18

Recently I refereed a paper in pure mathematics. Although I recommended a "major revision", I was quite impressed with the paper, and indeed I would like to: (1) describe his results in a grant proposal I am writing, and (2) eventually apply the author's techniques in my own work.

The author has not yet made his paper publicly available (e.g. on the arXiv or on his personal website). Ethically speaking, may I now freely refer to his paper in my proposal, and later in my work? Or am I bound to wait until the paper is published or until the author has otherwise made it available to the public?

I e-mailed him recently, let him know that I was a referee for his paper, and asked his permission to do this. If he writes back to offer (or deny) me permission, then that settles the issue. But what if I do not hear back?

Anonymous
  • 26,161
  • 6
  • 74
  • 121
  • If you do not hear back => What (if) you refer to the paper, submit the grant and by the time your grant is being reviewed, the paper is not publicly available. Is this going to affect your chances of securing grant? – The Guy Jul 18 '16 at 15:30
  • 30
    I'm not in mathematics so I'm not going to answer... but in my field, papers under review are strictly confidential. See e.g. this answer. – ff524 Jul 18 '16 at 15:40
  • 2
    @TheFireGuy: I highly doubt it. ff524: that looks pretty definitive and I'm guessing it is probably the "correct" answer for math as well. – Anonymous Jul 18 '16 at 15:43
  • 19
    I also think it's slightly unethical for one to reveal he/she was the reviewer. The process of blind reviewing is done with the expectation of anonymity. A lot of times people seem unable to hold their temptation to "boast" they were the ones who review your paper. I always feel uneasy about this. – Dilworth Jul 18 '16 at 17:23
  • 8
    Even contacting the author lies in a grey zone, in my eyes, and could well displease someone submitting to a blind review. So I'd preface this by saying something like that you apologize for even doing so, but consider the results/methods seen as so important/inspirational that you meant to ask if you could do as you say you would like to. It's too late for changing however you did this, but I would take no further action before hearing back (if you do hear back). When you put yourself in your colleague's shoes, I could easily see how you, too, would feel uncomfortable - of course, maybe not. – gnometorule Jul 18 '16 at 17:24
  • 2
    Side note which does not answer the question but may resolve your situation. Regardless of how you came to see it, if the paper is on the arXiv, you can read it and use it and so on. I know I always submitted papers to the arXiv around the same time as I submitted them to journals, although not everyone does this. – Richard Rast Jul 18 '16 at 18:51
  • @Dilworth: Actually I signed my original review -- I explained in my review that I'd personally done a lot of work closely related to the topic of the paper and even so had not realized that what the author accomplished might be possible. So the author knew who I was already, and I decided to e-mail him. (Also, reviewers for two of my own papers have revealed themselves in the past, so I took it from them that doing so was okay.) Do you believe that at some step I did something unethical? I confess that the possibility had not occurred to me. – Anonymous Jul 18 '16 at 22:33
  • 2
    Ideally, whether you signed your review or not, the editor would remove that identifying information. I think you don't understand the idea of the anonymity involved, and what it entails. – paul garrett Jul 18 '16 at 23:17
  • @paulgarrett: I prepared my review as a PDF, so unless he has some specialized software for editing PDF's, I don't think he could have done so. I explained at the top of my report why I signed my review, and also mentioned this in the "separate remarks to editor". Arguably this was not enough, but I did specifically have in mind that I was giving the editor an opportunity to object. – Anonymous Jul 18 '16 at 23:35
  • 1
    Depending on the amount of math in your pdf, the editor may just have copied and pasted your review from the PDF to plain text. Anyway, contacting the author directly, thus circumventing the editor, looks highly dubious to me. You might even have a conflict of interest now which might prevent you from being an unbiased reviewer for the next revision, given your interest in getting the paper published, and the fact that the author has some indication on who to blame when the paper does not get accepted at a later stage. – bers Jul 19 '16 at 00:06
  • In my field, it is not unethical for a reviewer to give up their own anonymity. (I'm very confident about this.) However, it is disturbing that you knew who the author was. That would be considered extremely poor practice. Neither reviewers nor authors know each others' identities by default. Gold standard: the editor doesn't know who the authors are either. If a reviewer de-anonymises, that's OK. But they don't get to know who the author is. (Obviously, if the paper is published, then they know.) – cfr Jul 19 '16 at 00:32
  • I should clarify that it would, I think, be unethical for a referee to contact the author directly. (Even if they could be certain who the author was.) When a reviewer has revealed their identity to me, it has been via the editor. For example, the referee has added to their comments 'I'm not worried about remaining anonymous'. It is then the editor who passes on the referee's identity, if they are OK with that. The author remains anonymous. – cfr Jul 19 '16 at 00:40
  • 4
    unless he has some specialized software for editing PDF's, — You mean like Acrobat? – JeffE Jul 19 '16 at 00:57
  • 2
    @cfr When I've reviewed (physics) the authors' names have been on the review copy. This varies a lot with the field. – Chris H Jul 19 '16 at 09:45
  • @ChrisH That is unfortunate. – cfr Jul 19 '16 at 10:24
  • 2
    @cfr it's also the norm – Chris H Jul 19 '16 at 10:27
  • 2
    @ChrisH It doesn't alter the fact that it is extremely unfortunate. If names are on the paper, it is essentially impossible for a referee not to be influenced by them, however well intentioned. – cfr Jul 19 '16 at 10:32
  • 3
    @cfr, you are confusing different standards of different fields. Many scientific communities have only a single-blind reviews. Not double blind as you wrongly assume. Whether this norm is appropriate is of course a completely different discussion. – Dilworth Jul 19 '16 at 12:58
  • Even if the author has not posted to ArXiv, he may still have a preprint on his own web page. So check there. – GEdgar Jul 19 '16 at 13:30
  • @Dilworth I clearly said 'In my field ...'. That was purely descriptive and is entirely compatible with it not being the case in any other discipline whatsoever. However, I stand by the claim that it is unfortunate that double blind is not the norm elsewhere. Indeed, triple blind is ideal. Those claims are, of course, normative. Probably I should have been clearer because you seem to be confusing my descriptive and normative claims which I intended to be quite distinct. (And I think my first comment was not sufficiently clear about this.) – cfr Jul 19 '16 at 13:54
  • Please take extended discussion to [chat]. – eykanal Jul 19 '16 at 18:11

4 Answers4

45

No. It's a standard expectation that you'll keep confidential anything in the papers you referee. Once the paper is published (or even made available online as a preprint) then you'd be free to refer to what has been made public.

Nobody
  • 13,796
  • 29
  • 57
  • 134
Brian Borchers
  • 37,106
  • 4
  • 72
  • 130
  • 2
    Exactly. The paper does not exist for a referee. It is a slightly different story if the paper was given to you to read by a colleague, and then your request is reasonable. But it still depends on them whether they accept you quoting their (yet unpublished) result. In math, I am aware of a close friend's result first being presented by a 3rd party, and despite their originator being mentioned, it seemed to have been considered killing novelty, so be aware that one can damage someone considerably by just mentioning the results, even if one gives them proper credit. – Captain Emacs Jul 18 '16 at 18:58
12

Ethically, I would say until the author endorses public discussion of his work (either by express permission, or the author himself talking about his work in public) don't discuss it. In math, it's common to post public preprints when you're ready to publicly announce your work, so not doing it could be a sign that that the author doesn't want this work public knowledge until after being refereed, and possibly having appeared. Possible reasons for such attitudes are: you don't want your work public until someone else checks it, or you are working on follow-up work and you don't want anyone to scoop you.

(The ethics here being to respect the author's wishes about their work. In any case, you don't want a reputation for not respecting other peoples' privacy. You might burn some bridges and people will be less willing to trust you with sensitive things.)

Note: I probably would not have emailed the author, who may or may not mind (and the editor might also mind), but it seems that is already done.

Edit: I agree with Brian Borchers that another issue is the violation of (in math, usually tacit) understanding that refereed papers are to be treated with a measure of confidentiality. I intended to mention this also, but it was late and I must have edited it out.

Kimball
  • 32,248
  • 7
  • 64
  • 169
  • You claim that the ethics are "to respect the author's wishes about their work". But it is not the work, which is the paper written, that the questioner intends to make public, is it? It is the results proven and the techniques used. The author doesn't have a patent on those results or those techniques, and anybody who stumbles upon them is free to use them in whichever way they wish, Are you claiming that if the author suddenly died, then the results obtained should never ever be used? Surely not. Obviously anyone who uses them needs to give due credit, but other than that, all bets are off. – Jaood Jul 18 '16 at 16:10
  • I will add that your sentiments are very common in the academic community, so to the questioner, you probably shouldn't do it. Not because it's wrong to do so, but because people whose opinion sadly matters, will think its wrong to do so. – Jaood Jul 18 '16 at 16:12
  • Thanks. That said, I might add that the author is not from the same country as myself, and he has not published much, and one possibility seems that he might be unaware of the arXiv. I wouldn't be surprised if he responds that he is quite happy to have his work publicized, although of course there is the possibility that he wouldn't be. – Anonymous Jul 18 '16 at 16:14
  • 4
    @Jin5 If someone told you about some progress they made, but asked you to keep it private, would you think it's okay to talk about their work with other people? I am saying it's possible they don't want the existence of their work yet known. – Kimball Jul 18 '16 at 16:53
  • Move away from your conveniently constructed examples and look at how it actually works in reality. The paper described here is in need of a major revision. It is likely that it'd take months for the paper to finally be published (if at all!). You want to wait that long before using results and techniques that would drive the community forward? If everyone did that, academia wouldn't be what it is (and certainly not what it can be in a better world). Sharing ideas unconditionally is a functional imperative for academia. You want to keep secrets, stroke your own ego? Try an industry. – Jaood Jul 18 '16 at 17:17
  • 3
    @Jin5: "Are you claiming that if the author suddenly died ... Move away from your conveniently constructed examples". Hmm. Most likely the questioner has been asked to keep it private, but didn't read or doesn't remember all the details of their duties as a reviewer. If there's no such expectation of confidentiality for reviewers for this journal, as the editor could readily tell them, then fair enough, but Kimball isn't raising a rare or highly speculative case here. – Steve Jessop Jul 18 '16 at 17:48
  • Firstly, your insinuation of a degree of hypocrisy in my posts falls flat. The one who asserts a general hypothesis (that'd be Kimball) cannot validate it by applying it to such conveniently constructed examples, yet the one who attempts to falsify an hypothesis (that'd be me) only ever needs that one single falsifying example. See e.g. Popper's dissatisfaction with the logical positivism that reigned supreme prior to his work. Secondly, there's a difference between confidentiality as a duty and as a request. A request implies that the original author will himself/herself.... – Jaood Jul 18 '16 at 20:42
  • .... make the information publically available as soon as certain trivialities are taken care of, e.g., writing it up neatly and submitting it arXiv. In such cases, I'd reason that it is in fact even optimal that the original author writes up his/her piece, since it'd be a more complete work on the topic written by somebody much more invested in it. In such cases, waiting for the author to disclose this himsefl/herself is actually what'll benefit academia the most. – Jaood Jul 18 '16 at 20:44
  • But when it is merely a duty, the probability of such an event occuring needs to be reconsidered on a case by case basis. In this case, since the paper is in need of major revision, and the author has yet to reply to a message (in which he/she could e.g. have reveal that he/she was about to publish the work on arXiv very soon) I wouldn't hesitate to use the technique used in the paper in my own work, for the benefit of science as a whole, since that is my priority. Everything else is at best pseudo-science and at worst anti-science. – Jaood Jul 18 '16 at 20:49
  • 4
    @Jin " I wouldn't hesitate to use the technique used in the paper in my own work". In advance of the paper being published? Would you care to share which journals you review for? I ask because I would now hesitate to submit to them. – NietzscheanAI Jul 18 '16 at 21:43
  • 2
    @Jin5, I think you are over-interpreting some otherwise-plausible principles. In reality, academics must prove to their dept heads, deans, and funding agencies that they have a sufficiently substantive research program... To insist that everyone disclose every idea they have at the earliest possible moment creates a huge danger for less-than-heroic people who're doing serious work. I don't wish for you the "scoops" that you inadvertently, accidentally, are in fact wishing for others. – paul garrett Jul 18 '16 at 23:21
3

As a supplement to other useful answers and comments: as I've said a number of times on these sites, the apparently immediate idea of open-ness has some issues, the idea of direct communication with authors of submissions one is refereeing has issues, and the idea that there is no impulse to game the system is pathologically naive. Think of retaliation on whistle-blowers, retaliation on sexual harassment complainants, and so on. My point is that it is injudicious to operate as though human beings would be dispassionate and rational... even while, or perhaps "ironically while"... we're talking about pseudo-rational issues such as scholarly merit.

When I referee papers (ok, math is my biz), I try to be positive, especially for junior people who need to get tenure ... or a job, but I do often also suggest substantive changes, and point out substantive infelicities. By this point in my life, maybe I'd be willing to bear a sort of public flak about my critiques/recommendations/edits, but I'd really rather not. So, if a publishing entity cannot commit to my anonymity as reviewer/referee/critic, I'd probably demur.

Lest people think I'm just being a baby, a coward, etc., I'd note that a few years ago I tried to help some good, young people edit their paper so that it was ... um... not literally fallacious. One of those situations where no one doubts the conclusion, but, ... srsly... the proofs should be genuine. ("Or is it just me?!?" ...) The authors did not understand the issue, got angry, etc. Now, while I am disappointed that they did not take my advice, or understand it, ... especially given that misunderstanding, I would not want to be known as the person who (to their minds) "did not understand their paper, and was a b*tch"...

That is, I don't mind giving other people the option to misunderstand what I'm saying, but I'd prefer to be out of the "sights" of their unhappiness.

paul garrett
  • 88,477
  • 10
  • 180
  • 343
1

journal submissions are generally confidential. The blindness of the referee is generally to protect the referee so you are entitled to waive it if you wish.

So you can't use results in the paper until it's placed in the public domain by the author or published or you get his permission.

I suspect the author doesn't want to be scooped.

One approach might be to emphasize to the editor the topicality of the work and ask that the paper's handling be expedited.

Mark Joshi
  • 2,312
  • 11
  • 14
  • 4
    The anonymity of the referee is not only to protect the immediate referee, but to (in effect) promise that future interactions will be of a similar structure. Thus, if referees are expected/allowed/possibly-to interact directly with authors, the rules of the game are very different than otherwise. E.g., if a referee refuses to admit their identity, are they ... "bad"? It's not just the literal fact, but the presumption of, or the possible, fact that can influence thinking. – paul garrett Jul 18 '16 at 23:06
  • 2
    @paulgarrett: Thanks for your comment. (I am not at all being sarcastic!) Several people here suggested that it was problematic that I had even e-mailed the author (which I wasn't expecting), this is the first comment addressing why. I see what you are saying. – Anonymous Jul 18 '16 at 23:19
  • 2
    I agree with @paulgarrett that it is not solely up to the reviewer to waive their anonymity (not even in single-blinded situations). However, as the confidentiality agreement is with the journal, I'd think you could email the editor, explain your case and ask them whether they could put you in contact with the authors. Or whether the editor could ask the authors to upload the pre-print to arXiv so it can be cited. – cbeleites unhappy with SX Jul 19 '16 at 15:33
  • 1
    i've never had anything in the materials received when refereeing saying that I can't waive my confidentiality. I have refereed around 100 papers. I've certainly had referees tell me they were the referee. – Mark Joshi Jul 19 '16 at 21:30