup$ Apr 12, 2010 at 5:24
  • $\begingroup$ I've seen many institutions, especially in Europe, that produce preprints for their researchers. I believe that's what the report number refers to, a way to give proper credit the way you do it by adding the journal info. Of course, these kind of reports are remnants of a bygone era, but I bet you'll see this field for a while still. $\endgroup$ Aug 21, 2010 at 3:42

Normally you need an endorser for arXiv.org but that is not always necessary or sufficient. Try writing to the administrators if you are stuck. If all else fails and you need it archived with a datestamp use viXra.org which was set up for exactly this purpose.

  • $\begingroup$ Thanks - I did not know about viXra.org - so +1 for that alone. In fact, the datestamp is what I want most. One strategy would be to (1) get a datestamped PDF online with viXra.org, (2) use the online PDF to seek an endorser for arXiv.org. Another strategy would to be support viXra.org's principled support of open science by not bothering with step (2). viXra.org plus a trusted reputation system (like mathoverflow) layered on top would be pretty compelling in my view. Food for thought anyway. $\endgroup$ Apr 12, 2010 at 5:23
  • 9
    $\begingroup$ I would suggest against using viXra -- I don't know how widely known it is, but I think you should worry much more that associating with it gets you ignored as a crank than that your work is so brilliant that the mathematicians you send it to try to steal it. As Phil said in the thread you linked: "Perhaps viXra will catch on and show that an open submission policy is not something to fear. At first they will probably laugh at some of what it contains, but when people find themselves citing e-prints in viXra they may finally get the point." I think we are still in the "at first" stage. $\endgroup$
    – JBL
    Apr 12, 2010 at 13:06
  • 4
    $\begingroup$ @Paul, I'm not sure it what way your first comment is a response to me. I encourage you to continue your mathematical work -- I just think that you should avoid viXra if you don't want to be labeled (incorrectly) as a crank. $\endgroup$
    – JBL
    Apr 13, 2010 at 13:19
  • 4
    $\begingroup$ There is no reason why anyone should consider you a crank because you use viXra.org. There are many papers there that have been published in peer-reviewed journals. There are also many cranky papers on arXiv.org. People post on viXra.org because for some reason they can't or don't want to use arXiv.org, not because they are cranks. Some people use both. $\endgroup$
    – Phil Gibbs
    Apr 14, 2010 at 18:35
  • 1
    $\begingroup$ I never saw a high-quality article on viXra.org, maybe someone can give an example. I saw article on arXiv, though, which was so wrong that you can literally give it to your "real analysis" students and ask them as an exercise to find the mistake. $\endgroup$ Mar 14, 2015 at 15:16

Official Report Number tends to stand for internal numbers assigned by the University or Research Institution for Technical Reports. For example, ATT Bell Labs often published their internal research documents as Technical Reports, and the Computer Science department at M.I.T. also tends to publish internal research findings, often supported by government sources or sponsored-research funded by corportations, as "white papers" or technical reports.

MIT's library is pretty good at finding these things: http://libguides.mit.edu/techreports, and their webpage defines technical reports as

What is a technical report?

Technical reports:

  • Are written to convey new developments or final results of scientific and technical research.
  • Are usually funded by government departments or corporate bodies.
  • Deliver technical information to the funding organization.
  • Provide a forum for peer information exchange.
  • Are not easy to find

Also, a lot of research funded by the military or DARPA or the OSD in the United States has a Final Report as its ultimate end-result and the means of disseminating the findings and recommendations, rather than a peer-reviewed journal article.

It is a shame that have an educational e-mail address for affiliation gets you a bye for submitting to arxiv; perhaps even academics should need an endorsement prior to being allowed to commit an article to arxiv. The seed recommenders would have to have been planted earlier, anyway.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.