Sometimes, a mathematician dies suddenly, leaving behind very good mathematics that didn't make it through the publication pipeline. For example, it is possible they had a paper entirely ready to submit (maybe even already shared with their network, or on arxiv). Or they might even have submitted the paper and then died before the referee report came back (as referee reports seem to be taking longer and longer, this scenario is becoming increasingly likely, sadly). Rather than have important work hang around forever as unrefereed preprints, perhaps mathematical friends of the deceased would like to see their final work published. Assume that there is reason to believe the recently deceased actually wanted the work published (e.g., a preprint they shared with friends and discussed submitting).

What is the algorithm for getting the last paper of a recently deceased person published?

If the mathematician is a huge deal, someone might archive or publish their nachlass, as happened to Gauss. But that tends to happen more in other fields like philosophy, whereas I'm asking about submissions to normal math journals, where the paper would be reviewed by a professional mathematician and checked for correctness (and presumably someone, perhaps a student or co-author of the deceased, would make minor changes and corrections).

I know a few people have done this. For example, Georges Maltsiniotis has gotten a lot of Grothendieck's stuff published, including Pursuing Stacks. But I imagine some of this was contentious, and I imagine Maltsiniotis already gets too many emails about Grothendieck, so I figured I'd ask here instead of emailing him. A few examples of mathematicians who died suddenly in my field and left important mathematics behind are Bob Thomason, Jean-Louis Loday (I note that there are several preprints still on his webpage, and I believe his work with Vallette was finished after his death), Gaunce Lewis, Mark Steinberger, and Mark Mahowald (who, despite being 81 when he died, still had work in progress that was completed by his co-authors after his death) just to name a few.

I imagine someone has to get the rights to publish the paper. Do those normally pass to next of kin by default? What other steps are necessary that I'm not thinking of? Do people know of journals that have done this before?

Long term, I'd love to see a journal designed to publish papers of mathematicians who died before they could see their last works through the publication process. Depending on how much work was required to get the paper in shape, you can imagine partial credit going to a non-anonymous helper who corrected errors, rewrote proofs, added an introduction, etc. But I'm asking a simpler question now, about how to get it done in the current math publishing world. Because of the nature of the mathematical peer review process, I think this question is a better fit for MathOverflow than Academia.SE. A related but very different question was asked on MO last month.

  • 4
    $\begingroup$ Eric van Douwen died in 1987. He has about 40 papers on his sole name published in the period 1989-1993 (for the one I knew, "The automorphism group of 𝒫(ω)/fin need not be simple", there is explicit mention that this paper circulated years before his death, and that the process of finishing the paper and publishing it was done posthumously) — these papers are mostly short. $\endgroup$
    – YCor
    Commented Jun 21, 2022 at 16:52
  • 6
    $\begingroup$ I once finished a paper for a friend and colleague, but in this case he knew he was dying and so he sent me what he had and a sketch of where he wanted to go with it and I mostly did some minor editing. $\endgroup$ Commented Jun 21, 2022 at 17:24
  • $\begingroup$ There are a number of publications posthumously coauthored by Mark Mahowald (the most recent one published in 2020). I don't know anything about how they were handled. $\endgroup$
    – Dan Ramras
    Commented Jun 23, 2022 at 0:47
  • $\begingroup$ "Long term, I'd love to see a journal designed to publish papers of mathematicians who died before they could see their last works through the publication process." Mathematica Mortuorum. $\endgroup$ Commented Jun 23, 2022 at 1:07
  • 2
    $\begingroup$ In a case I was involved in, an author died just before the journal received the referees' reports. Since the reports were very positive and only requested minor changes, we asked one of the referees to make the changes and then published it with a footnote explaining the circumstances. I think we discussed it with colleagues of the deceased author first, but it was a long time ago and I don't recall the details. The journal leaves copyright with authors, so we didn't require a copyright transfer from a legal heir (though a lawyer could doubtless argue about our legal requirements). $\endgroup$ Commented Jun 24, 2022 at 0:46

5 Answers 5


The OP asks for a specific "step by step process to get a paper published by a deceased person who wrote the paper alone".

A statement on this was made a few years ago by COPE (Committee on Publication Ethics). This involved a case where the manuscript was submitted by the author, who passed away before it was accepted.

  1. The first step is to ascertain whether the author intended their work to be published in the present form. Ideally the author has had a chance to respond to referees, but one might make the case that submission to arXiv suffices.

  2. If this is ascertained, then one needs to obtain permission from the legal heirs of the deceased. Which documents are needed would depend on the journal, it should include at least a copyright license and for some journals also a statement of "no conflict of interest".

Point 2 should not pose an obstacle, point 1 is more problematic. Journals generally require a statement to the effect that "all authors must have read the paper and approve of it". It would be unusual for a journal to publish a manuscript posthumously if it was uncertain whether or not the author was ready for publication.

  • 3
    $\begingroup$ This is helpful. I'm going to hold off on accepting, hoping that others might add further steps in the algorithm that I'm not thinking of (e.g., if a lawyer has to be involved). I didn't know about the COPE statement and that helps a lot. I guess permission from legal heirs must be a somewhat formal document and possibly input from a lawyer will be needed there. Another point is that manuscripts of the sort I ask about might require minor edits or corrections, making (1) a bit stickier. $\endgroup$ Commented Jun 22, 2022 at 9:14
  • 1
    $\begingroup$ Nowadays, most papers that one might consider publishing exist in electronic form, so the topic of digital inheritance is relevant. Specifically, one would need to determine who has the rights to the paper; often this is easy, but not always. As for lawyers, a general principle (no matter what the issue is) is that almost never is there a legal requirement to get a lawyer involved, but it can be helpful if the situation is complicated or delicate. For example, in simple cases, permission from legal heirs need not be a "formal document." $\endgroup$ Commented Jun 22, 2022 at 12:01
  • 4
    $\begingroup$ (continued) Formalities would be required only if there is reason to believe that someone (e.g., the publisher) will be skeptical of who has the rights to the paper, or skeptical that the person who has the rights really gave permission. Even in such cases, what you'll most need is not a lawyer, but proof of who the rightful heir is (e.g., in the form of a will, which you can probably get from the probate court), and proof of the rightful heir's intentions (e.g., emails). $\endgroup$ Commented Jun 22, 2022 at 12:06
  • 1
    $\begingroup$ This decision concerns a very specific case of an author dying after the paper was accepted, but before paperwork was completed. The editor told the Forum that the author had seen the final accepted version. Hence the Forum agreed that it seems reasonable that the author should remain on the byline. This "hence" seems to imply that had the deceased not seen the final version, publication with his name would have been inappropriate. $\endgroup$
    – Kostya_I
    Commented Jun 22, 2022 at 20:06
  • 3
    $\begingroup$ @Kostya_I I disagree. I interpret the "hence" to mean that "the author had seen the final accepted version" was a sufficient condition, but this does not mean it's a necessary one. $\endgroup$ Commented Jun 23, 2022 at 4:11

This is delicate. First, someone who mentioned a result to a famous mathematician, and received some feedback, might decide to add the famous mathematician as a coauthor after their death, without giving that person a chance to protest. It is hard for journals to distinguish such situations from genuine collaborations.

Second, an author might decide to keep a manuscript in their drawer because they are not quite sure it is right. If it is then published after their death, it can lead to further contradictions. I was told that a story like that is behind the paper

Pontrjagin, L. Characteristic cycles. C. R. (Doklady) Acad. Sci. URSS (N. S.) 47, (1945). 242–245.

The reviewer (Hassler Whitney) wrote in https://mathscinet.ams.org/mathscinet/search/publdoc.html?r=1&pg1=MR&s1=13317&loc=fromrevtext that ``Both theorems contradict a statement of the reviewer [Proc. Nat. Acad. Sci. U. S. A. 26, 148–153 (1940), near the end of §4; MR0001338]. It is not easy to find who is wrong where.''

I was told (but have not verified this) that the mistake in Pontryagin's paper was due to reliance on a posthumously published paper by another mathematician. (Edit: As noted in the comments, this could not be E. Cartan.)

  • 3
    $\begingroup$ Thanks for the post. These are interesting stories, but I don't think they answer the question. As I wrote, we may assume the deceased person wrote the paper alone, and intended to submit it for publication. So it's neither of the situations you describe. Plus, I'm asking for a step by step process to get such a paper published. Anecdotes are interesting, but not the point of the question. Thanks again for your interest. $\endgroup$ Commented Jun 22, 2022 at 7:42
  • 1
    $\begingroup$ "... someone who mentioned a result to a famous mathematician, and received some feedback, might decide to add the famous mathematician as a coauthor after their death, without giving that person a chance to protest." And until now I thought nobody new would ever get Erdos number 1. $\endgroup$
    – KConrad
    Commented Jun 22, 2022 at 15:38
  • $\begingroup$ @KConrad At the time of Erdos's death, there were certainly quite a few papers in the pipeline with Erdos as a co-author. I remember a colleague joking that he would never publish as many papers as Erdos, even if we were to handicap Erdos by counting only Erdos's posthumously published papers. I'm not aware of anyone who added Erdos as a co-author illicitly, but I've certainly heard people joking about doing so. $\endgroup$ Commented Jun 22, 2022 at 22:13
  • 1
    $\begingroup$ @TimothyChow the "nobody new" I had in mind were people becoming a co-author now rather than around the time he passed away. On the general topic raised by the OP, the first example that I thought of was Kornblum's paper proving Dirichlet's theorem in $\mathbf F_p[x]$ (see eudml.org/doc/167539 and hsm.stackexchange.com/questions/11771/who-was-heinrich-kornblum). It appeared in 1919 under his name but was prepared by Landau since Kornblum had died in 1914 as a soldier during WWI. I am sure the way things were done legally 100 years ago offer no lessons for today. $\endgroup$
    – KConrad
    Commented Jun 22, 2022 at 23:48
  • 1
    $\begingroup$ Elie Cartan : 1869-1951. 1945 is not that posthumous. $\endgroup$
    – username
    Commented Jun 23, 2022 at 7:24

My student Daniel Singh sadly died in 2020, having left academia and never having published his thesis (which was completed in 2004). At the request of his family, I uploaded the thesis to the arxiv (https://arxiv.org/abs/2205.06875). I contacted the arxiv moderators about this, and their swift and efficient response suggests that this is not the first time that they have encountered this situation. I sent them a letter signed by Daniel's brother granting me the relevant permissions, then uploaded the thesis using my own arxiv account; it is possible to do this but still have Daniel listed as the sole author. Both for submitting another author's work, and for uploading PDF without LaTeX source, it is necessary to be granted permission in advance by the moderators.

  • $\begingroup$ This is helpful, thanks. And sorry to bring up potentially painful memories. Did you think about trying to get the thesis published in a peer-reviewed venue? People might be more likely to cite it and use it if it was peer reviewed. $\endgroup$ Commented Jun 23, 2022 at 15:00

There is a related point, although it is a bit more in line with the referenced related question, but it was too long for a comment:

Write your will! Do it right now, not just before you need it. Personally I fully expect to still be around for many more decades and I don't think my family will squabble much over what they will inherit. Still, I have some details written out for them, because right now I can think up the things that will need to be done in one afternoon, while piecing them together from my unsorted stacks of paperwork would take them weeks, if not longer.

So to get back to the question, as with any mathematician, unfinished papers will very likely be part of my legacy. So the will includes a provision for that, specifically to send the folder on my pc that normally includes all the current drafts and ideas to one of my more frequent co-authors, giving him the express permission to continue and publish whatever he deems salvageable, or to pass it on to someone better suited, as well to include any co-authors if warranted.

This may not avoid all the problems. But giving a written, broad permission should avoid any legal objections to publishing or questions about my intent, while at the same time, explicitly giving the task to someone I trust both mathematically and morally should hopefully reduce the chance of publication of mistakes or of misattributions.

Although it should be said that personally, I would rather risk having posthumous publications of questionable quality, than leaving potentially good results unpublished in order to shield my legacy.

  • 1
    $\begingroup$ Of course, I agree with this answer. Before setting out on a rather dangerous extended hiking trip in the Himalayas, I wrote a will including the password to my Google Drive, an exact location where my work in progress could be found, and one of my co-authors who I wanted my family to email with the files therein. Funnily enough, I didn't mention it to that co-author, so one day he's in for a big surprise! Rather than hoping he'll work out every idea, I'd be happy with just having the ideas and partial solutions posted somewhere online, like my advisor Mark Hovey's open problem list. $\endgroup$ Commented Jun 22, 2022 at 15:34

To state the obvious:


The only way to know for sure that the author intended a paper for publication in a particular venue is if they submitted it there, or explicitly given you their blessing or "power of attorney". Even if it is the case, the paper must be camera-ready or accepted after (trivial) minor revisions; substantial edits are unthinkable. Many times I rewrote proofs by my co-authors in a way that felt more natural to me, but they hated it, or vice versa. Do you want to do that to a dead person? Not to mention the possibility of actually introducing errors.

We are in the age when publishing in a journal and dissemination have little to do with each other. Most of the reasons to formally publish a paper do not apply to a dead author:

  • they don't care about promotions and grants;
  • as argued above, the paper cannot be improved based on reports.

The only remaining reason is that the journal vouches for correctness. But in practice, referees often don't check proofs carefully, there are plenty of wrong papers published, and in the end of the day it is the reputation of the author that gives most credibility to the work, not formal publcation. Again, imagine there's a serious error in a posthumous paper, the referee lets it slip, and the paper is published. That would be very bad, huh?

Instead, you can do some of the following:

  • with the permission of the heirs, archive the manuscripts and notes online "as is",
  • write your own exposition, séminaire Bourbaki style,
  • assign a student to write a master thesis.

This will solve the dissemination problem while avoiding the above pitfalls.

  • 1
    $\begingroup$ I don't think anyone was imagining that an edited version of the decedent's work would be published/disseminated without the unedited version also being published/disseminated. If the concern is to respect the decedent's wishes, then how do we know that the decedent would be okay with having the manuscripts and notes placed online for public viewing? $\endgroup$ Commented Jun 22, 2022 at 22:23
  • $\begingroup$ A small Twitter thread on a related case: twitter.com/AndresECaicedo1/status/… $\endgroup$ Commented Jun 23, 2022 at 1:45
  • $\begingroup$ How is that fair to the co-authors, who presumable devoted their work to the project in the expectation that the results would eventually be published? I suppose that, before engaging in any collaborative work, we should make a written contract with provision for eventualities such as death of incapacitation. $\endgroup$
    – bof
    Commented Jun 23, 2022 at 2:34
  • 1
    $\begingroup$ If a co-author dies after the paper is submitted for publication, and then the referee's report comes proposing substantial changes, we seem to have an impossible dilemma. We have no way of knowing for sure that our dead colleague would have agreed to any changes we might make, but neither have we any way of knowing for sure that he would have agreed to withdraw the paper. $\endgroup$
    – bof
    Commented Jun 23, 2022 at 2:39
  • 5
    $\begingroup$ @bof The math community seems to have largely taken the stance that the surviving co-authors can faithfully make decisions for the whole group. For example, in my post I mentioned the papers Mark Mahowald's coauthors completed and published after he died. For sure he never saw the final version. Paul Erdos is another example. He's co-author on many papers even more than a decade after he died. Ron Graham gave a talk about one where the third co-author was very young, probably a child when Erdos died (1997), and only worked on the project after 2010. $\endgroup$ Commented Jun 23, 2022 at 4:22

Your Answer

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

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