It’s been almost two years since Richard Poynder published his in-depth assessment of the current state of affairs in the open access movement. Has that been enough time for us to further evaluate his critique and analysis? Does anything look different now than it did then?

A week ago [18 November 2019], Richard Poynder, a well-known and widely respected observer of the scholarly communication ecosystem whose blog Open and Shut? is generally considered a must-read source on the topic, published an extensive commentary on the current state and future prospects of both open access (OA) and the open access movement. Titled “Open Access: Could Defeat Be Snatched from the Jaws of Victory?,” it is an important contribution to the ongoing discussion of the future of scholarly communication.

Before I proceed to summarize and respond to some of the points he makes in this wide-ranging and frankly magisterial document, I should point out that the distinction I’ve made above — that between OA itself and the OA movement — is important. Both the constellation of OA publishing models and the global social movement that seeks to promote them are complex and multifaceted, and the strengths and weaknesses of one are not necessarily commensurable with those of the other. The importance of this distinction frequently becomes obvious both in Poynder’s paper and, I hope, in my response here.

Abstract Technology background.Security concept with padlock icon

Two Fundamental Points

As I understand it, Poynder is making two fundamental points in his analysis, each of which is summed up conveniently in a sentence that can be quoted directly:


We have re-discovered the truth that there is no such thing as a free lunch. Providing free content and services inevitably requires some form of revenue from somewhere.

Interestingly, this is not actually a controversial statement; when the issue of inevitable costs is raised with OA advocates, the response is usually irritable impatience: “Of course OA isn’t free; we get that.” But according to Poynder, in the early years of the movement, “OA advocates gave little or no thought to how the free online content and services they were demanding would be funded.” He provides a number of examples of early advocacy rhetoric that exposes this lack of consideration for the problem of cost allocation, or that suggest a fatal naïveté as to the real costs of providing publishing services — a naïveté that continues to be demonstrated whenever someone characterizes the work of publishers as “taking content for free and selling it back to us at an enormous profit.” Serious open publishing initiatives from arXiv to PLOS to Knowledge Unlatched have clearly shown the fundamental unseriousness of such characterizations; systems always look simpler and cheaper from the outside than they do from the inside, and scholarly publishing is no exception to that rule.

Poynder’s second fundamental point is the one on which he spends the bulk of his time, and it is also by far the more controversial:

We have learned that openness is by no means an unmitigated good.

This is the statement that will likely be met with the most consternation by many in the OA and open scholarship movements. But Poynder supports his assertion extensively, evoking, among other concerns, the issue of what has come to be called “surveillance capitalism” — a phenomenon that thrives (and can only thrive) in an environment of free and open information sharing. As Poynder points out, “the Internet’s foundational ethos of free sharing led web companies to devise business models that are now seen as both deceptive and predatory.”

To be clear, he’s not talking about predatory/deceptive journals here; he’s talking about Facebook and Google, which operate on models under which “the user has become the product.” Of course, it’s important to point out that the “user as product” phenomenon has existed for as long as there have been newspapers and magazines — not to mention free TV broadcasting — which relied on a business model that lured consumers with artificially cheap access to content, thus corralling their eyes on behalf of advertisers whose payments underwrote the bulk of the publishing or broadcasting costs. The major difference between “surveillance capitalism” as it exists now and as it existed in the print-and-broadcast era is one of both scale and effectiveness: in the networked online era, harvesting useful data about information consumers is much, much easier and more efficient than it ever was before — thanks entirely to the radical openness of the Internet.

Unintended Consequences and Unexpected Outcomes

In broader terms, though, when Poynder says that “openness is by no means an unmitigated good,” he is referring both to the inevitable emergence of unintended consequences, some of which will necessarily be negative, and the inevitable failure of some intended consequences to be realized. Among the outcomes that he believes OA advocates have failed to anticipate are these:

  1. The fact that an open and online world creates new tasks and costs in addition to obviating old tasks and costs.
  2. The ability of legacy publishers to adapt to the new environment in ways that would “allow them to maintain their power and… to increase it.”
  3. The reluctance of researchers to embrace Green OA. (In this respect, Poynder argues in particular that “physicists were not typical,” in that they flocked to embrace online distribution of preprints, which itself was the natural extension of long-established print-based practices in the discipline. Though even here, preprint distribution falls well short of most definitions of either Gold or Green OA, and of course does not involve the institutional repositories that have generally been the movement’s preferred locus of Green OA deposit.)
  4. The continued attachment of both researchers and their host institutions to off-the-shelf evaluation tools such as the Impact Factor (IF). (In this regard, Poynder shares a notable statistic: “over 90% of Berkeley faculty still consider high Impact Factor an extremely important criterion when determining where to publish.” And here one has to wonder how these faculty feel about the University of California’s recent decision to cancel access to some of the most high-IF titles in their disciplines. Whether this attachment to the IF is wise is an important, but obviously very different question.)
  5. The potential for damage caused by unrefereed preprints, which can (for example) be posted by drug companies in the knowledge that journalists and others will likely cite them as having been “published” — the thoroughly debunked but widely cited cellphones-and-brain-cancer study (still available in bioRxiv*) being just one of the more egregious examples of this danger. (It’s important to note here that preprints also pose the potential for significant public and scholarly benefit.)
  6. The possibility that “geowalls” will take the place of “paywalls.” (This would be a natural, though unintended, consequence of the popular argument that “taxpayers deserve access to scholarship arising from the research they have funded.”)
  7. The further danger that ideological, social, and economic conflicts between nations could lead to the splintering of the Internet itself, with countries walling their citizens off from the larger world (and walling certain elements of the world out, as well). We see what could be the beginning of a global trend with China’s infamous “Great Firewall,” which some have characterized as the “world’s biggest non-tariff trade barrier.” It would be easy to assume that this can’t happen; that the Internet will (because it inevitably must) remain both open and global. And yet it already is substantially less than globally free and open; many millions of the world’s citizens are actively restricted from accessing it freely — and more than 50% of people (mostly in the Global South) still have no access at all. The argument from inevitability does not have a particularly distinguished career of success in human history.

A Tipping Point? Maybe, and Maybe Not

Despite all of the above issues, however, Poynder argues that two more recent developments have convinced OA advocates “that a tipping point has been reached and the war won.” The first of these is librarians’ growing enthusiasm for transformative agreements (as witnessed, for example, in Elsevier’s abortive negotiations with the University of California system and its successful negotiation with Carnegie-Mellon University), and the second is the slowly-growing trend of governments and funders creating “ever more coercive mandates to compel researchers to embrace OA.”

Poynder, however, suspects that the “tipping point” is illusory, and he sees a number of reasons why the battle for OA could end up being lost rather than won. He outlines several serious challenges to the ultimate success of universal OA. He does so not for the purpose of discouraging its advocates, but rather to help them “anticipate potential problems and try to mitigate them.” Some of the potential problems he outlines are organic (originating from within OA itself, in its various manifestations), and some are external (originating in other social and geopolitical dynamics). They include:

  1. Pushback/Counterrevolution. While many advocates see open access as a moral imperative, government funders tend to see it in terms of potential financial benefit. But if OA fails to yield the anticipated economic benefits, government support could erode quickly. (It also bears pointing out that the moral-imperative argument for OA is hardly universally accepted, even within the scholarly community.) And, of course, OA journals can be flipped back from open to closed in response to pushback from authors or funders — this has happened already in a notable number of cases. Privacy concerns, worries about the mismanagement or nefarious application of free information (by poachers and traders in human remains, for example), the potential for open access and open data to contribute to dangerously uncontrolled development of artificial intelligence, and concerns on the part of researchers about others exploiting data they have labored to generate may also create pushback. Concerns are also already arising about the redirection of research money to support the free provision of access to content.
  2. Populism/Nationalism. In recent years we have seen growing attacks on academic freedom from governments throughout the world — including in the US, where the Trump administration has taken an aggressive approach to discouraging research into topics such as climate change, while actively seeking to expunge research data on such topics from the public record. At the same time, we are also seeing an increasing tendency of democratic processes to yield illiberal electoral outcomes. In this context, there is growing concern that making access to research results more broadly available does not necessarily increase the public’s understanding of those results, nor does it necessarily inoculate the public from the predations of scientific hucksters.
  3. Economic Protectionism. Trends toward economic isolationism and rivalries between powers both great (China vs. the US) and somewhat-less-great (North and South Korea; Saudi Arabia and Yemen; Iran and Iraq) are likely to lead not to more global openness and collaboration, but rather in the opposite direction. OA is, by its fundamental nature, international, but blockades on currency exchange, for example, make it impossible for authors in some countries to pay APCs to foreign journals; similarly, scholarly exchange between scientists in some countries is restricted by government policy (this was true in the US under President Obama, and is even more the case now, under President Trump). Furthermore, there is no question that the free sharing of information, both for reading and for reuse, will inevitably have uneven effects in the world, offering relatively less benefit to the most scientifically and technologically advanced countries and relatively more benefit to less-developed countries. While many of us might argue that this sounds more like a feature than a bug, it may seem just the opposite to those in positions of power in an advanced country whose job it is to help that country stay dominant. Poynder points out the potential implications of this dynamic for (to take just one example) the US National Institutes of Health—the “largest public funder of biomedical research in the world.” At the same time, of course, the massive and systematic cyber theft of intellectual property is a growing international problem, and geopolitical competition and conflict — notably between China and the US — make resolving that problem harder rather than easier.
  4. Naïveté. Poynder argues that a seemingly willful ignorance of the potential for unintended consequences has been a hallmark of the OA movement since the beginning; indeed, even today (as many of those consequences are becoming painfully apparent) attempts at discussing them are regularly dismissed in the community as “fear-mongering.” Such consequences include:
    • the APC model leading to the problem of predatory journal publishing;
    • OA mandates leading to a backlash among researchers;
    • insistence on Creative Commons licensing creating not only resistance among authors but also unintended IP consequences of its own;
    • the potential for OA to lead to decreased funding for libraries.

In one telling example, Poynder reports that this very naïveté about open science and scholarship practices in general has led the FBI to travel to research universities to brief their administrators on best practices regarding information security.

Poynder devotes many pages in his report to an extensive analysis of current global ideological divides and great-power struggles (particularly between China and the United States), most of which is interesting but much of which seems less centrally related to issues of OA and scholarly communication; I let it pass without comment here not because it’s uninteresting or irrelevant, but mainly because I don’t want this response to be as long as the document to which I’m responding. For our purposes, suffice it to say that Poynder sees significant negative implications of these ideological and geopolitical struggles for the future of openness in science, scholarship, and scholarly communication, and I believe he makes a compelling argument for that position. Two of his observations in particular seem especially worth quoting: “The future of scholarly publishing will surely depend to a great extent on what China does — not least because it is now the second largest publisher of research papers in the world and expected soon to overtake the US as the world’s top economy… It seems logical to ask whether China’s interest in OA demonstrates a commitment to openness or simply a desire to have access to research produced in other countries.”


Richard Poynder’s full 84-page document is well worth the time and energy required to read and digest it. Not only does he offer sharp and often trenchant analysis of the state of open access itself, but he also provides wide-ranging geopolitical and economic context for his analysis of the current state and possible future(s) of OA, and of the movement that is dedicated to promoting it.


*Disclosure: I serve on the advisory board of bioRxiv.

Rick Anderson

Rick Anderson

Rick Anderson is University Librarian at Brigham Young University. He has worked previously as a bibliographer for YBP, Inc., as Head Acquisitions Librarian for the University of North Carolina, Greensboro, as Director of Resource Acquisition at the University of Nevada, Reno, and as Associate Dean for Collections & Scholarly Communication at the University of Utah.


14 Thoughts on "Revisiting — The Tyranny of Unintended Consequences: Richard Poynder on Open Access and the Open Access Movement"

You’ve got to hand it to open access advocates. They (warning: gross over generalization) hate the big publishers but rush to pour millions of dollars & euros into their coffers (TAs). They hate subscriptions but vehemently defend policies designed to keep those journals in business (green OA, RRS). They wax lyrical about academics “taking back” publishing but imagine it will be done without spending a cent (at least without forming another committee first). All the while publishing their takedowns of the global publishing system in journals published by the very companies they hate.

The above poster prefers to ignore the significant work done by the Library Publishing Coalition and others to bring (OA) publishing to the academy rather than commercial publishers. Last I checked, no one in their cohort ignores the issue of cost.

I also ignored PKP, SciElo, Open Library of the Humanities, Erudit and more.

There are good, principled and competent people working to bring about a fair open access future. But sadly I don’t think they are the most prominent OA advocates and none of them seem to have the support of coalition S, who would prefer to dump tons of cash on Springer Nature and Elsevier’s doorsteps and then complain about how horrible it is that all the best journals are published by them.

Interesting post. However, “information wants to be free”–not. (Did your original post say this?) Information, whatever it is, is not an entity with feelings of any sort. Maybe its creators want it to be free, but its corporate distributors surely don’t.

The basic point that Richard Poynder is well taken: no word, including “open” can convey unambiguous meaning alone. The same can be said of “freedom” or “democracy”. The battle that has been going on for a long time is about scholarly publishing would evolve. The first victory is that many publishers have now, grudgingly perhaps, admitted that some version of open was an option difficult to neglect. The next step will be to help define “open” in a way that reconciles the processes of scholarly publishing with the objectives of scholarly communication.

On a final note, Rick, I did not know that Trump was still president of the USA. Here, in Canada, we have heard that this was no longer the case. Is it different in Utah? 🙂

The next step will be to help define “open” in a way that reconciles the processes of scholarly publishing with the objectives of scholarly communication.

Built into this statement is the assumption, of course, that the current processes of scholarly publishing are not in harmony with the objectives of scholarly communication–an interesting proposition, but not one that I think we can simply take uncritically as given.

I did not know that Trump was still president of the USA. Here, in Canada, we have heard that this was no longer the case. Is it different in Utah?

As the title, introduction, and first paragraph all indicate, this is a repost of an earlier piece, one that was written during the Trump presidency.

Sorry about missing the markers of a repost…

As for the disconnect between scholarly communication and scholarly publishing, it is indeed a thesis that needs to be debated. The first step is to see how scholarly communication would proceed if scholarly publishing took place with n reference to journals, their reputation, rankings, etc… For starters, there is plenty of literature on the distortions on research created by the impact factor.

Thank you for reposting your summary of OA’s various stripes and colors and some unintended outcomes. CSE’s 2004 annual conference theme was “access to scientific content” with the only OA contributor, PLOS, represented. Past-President Michael Held, Joy Moore, and I laid down a few tracks in Vancouver to open doors for dialogue. It is good to distinguish OA itself from its movement and to elevate it to more than a sidenote of the movement. Objectives and processes don’t always run in tandem, so parsing and chronicling any gaps of intent remains critical to discussing and formulating ways to achieve best possible outcomes. Education and communication are ever evolving.

Unfortunately, there is a recurring but inaccurate narrative about the cell phone study referred to in this post (previously pointed out in comments on an earlier Scholarly Kitchen post – In this incorrect narrative, a group of irresponsible authors claim that cell phone radiation causes cancer, the claim is debunked, and this somehow reveals a deep flaw with preprints. The reality is more complicated and provides important lessons about experimental science and its reporting.

What the preprint presented was early results from an intramural study of the effects of exposure of rats to cell phone radiation in experiments performed by the National Institute of Environmental Health Sciences (NIEHS) as part of the US National Toxicology Program (NTP). The paper reported an elevated incidence of certain tumors in some instances but not others. Confusingly, male and female rats reacted differently, there were unusually low levels of cancer in the control group, and there were a number of other aspects of the study that would limit the inferences one should draw. Importantly, there has been no suggestion the results were in any way fraudulent or inaccurate, though one can reasonably criticize the experimental design.

The study was discussed and reviewed internally at NIEHS and the decision was taken to post the work as a preprint. Part of the motivation may have been that if they did not make the work public and reports of it nevertheless emerged, they would have been accused of suppressing the findings – “Government buries controversial research showing cell phones cause cancer” would have been a far more problematic headline. For transparency, the authors took the unusual step of submitting the work to bioRxiv along with peer reviews commissioned by the NTP and NIH that illustrated the debate among experts over what could actually be concluded. Importantly the title and abstract made no bold claims, merely noting that the finding of glioma and schwannomas were “of particular interest” (that at least proved correct…).

When the paper was received at bioRxiv it underwent screening. As is the case for a small number of papers, it was escalated internally for further debate to decide whether it constituted material that might be harmful to the public. The decision taken was that this might cause some alarm and perhaps a few people might choose to use their cell phones less but that ultimately this was not going to cause a dangerous change in behavior that would threaten public health in the way HIV denialism, vaccine challenges, or a biosecurity risk might. So the manuscript and its reviews were deemed appropriate for posting. The existence of the preprint would afford the scientific community an opportunity to debate the work.

Inevitably there was significant media attention when the preprint posted, and it is important to stress that strong conclusions about what the work showed were made by a few media outlets not in the paper itself. Indeed, more responsible journalists pointed out the results were hard to interpret and ambiguous. As always, the media moved on quickly, and since then the only people who continue to be interested seem to be a handful of science publishing commentators. It’s hard to see how anyone could conclude there was any “damage” done.

Preprints occasionally draw sensational media coverage but the issue is far from confined to preprints – as we have seen with studies claiming 5G towers and comets cause COVID-19. Meanwhile, the infamous arsenic life paper provides just one illustration in which conclusions that everyone knows to be wrong remain in an unretracted paper because the results themselves are not deemed to be flawed, just the inferences.

All of this reminds us that science does not move forward in a straight line. It proceeds in fits and starts; there are wrong turns and dead ends and the occasional baffling result. Preprints provide an opportunity to show this course transparently and enable broader scrutiny of confusing findings in particular. We should remind ourselves that science proceeds based on consensus drawn from many studies not just one result. As a publishing community we certainly need to do a better job of providing context and forward linking around individual papers, but we should also avoid perpetuating and amplifying oversimplistic and incorrect narratives.

Your story about this cell phone paper isn’t accurate, nor is it defensible. Please see this: for a more accurate depiction of what occurred. And here you are, months later, regurgitating the same inaccurate story. It doesn’t stand up to scrutiny . . . again.

Under duress, preprint server managers either bring an ad hoc and intense review process to bear to cover their previously lax practices, or suddenly reveal a thoughtful and careful process never made clear to consumers of the information otherwise. What is the standard? What are you doing? Just whatever you make up at the time, while never disclosing that review levels and practices vary, processes are customized to circumstances, and that you’re just making it up as you go along, and that it’s pretty idiosyncratic and there’s not much of a record of what you’ve done?

Science doesn’t proceed in a straight line, but preprint servers are scribbling and spilling all over the paper. That’s not progress, and it’s not helping. Privileging laxity and affordances isn’t benefiting us. Privileging rigor and more consistent standards of evidence and review does help. Science matters. If you take it seriously, act like you take it seriously.

Well said, as always, Kent! We should all be scrutinizing Preprint Server Mangers and their agendas.

“The continued attachment of both researchers and their host institutions to off-the-shelf evaluation tools such as the Impact Factor (IF).”

This is the major problem. Vain researchers are hopelessly addicted to IF and they only trust certain dealers know how to best supply it to them.

Comments are closed.