## Beyond the zombies: How we might get out of the science publication disaster

This is a guest post by Martin Haspelmath, building on our continuing coverage of Open Access in linguistics.

By now, everyone knows that scholarly publication is serious trouble. The actual costs of disseminating content have plummeted drastically, and yet academic institutions are paying more and more to the commercial publishers. This feels deeply wrong — as if Facebook charged us for posting cat videos. In some fields such as linguistics, there has been a lot of discontent for quite a while. Johan Rooryck's efforts to take the old "Lingua" away from Elsevier have been widely publicized, and Elsevier's handling of the situation, as well as the continuation of "Zombie Lingua", are regular talking points among linguists.

But while it's easy to agree that Elsevier's behaviour is outrageous, it is not so clear what a good general solution will look like. Some of us had hoped that the Lingua/Glossa example would be taken up by more and more other journals, and linguists (as well as academics in other fields) would increasingly move away from the commercial giants and adopt "fair open access" publication models. Authors would increasingly submit to the good journals (e.g. those on George Walkden's platinum open access journal list), and there would be more book-publishing brands like "Language Science Press" which are based on community collaboration and charge neither authors nor readers.

Is this happening? I'm not sure. Most of our colleagues seem to have sympathies with what we activists are doing and saying, but at the moment, it takes quite a bit of extra effort to find an alternative model for one's journal, or to typeset one's book manuscript in LaTeX (as is required by LangSci Press). And with all the other pressures on academics, few people have these extra resources. There are over a hundred well-established linguistics journals, and the great majority of them show no signs of wanting to change their publication model.

One can understand this, because individual scholars are not really affected by the disastrous situation. While paywalls are sometimes a nuisance, the access we have to the literature has actually been getting better and better. This has largely happened thanks to developments that had nothing to do with our traditional literature providers (our academic libraries): First, more and more colleagues uploaded their papers to their personal websites (a development of the early 2000s), and then, Academia.edu and ResearchGate made it very easy to share one's research with one’s colleagues. Again, it's a nuisance that often one has access only to the manuscript version of a paper (because only the manuscript is truly public, while the "published" version is paywalled), but we've learned to live with this. We know that our institutions waste a lot of money on hugely overpriced content, and we feel that our libraries are less and less needed. But these problems do not affect us directly.

For a long time, open access seemed to be the solution to the problem of high subscription prices and lack of access — and it was taken up eagerly by politicians in some countries, because it seemed reasonable that the results of publicly funded research should be publicly accessible. But these politicians did not understand the logic of academic publishing: They saw research papers as providing insight that might benefit the society, whereas for an academic, a branded publication is a way to measure their professional success. And of course the default funding model of open access publication has always been APCs (author fees), so if a linguist is told that they should pay €10.000 to publish an open-access book with De Gruyter, they naturally prefer the traditional paywall route, and share the PDF privately with all those colleagues that are actually likely to read the book.

So the "science publication disaster" is a nuisance for ordinary academics, and a true disaster only for science funders, who have to spend more than US$10 billion per year on publication (i.e. about US$5,000 per article). Publishing an article does not cost more than a couple of hundred dollars, so one billion should be enough, and science would gain US$9 billion per year if we could "take back control". Probably the most successful academic-led effort to disrupt the disastrous system is Sci-Hub, which now gives access to 77% of all research articles (including 97% of Elsevier's articles). This is not a legal site according to Western standards (although using Sci-Hub is at least partly legal, according to an expert opinion from Germany), but it is widely used by scientists from around the world, even for access to articles that are available through their institution — naturally enough, because Sci-Hub’s interface is as easy to use as Google. Sci-Hub (or its successors) will not go away, so it seems clear that the transition to full open access will be very quick. Thanks to Sci-Hub, all paywalled journals are now zombies for the publishers. Even small, family-owned linguistics publishers like De Gruyter and Benjamins will charge linguists for publication in the future, maybe even the near future — there will not be any alternative business model left. But will this solve the problem that science funders spend$9 billion too much for publication? No, because the publishers have the same costs and want the same profits. If they can get away with it, they will charge the same amounts in author fees that they have traditionally received through subscriptions. And since we need their brands, they may well get away with it.

The only long-term solution that I see is for scientists to implement the principle of non-profit publication (as recently argued forcefully by Jefferson Pooley on the LSE blog). We take the academic freedom that comes from the non-profit nature of our research for granted — and in the same way, we should demand non-profit publication to free ourselves from the publication zombies.

If publishing were funded in the same way as research (e.g. by giving 20-year tenure to a journal after it has been evaluated positively), then it could be vastly cheaper, because the funders/publishers (mostly universities, but maybe also charities like the Gates Foundation) could buy the technical services from the cheapest bidder. The value of publication brands (like journal titles or book imprints) would be used in the same way as university brands are used at present: To attract the brightest minds and to motivate people to do the best science, rather than to extort money from science budgets.

This will not be easy, because our evaluation system relies on prestigious journal labels and book imprints, which are mostly owned by for-profit organizations. A book publication with Cambridge University Press still carries a lot of weight on everyone's CV, and a paper in Springer's NLLT is much better than a paper in the platinum journal Catalan Journal of Linguistics. The brand owners will do everything they can to preserve the value of their brands (this is what we see with Lingua, a brand whose value has been severely compromised, but that Elsevier doesn't want to give up on).

Thus, even if we had enough funders willing to redirect money from their libraries to journal publication, there would still have to be a shift in the behaviour of the community. We would have to distinguish between "best-practice publishers" and "legacy publishers", and there would have to be some mechanism for rewarding best-practice publication and penalizing legacy publication, to counteract the prestige of the old-style labels.

Thus, I see no easy way out of the current predicament. But maybe a disruptive development will come from a direction where nobody is expecting it now. One thing is for sure: The present system is a disaster, and is ripe for major change, going far beyond individual journals and individual disciplines.

This is a guest post by Martin Haspelmath, building on our continuing coverage of Open Access in linguistics.

1. ### Mark Meckes said,

August 28, 2017 @ 3:43 pm

Readers of this blog interested in seeing a similar perspective from another field may be interested in this post by mathematician Timothy Gowers.

2. ### Tim Leonard said,

August 28, 2017 @ 3:50 pm

The big costs are borne by academic libraries. Encourage them to force a solution. A union of libraries could fund a high-prestige non-profit publisher, using money they'd otherwise spend on zombie journals.

3. ### Matthew T Bradley said,

August 28, 2017 @ 4:40 pm

This interview from a 2014 issue of ‘Cultural Anthropology’ might be of interest.

https://doi.org/10.14506/ca29.2.04

4. ### Paul Garrett said,

August 28, 2017 @ 6:48 pm

Btw, at least in comparison to other publishers, Cambridge University Press is willing to negotiate very sane deals with authors (such as myself): retention of copyright, "permission" to keep PDF on my personal/professional website without delay/embargo (under mild conditions about referring to the existence of the physical book), etc.

But, otherwise, yes, I've found ever less reason (apart from documentability of status-acquisition, in mathematics), to go the usual route of "publication" (really meaning "gate-keeper/status-controlling games", since the delays are silly, referees often behave inappropriately toward "competitors", and the current standards for "publishability" have been wildly corrupted by the ever-more-pressing need to commodity-publish in ways that do not necessarily benefit anyone… except the author, who can then suitable score points with dept and dean, and get a less awful pay raise, etc.)

"Don't get me started!" :)

Put things on your own web page.

5. ### James Wimberley said,

August 29, 2017 @ 5:01 am

Gowers' use of "flip" for journals going open-access should be commended and followed. How long before it makes it into the OED?

6. ### A. Jones said,

August 29, 2017 @ 7:24 am

Does a preprint archive exist for this field, and would it be used if it existed? In physics, essentially everyone posts their work to arxiv.org, usually at the same time as it's submitted to a "legacy" journal. In practice, when I'm looking for papers from the last 10 years, I don't usually even bother looking at the "published" versions, and there's no need to hunt for papers from people's websites.

7. ### Vilinthril said,

August 29, 2017 @ 12:16 pm

Adding to what A. Jones said, the same is true for (at least my subfield of) mathematics. If it exists digitally, it most likely is on arxiv.

8. ### howagreg said,

August 29, 2017 @ 1:02 pm

As a researcher from a health-oriented profession, let me sum up my concerns about some of these solutions. The idea of arxiv makes a ton of sense in physics and the natural sciences (in which I used to work) — we shared articles by preprint long before arxiv. But in the health fields, the stakes are immeasurably higher. These are not generally collaborative fields like math and physics, but highly commercialized fields where a single paper on a major product (for example, glyphosate) can make or break a billion-dollar industry. There is a constant and often subtle agenda of industry trying to break down academic opposition to their products. (Look at the atrazine controversy, or the lead industry's decades-long attempt to ruin the career of Herbert Needleman, who demonstrated loss of IQ due to lead in kids.)

In cases like this, peer review is absolutely crucial, and in fact needs to be greatly strengthened. The idea of using results from a preprint system — or even having those results widely available for naive or malicious individuals to wave about — makes me cringe in fear. There's enough weak science and low-quality science reporting.

I love the idea of academics coming together to do real free open-source journal work, but doing unfunded work as an academic is a difficult proposition (especially for junior faculty). A great example was Open Medicine, which split from the Canadian Medical Association Journal, and ran as open and more or less DIY. Unfortunately, it didn't last long.

My point is that this is a huge problem, and the solutions may be very different in different fields, since they have different motivators and challenges.

9. ### Eric Baković said,

August 29, 2017 @ 6:26 pm

@howagreg: I don't think anybody is arguing that we ought to substitute peer review for "free open-source journal work"; the two are not in any way incompatible. You may be right that some fields will require more stringent solutions than others, but that's already true of the current situation: indeed, with profiteering publishers out of the equation, "free open-source journal work" stands a chance of being *less* beholden to "naive or malicious" forces in the form of corporate interests.

10. ### Lauren Collister said,

August 30, 2017 @ 9:55 am

Thanks for this excellent post. I am a linguist who works in publishing and scholarly communication in academic libraries, so I love to see this conversation happening. I want to point out a couple of academic library publishing endeavors that readers may be interested in.

Many libraries are already working on open access publishing initiatives. I encourage those interested to ask your library (look for someone with a title like Scholarly Communications Librarian, or Digital Scholarship Librarian), or look at the listings and members of the Library Publishing Coalition: https://librarypublishing.org/
For example, the library where I work at the University of Pittsburgh currently publishes 39 open access scholarly journals, and not one charges APCs. Some are edited by Pitt faculty, but many are edited by external partners or at scholarly societies. If anyone has a journal looking for an OA publisher, I know that we'd be delighted to talk to you. More info and our journals here: http://library.pitt.edu/e-journals

Libraries are also diverting some of their budgets to innovative open access publishers who follow these best practices. One that may interest linguists is the Open Library of Humanities (OLH), which is a megajournal platform that publishes content across the humanities, but also publishes several journals on specific topics. The editorial board of Lingua started their journal Glossa on the OLH platform. OLH is supported using a library subsidy model, where academic libraries pay a (very small!) fee to be a supporting member, and no APCs are charged. The fee is much less than we'd pay for a traditional journal subscription; many academic libraries support this initiative. Here is their website: https://www.openlibhums.org/

In close, you are right that "there would still have to be a shift in the behaviour of the community". We have these tools and services available and are always building more, and a few people are taking advantage. We'd love to see more editors of journals come to their libraries to start a new journal or to flip the one that they already edit from a subscription model. I invite any linguist who is interested to contact me and I will be happy to talk with you about the process and try to connect you to someone who could help you locally.

11. ### Peter Gerdes said,

August 30, 2017 @ 1:21 pm

I suspect the eventual disruptive development for journal articles will come from the collaborative online work at discipline specific websites. Think mathoverflow (or whatever overflow) meets arxiv meets yelp.

At some point the tools to let us conduct discussions around papers online and rate them will become attractive and useful enough to make them regularly used. Such tools may need to restrict access to bona fide members of the discipline or otherwise weigh different people's ratings differently but in the long run the whole idea of a paper being looked at by a few colleagues and that being the end of it will seem quant.

Instead, paper publication will become nothing but posting the paper to the appropriate academic website and ratings by colleagues will be used to rate the importance of the contribution.

Journals also represent division into interest groups and we will have that as well. Just as we now have groups of academics get together to write blogs we will have more and less influential groups who simply share lists of paper links they found good this month on a certain topic.

In the long run peer review once and done model just doesn't seem like it can stand up to the 'lets see if anyone suggests any issues with this model and have a public discussion about how important they are over time' model which the internet is already bringing us closer to (e.g. Gellman's influence on social sciences).

Yes, there are some things we need to figure out first to get there, e.g., how do we appropriately incentivize reviewing of papers by new entrants to the field and disincentivize people from submitting lots of low quality papers. These solutions may be social or algorithmic but I think in the long run science will look more like a really really fancy blogging community with formal 'final' versions and revisits and with direct measures of community appreciation like ratings rather than indirect ones based on which journal you get published in

12. ### howagreg said,

August 31, 2017 @ 10:18 am

Eric — I'm not saying that anyone here is suggesting abandoning peer-review, but arxiv gets brought up as a model a lot and it's a lot like publishing before peer review. Articles about new arxiv publications sometimes, but not always, make the point that they are not peer-reviewed, and sometimes trumpet them rather dramatically.

Of course, I would absolutely love to see the open journals flourish. But it's hard enough for professionals in the field to figure out which minor journals are legit and which are scams. I don't think a journalist or layperson could be relied upon to distinguish between, say, The Journal of the Science of Important Things (run openly by top scientists in the field, and not beholden to special interests) and The Journal of Important Things Science (brought to you by the American Council for the Support of Polluting Industries, and run entirely by special interests). There is (unfortunately) still a lot of value in having gatekeepers.

Incidentally, in my field the top journal is Environmental Health Perspectives, which is largely funded by NIEHS/NIH, is completely open access, and has no page fees. That's really by far the best model I know of.