What really happened when two mathematicians tried to publish a paper on gender differences? The tale of the emails


Retraction Watch readers may be familiar with the story of a paper about gender differences by two mathematicians. Last month, in Weekend Reads, we highlighted an account of that story, which appeared in Quillette.

The piece, by one of the paper’s authors — titled “Academic Activists Send a Published Paper Down the Memory Hole” — touches on issues familiar to those who follow the culture wars, which isn’t all that surprising given the controversial topic, one once discussed by then-Harvard president Larry Summers.

The piece has generated a great deal of conversation, some of it quite heated, and a number of participants in those conversations have suggested that seeing the emails that the author of the Quillette piece says support his account would be useful. That’s what this post is mostly designed to do: Surface those emails.

But first, some background. In a nutshell, the paper — by Theodore P. Hill and Sergei Tabachnikov — was accepted last year by one journal — The Mathematical Intelligencer — and then that acceptance was rescinded. It was then accepted — and published, on November 6, 2017 — by another journal, The New York Journal of Mathematics.

That status, however, turned out to be short-lived. On November 9, 2017, the paper disappeared, eventually to be replaced by a completely different paper. The story goes on from there, as Hill details in Quillette, and a preprint version of the manuscript is available on arXiv.

Now, we will stipulate, as we sometimes feel the need to do, that we are not experts in this subject matter, whether one considers the subject math, human variation, or something else, for that matter. We will leave the analysis of its merits to mathematician Tim Gowers — here and here — and statistician Andrew Gelman, to whom, we should point out, one player in this episode suggested sending the manuscript for review. (You’ll see what we mean on page 17 of the document we promised we’d link to.) Other analyses are welcome in the comment section, as links, or text.

Where we are on firmer ground as experts is in scientific publications, particularly how they are retracted, and to some extent how their acceptances are rescinded. Let’s take the latter situation first: Acceptances are sometimes rescinded, and while such moves are inconvenient — to say the least — for authors, there is usually some strange or even troubling story behind it. We’ve seen several cases recently, for example, when editors rescind papers after showing it to publishers’ lawyers. But these papers aren’t retracted per se, because they were never published. That — sans the lawyer part — is what happened to Hill and Tabachnikov’s manuscript in The Mathematical Intelligencer (see page 14 of the document).

The lightning-fast retraction of the NYJM paper, however, is even more squarely in our wheelhouse. Following the publication of Hill’s account in Quillette, in a statement that studiously avoids the word “retracted,” substituting instead “pulled” and “rescind,” Benson Farb, of the University of Chicago — whom Hill asked the university to reprimand for “conduct unbecoming” of a university professor for his role in the affair (the university declined) — writes

I believe that the editor-in-chief should have added a statement about why this was done, but he did not.

Farb is correct. Failing to include a retraction notice or provide any other explanation of a retraction — and that’s what this is, no matter what the journal wants to not call it — violates Committee on Publication Ethics guidelines and might just, well, give ammunition to the idea that this paper fell victim to the culture wars.

We have learned that Mark Steinberger, the editor of NYJM, became ill in March, so he could not respond to our request for comment about why the journal hadn’t included a retraction notice. (Update: Steinberger died on Saturday, September 15, 2018.) Interim editor Kehe Zhu tells us:

I became an editor in late June, several months after the Hill article incident, and I haven’t even heard about it until recently. So obviously there isn’t much I can say here.

So much remains a mystery about this story. But in an attempt to shed some light, we asked Quillette and Hill if we could publish the documentation — mostly emails — that Hill said backed his version of events. He agreed, as long as it was a read-only version. You can find that here, with certain redactions of email addresses and phone numbers for privacy.

We look forward to a lively discussion in the comments.

Like Retraction Watch? You can make a tax-deductible contribution to support our growth, follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up for an email every time there’s a new post (look for the “follow” button at the lower right part of your screen), or subscribe to our daily digest. If you find a retraction that’s not in our database, you can let us know here. For comments or feedback, email us at team@retractionwatch.com.

29 thoughts on “What really happened when two mathematicians tried to publish a paper on gender differences? The tale of the emails”

  1. It’s odd that you would call Farb’s choice of the words “pulled” and “rescind” in his statement a studious avoidance of the word “retract.” The word “pulled” reflects the actual word he used in his email to Steinberger, and it’s clear he wasn’t being particularly careful in his choice of words there. The word “rescind” is used several times by different people in Hill’s file, and the word “retract” isn’t used once. His choice of words seems faithful to the way people were talking about it in the emails Hill has released, while you seem to be suggesting there’s something sinister in it.

  2. Rivin sends an email requesting the article 10pm on a Friday. Hill receives the referee report before 8am on the Sunday.

    36 hours is a quick turnaround for a paper that is very much multi-disciplinary. The referee report is only four paragraphs too.

  3. As far as I can see, the manuscript in question is not about gender, but about sex differences. It is also interesting to read how the mathematician Tim Gowers immediately associates the mere idea of biological sex differences with right wing ideology…

    1. The mathematician Tim Gowers is extremely clear and upfront about the political perspective he brings to the discussion, and how it informs his views. His posts are thoughtful and considered, unlike your garbage comment.

      1. What part of my comment was garbage? Tim Gowers describes two hypotheses about sex differences and assigns them to the “liberal” versus “right wing” political spectrum.

        Firstly, I somewhat disagree on the political terms (what happened to “conservative”?), but secondly I would rather like to keep politics out of the equation, but look at data and/or hypotheses. If your mindset is fixed due to political association, then all scientific debate is over already. And I guess, this is what the debate is about.

  4. …His choice of words seems faithful to the way people were talking about it in the emails Hill has released, while you seem to be suggesting there’s something sinister in it….

    Indeed, there does seem to be something sinister there. I agree that the choice of words used does provide an indication of thought processes, and that is what is sinister here. We are talking about a paper which is part of the academic record, and which should be being formally tracked. To ‘pull’ something suggests an offhand approach which minimises the importance of what is being done – possibly due to embarrassment, similar to the words ‘whack him’ spoken by a gangster….

    1. Farb used “pull” in the context of the temporary removal. Nothing dodgy or sinister about that. “Retract” would obviously have been the wrong word.

  5. It’s curious why the author claimed in email that the NYJM’s actions would result in his being unable to publish the paper elsewhere. Indeed, it’s currently published in ArXiv today, which requires authors to still hold the license to their work. It actually sounds like the NYJM was trying to do him a favor, by not leaving the paper in any sort of “published” state so that he could reasonably assert rights to publish elsewhere without having to explain.

    Like it or not, journals can do as they please, . Although detailed retraction notices are generally good, would the author in this case honestly have preferred that option? To have a statement attached to the manuscript publicly stating that it did not meet the standards of the journal and was published in error, instead of nothing at all? The journal may have handled this the wrong way, but the author certainly isn’t helping himself by pushing the issue and making absurd allegations of academic misconduct.

    1. No, a journal cannot do as it pleases. First off, it loses its scholarly reputation if it pulls out articles for being politically incorrect, or lets a majority of its board bully an individual editor for political reasons. Second, when a journal makes a contract, it is legally bound. It is, I think (I could be wrong, but I know a lot of law), a contract if the author makes an exclusive submission to one journal and that journal then agrees to to publish the article. Damages for breach may be slight, to be sure, but it is legally binding nonetheless— there is “consideration” on both sides, a meeting of minds, offer and acceptance, etc.

      1. To your first point, if a journal wants to risk damaging it’s reputation, that’s their call. They’re a private entity who can calculate that risk for themselves. They can do as they please.

        To your second point, I recommend you read some author agreement forms. And don’t skim the fine print. You’ll be surprised at what journals are completely within their rights to do. That said, the NEJM policies appear oddly short and informal, very much unlike more prominent journals. Those policies do appear to at least imply that the journal will publish the paper “on their web.” There is no specification for how long.

        There are several circumstances that can render any contractual agreement void, and it’s not clear that one may not apply here. I’m not even certain the remarkably brief and shortsighted policies (on a website that reminds one of Geocities) would even meet an accepted industry standards test.

        Further, Dr. Hill himself specifically states in one email that he is not appealing for (re)publication of his paper, but instead bemoaning the effects that the journal’s actions will have on the reputation of the mathematics profession. His long-winded, pompous emails read very much like someone whose goal is not simply to publish original research, but to use the opportunity to stir up drama with small journals who lack the resources to defend themselves as effectively from such bullying, and serve his own sense of self-satisfaction. It’s very much reminiscent of the arguments and language used by prominent right-wing extremists after being dis-invited from college campuses recently. It would probably be a difficult case to prevail on, but I imagine a skilled attorney could assemble enough of a “bad faith actor” argument simply from Dr. Hill’s own words to justify the journal’s actions (and it would be at least indirectly supported by the findings of the UC inquiry).

  6. “Indeed, it’s currently published in ArXiv today”

    No. ArXiv is a preprint server. Putting a paper there does not count as a ‘published’ paper.

      1. Marco, your comment is contradicted by the wiki page you link to. It says

        “Journals focusing on physics and mathematics are excluded because they routinely accept manuscripts that have been posted to preprint servers.”

        1. I noted that some journals do consider a preprint a publication. I did not make any specific references to the fields of physics or mathematics. There thus is no contradiction.

    1. Sorry, but you misunderstand. If your paper has been published in a journal, you have assigned your publishing rights to that publisher. If you no longer hold the rights to your paper, you cannot submit it to ArXiv. Therefore, by publishing this paper in ArXiv, the authors asserted at that time that they held the license to their work. If they attested to that in compliance with ArXiv’s policies, they could have done the same with any journal.

      The point is that NYJM’s actions did not preclude the author from later attesting that he still held the publication rights to the work, which would have been necessary to publish anywhere, at ArXiv or a journal. Therefore, the statement made by Dr. Hill in his email of 11/14 is demonstrably false.

        1. Thanks for pointing out. Establishes without at doubt that the victimhood claimed by Dr. Hill by that particular statement is without merit.

      1. It is routine in mathematics and physics to post things on arXiv and publish them in traditional journals. In fact, it is very odd if a paper in mathematics published this century would not also be on arXiv. This is regardless of whether or not the publisher actually explicitly allows this.

  7. No again. The vast majority of Maths journals accept papers that have been posted on ArXiv. It’s been standard practice for many years.

    And putting a paper on ArXiv does not count as a publication, for example for REF purposes.

    1. That is true, but it has absolutely zero relevance to the point I’m making. When you publish in ArXiv, you aren’t assigning rights, but you still have to confirm that you (and not some other publisher) hold them.

      Please at least try to understand the point being made before making another flippant non-sequitur in response.

      1. When you publish in ArXiv, you aren’t assigning rights, but you still have to confirm that you (and not some other publisher) hold them.

        It’s been 3 years since my last posting to ArXiv, but as best I remember neither then nor at any time in the preceding 17 years did a submitting author have to “confirm” rights by any positive act (even as small an act as checking off a box on a form, which—or some electronic equivalent—is all some mathematics journals request, in my experience). No doubt there’s language, somewhere, warning submitters to keep their skirts clean, which might function to protect the ArXiv’s interests in case a journal decided to get nasty about pre-submission ArXiv posting; but I have never heard of a (mathematics) journal doing so.

  8. I would like to know what the second referee report stated. Rivin claims he got two, though the second one was only seen by the rest of the editors three months later. Did Hill ever receive a copy?

  9. The same is true of economics, law and political science—preprints are fine, and universal. Most journals don’t even ask that preprints be pulled down after the journal publishes the article. What fields do exist whose journals won’t publish papers that were posted as preprints?

  10. There have been several cases, some commented on here, where editors jackhammer papers into inappropriate journals to satisfy their prejudices. These often do not end well. This seems to be a prime example

  11. I am not sureif the text was retracted for political reasons (cultural war), or because it was considered weak…while the premice seems an amusement for the mathematicians…it lacks in biology knowledge to make it an interesting read…Mathematical juggling not being enough. Two important consideration shouyld have the authors take the opportuity to revisit their subject. 1) Status: People are selective in mating individually and communally. The more primitive the organisation of the community, the more mates were chosen not by the individuals themselves, but by the community, this community approving or rejecting the associations. Further, females seek status in their mates, and so do males. Further, the occupation of women in parenting had them work in community doing very similar chores. For this reason, similar status women must have had more similarity. While similar status men could endure greater differentiation since their chores could create competitive advantages around differing physical-biological characteristics. 2) Recently, it has been determined that high immuno-resistant individuals have a tendency to mate instinctively between themselves.

    Considering these 2 points could increase your chance of finding more authority in your current article.

    (I am not an academic, just an interested reader)

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.