95

I am from country A and am studying in country B. A researcher from country C would like to collaborate with me. However, my home country (country A) has poor (hostile) relations with the researcher's country (country C). I'm concerned that if I accept this collaboration offer, which will be done in country B, I may get into serious trouble when I return home (should officials in my home country realize what I've done).

I do not want to lose this offer, as it may result in a paper which would be helpful for my career. Is there a way that I could use a pseudonym on any paper but still be able to claim authorship in the future when applying for jobs?

Edit

I am a PhD student and I do not care if I don't get the credit of the possible paper when I get retired in coming decades! I just need it for my first post doc position. I hoped others who had similar situation would write about their experience. I am sure I am not the first person in country A who is in this situation.

I hoped that I might use a fake name but use my original ORCID ID, so that when needed I can expose the paper in my profile for a limited time. I also thought that a recommendation letter from the coauthors would be good too even if I choose not be a coauthor of the paper.

  • 5
    The answers to this question may be partially relevant to your situation. – Dan Romik Oct 26 '15 at 21:04
  • 13
    Find two suitable very large prime numbers. Multiply them together. Publish the resulting semi-prime along with the paper. When claiming authorship, show that you can factor the semi-prime published with the paper. Luckily, modern cryptographic software will do the hard work of finding two suitable large primes for you! The fundamentals of the RSA public-key encryption system to the rescue again! – recursion.ninja Oct 28 '15 at 13:53
  • 5
    @SnakeDoc USA and North Korea? Russia and Ukraine? Israel and Iran? Especially if the research has military applications. – OrangeDog Oct 28 '15 at 18:43
  • 14
    @SnakeDoc 'Get a better home country' is easier said than done, even if the OP is prepared to permanently sever all ties with family and friends. – cfr Oct 29 '15 at 1:29
  • 8
    @SnakeDoc Other academics won't care that the asker worked with somebody from country C. But the question expresses the concern that the government of the asker's country of citizenship might care about him working with somebody from that country. – David Richerby Oct 29 '15 at 12:47

10 Answers 10

59

You are clearly in a tricky situation that the usual customs of academic publishing are ill-prepared to deal with. There definitely doesn't seem to be a standard, off-the-shelf solution to your problem like some kind of registry for anonymous authors. However, with a bit of creativity there might be a path forward. I suggest the following:

  1. Agree to collaborate with the researcher from country C, but make clear to him/her that you would like your identity to be kept secret for an indefinite period. Make sure that the collaborator is okay with mentioning you as an anonymous coauthor not just in the paper you end up publishing but also in talks, presentations, posters etc. Make sure this is a person reliable enough to trust on such a serious matter, and keep in mind that as with any secret, there is always an unavoidable risk that the truth will come out and somehow make itself known to officials in your country.

  2. If/when you write a paper on the project, have the collaborator submit it to a journal, asking the journal to list you as an anonymous coauthor, under a name such as "John Doe", with a note explaining that this author chooses to remain anonymous due to fear of political persecution. I cannot be sure, but this strikes me as a request that a reasonable journal is likely to agree to. The journal may ask to have your identity disclosed to them so that they can at least have it on record for legal reasons. It may be okay to reveal your identity to them (which could actually help later on in your career when you wish to get the credit for the work), but keep in mind that this again increases the risk that the secret will come out before you intended it to or will become known to the wrong people.

  3. If/when you ever wish to claim credit for the work in connection with job applications or other career-related reasons, make sure you have a reliable promise from your collaborator to vouch that you are indeed the coauthor, as well as some supporting evidence like email records, files, notes, etc. E.g., you and the collaborator can plan to have him/her send a letter of recommendation explaining the situation if and when you need it. Or, you can have the collaborator publicly disclose your authorship, e.g. in a signed letter he/she will send to you that you can post on your website. If the journal knows your identity you can ask for similar confirmations of your authorship from the journal.

    Keep in mind that it may be good to have a backup plan in case something goes wrong, e.g., the collaborator dies or turns out to be unreliable, or is worried about political repercussions for him/herself in country C. Having supporting evidence as I mentioned above, and having a few other people you trust know about your authorship, will all make it easier for you to confirm your authorship if and when you want to. Of course, the more people know your secret, again the higher the risk that the truth will come out in an unintended way.

The thing to remember in this analysis is that there is no approach that is entirely without risk. Given that political persecution is no joking matter, you would be well-advised to give very serious thought to the risk you are taking versus the possible reward. I don't know which countries you're referring to in the question, but I can think of examples where if your actions became known you could easily be accused of espionage and thrown in jail for many years.

  • 22
    +1, however, I think that it may be wise to choose a unique pseudonym and not John Doe. This way, it should be easier to claim all the work published under the pseudonym. – Wrzlprmft Oct 27 '15 at 9:17
  • 9
    While the OP may prefer to use a unique pseudonym, I was thinking of the journal editor's perspective when I proposed John Doe. As an editor, I would consider the journal's duty to be maximally transparent to its readers, which means: 1. disclosing the reason for the author remaining anonymous (see this discussion), and 2. making the choice of pseudonym convey clearly to the readers that this is not the author's true name. So a generic pseudonym like John Doe is more honest IMO. – Dan Romik Oct 27 '15 at 19:32
  • 3
    I understand that this is a serious problem for OP, but what a clever marketing strategy :-). A paper with a mysterious co-author who hides himself due to the fear of political repression. Some people would read the paper/attend presentation only to see what exactly has been done by a mysterious author. – Salvador Dali Oct 29 '15 at 5:25
  • 2
    +1 for an answer that does not rely on cryptography... – Andrew Oct 30 '15 at 12:12
  • 3
    Instead of "John Doe", how about "John <unusual name> Doe"--it's still clear it's anonymous but it won't get confused with any other John Does. – Loren Pechtel Nov 1 '15 at 3:45
29

I would like to somewhat second David Mulder's idea to use cryptography to support your authorship claim in case it is ever needed. However, there is a simpler approach.

  1. Use a pseudonym in the paper as your author name. As suggested before, it should be roughly unique (so John Doe is not a good name). For simplicity, have the pseudonym match your gender.
  2. Add a footnote to the title page of the article stating something like the following:

    Due to a strained relationship between the authors' countries, the (first/second) author used a pseudonym. Should at some point the (first/second) author be able to disclose its identity to some individuals or the public, she/he will do so by providing an explanation document with the SHA256 checksum A4D3.....F1.

This approach has the advantage that it is compatible with the "archival" idea, and is minimally invasive to the standard way of handling things in journals. Obviously, the document will need to be prepared and stored safely before submitting the final version of the paper, and its contents should be presentable to the public and hiring committee members, i.e., to every person whom you may ever want to prove authorship of the article.

  • Related to the second point, do journals usually allow pseudonyms? – Taladris Oct 27 '15 at 10:39
  • Yeah, considered exactly this (with a string instead of document) as well after I had finished writing 90% of my answer, and then I was like 'in general the idea of people owning a private key is better understood then having a string with a public hash', but yeah, same concept in the end, just depends a bit on how much attention you wish to draw. – David Mulder Oct 27 '15 at 20:50
  • 5
    If you wish to go the crypto way, get a signed statement from your co-author anyway. Some people, especially outside computer science, might not be familiar with this whole hash thing, and will probably insist in seeing a piece of paper. And get it now, just in case your co-author gets hit by a bus. – Federico Poloni Oct 27 '15 at 22:30
  • 1
    @FedericoPoloni Even if people might not be familiar with it, if it's about applying for a job I am pretty sure they could check with the IT department in no time if they have any doubts. – David Mulder Oct 28 '15 at 8:37
  • 1
    @DCTLib, but there's no need for that -- just use the same unique name in all papers; another alternative would be to just sha256 the name itself, but then it would be vulnerable to have the name guessed and independently confirmed by any really interested party. :-) – cnst Oct 30 '15 at 12:38
27

Publishing under a pseudonym and receiving credit later has been possible under rare circumstances. For example, William Sealy Gosset published under the name "Student" because of restrictions imposed by his employer, and his discovery of Student's t-test wasn't publicly credited to him until after his death. Dan Romik's excellent answer outlines a reasonable path to trying to accomplish this.

However, it's worth thinking about ways this could go wrong. If someone in my department had a pseudonymous coauthor, it would be fascinating news, and some of the grad students would surely jump at the challenge of figuring out who it might be (on the grounds that there must be something exciting going on). Maybe a friend or visitor? Maybe someone else who recently attended a workshop on this topic? The grad students would chat at length about possible candidates and reasons why they might have wanted to keep their identities secret.

Of course it's unlikely that they could prove it was you, and the discussion would probably never reach your government in any case. However, it's possible that your name would come up in these discussions as a candidate (and conceivable that these discussions would even appear in web searches for your name, although I think they would mostly be in person rather than on the web). If this would freak you out, then I'd be wary of publishing under a pseudonym, since it's noteworthy enough that it's pretty much guaranteed to attract speculation. However, it could be worth trying if unproven speculation wouldn't bother you.

  • 6
    +1, but there is an additional point I see so far not addressed in any answer: the situation must be be reciprocal. The potential co-author in country C likely faces the same hostile reactions if cooperating with someone from country A. If they begin to work with an anonymous co-author, it takes only little guessing of the local thought police to at least turn suspicious. I don't see how this is workable at all (the issues with anonymity aside - student published only as student, here the desire is to be anonymous for this collaboration only (see linked old question for issues with that). – gnometorule Oct 27 '15 at 4:30
  • 10
    @gnometorule: Good point, you are right that reciprocity may be a difficulty. (However, I think there exist pairs of countries that might react in asymmetrical ways, with one being more oppressive than the other.) – Anonymous Mathematician Oct 27 '15 at 4:39
  • 5
    @AnonymousMathematician that's very true about the lack of symmetry. On the other hand, there's also the future to consider: political relations might sour even further and country C's now-permissive attitudes may turn less liberal at some point. – Dan Romik Oct 27 '15 at 5:08
  • Bear in mind that the example of William Gosset is now over a century old and may or may not be applicable to current academia. – Daniel R. Collins Oct 31 '15 at 3:38
22

Honestly, depending on just how sour the relationship between those two countries is it might not be a good idea to openly in your paper mention that the name of one of the authors was withheld due to political tensions and the like. The cool thing however is that modern cryptography provides a solution here that I think is pretty good:

  1. Use a fake (realistic) name without any special meaning
  2. Place in the article links to the homepages of all the authors, creating a special homepage for your pseudonym author.
    In some fields this is pretty common, in some less, but I think that most editors would be quite understanding if you explain the reasons... provided you don't work in a field where it's common to have lots and lots of co-authors.
  3. On 'your' homepage (the homepage of the pseudonym author) list only the one article and a public key from a public and private key pair

Now, if you ever apply for a job and you wish to prove that you were the author of the article all you need to do is tell them to follow the link in the article, download your public key and verify that you are capable of signing a message with the corresponding private key. Now, I wouldn't be surprised if this sounds like magic to you, but here is a wikipedia page and if you Google around there are a lot of ways you could set this up (for example, a lot of sites explain how you can set up this kind of scheme for email). If you want you can go even one step further and use your public key to encrypt a message which you place on your site explaining your circumstances.

The nice thing about this entire set up is that:

  • Anybody looking at the article will not be suspicious
  • If you apply for a job you can simply state 'Articles published under the pseudonym X due to political tensions (can proof ownership if need be)'. You do not need to involve your co-authors or editor at this point (for example if they die) and there is virtually no way to forge this (especially if the homepage is hosted on a third party system with logged history).

Oh well, it's a bit unorthodox and slightly harder to set up initially, but I just thought I would add this solution as it's pretty robust.

  • 2
    You should make more clear that the homepage you suggest to be using is for the pseudonym author. – Wrzlprmft Oct 27 '15 at 9:23
  • Web page seems a bit unreliable: websites do go down, sometimes permanently, and also it's a possible way to deanonymize the author. It seems DCTLib's way with a hash is more reliable. – Ruslan Oct 28 '15 at 11:17
  • @Ruslan There are two things to consider there, but most importantly DCTLib's approach is definitely better if you're fine drawing attention to it. Hosting a public key however doesn't draw attention (normal is a big word, but it's far from uncommon) and there are a lot of databases that collect public keys you can refer to as well. Setting up a site on one of the academic social networks will of course be no guarantee it stays online, but the chances that it will are quite big. – David Mulder Oct 28 '15 at 11:21
  • 2
    @Ruslan +David: if you make sure your website/page gets included in the "wayback machine" archive.org/web you're probably okay even if your hosting arrangement fails later, although deanon is still an issue. Or use specifically a PGP key (convenient anyway for signing proofs with e.g. gnupg) uploaded to the public keyserver network and put your keyid in the article; millions of people use that network so it is very 'normal' and even if PGP fades in the future I'd be very surprised if such a large block of keys don't get preserved in usable form. – dave_thompson_085 Oct 29 '15 at 18:48
9

If your common project is sufficiently large, you can agree in advance to split it into two publications. You submit paper #1 under your real name and your coauthor submit paper #2 under his real name. Each paper cites the other paper (I hope this is not a problem with the authorities in your country to cite an author from a hostile country). This way, each of you receives proper credit and even a citation.

If your papers become sufficiently important, it is likely that future researchers will cite both papers, so you won't lose anything in terms of citation ratings.

A disadvantage of this approach is that MAY be considered "salami publishing", defined as: "the situation that one study is split into several parts and submitted to two or more journals". This may be considered unethical. However, if your project is sufficiently substantial, it may be too long for a single paper anyway, so the splitting may be justified from an academic perspective. As an example, see these two papers by Gul and Stacchetti, which are on the same topic and were published in the same journal in two consecutive years. Each paper cites the other "companion paper".


Just for fun, I would also use ideas from Livings' answer: "you could embed a secret that only you know, e.g. a hash of a secret text, a product of two big primes or similar. Extra points for embedding those in a way that no one recognizes"

  • 5
    I really like this solution. I was focused on the question of publishing under a pseudonym, without even considering the possibility that maybe it's not the best solution in the first place. – Anonymous Mathematician Oct 27 '15 at 14:39
  • 4
    Sorry, but this solution makes no sense. Either the two papers contain substantially different material, or they don't. If they do, then both collaborators must sign them as coauthors, otherwise they are both foolishly giving up credit for work they have done, and unethically and misleadingly pretending that they did not contribute to a paper to which they did contribute. And if the papers are really so similar that the previous clause doesn't apply, then publishing them as two separate papers is a form of (group) self-plagiarism. – Dan Romik Oct 27 '15 at 19:20
  • 7
    I was interpreting this as splitting the results along some natural fault line. E.g., if one coauthor did X and the other did Y, then instead of publishing X+Y together they could publish them separately with cross references. Sometimes the work might split up this way (e.g., theory and experiments, or something combinatorial followed by some asymptotic analysis), and even if it would be more natural to publish in a single paper splitting it might help here. But I agree that if it's inseparably joint work, then publishing it in papers with disjoint authorship is not an option. – Anonymous Mathematician Oct 27 '15 at 23:10
  • They haven't started collaborating yet, so unless the research agenda is very fixed, I think it should be possible to plan it in a way that creates a two separate parts. If it's possible to do it this way, it's definitely the most practical solution. – michau Oct 30 '15 at 13:07
5

I would keep it very very low key, don't inform the editor, don't put in a note about political prosecution. That just makes people dig for more information. Some reader is going to inform your authorities, there are not that many possible countries.

The editor is not your friend either, you don't want him to mention your case when he gives a talk about his interesting life, maybe even in your country.

Take the name of your cat, something that is slighltly unusual, that will increase yhe chance of people believing you later. If you decide to claim the article, put a note in you cv that it was published under a pseudonym, if people don't believe you, they won't hire you either. No one will ever ask you for a formal proof of authorship.

Keep all the drafts, and higher resolution images than used in the print version. Just for the fun of it, you could embed a secret that only you know, e.g. a hash of a secret text, a product of two big primes or similar. Extra points for embedding those in a way that no one recognizes.

And think whethe it is really worth the risk, most articles are probably not.

Edit: In many cases, editors have not kept the names of reviewers anonymous, also, it is common knowledge that interesting articles are often shared pre publication with competing groups. One of my friends had an editor steal her idea, delay the article, and then write a better version himself. Of course, most editors are honest, but you don't want to trust someone like El Naschie, former editor of Chaos, Solitons and Fractals.

  • 1
    Nearly..., but I promised to protect his privacy. – Livings Oct 27 '15 at 6:09
  • 1
    More seriously, how do you propose that OP publish the paper pseudonymously without informing the editor and obtaining the journal's consent? Are you proposing that he/she masquerade as his cat? This does not sound like a very practical suggestion and raises serious ethical and maybe legal issues. – Dan Romik Oct 27 '15 at 9:01
  • 1
    If you have any images, it's likely not that hard to embed an encrypted secret into them. – Christian Oct 27 '15 at 10:57
  • 2
    @Christian: Somewhat harder is assuring that the publication process preserves your steganograph. – Ben Voigt Oct 27 '15 at 13:28
  • 1
    @danromik I don't think any editor will check the affiliation and name. I know a couple of guys who have a fictitious affiliation, because they are retired or their companies don't allow to publish under the company name. Something like Santa Monica research center. No problems. As long as the op is not the lead author, I don't expect much checking. Of course his collaborator should do the submitting. Your concerns are valid of course, but in this tricky situation it seems the least worst solution. Personally, none of my papers would be worth interrogation or worse to me. – Livings Oct 28 '15 at 1:46
2

It's as easy as 1, 2, 3!

Step 1: Cryptography

Using the fundamentals of the RSA cryptosystem, you can find two sutible, very large prime numbers P & Q and multiply them together to get a resulting semi-prime number N. If the prime numbers are chosen correctly (software will do this for you), factoring this semiprime number N to determine P & Q will be computationally infeasible. The only feasible way to factor the semiprime N will be by saving the original two prime numbers P & Q used to produce the semiprime N. However it is trivially easy to check that N = P*Q if P & Q are given to you.

Step 2: Publishing

Publish anonymously as John Doe or a similar pseudonym and include the semiprime N you calculated earlier with the published paper. Keep the prime numbers P & Q used to create semiprime N secret! Store them in a password protected file, or write them down and put them in a trusted location such as a safe or with a friend outside Country A. How you safely store the prime numbers is up to you but you must have confidence that they will not be disclosed, either by betrayal of trust from a third party or from being coerced yourself, before you are ready to claim authorship.

Step 3: Profit

When you are finally ready to claim authorship either publicly to the whole world or privately to an individual on an post-doc selection committee you will do so by proving that you can factor the semiprime N included with the publication. You announce that:

I user43259, am in fact the author of Paper X. I remained anonymous in publication for fear of political reprisal. As proof of authorship I show that factorization of the semiprime N in Paper X is N = P*Q.

  • 1
    This is dangerous. Snowden teaches us that government agencies don't crack cryptosystems: they bypass them. In this case, the bypass is simple. The security agencies in the author's country notice that somebody has published a paper with Recursion Ninja, a citizen of the Great Enemy, while making a big song and dance about how they have to keep their identity secret. So, that anonymous author is probably a citizen of the Great Motherland. Now they hack the co-author's computer, find lots of incriminating emails between the author and co-author and now they know who to throw in jail. – David Richerby Nov 1 '15 at 10:26
  • @DavidRicherby, if the author is concerned about the integrity of his computing device under the scrutiny of Great Motherland, why would he need to store the key or other implicating material within Great Motherland's sphere of influence? I did mention storing the key outside Great Motherland with a trusted third party. The strategy is sound and not as dangerous as you imply. – recursion.ninja Nov 2 '15 at 17:15
  • You seem to have missed my point. The Great Motherland doesn't need to find the key. They just need to find the drafts and the emails which are, of necessity, on an internet-connected computer. – David Richerby Nov 2 '15 at 19:41
  • @DavidRicherby I understand exactly what you are saying. However, securing communications from Great Motherland isn't the scope of this question, as there are different strategies for achieving that goal. A strategy for ensuring current plausible denyability and a future claim of authorship that can be verified with a high degree of confidence is the scope of the question. This is what I have provided. – recursion.ninja Nov 2 '15 at 19:46
  • Securing communications from the Great Motherland is entirely within the scope of the question. The asker says they fear the consequences of their government knowing they're collaborating with a researcher from a particular country. Therefore, they need to take appropriate steps to protect against all ways that their government could find out. – David Richerby Nov 2 '15 at 21:21
2

Without knowing which countries are involved and the research discipline, this is impossible to effectively answer. In some cases, such collaborations could run afoul of laws governing such things as export control of restricted technologies/dual-purpose technologies or international sanctions. It depends on the project on which you are working, the source of funding, the regulations about technology export in country B, as well as the laws governing technology export to countries A and C from B as well as relationships and laws between countries A and C. (Aside: If one of these countries is the USA, this is a non-trivial problem and a potential risk that deserves thorough investigation. Anonymous international research in an area of technology that even remotely connected to anything export controlled could easily trigger an investigation).

So, while some of the strategies suggested here for how to publish anonymously may work in theory, without knowing the countries involved and the area(s) of research, the advice given in some of the answers could carry penalties and risks as big (or bigger) than those you seek to avoid.

Having said all that: assuming there are no specific legal risks (and your only concern is reputation or political fallout from publishing your collaboration), you need to find out if the journals you intend to publish in will even allow anonymous authorship. Until you do that, strategies for how to protect your anonymity are moot.

2

If you are genuinely worried that the government of your country of citizenship will persecute you, your friends or your family for collaborating with a researcher from an "enemy" country then the only way to stay safe is to avoid that collaboration.

All of the answers that propose ways to maintain your anonymity require you to trust a number of complete strangers (your collaborator, journal editors, employers, etc.) with a secret and, if any of them reveal that secret, deliberately or accidentally, you believe you will be in serious trouble.

Even if none of these people reveals the secret, you're still at risk from any of following, which could occur if your government observes that you left the country to study differential widget theory and a paper about that has appeared authored by somebody from the enemy country C and somebody who's probably from your country because he's made a big song and dance about not wanting to be identified.

  • They ask you about it when you next go to visit your family.

  • They hack your computer and find lots of drafts and emails to and from your co-author.

  • They hack your co-author's computer and find lots of emails to and from you.

If you think the risk of your government knowing you worked with this person is genuine, none of the solutions offered will keep you safe from that risk. The only safe option is to decline the collaboration. Whether or not you give the true reason for declining is up to you.

1

I'd like to expand on an earlier answer https://academia.stackexchange.com/a/56983/6026 by DCTLib.

When publishing the paper, do it under a unique name, but DO NOT include any explanations about persecution, especially if you fear that an early investigation might be launched to connect the dots.

Instead of thinking up of a good explanation to sha256, just do it to a final or near-final version of the paper, but with your full name and attribution fully present.

I would also suggest that you might want to do it to a text-only version of the paper, as opposed to a binary PDF, because, ultimately with binary files, it may eventually be possible to generate hash collisions if the attacker is allowed to change the content (since the attacker could then change one of the hidden parts of the binary file to engineer a collision), whereas with a text-only version of the file, with just one of the author's name being different (and with the requirement of having to link it to an existing identity of the perpetrator, potentially over more than one paper), generating a collision is simply nearly mathematically impossible, since there's only so much of the visible information that you could amend without notice (even if you do use a hash on a non-final version of the paper).

Subsequently, once you have the sha256 in question, use the procedure outlined at http://cr.yp.to/bib/documentid.html to incorporate the mention into the paper, effectively without arising any suspicion! (Well, at least it wouldn't arise any if your field is anywhere near mathematics and computer science!)

When I write a new paper, I create a new ``document ID'' for that paper, and list the ID inside the paper:

Date: 2004.04.02. Permanent ID of this document: …

Subsequently, when making references in your resume, provide an brief explanation of the situation, along with the original files used to generate the checksums in question.

protected by eykanal Nov 1 '15 at 2:52

Thank you for your interest in this question. Because it has attracted low-quality or spam answers that had to be removed, posting an answer now requires 10 reputation on this site (the association bonus does not count).

Would you like to answer one of these unanswered questions instead?

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