Two steps to a revolution in scholarly publishing – a thought experiment

During the Berlin 10 conference on Open Access, the first instance of the Berlin Conference held in Africa, some of the most compelling speeches came from those who advocated a much more radical approach to breaking the hold over academic publishing currently exercised by commercial firms.  Especially from Dr. Adam Habib, the Deputy Vice-Chancellor for Research from the University of Johannesburg, the delegates heard a call for immediate and direct action, even if it provoked a fight with commercial publishers.  In his opinion, at least, that would be a fight worth having, particularly for an African university.

These calls for radical solutions got me thinking about which stakeholders had the power to break open the current structures, which were called feudal and iniquitous by one speaker, and what actions they might take.  The result was the following thought experiment.

First, it seems to me that only two stakeholders really have both the will and the power to radically alter the terms and conditions of scholarly publishing.  One set of stakeholders are the academic authors themselves, who seem to largely recognize that the current system is grossly unfair to them and does not serve the best interests of scholarship or of scholarly authors.  Nevertheless, academic authors are limited in the kinds of action they can take, both because they are not well-organized and because of their dependence on the promotion and tenure system.

That leads to the other set of stakeholders who could shake up the system – the universities themselves.  Are there steps they could take that would reset the conditions for scholarly publication?  I think there are, and it seems like universities, if they were so minded, could create a revolution in scholarly communications with two simple, but by no means easy, policy changes.

First, universities could alter their intellectual property policies to assert that faculty scholarship is, in fact, work made for hire.  The legal argument here is simple and persuasive, that faculty work is created by regular employees within the scope of their employment.  Courts have recognized this argument for years, but universities have rightly been unwilling to press the case, for fear of doing harm to relations with their faculty members.  But as the scholarly publication system increasingly fails to adapt to the radical new conditions created in the digital environment, it is possible to imagine a policy change like this undertaken with the cooperation of the faculty authors themselves.

To make this work, universities could assert ownership, under work for hire rules, and at the same time grant back to faculty authors a broad license to make customary scholarly uses of their own works.  This would actually be a similar situation to the one that now obtains, where authors give copyright away to publishers without compensation and receive back fairly limited licenses to use the works they created.  Universities would almost certainly prove better partners in this approach to intellectual property because they could extend much broader licenses to the original authors and because their values are much more congruent with those of academic authors in the first place.

If universities owned scholarly writings as work for hire, however, they would have control over the means of publication.  They would have the power, and the incentive, to refuse to have work published in commercial journals.  They could give complete preference, if they wished, to open access journals and specifically to those open access journals that are run by scholarly societies and university presses.  Almost overnight universities could put an end to the subscription model that worked so well for 300 years and has now become an obstacle to academic research and scholarly communications.

Of course, if universities were to cut out the commercial publishers, the second step in this revolution is obvious – the system of evaluation for academics, the standards for promotion and tenure review, would have to change.  And this would also be a very good thing.  The system we now have, dependent as it is on the impact factor of the journals in which articles are published and the reputations of the presses who publish books, is inefficient and under-inclusive.  It assesses a work of scholarship on only one type of impact, citations in other scholarly journals, and evaluates that very limited metric at the journal level rather than article by article.  Once we stop confusing what was the only thing we could measure, in the past, with what is truly important, we will see that in today’s scholarly environment we actually can do much better.

So the second policy change that universities would need to undertake to affect this revolution would be to require that all assessments be based on article-level metrics applied to openly available works.  This change sounds very radical, but some institutions are already moving towards it.  At the University of Liege, in Belgium, it is already the case that faculty assessment is done only for articles that are in the university’s open access repository; this was the way Liege decided to put teeth into their open access mandate.  But universities could require open access and article-level evaluation measures while still supporting a variety of publication models.

And that is the final point to be made about this make-believe revolution.  If universities carried it out, it would free up more money than it would cost.  Once academic publication in commercial journals was halted, library collection budgets could be redirected.  Instead of a long transition period during which costs would be expected to rise because both subscription models and open access based on article processing charges would have to be supported, which is what the Finch Report predicted in the U.K., this suggestion would allow for wholesale cancellation of commercial publications.  The money saved would then be available to build up the infrastructure for repositories and to support APCs for gold open access publication.  Authors would have a choice – they could publish in an OA journal or they could publish directly to the institution’s repository.  Peer review could be preserved in a distributed model;  OA journals would continue to support traditional peer review, while some of the money saved from commercial subscriptions could be redirected to a more independent, discipline-specific system of peer-review.  This would provide an important role for scholarly societies, and subventions provided to support such society-run peer-review would help protect those organizations from any negative consequences of this radical re-visioning of the publication system.  Societies and non-profit presses, of course, could also find support through the publication of gold OA journals and even monographs.  University funds from library collection budgets would be more distributed than they are now, able to be used more efficiently to support activities genuinely central to the academic mission, and they would, I believe, be more than adequate to the task.

So that is the thought experiment that began at the University of Stellenbosch while listening to calls for radical action.  I don’t know whether I think it would be a good idea to implement or not.  But I believe it would be the fastest way to dramatically fix the current, broken, system of scholarly communications.  There would be many obstacles to these two policy changes, and the medicine might be worse than the disease.  But at the very least, thinking through this experiment in revolution has given me a better perspective on the power dynamics of the current system.  In libraries we are accustomed to thinking that the huge commercial publishing firms hold all the power and that there is nothing we can do to break their stranglehold over scholarship.  But upon reflection we can see that the real power over the system, which can either perpetuate it or revolutionize it, really does reside on our own campuses.

Another fair use victory for libraries

Note to readers — a commenter has correctly pointed out that I was a bit over-enthusiastic in this post.  It is not strictly true that no infringement has been found in any of the cases against libraries.  In the Georgia State case the Judge did find five instances of infringement among the seventy-five digital excerpts that were challenged.  But the Judge also held that the University’s fair use policy was a good faith effort to comply with the law and, for the purpose of awarding costs and attorneys’ fees,  determined that GSU was “the prevailing party.”  GSU was simply ordered to revise its policy to reflect places where the Judge felt it had not adequately defined fair use.  To be adequately cautious, however, my comments at the end of the first paragraph and of the full post should have said that fair use has been an effective tool for academic libraries that has been upheld in each case and has prevented any significant liability.

We knew some time ago that the second complaint filed in the copyright infringement case brought against UCLA by the the trade association AIME over streamed digital video had been dismissed.  But last week Judge Consuelo Marshall filed her order that explained the grounds of that dismissal.  What we have learned is that this case is a slight victory for fair use in libraries.  On the specific issue we do not have clear guidance, just an affirmation that fair use arguments for streamed digital video are not unreasonable or obviously wrong.  But it is helpful to see this ruling as part of an overall picture, one in which all three cases claiming copyright infringement by academic libraries which were defended on the basis of fair use have now been decided at the trial court level and NO INFRINGEMENT HAS BEEN FOUND.

In the AIME v. UCLA case, the reasons for the dismissal are primarily that the plaintiffs lacked standing  — the legal right to seek judicial redress of a claim — and that the defendants have either sovereign immunity or qualified immunity based on their status as state officials performing their jobs.  A lot of the interest in the ruling is in the part that discusses qualified immunity, because that holding involved deciding whether or not a reasonable person in the position of the official charged would have know that his or her actions would be “clear violations of established copyright law.”

As for the standing issue, the Court says pretty much the same thing I said at the time the second complaint was filed — the plaintiffs have not shown any facts that make this complaint different from their first one, which was dismissed earlier.  These plaintiffs simply are not the right people to bring this case, and the folks who would be in a position to claim copyright infringement have not stepped forward to support the complaint.  Perhaps those other parties understood the risk that there would be another positive ruling for fair use, and did not want to take the chance.

But as I say, the real action is on the issue of qualified immunity, where a plaintiff must prove, to overcome such immunity, that the defendants either knew or reasonably should have known that they were infringing copyright law.  In the light of what the Judge calls the “ambiguity” of the fair use issue, she declines to so rule, and hence dismisses the complaint against these defendants (including library and IT staff) because of their qualified immunity.

Judge Marshall’s fair use analysis is very cursory, but still quite interesting.  On the first factor, the purpose of the use, the Judge simply states that it favors fair use, presumably because the use is educational and non-profit.  This is no surprise.  On the second factor, the Judge rec0ognizes that the works involved, performances of Shakespeare plays, are creative, but also that they are being used in an educational context.  She finds this factor to be neutral.  As to amount, Judge Marshall thinks this factor weights “slightly” against fair use — the entire films are being streamed, of course, but she finds “compelling” the argument that this streaming is analogous to the “time shifting” upheld by the Supreme Court as fair use in the Sony video recorder case.  Finally, as to the impact on the market factor, the Judge says this factor FAVORS fair use because there is no difference in a streaming situation as to market harm than would be the case in the clearly permitted situation where a classroom full of students watched an entire movie together.  The recognition of this essential similarity as to market harm strikes me as fundamentally important.

So the Judge holds that “there is, at a minimum, ambiguity as to… fair use.”  In fact, her analysis would be very likely to uphold fair use, but she does not need to reach such a ruling in order to find in favor of qualified immunity and dismiss the complaint.  Judge Marshall then goes on to examine the licensing terms under which UCLA obtained these videos, to see if those terms state unambiguously that streaming is not allowed.  Here too she finds ambiguity, and thus upholds qualified immunity regarding that assertion as well.

Legally, the bottom line in this case is that all of the claims are dismissed with prejudice, which means that they cannot be refiled, although, of course, the dismissal could be appealed, if AIME and Ambrose Video are inclined to waste even more money.

For libraries, I think there are several takeaways from this case:

1.  The fair use issue as to streamed digital video was not decided.  This is not a blanket authorization for schools to proceed with such projects; they still require careful thought and a local decision that balances the risk of copyright litigation with the pedagogical value of proceeding.

2.  It is very important to look at licenses for the videos in question, where they exist.  It seems clear that an unambiguous license would have overcome qualified immunity in this case.

3.  The grounds for this dismissal apply only to the state officials involved; the same analysis would not apply to a private institution, although the fair use analysis, of course, could be raised by such a college or university.

4.  Overall, the atmosphere for fair use on campus is a lot better today today than it was six months ago.  We have seen a pretty convincing victory in the Georgia State e-reserves case, a sweeping one in the HathiTrust case, and a tepid affirmation of fair use (probably!) in this streamed video case.  Although the first two of those cases are being appealed (we do not know if AIME will appeal this latest decision), we now know that courts are quite sympathetic to fair use by academic libraries and on college campuses.  We know that most of the arguments that content providers have long offered to discourage reliance on fair use have been rejected.  We know that no campus has yet been found guilty of infringement when it makes reasonable fair use of lawfully acquired materials for limited teaching purposes.

A foray into politics

This site has been quiet for about three weeks; for much of that time I was in South Africa, first as a tourist and then to attend the Berlin 10 Conference on Open Access that was held in early November at the University of Stellenbosch.  I hope to write about the conference and the South African context in the near future.  But today I could not resist a foray into politics.

On Friday, the U.S. House Republican Study Committee, which is described as a policy development group for the more right-wing Republican members of the House, issued a memo on copyright reform.  I admit that I flipped out when I saw it; I have never been a one-issue voter, but this memo had the potential to turn me into one.  As I will explain in a minute, it was, in my opinion, a home run explanation of what is wrong with copyright policy making in the US today and what solutions would look like.  It suggested a radical rethinking for the Republican party about its approach to intellectual property law, one that is actually much more in step with other values that are supposed to motivate the GOP.

Alas, however, it was not to be.  By Saturday afternoon the Executive Director of the Study Committee issued a memo withdrawing the document and apologizing for its release.  His explanation was that it had not been properly “reviewed,” but the far more convincing explanation, offered by Mike Masnick of TechDirt, was that the lobbyists for the recording and motion picture industries had gotten very busy very fast and threatened the GOP with the loss of lots and lots of campaign donations if the memo was not retracted.

There are two ironies in this story of realpolitik over reason.  The first is that Derek Khanna, who wrote the memo, really got the analysis right.  The other is that sticking to their guns on the topic might have been a very fruitful political strategy for the GOP as they seek ways to capture the hearts and minds of younger voters.  The politics of the situation is nicely explained in this article from Slate (which has its own take on the politics, of course).  But I would like to look at the substance of the memo for a minute.

Khanna’s memo is built around three “myths” about copyright that he debunks.  The first is that copyright is primarily intended to compensate creators.  The second is that copyright is all about free market competition.  And his final myth is that the current copyright regime is the best way to foster innovation and productivity.  To me, it is very obvious that the first and three of these are, indeed, myths.  A quick glance at the clause authorizing Congress to pass copyright laws makes it clear that the purpose for those laws is the public good, and that compensation for authors is an instrumentality that is justified when, and only when, it helps serve that public purpose.  And there are few people who are not employed as lobbyists for the content industries who would seriously maintain that our current regime is actually the best balance of protection and access to foster innovation.  Take a look, for example, at this blog post by federal Circuit Court Judge Richard Posner for a concise explanation of how badly our current patent and copyright laws over-protect and therefore hamper productivity.

To me the real surprise was that this policy strategist even had to address the myth that says that “copyright is free market capitalism at work.”  Copyright, of course, is a government-granted monopoly; the very antithesis of free-market capitalism, it is government protectionism.  Indeed, one of the reasons that copyright has to be so carefully balanced is because it can so easily have anti-competitive effects, lending government support to favor one industry or type of business over others.  This is why our anti-trust laws have actually developed special rules to address claims of restraint of trade in industries that rely on the artificial monopoly conferred by IP laws.

For a political party that advocates free-markets and the reduction of government interference, copyright reform seems like a perfect platform, especially as it would appeal to younger, tech-savvy voters. It was refreshing to see that at least one policy strategist recognize that this was not just an important issue, but one that could capture the attention of the electorate.  Unfortunately, the strength of the content-industry lobbyists appears to have won out, at least for now.  But I am hopeful that the Republicans, and maybe even Democrats (although they have traditionally been even more dependent than the GOP on money from Hollywood) will continue to wake up to the fact that our current laws create an artificial thumb of the scale of markets in the US, putting the government in the position of supporting legacy content industries at the expense of new and innovative businesses.

Is the Web just a faster horse?

On Monday the Duke Libraries celebrated Open Access week with a talk by Jason Priem that was ostensibly about alternative metrics for measuring scholarly impact – so-called AltMetrics.  Jason is a Ph.D. student at the University of North Carolina School of Library and Information Science, a co-author of the well-regarded AltMetrics Manifesto, and one of the founders of the Web tool called ImpactStory.  In addition to his public talk, Jason gave a workshop for a small group of people interested in how ImpactStory works; you can read about the launch of that tool in this article.

So Jason is as qualified as anyone to talk about AltMetrics, and he did so very well.  But his talk was much broader than that subject implies; he really gave a superb summary of the current state of scholarly communications and a compelling vision of where we could go.  For me the most memorable quote was when he said, toward the end of his talk, that the Web had been created to be a tool for scholarly communications, yet while it had dramatically changed many industries, from bookselling to pornography, it had not yet revolutionized scholarly publishing as it should.  The problem is that publishers, and, to some extent, authors, are treating the Web as simply “a faster horse” and not truly exploiting the possibilities it offers to change the way scholarship is done.

Jason began with some history, pointing out the the earliest forms of scholarly communications were simply direct conversations, carried out through letters.  The first scholarly journals, in fact, were simply compilations of these sorts of letters, and you can still see that venerable tradition reflected in some modern journal titles, like “Physical Review Letters.”  But the modern journal, with its submission process, peer review, and extremely delayed publication schedules, was a revolution and a dramatic re-visioning of how scholars communicated with one another.  Certainly there were significant gains in that new technology for communication, but things were lost as well.  The sense of conversation was lost, as was immediacy.  Now, according to Jason, we have the ability to recapture some of those values from the older model.

Another dramatic change in scholarly communications was the movement called “bibliometrics,” which led to the creation, in the early 1960s, of the citation index and the journal impact factor.  Like the journal itself, the impact factor is so ingrained in our current thinking that it is hard to remember that it too was once a new technology.  And it is a system with significant problems.  As Jason said, the impact factor can track only one kind of person, doing one kind of research by making one kind of use.  The impact factor cannot track non-scholarly uses of scholarly works, or even scholarly uses that are not reflected in another journal article.  Also, true social impact , the kind of policy-changing impact that many scholars would see as an important goal, is seldom reflected in an impact factor.  The problem we face, Jason argued, is that we have confused the kind of use we can track with use itself.  In the process we often miss the real heart of scholarship, the conversation.

In the digital age, however, we can begin to track that conversation, because much of it is taking place online. AltMetrics, by which we teach computers how to look for a variety of article-level citations and discussions, offers the chance to analyze the scholarly environment much more thoroughly, and give individual scholars a much clearer and more comprehensive story of their real impact.

One connection that was hard for me in Jason’s talk, but ultimately persuasive, was his discussion of why Twitter is important.  I admit to being a reluctant and unenthusiastic Twitter user.  This blog post will be distributed via Twitter, and most of my readers seem to find what I write that way.  But still I was startled when Jason compared Twitter to that earliest form of scholarly communications, the conversation.  What was new to me was to think of Twitter as an opportunity to have a preselected jury suggest what is important to read.  If I follow people whose work is interesting and important to me, and they are all reading a particular article, isn’t it extremely likely that that article will be interesting and important to me as well?  And isn’t that peer review?  We sometimes hear that peer review is professional evaluation while Twitter is merely a popularity contest.  But Jason challenged that distinction, pointing out that if we follow the right users, the people whose work we know and respect, Twitter is a scholarly tool in which popularity becomes indistinguishable from professional evaluation.  Since many scholars already use Twitter, as well as blogs an e-mail lists, in this way, it is fair to say that new forms of peer-review have already arrived.  The AltMetrics movement aims to track those other forms of scholarly impact.

Jason ended his talk with a proposal to “decouple” the scholarly journal, to recognize that journals have traditionally performed several different functions — often identified as registration, certification, dissemination and archiving.  Some of those functions are now trivial; why pay anyone for dissemination in an age when an article can be distributed to millions with the click of a mouse?  Other functions, especially certification (peer-review) are changing dramatically.  Jason suggested that peer-review should be a service which could be offered independently of how an article was to be disseminated.  Scholarly societies especially are in a good possession to provide peer-review as a service for which scholars and their institutions could pay when it was felt to be necessary.  but in an age when so much peer-review is already happening outside the structure of journal publication, it is clear that not all scholarship will require that formal service.  So in place of the rigid structure that we have now, Jason suggests, illustrates, and enables a more flexible, layered system of scholarly networks and services.

As should be obvious by now, I found Jason’s talk for Open Access Week provocative and thought -provoking.  I hope I have represented what he said fairly. I have tried to indicate where I am paraphrasing Jason directly, and he should not be blamed for the conclusions I draw from his comments.  But for those who would like to hear from Jason directly, and I highly recommend it, he and several other leaders in the area of AltMetrics will take part in a webinar sponsored by NISO on November 14, which you can read about and register for here.  You can also finds slides and a video from a presentation similar to the one he gave at Duke here.

Coming clean on technological neutrality

It is not a case that draws much attention from higher education circles, but the case of WNET et al. v Aereo has drawn an amicus brief that should worry anyone who is interested in how copyright law serves or inhibits innovation and competition.  What is most disturbing is that the gross misunderstanding of how copyright law is supposed to work has come from former U.S. Registrar of Copyright Ralph Oman, now a “Professional Lecturer” in IP at the George Washington law school.

The case is brought by television broadcasters against a service that offers to take broadcast TV off the air and stream what it receives to its online subscribers.  I have no opinion on the merits of the case, although it is interesting to note, as this outraged post about Oman’s brief does, that the service was carefully structured to try to remain within the boundaries of the law as the courts have interpreted it.  And, as someone who has long wanted to “cut the cable,” I might be interested in the Aereo service if it is upheld.  But it is not the arguments or the merits that should concern us, it is the approach that former Registrar Oman takes in his brief on behalf of the plaintiffs.  Here is the central quotation that is so troubling:

Whenever possible, when the law is ambiguous or silent on the issue at bar, the courts should let those who want to market new technologies carry the burden of persuasion that a new exception to the broad rights enacted by Congress should be established. That is especially so if that technology poses grave dangers to the exclusive rights that Congress has given copyright owners. Commercial exploiters of new technologies should be required to convince Congress to sanction a new delivery system and/or exempt it from copyright liability. That is what Congress intended.

This is an extraordinary statement, suggesting that the Copyright Act was intended to force all innovators to go to Congress before beginning any service that might threaten some established form of exploiting the rights of copyright holders.  It is a recipe for economic suicide in a digital world, apparently willing to sacrifice the gains we can make through rapid innovation, new markets, and online opportunities to the goal of protecting the legacy industries from any need to adjust their business models.

More to the point, the law is not silent or ambiguous on how new technologies and business that are built on them should proceed.  Registrar Oman starts his brief by stating that,

There can be no serious dispute as to whether rights under the Copyright Act are broad, subject only to specific, narrow limitations enacted by Congress and that new developments in technology are not supposed to be able to truncate those rights.

To make this statement, Registrar Oman conveniently forgets fair use, which is not mentioned once in his brief.  Now I do not know if a serious fair use defense was raised on behalf of Aereo (they are apparently in the Appeals Court over the refusal of the District Court judge to grant a preliminary injunction).  But fair use is in our law, even if Registrar Oman (who helped draft the 1976 Act) apparently does not like it.  In fact, it is the first of the exceptions to the exclusive rights that is listed in the Act, and it is not specific and narrow.  It is deliberately open-ended and flexible, giving our law space to accommodate new technologies (like the VCR) and even new business models (like Google).  Apparently Registrar Oman wants to ignore fair use, and would prefer to force each of those new technologies and businesses to go to Congress for explicit permission before offering their products and services.  How much economic growth and how many new jobs would be lost if Congress and our courts took this approach?

The real problem with this brief, or perhaps the place where it is most honest, is that it seems to completely jettison the argument that copyright law should be technologically neutral.  It is ironic because the content industries often use that assertion to suggest that precedents from the print environment should be applied, when they are favorable to those industries, directly to digital activities.  For example, the plaintiffs in the Georgia State case used the claim that copyright is supposed to be technologically neutral to support its analogy between what GSU was doing with electronic reserves and the earlier cases about printed coursepacks.  In fact, the plaintiffs in that case even complained that the judge had applied a different standard to digital copies, which was nonsense.  At that point, technological neutrality was a battle cry.  But now, when such neutrality would seem to favor Aereo, Registrar Oman argues on behalf of the rights holders that every new technology should be treated differently, presumed to be illegal, and subjected to legislative scrutiny before being allowed.  What has happened?

The most obvious difference is that in the GSU case, the rights holders thought that the argument about technological neutrality would help them.  They hoped the Judge would ignore the distinction between the coursepack cases, in which a commercial intermediary was involved, and the entirely non-profit, educational situation at GSU.  By asserting that “the same law” applied in print and in the digital realm, they were hoping to obscure a substantive difference behind a rhetorical smoke screen.  As I have noted before, talk of technological neutrality is often more of a rhetorical strategy than an actual commitment.  In the Aereo case, the law as it stands looks more favorable to Aereo’s business model, so this brief on behalf of the rights holders takes a new and shocking tack by asserting that every new technology should be presumed to be infringing until proven otherwise.  Registrar Oman seems to be saying that we should not even try to apply the current law to new situations, but to presume that new technologies are always bad things and make their proponents argue their case to Congress before any new innovation is allowed.

The problem, of course, is that the GSU plaintiffs were right; the law really is supposed to be technologically neutral.  The same law is supposed to apply in both the print and the digital realm.  Where the GSU plaintiffs were wrong was in the application, not in the principle they stated.  And where Registrar Oman is wrong is in suggesting that we cannot even try to apply the law to new technologies.  The presences of fair use in our statute is testimony to the intention to give US copyright law the flexibility to adapt to new technologies, so it is no surprise that the former Registrar can only make his argument by ignoring its existence.

Technological neutrality is something that copyright laws need to strive for, because it is what avoids the need to make large-scale changes in the law every time a new technology — the player-piano, the radio, or the video recorder — comes along.  But it is a goal from which the law often falls short.  As Tomas Lipinski argues in this article, the control that the content owners want to assert over digital uses has had the effect of undermining the principle of technological neutrality.  It is a compelling, and distressing, argument; if one needs additional evidence for it, one need look no farther than the radical proposal made by former Registrar Oman to treat each new technology with suspicion and presumptive condemnation.  Presumable Oman’s brief is just another strategy, made to assist a particular plaintiff and with no thought about the consequences his argument would have as policy.  Because those consequences would be disastrous.

Three things open access is not

Lots of news stories and emails flying around about open access in the past few weeks, and as I tried to think what theme might bring them together, I realized that I wanted to talk about three things that open access is not.  Here they are:

First, open access is not more prone to abuse than other types of publishing.  We hear a lot about “predatory” open access journals, and recently we have also heard a lot about fraud and retracted articles from traditional journals.  We need to connect the dots and realize that both systems can be abused, just as all systems devised by human agents can be.

Consider, for example, this story from the Chronicle of Higher Education about a researcher who apparently faked nearly 200 scientific studies.  The journals he published in were not top-of-the-line, but they were respectable, traditional subscription-based journals that libraries all over the world pay for.  The principal journal mentioned in the article, Anaesthesia, is published by Wiley-Blackwell and is part of the journal package that my library buys.  Were we the victims of a predatory subscription journal, or just of an increasingly slipshod system that often fails to live up to the claims made for top-notch editorial work?  Either way, the difference between the failures at Anaesthesia and those for which OA journals are sometimes blamed does not seem all that significant.

Indeed, reform of the system by which scholars are evaluated and rewarded is exactly the recommendation of this New York Times article about the rise in scientific retractions, and that accords nicely with some of the changes OA journals can facilitate.  After detailing the really alarming rise in retractions, the article quotes suggestions that one way to combat fraud is to stop evaluating scholars simply on the number of papers they publish and the reputations of the journals in which they are published.  They do not directly address the question of why these journals, with all of their professional editorial staffs, do not initially catch the fraud, which is increasing at a faster rate than the number of scientific articles is.  But their recommendations point in the same direction the OA movement has been going for a number of years — away from a focus on impact factors, which only rate journals, and towards a more flexible and article-based set of metrics that actually relates to the specific article and to the nature of its impact.  More about this in a minute.

The second thing that open access is not is just one thing.  Recently I have seen a lot of debate about what is and is not open access.  Much of this debate has centered on the Finch Report in the UK, which recommended a rapid transition to open access publishing of research results, but put a heavy emphasis on gold OA, and specifically that subset of gold OA in which publishers are supported by article processing changes paid in advance of publication.  You can see one example of the debate this has caused here.

Note that I said that charging article processing fees is a subset of gold open access.  Much gold OA happens without such fees.  Some fully OA journals are simply supported by organizations, whether those are published by scholarly societies, by libraries (Duke’s are here) or by major funders such as those that support eLife.  Also, the new journal PeerJ is trying a different experiment in gold open access, financed by memberships.  Finally, the recent announcement about the SCOAP3 effort to flip the financing for all of the prominent journals in high-energy physics shows that radical new experiments somewhat different from those on which the Finch Report was focused can succeed.

The diversity of these forays into open access are important, and in my opinion there will be lots of different styles of open access for some time to come.  After all, the most effective effort at improving access to funded research in the US so far, the public access policy of the National Institutes of Health, relies to some extent, at least, on green open access, and never depends upon or requires the payment of article processing charges.  One of the reasons I especially like this new tool from SPARC, called “How Open Is It,” is because it allows one to evaluate openness on a variety of factors, and seems to recognize that openness is an ambition and a process, not one specific definition.

Failing to keep the diversity of open access efforts in mind leads to some unfortunate conclusions.  In a recent statement the American Historical Association “voiced concern” that open access would not work for the humanities and social sciences the way it has for the natural sciences.  This is a common complaint, but the myopia on which it is based is especially obvious in the AHA statement, which cites the Finch report and is focused entirely on “author pays” models of open access.  Nowhere does the AHA statement consider that green open access would produce significant benefits for historians without the difficulties about which the AHA is wringing its collective hands.  Perhaps this failure to see the whole picture is part of the reason that the AHA’s own flagship journal, the American Historical Review, does not facilitate author self-archiving, allowing only the pre-print versions of articles to be made accessible, which is nearly never an acceptable option for historians.   If the AHA really wants to keep up with the inevitable future, it needs first to change the policies at AHR.  The results, I predict, would help facilitate the transition to other models for history publishing.

The third thing that open access is not is just a business model.  In all the debates about which form of OA is best and how each form can be financed, we can lose sight of the fact that more than how we divide up a pot of money is at stake here.  Open access is also a statement about the values of scholarship; an attempt to introduce more transparency into the process of research and to encourage greater participation in its creation, financing, and evaluation.  Which gets me back to better metrics for assessing the quality and impact of scholarship.  The movement called AltMetrics is one of the most exciting thing about open access; it is a chance to use new tools to study the impact of specific articles on a more granular level, yet across a much wider field.

At Duke, we are very excited to have one of the pioneers in the move toward AltMetrics, Jason Priem, who co-founded the ImpactStory project, coming to speak in our Library for Open Access week.  Jason will talk on October 22 on “Altmetrics and the decoupled journal: an endgame for Open Access.”  His talk is described here.  Jason’s lecture is open to the public, and we hope many people — anyone with an interest in where the future of scholarship is headed and how we can break out of the cumbersome and unreliable system of evaluation that has accreted over the years — will attend.

A big win for fair use and libraries

What a stretch of four months it has been!

First, in July, we had the ruling in the Georgia State lawsuit affirming that most of the excerpts for teaching that were challenged as copyright infringement were actually fair use.  The decision is being appealed, but libraries go in to that appeal on the winning side, which is always the better place to be.

Then, last  week we learned that the lawsuit filed against UCLA over digital streamed video had been dismissed for the second time.  The dismissal, so far, is only noted in the minutes taken by the Clerk of the Court during a hearing, which are filed in the docket.  So we do not know why the case was dismissed, although it seems to be the same reasons the first one was — because the groups bringing the suit either do not own the rights they are trying to enforce or already granted a license broad enough to cover the use they now object to.  More importantly, we do not yet know if the dismissal was with prejudice or without, meaning whether or not the plaintiffs will be allowed to refile the case again.

But tonight is the big news, in the lawsuit brought against the HathiTrust and five of its university partners by the Authors Guild, several foreign associations similar to the AG, and a handful of named authors, including Fay Weldon and J.R. Salamanca.  In that case, Judge Harold Baer of the Southern District of New York has ruled on various motions for judgment on the pleadings and summary judgment, and he has completely vindicated Hathi and the libraries.

A copy of the ruling is here.

A lot is happening in this opinion, which I will try to sort out briefly.

First, the plaintiff’s motion asking the Court to rule that fair use was not available to the libraries as a defense, alleging that because libraries had the benefit of the specific section 08 exceptions they should not also be able to assert fair use, was denied.  The Judge spent comparatively little time on this ridiculous claim, which flies in the face of the clear language of section 108 itself.  He entirely rejected the argument and dismissed the claim.

Second, the Judge ruled on a motion for judgment on the pleadings (also called judgment as a matter of law) raised by the defendants (the libraries and HathiTrust) that asserted two things.  First, it argued that the copyright law does not allow associations like the Authors Guild to sue on behalf of its members.  On this argument, the Judge agreed as to the US association (the Authors Guild), but held that he would not foreclose a foreign association from bringing a suit because of our treaty obligations.  The second argument was that the whole issue about the Orphan Works project proposed by HathiTrust was not “ripe” for a decision because nothing had been done as of yet.  The Judge declined to decide that there was a risk of infringement until and unless the project got underway.  So on this motion, the defendants won two-thirds of their arguments and the one they lost, about associational standing for foreign organizations, turns out not to matter.

Because the real news is that the Judge granted the defendants motion for summary judgment on the issue of fair use.  Judge Baer found that the arguments made that the mass digitization project and the Hathi Digital Library was fair use were so strong that he made the decision without a trial, on summary judgment.

The Judge did a four factor analysis to arrive at his conclusion, and it is worth looking at that analysis.

On the first factor, Judge Baer first held that the purpose of the use was research and scholarship, which are favored in the fair use statute.  But he went on to hold that the use of these copyrighted materials in HathiTrust was also a transformational use.  Unlike Judge Evans in the GSU case, Judge Baer cited case law that has determined that a use can be transformational because it has a different purpose, not only when an actual change in content has been made.  And providing a searchable database of books, within copyrighted works only available to the visually-impaired, was, in the Court’s opinion, transformative.

Because the purpose was transformative, the Judge held that the second factor about the types of works involved, was not dispositive, since transformative works of even very creative productions (like “Oh, Pretty Woman”) have been held to be fair.

On the third factor, the Judge said that the amount used, even where entire works ware scanned, was reasonable to accomplish the purpose.

Finally, the Court found that the issue of market harm did not undermine fair use.  There is no direct market competition between the HathiTrust and any existing market for the works in question.  And as for the argument that HathiTrust foreclosed a potential licensing market, the Judge said that rights holders were not allowed to preempt transformative uses because of a potential market.

At the end, Judge Baer concludes with this sentence:

I cannot imagine a definition of fair used that would not encompass the transformative uses made by the defendants and would require that I terminate this invaluable contribution to the progress of science and the cultivation of the arts that at the same time effectuates the ideals of the ADA.

As the last part of this comment indicates, the Judge also upheld the provision of digital files to persons with visual disabilities to facilitate adaptive access, using a combination of fair use and section 121of the copyright law.  Hard to believe that the AG thought it was a good idea to challenge that practice, but they did.  So overall this is a comprehensive win for the libraries and for the important public interest that they serve.

This opinion follows a clear line of reasoning in fair use cases over the past three decades, and it applies that reasoning squarely to library services.  I have bemoaned these lawsuits in the past, but I have to admit that I am beginning to feel grateful for them.  The string of opinions that is now taking shape ought to give librarians a great deal more confidence when they are making reasonable applications of fair use.  Where once I feared a chilling effect, I am now sensing a warming glow.

Questions about the AAP Google settlement

There have been quite a few stories about the news that Google has settled with the Association of American Publishers in its long-running lawsuit alleging copyright infringement in the Google Books project.  There are stories in the NY Times, Chronicle of Higher Education and Inside Higher Ed.

As all of the stories note, the terms of the agreement are confidential, so we really do not know much about what has happened.  So there is very little to say, and most of it is grounded on uncertainty.  To a sensible person that would suggest reticence, yet here I am writing a blog post about the settlement.  My modest purpose, however, is to address three questions that I have heard, in order to lower expectations about what has happened.

First, How is this settlement related to the earlier one that caused such a fuss and was ultimately rejected by the Judge?

The answer here is that it is not really related.  That proposed settlement was a “class” agreement that would have bound essentially all authors who owned a copyright.  It was because of its sweeping scope that it had to be approved by the Judge, and because that scope was too sweeping that Judge Chin rejected it.  This settlement, on the other hand, is just between the named plaintiffs — several large publishing houses — and Google.  It makes the publisher half of the lawsuit go away (the Authors Guild is continuing with their part of the litigation), and it sets terms for how Google will deal with books from those publishers.  It does not have any impact on relations between Google and any other authors or publishers, although it acknowledges what has really been the case all along — that Google will withdraw materials if a publisher asks and will enter in to agreements with individual publishing houses.

Second — What does this settlement say about fair use?

As far as we know, the answer is that it says nothing about the key issue in the lawsuit — whether or not the Google digitization for the purpose of creating an index and providing a “snippet view” of in-copyright books is fair use.  That issue is still being litigated, with the Authors Guild as the plaintiff.  For the publishers, there apparently is some language about how much of a book Google can display; it is more than mere snippets, but it is not clear to me how that permission relates to other provisions of the settlement.  In any case, the settlement does not, and cannot, decide the fair use assertion.  In terms of fair use, we know no more today than we did yesterday, before the settlement was announced.

Third, how is this settlement related to the Georgia State decision?

Once again, the answer is that there is no direct relationship.  The settlement is not really about fair use, but rather avoids the question that was addressed directly in the GSU case.  The only relationship is that the AAP, which was a plaintiff in the Google case and has now settled to avoid continued litigation, is also bankrolling half of the litigation costs in the ongoing GSU case.  One could speculate that the the AAP decided to settle with Google because it wanted to avoid another case where fair use was affirmed, or that the cost of the GSU case, especially after the plaintiffs have been order to pay nearly 3 million in fees and costs incurred by GSU and where an appeal has just been filed, caused the AAP to decide on a strategic retreat from the other, probably more expensive, case.  But both of those possibilities are pure speculation.  In fact, the length of time that settlement negotiations usually take make it unlikely that recent events had a direct causal relationship to this settlement, but it is certainly possible that the overall trajectory of the GSU case over the past three months was a partial motive for the AAP move as quickly as possible to settle with Google.

In short, I think there is very little to see here.  I am glad when litigation is settled, as a general principle, because I think agreements are usually better, and more efficient, than lawsuits.  And the educational benefits of the Google digitization efforts are important, so I am glad a little of the pressure has been taken off of those efforts.  But this news of a settlement in the publisher half of this lawsuit is worth only a smile and a nod of approval, I think, before we turn back to the day-to-day struggles of higher education in a digital age.

The six million dollar fair use standard

The trial judge in the Georgia State copyright infringement lawsuit filed her final judgment in the case yesterday, bringing that portion of the lawsuit to a close.  The only news left for this final order was the amount of money that the plaintiff publishers would be forced to pay to Georgia State.  Judge Evans had already ruled that GSU was “the prevailing party” and therefore entitled to have the other side pay their fees and costs, and a lot of motions were filed arguing over what those numbers would be.  The final amount (including both attorney fees and court costs) is $2,947,085.10.

Based on the way these things usually turn out, Georgia State and its lawyers probably think this number is too low.  The publishers and their lawyers undoubtedly think it is too high.  The result is that it is probably a reasonable estimate of what it actually cost one side to litigate this case.  So if we want to estimate how much it cost to get us where we are today, that estimate would be about six million dollars, assuming that the plaintiffs spent approximately the same amount that they must now pay to the defendants.

And please do not forget; half of that money is coming from the Copyright Clearance Center, a corporation that says it acts on behalf of scholarly authors and to whom universities pay large sums of money.  Our own dollars — lots of them — are being used to bring this lawsuit to squeeze more dollars out of us.

The result of all this expenditure, thus far, has been a judgment that gives us what the publishers originally said they wanted.  The Judge has articulated a reasonable fair use standard.  It is too rigid for my taste, and too permissive for the publishers, but it is not unreasonable.  If it was left to stand, nothing much would change, except that we would have a greater degree of confidence if we confined our fair use assertions regarding the provision of digital course readings to the boundaries Judge Evans has set.  Doing that would cause some institutions to pay less for permission fees, and some to pay more.  Some schools might push the boundaries a little further, but we probably would not see the widely different interpretations of fair use in this context that we sometimes find today.  Other schools would still find that this risk is too great and continue to stay out of the e-reserve business all together.

To my mind this would not be a terrible result, although it would have cost a terrible amount of time and money to get there.  We could all probably live in a world where Judge Evans’ ruling was the last word in this case.

Unfortunately, the publishers are unwilling to live in such a world; they have already announced their intent to appeal, and they now have thirty days to file that appeal in the 11th Circuit Court of Appeals.  The only excuse for their decision is the desire to force universities to pay even more money than the already do to publishers.  Prices are not rising fast enough, apparently, so greater income from permissions is required.  If other parts of the educational mission of universities have to suffer, that too is price the publishers seem willing to pay.

We can no longer preserve the illusion that all this was about was to provide some certainty about fair use for digital course content.  The publishers spent 6 million and now could walk away with a workable, if unpopular, standard.  Instead the battle against universities and higher education will continue.  How sad.

How do you recognize a catastrophe?

When a financial analyst uses the term “catastrophic” in regard to the impact of open access on publishing giant Reed Elsevier, it is bound to attract some attention.  In a September 10, 2012 report from the firm Sanford Bernstein, senior analyst Claudio Aspesi did just that, and the reaction has been predictable.  “Catastrophe” has been in every headline (including mine!), from those that seem a little gleeful at the idea of bad things happening to the publishing company so many academics love to hate, to those that wish to defend the conglomerate.

Thanks to the kind permission of Mr. Aspesi, the complete report can be accessed here.

It is worth noting how the word is actually used in Aspesi’s report.  In assessing the overall prospects for Reed Elsevier, Aspesi notes that they are very dependent on the high profit margins at Elsevier, the academic publishing arm, so that a reduction in those profits might be “catastrophic” for the parent company.  He then goes on to explain why open access, specifically gold open access with the anticipated shift from subscription-based revenues to article processing fees (APCs), seems to threaten that high profit margin.  In the past, Aspesi has recommended that Reed Elsevier should spin off other parts of its company to reduce its exposure; a call that the company has rejected.

The bottom line for investors in the Berstein report is that the company rates Reed Elsevier to “underperform,” and pegs the target stock price at $6.45.  As this comparison report shows, other analysts put the target stock price somewhat higher.

To me the most interesting part of the report is its discussion of where article processing fees for gold open access journals are likely to go.

The fundamental conclusion in Aspesi’s report is that for Elsevier to maintain its current revenue through a complete transition from subscriptions to APCs — for that transition to be “revenue neutral” for Elsevier — the APCs charged to each author for publishing in an Elsevier journal would have to increase about 70%,  from current levels (for open access) at about $3000 per article to about $5,145 per article.  Aspesi does not think the market would bear this increase, and in that I think he is surely correct.

This discussion of APCs is fascinating, and there are several nuances we can add to it.  First, Aspesi notes that an option for Elsevier to maintain revenue in the OA transition without raising APCs this high would be to publish more articles.  The move to online certainly makes this possible, but the option has significant drawbacks, especially for a firm that depend heavily on its “brand” (as do university tenure and promotion committees).  Another possibility would be to recognize that not all journals are equal and charge differential APCs, based on the desirability of publishing in that particular title.  This raises the possibility that scholars would begin to shop around for journals with lower APCs for those articles that are not accepted into the top-tier journals.  We have not had much competition in the area of scholarly publishing in the past, but as different publishers charge different APCs, and those fees might even, perhaps, vary within one publisher’s titles, this possibility of competition, where cost and reputation are balanced together in the decision about where to publish, could grow.

In explanation of his conclusion that even with higher APCs it would be difficult for Elsevier to maintain its high revenues, Aspesi cites a simple fact — in the OA world, journal bundling would no longer be possible.  At present, Elsevier and others are able to sell lower-readership journals as part of “big deals” and at no additional cost to the company.  The OA world, however, imagines that each journal would have to stand on its own bottom, with revenue dependent on how desirable publication in that particular title is.  Aspesi’s discussion of what OA “big deals” might look like — various models for offering universities better pricing or “all you can eat” agreements for open access publishing — is fascinating but concludes that it would be very difficult to reach such deals and that they would not solve the problem of needing to increase APCs to cover costs.

One additional problem with increasing APCs for Elsevier journals, which returns to the issue of competition among journals, is the likelihood of downward pressure on APCs just as the transition would seem to force established traditional publishers to raise them.  The growth of OA megajournals like PLoS One, which already charges a much lower APC for articles it publishes than Elsevier does for its “sponsored articles,” may create such downward pressure.  So might some of the new models for open access, like eLife, which is supported by research funders and will not charge APCs, or PeerJ, which is building a membership model for OA.

The key to this analysis, and the problems it poses for Elsevier — whether catastrophic or not — is the cost per article.  Aspesi bases all of his projections on the assumption that it costs about $3,200 for Elsevier to publish each article, or, more significantly, $1200 per article submitted (850,000 submitted articles versus 316,000 published).  Many fully OA journals do not have such high costs, which is the reason they can charge lower APCs.  Aspesi acknowledges that Elsevier may be an expensive organization to run (not run on a “shoestring,” he says), but does not explore how or if those costs per article could be reduced.  Yet it seems to me that that is the key for survival for traditional publishers if they are to weather the transition to open access.

We can experiment with as many different pricing and funding models as we like, and I hope there are a lot, but we fundamentally need to have a conversation about how much it costs to publish a single scholarly article.  Those costs have been hidden in the byzantine structure of subscription prices and bundling, especially from the larger publishers.  That model has allow great opacity about actual costs, and for years universities have accepted that situation, which has now become clearly unsustainable.  But as we make the transition to open access, and gold OA plays a large part in the new paradigm, it will be harder and harder to avoid being open and honest about how much scholarly publishing really costs, and how some groups seem to be able to do it for so much less.

Discussions about the changing world of scholarly communications and copyright