Why can't we just use platforms such as arXiv for proper peer-reviewed publishing?

Question

Many of you will know that, whether per subscription fees or via open access models, there is a ridiculously absurd amount of money flowing to many big (or medium-size) publishers for getting our research outputs (i.e. papers, etc.) published in high-quality venues in the field.

The problem: Content and peer reviews (I think, the most important artifacts from a scientific point of view) are delivered largely for free and voluntarily. In some fields, such as math, physics, CS, EE, and related, even proper camera-ready type setting (e.g. via LaTeX) is also very often done by the researchers themselves. Moreover, in fact, the downstream (i.e. after-acceptance) editorial process in traditional publishing companies is nowadays (for a number of reasons I don't want to elaborate on here) more likely to introduce faults through the various conversions and process stages that are involved. I'm speaking not only of my own experience but have even seen publications that have been corrected using post-prints, personal copies of authors, etc.

So, my question again: Can't we just extend well-known and widely-used archival platforms such as arXiv, biorXiv, zenodo.org, etc. with a light-weight infrastructure to perform peer review (e.g. via an integration with EasyChair or the like and building up editorial teams) and transfer the whole peer review life-cycle from established publishers (with over the years questionably evolved business models) to such platforms? A tiny fraction of the (mostly tax payers') money currently flowing to such publishers would be invested in running the mentioned infrastructure. My naive assumption is that no relevant quality loss (if not even a gain in quality) is to be expected in certain fields of science after an appropriate ramp-up phase.

The only relevant issue, I can see for now, is the issue of a lacking reputation or trustworthiness of such a platform as a publishing venue where, I suppose, it will be a matter of time for this issue to be overcome.

One of the reasons why I am asking is because I think that the mentioned amounts of money can be better invested in science for the public than it is now. A question (totally obvious to me) that I've pondered over for a long time but was afraid to ask. Thanks for your thoughts on these, I apologise, probably too progressive and naive ideas. Fortunately, similar concerns are shared elsewhere.

In response to some of the comments: Archival has very little to do with publishing. This question is about modern peer-reviewed publishing. Please, also note that with this question I very much value professional editorial work and am looking for solutions in support of EiCs, editorial committees, and reviewers. By no means is this question to be understood in any way as a critique of any of the work done by EiCs, editorial committees, and peer reviewers, not at all. If so then it was definitely not my intention.

Answer 1

There are Diamond open access journals (which are free to both author and reader) run as arXiv overlay journals. An example in my area is Logical Methods in Computer Science, which is generally well-regarded in the field (and indexed whereever it needs to be).

Besides the already existing journals, there also seem to be plans for additional journals of this type to better cover the various disciplines. However, founding a new journal takes a lot of effort from a team of respected (and thus very busy) academics. Early on, it will not be on all indices and thus less attractive to the unfortunate amongst us whose administrations force these issues. Taking an established journal away from a publisher to move to this framework is, even where possible, a radical act and usually pure inertia will stand in its way.

Nevertheless, my sentiment is that Diamond open access (with arXiv providing the long-term storage) is the way to go; and I believe that this is a rather popular perspective.

Despite being lightweight, an arXiv overlay journal still has some monetary costs. They seem to be enough to be an issue, but also to pale in comparison to journal subscriptions or open access charges by commercial publishers. My best guess why institutions/governments aren't more eager to support them is that traditional publishers have a lot of lobbying power, while arguing pro-Diamond open access is more of a hobby.

Answer 2

tl;dr: we could, but we won't, because we're in a Nash equilibrium.

It would be the better outcome for all involved parties (except the traditional publishers) if we would find ourselves in the situation that you sketch. However, if a few individual parties make a move towards the optimum, while others all stick with the current strategy, the few movers lose. Hence, the change is unlikely to happen.

---

Suppose that you're a PhD student, who seeks a decent venue for their next publication. You can either go for arXiv, where science is free, or you can go for a traditional venue, where it is not. In a few years, however, you might want to apply for a tenure-track position at some top university, and the competition will be fierce. Most of your competitors will still go for the traditional publication venues, even if you move on. Current practice at hiring committees is to look for candidates with many publications at the top venues, which is typically measured by traditional journals and conferences. Why would you run the risk of having your best work not recognized by future hiring committees? Surely, you shouldn't run the risk of kneecapping your own future employability, so submission to traditional venues it is.

Suppose that you're a tenure-tracker, who seeks tenure. You want to apply for grants that bolster your tenure application. To get those grants, it helps if you can show that the community considers you an expert in the field. How do you establish your name as an expert in the field? For instance, by serving on the program committees of the top conferences, or becoming a member of the editorial board of the top journals in your field. You may consider reviewing for arXiv, but there is only one tenured spot at your university; the four other tenure-trackers with whom you must compete will all review for the top journals/conferences at the traditional publishers. Surely, you shouldn't run the risk of not getting tenure, so reviewing for traditional venues it is.

Suppose that you're tenured. This is the moment! You have freedom, so now you can finally be the change you want to see in the world! But hold on, you're writing research papers together with a new batch of PhD students who all would like to be tenure-trackers someday, and you apply for funding jointly with tenure-trackers who all would like to be tenured today. Will you run the risk of kneecapping your direct colleagues for the sake of your principles?

---

In all these examples, it's not impossible to advance in a scientific career while moving along the arXiv path. In all these examples, it's probably over-the-top to characterize choosing the arXiv path as kneecapping yourself and your coworkers; any real damage may be subtle and small. The point is, though, that the academic career ladder is murderously competitive, and every little bit just might help edge out your competitors. It stands to reason that not many people would allow themselves to risk any disadvantages in this fierce competition, and hence I see no path to get from the current situation to the situation you suggest.

Answer 3

If you restrict attention to areas like pure math, where papers are mainly text with equations, then Arvix overlay journals can work really well. However, different areas of science have needs that are not so compatible with the Arxiv. Of course, other preprint servers might work.

One problem is with supplemental files. Consider the supplemental videos and how they appear one this webpage of the Proceedings of the National Academy of Science. Nice descriptions, links to forms for copyright permission. Now consider the supplemental videos in this preprint of the arxiv. All you get is a title.

Images on the arxiv are another issue. A few of my papers I did not post on the arxiv as the images do not compress well and I did not want to fiddle with creating low-resolution images just to get the figures small enough for the arxiv. (The size limits used to be smaller.)

Recently the Arxiv has added the "papers with code" feature, but this mostly assumes hosting the code elsewhere.

I think the needs of journals are too diverse to count on overlay journals to cover too much of math and science.