There is a persistent problem with polemics. When writing to address someone else’s position with which one disagrees, it is easy to lose sight of the proverbial forest for the trees.
In my previous two posts, I was addressing a misunderstand that I am afraid might lead authors to be less attentive and assertive about their publication contracts than they should be. The specific issue was whether or not it is feasible to maintain that a copyright is transferred only in a final version of a scholarly article, leaving copyright in earlier versions in the hands of the author. I argued that this was not the case, that the distinction between versions is a construct used by publishers that has little legal meaning, and that author rights that do persist in earlier versions, as they often do, are created by the specific terms of a copyright transfer agreement (i.e., they are creatures of a license). These points, which I believe are correct, prompted a number of people to get in touch with me, concerned about how these specific “trees” might impact the overall forest of self-archiving policies and practices.
So now I want to make several points that all address one conclusion; this argument about the nature of a copyright transfer does not necessarily have any significant impact on what we do to enhance and encourage self-archiving on our campuses. Most of the practices I am aware of already take account of the argument I have been making, even if they are not explicit about it.
On the LibLicense list today, Professor Steven Harnad, who is a pioneer in the movement to self-archive scholarly papers, posted a 10-point strategy for accomplishing Green open access. Essentially, he points out that a significant number of publishers (his number is 60%) allow authors to self-archive their final submitted versions of their articles, and that those who have retained this right should exercise it. Elsevier is one such publisher, about which more later. Harnad argues that there are other strategies available for authors whose copyright transfer agreements do not allow self-archiving of even the final manuscript. One option is to deposit the manuscript in a repository but embargo access to it. At least that accomplishes preservation and access to the article metadata, and it facilitates fulfillment of individual requests for a copy. Another option is to deposit a pre-print (the version of the article before peer-review) in a pre-print repository, which is a solution that has long worked well in specific disciplines like physics and computer science.
All of these strategies are completely consistent with the point I have been making about copyright transfer agreements. Harnad’s model recognizes that copyright is transferred (perhaps improvidently) to publishers, and is based on authors taking full advantage of the rights that are licensed back to them in that transaction. This makes perfect sense to me and nothing I have written in my previous two posts diminishes from this strategy.
One of the questions I have received a couple of times involves campus open access policies and how they affect, or are affected by, copyright transfers. These policies often assert a license in scholarly articles, so the question is essentially whether that license survives a transfer of copyright.
It is a basic principle of law, and common sense, that one cannot sell, or give away, more than one owns. So if an author has granted a license to her institution before she transfers her rights to a publisher, it seems clear that the license should survive, or, to put it another way, that the rights that are transferred to the publisher are still subject to this prior license. There was an excellent article written in 2012 by law professor Eric Priest about this situation, and his conclusion is “that permission mandates can create legally enforceable, durable nonexclusive licenses.” The article provides an extensive analysis of the legal effect of this “Harvard-style” license, and is well worth being read in its entirety by all who are interested in the legal status of Green open access.
An additional wrinkle to the status of a prior license is provided by section 205(e) of the copyright law, which actually addresses the issue of “priority between conflicting transfer of ownership and nonexclusive license.” This provision basically affirms what I have said above, that a license granted prior to a transfer of copyright survives the transfer and prevails over the rights now held by the transferee, IF it is evidenced by a written instrument. Because of this provision, some schools that have a license that is created by an open access policy also get a document from the author at the time of OA deposit that affirms the existence of that license. Such documentation helps ensure the survival of a policy-based license even after the copyright is later trnsferred to a publisher.
Even when we decide that a license for Green open access exists and has survived a copyright transfer, however, we still have a policy decision to make about how aggressively to assert that license. Many institutional practices look to the terms of the copyright transfer and try to abide by the provisions found therein, usually relating to the version that can be used and when it can be made openly accessible. They do this, I think, to avoid creating an uncomfortable situation for the authors. Even if legally that license they granted would survive the transfer of rights, if a conflict with the publisher developed, the authors (whom we are, after all, trying to serve) would be in a difficult place. So my personal preference is to conform our practice to reasonable publisher policies about self-archiving and to work with authors to get unreasonable policies changed, rather than to provoke a dispute. But this is a policy matter for specific institutions.
Finally, I want to say a couple of things specifically about Elsevier, since it was Elsevier’s take down notices directed against author self-archiving that began this series of discussions.
Elsevier’s policies permit authors to self-archive the final manuscript version of an article but not the published version, and, as far as I know, all of its take down notices were directed against final published versions on institutional or commercial websites. So it is true that in my opinion, based on the analysis I have presented over the past week, that Elsevier is legally justified in this take down campaign. It may well be a stupid and self-defeating strategy — I think it is — but they have the legal right to pursue it. Authors, however, also have the legal right, based on Elsevier’s policies that are incorporated into their copyright transfer agreements, to post an earlier version of the articles — the final author’s manuscript(s) — in place of these final published versions. So I hope that every time a take down notice from Elsevier that is directed against the author of the work in question is received, the article that is taken down is replaced by a final manuscript version of the same content.
As many know, Elsevier also has an foolish and offensive provision in its current copyright transfer agreement that says that authors are allowed to self-archive a final manuscript version of their article UNLESS there is an institutional mandate to do so. As I have said before, this “you may if you don’t have to but not if you must” approach is an unjustifiable interference with academic freedom, since it is an attempt to tie faculty rights to specific policies that the faculty themselves adopt to further their own institutional and academic missions. Elsevier should be ashamed to take this stance, and our institutions that value academic freedom should protest. But based on what has been said above, we can also see how futile this approach really is. If the institution has a policy-created license, that license probably survives the copyright transfer, as Eric Priest argues. In that case, the denial of a self-archiving right only in cases where a license exists is meaningless precisely because that license does exist; authors could self-archive based on the license and do not need the grant of rights that Elsevier is petulantly withholding. I said above that institutions should consider whether or not they want to provoke disputes by relying on the prior existence of a license to self-archive. Elsevier, however, seems to have decided to provoke exactly that dispute with this provision, and they are even more unwise to do so since it is likely to be a losing proposition for them.