Journals and the arXiv

Over the past decade, a number of mathematical journals experimented with becoming “arXiv overlay journals”. This meant, essentially, that the final versions of all their articles would be available on the arXiv, and indeed that their online editions would simply be a list of links to arXiv hosted papers.

In this post, I want to collect what I know about the history of the arXiv overlays, talk a little about why the idea failed, and what the prospects might be for future integration between journal operations and the arXiv. I hope people will find analyzing this history relevant amongst the current surge of enthusiasm for transmogrifying mathematical publishing!

I’m hoping also that some of the “primary sources” for this history might jump in to correct and add to what I’ve found.

You can find the old 2001 announcement (by the arXiv advisory board) that Annals of Mathematics had decided to become an arXiv overlay journal, preserved in a sci.math.research archive. At the time of this decision, there were three other arXiv overlays: Advances in Theoretical and Mathematical Physics, Geometry & Topology and Algebraic and Geometric Topology.

Since then, it appears that all of these journals have reverted away from arXiv overlay status. I can’t find any trace on the internet of any actual announcements to this effect, but you can read through some comment threads (e.g. at the nCafe) where they try to track this down. Here’s a quote from Thierry Bouche at the IMU’s blog on mathematical journals:

even the Annals of math–everyone thought they could do whatever they wanted being The Annals–stepped down from arXiv overlay to a mere MSP charged access journal

My limited understanding (having spoken to Greg Kuperberg yesterday) is simply that being an arXiv overlay journal was financial suicide. Annals found that libraries were dropping their subscriptions, and simply could not be persuaded to continue forking over cash when all the content was guaranteed to be available online for free! (I think this is a very important point for open access advocates to keep in mind — funding a journal through library subscriptions is flatly incompatible with open access.)

On the other hand, some of these journals have maintained limited connections with the arXiv. Geometry & Topology, for example, encourages you to submit your paper by providing an arXiv identifier. Personally, I think this is great — it may significantly increase the likelihood that the paper actually makes it to the arXiv, and that subsequent revisions from the refereeing process appear as new versions.

There’s a quite outdated page about journals which accept arXiv submissions, hosted on the Front. It lists (incorrectly) several of those as being arXiv overlays, but also fails to identify the one journal I know of that’s actually currently an arXiv overlay, SIGMA.

My internet searches suggests that at Eureka Journal Watch there’s a page about arXiv overlays, but that wiki seems to be down at the moment.

Some questions for our readers:

  • What other journals accept submissions via an arXiv identifier?
    • (Could we ‘crowdsource’ an update of the page at the Front?)
  • Could more journals be persuaded to allow this? (Is it even a good idea?)
  • When exactly did Annals retreat from being an arXiv overlay? (Is there public discussion of this recorded somewhere?)
  • What is an “associated journal” at Mathematical Sciences Publishers? (e.g. Annals is now listed in this category)

22 thoughts on “Journals and the arXiv

  1. I believe G&T is planning to post the official version of papers to the arXiv after they become freely available. However, they have increased that time to 5 years, so it’s been a while since they’ve made any papers available.

  2. I guess the main problem with arXiv overlaying comes from the fact that most mathematical journals still insist on having a printed version. If they dropped the printed version, then the problem of “financial suicide” would disappear: After the author incorporates the final suggestions of editors and referees, he would put his paper on arXiv (marking it as the accepted version) instead of forwarding it to the publisher. This requires no financial expenses whatsoever.

  3. My internet searches suggests that at Eureka Journal Watch there’s a page about arXiv overlays, but that wiki seems to be down at the moment.

    It is — sorry about that! We’ve had server problems complicated by, well, complications, and are currently setting the wiki up again in what I hope will be a more stable home. Some much-needed software updates will be part of the process.

  4. @Dmitri,

    I think many journals have expenses beyond printing costs, and it’s a mistake to believe that eliminating paper copies makes everything free. Often there is a part time staff member at the institution of the managing editor who handles much of the administrative work. Their salary is paid out of the journal budget. (Sometimes, I hear, managing editors are paid “in kind” directly by their institution, in teaching reductions.)

  5. I would really like to see someone who knows give a breakdown of the actual costs of publishing a journal. People say there is administrative work etc. but I’d like to know exactly what that entails. What really is there to do, aside from the work of editors and referees, if there is no actual act of “publishing” except linking to an arXiv paper as “published by X journal”?

  6. @Scott: Could you please clarify the meaning of “administrative work” in your comment? Producing a printed journal definitely involves a lot of administrative work (printing, distributing, collecting subscription fees, etc.), but I cannot see how any of it would affect a purely electronic arXiv overlay journal. The pre-publication stage involves no administrative work (the author, the editor, and the referee exchange a bunch of emails), and the publication stage simply consists of putting one’s paper on arXiv (the editor of the journal might also update the list of papers published in the journal, but this is trivial).

  7. “People say there is administrative work etc. but I’d like to know exactly what that entails.”

    You guys are forgetting the most important stuff. The publisher’s CEO needs an administrator to guard his office and make sure no smelly hippies get in. He also needs an administrator to carry his golf clubs at the golf course. At home, he needs administrators to cook and clean for him. Dozens of administrators are needed to go talk to legislators to push for important laws like the Research Works Act or SOPA/PIPA. And then there’s the army of administrators needed to sue people for downloading and reading our mathematical papers without proper permission…

  8. Allow me to quote (from private correspondence) a senior mathematician on the differences between editing for Advances and for a major nonprofit journal:

    “Papers were submitted directly to me, I was responsible for logging them, assigning referees, keeping track of the time-table in terms of when to remind referees, writing all the correspondence to referees and authors, reminding referees again, corresponding with other editors, writing delicate rejections, etc. For Advances, I do only two thing: assign the referee and make decisions.”

    This is someone who has contributed immensely to the mathematical community through editing, mentoring and other forms of service. Making editing easier for people like that strikes me as a worthwhile expenditure.

  9. To head off a criticism, I don’t think this is where most of Elsevier’s subscription fees go to and, like Mike Shulman, I would love to see an accounting. But it is a cost, one which I think is worthwhile, and a journal with no revenue stream would have trouble meeting it.

  10. @David: Modern journal management software completely automatizes tasks such as maintaining timetables, sending papers to referees, reminding referees, sending acceptance/rejection letters etc.

  11. But Dmitri, that’s patently false. David’s example’s editor does all this by hand. It might be true that such software exists (could you cite some examples? — I only know, which is undergoing rapid development, but is still far from completely ready). But the fact that it isn’t currently used means there’s a gigantic problem.

    Maybe this problem has solutions; improving the software, educating editors about its existence, etc. But merely asserting the existence of such software doesn’t help much.

    Foremost, we need to hear the particular problems that editors face in adopting such software, and what needs the software can’t fulfill.

    As an example of something that obviously can’t be automated, David’s example mentions “writing delicate rejections”.

  12. @Scott: Here is an example of such a system:
    It is free and open source.
    Here is a list of journals that use OJS:
    Bulletin of the Australian Mathematical Society is an example of a mathematical journal in the list.

    >But the fact that it isn’t currently used means there’s a gigantic problem.

    It’s certainly a problem, but I don’t think it’s gigantic.
    And some mathematical journals do use such systems, as I pointed above,
    which means that other journals can follow them.

    >As an example of something that obviously can’t be automated, David’s example mentions “writing delicate rejections”.

    Writing individual delicate rejections requires the editors to get involved.
    I guess Advances has a standard delicate rejection,
    which is sent to everybody. This can be automatized.

  13. Thanks Dmitri!

    I found a handful more journals which publish pure mathematics, using OJS:

    Contributions to Discrete Mathematics,
    Discrete Mathematics & Theoretical Computer Science,
    Internationsl Journal of Mathematics and Soft Computing,
    International Journal of Pure and Applied Mathematics,
    Journal of Informatics and Mathematical Sciences,
    Le Matematiche,
    Studies in Mathematical Sciences,

    I wish we could get some sense of why almost no journals are using this software. It would be interesting to hear from the Bulletin of the Australian Mathematical Society about their experiences with this software.

  14. Good software solves some problems; however, it is not a substitute for administrative help. No matter how good the software is, a high profile journal has to deal with a veritable flood of questions and requests. A real live person is needed to filter these before they reach the editors. Also, you need support staff to maintain the computers and software and keep the website functioning. This all costs money, probably more money than most people realize.

  15. I am an editor at Math. Zeitschrift, a Springer journal. Surprisingly, we don’t use any software at all — I think that this is the choice of the editorial board, which I believe has resisted Springer trying to foist their software on us. We do have an administrative person who sends out emails keeping everything on track and she also takes care of the papers once they are accepted.

  16. If I may be gauche: here is the boilerplate from Advances (but as it says below, this is a pre-refereeing filter)

    Dear ___

    Following some consultation with the AIM Editorial Board the Managing Editors have now assessed your submission referenced above. Because of the length of the refereeing process, the limited number of good referees, and the large back log of papers, some decisions are made before the beginning of the formal refereeing process.

    The Managing Editors feel that your paper may be publishable, but suitable for a more specialized journal. It is therefore, with regret that the Editors have decided to reject your paper in order to allow you to seek another journal without further delay.

    They are sorry they cannot be of further assistance and hope this negative decision will not damp your enthusiasm for submitting future work to the Journal.


    Which, apart from using the word “damp” as a verb, seems fairly cordial. Certainly it would seem to serve as an automated reply, supporting the last part of Dmitri’s comment at 14.

  17. I stand corrected (but grumbling nevertheless). I’m now mildly curious which other verbs ending in -en have parallel/synonymous forms without. My impression (based on no formal knowledge of grammer/linguistics whatsover) is that the suffix has connotations of “to make [noun]”, viz. embolden, deaden, hearten, sadden…

    And don’t get me started on “momentarily” or “alternate” …

  18. Pingback: Fourteenth Linkfest ~

Comments are closed.