According to today's SHERPA/RoMEO statistics, 36% of the 308 included publishers are green ("can archive pre-print and post-print"), 24% are blue ("can archive post-print (i.e. final draft post-refereeing)"), 11% are yellow ("can archive pre-print (i.e. pre-refereeing)"), and 28% are white ("archiving not formally supported"). Looked at another way, 72% of the publishers permit some form of self-archiving.
These are certainly encouraging statistics, and publishers who permit any form of self-archiving should be applauded; however, leaving aside Creative Commons licenses and author agreements that have been crafted by SPARC and others to promote rights retention, publishers recently liberalized author agreements still raise issues that librarians and scholars should be aware of.
Looking deeper, there are publisher variations in terms of where e-prints can be self-archived. Typically, this might be some combination of the author's Website, institutional repository or Website, funding agency's server, or disciplinary archive. Some agreements allow deposit on any noncommercial or open access server. Restricting deposit to open access or noncommercial servers is perfectly legitimate in my view; more specific restrictions are, well, too restrictive. The problem arises when the agreement limits the author's deposit options to ones he or she doesn't have, such as only allowing deposit in an institutional repository when the author's institution doesn't have one or only allowing posting on an author's Website when the author doesn't have one.
Another issue is publisher requirements for authors to remove e-prints on publication, to modify e-prints after publication to reflect citation and publisher contact information, to replace e-prints with published versions, or to create their own versions of postprints. Low deposit rates in institutional repositories without institutional mandates suggest that anything that involves extra effort by authors is a deterrent to deposit. The above kinds of publisher requirements are likely to have equally low rates on compliance, resulting in deposited e-prints that do not conform to author agreements. To be effective, such requirements would have to be policed by publishers or digital repositories. Otherwise, they are meaningless and are best deleted from author agreements.
A final issue is retrospective deposit. We can think of the journal literature as an inverted pyramid, with the broad top being currently published articles and the bottom being the first published journal articles. The papers published since the emergence of author agreements that permit self-archiving are a significant resource; however, much of the literature precedes such agreements. The vast majority of these articles are under standard copyright transfer agreements, with publishers holding all rights. Consequently, it is very important that publishers clarify whether their relatively new self-archiving policies can be applied retroactively. Elsevier has done so:
When Elsevier changes its policies to enable greater academic use of journal materials (such as the changes several years ago in our web-posting policies) or to clarify the rights retained by journal authors, Elsevier is prepared to extend those rights retroactively with respect to articles published in journal issues produced prior to the policy change.
Elsevier is pleased to confirm that, unless explicitly noted to the contrary, all policies apply retrospectively to previously published journal content. If, after reviewing the material noted above, you have any questions about such rights, please contact Global Rights.
Unfortunately, many publishers have not clarified this issue. Under these conditions, whether authors can deposit preprints or author-created postprints hinges on whether these works are viewed as being different works from the publisher version, and, hence, owned by the authors. Although some open access advocates believe this to be the case, to my knowledge this has never been decided in a court of law. Michael Carroll, who is a professor at the Villanova University School of Law and a member of the Board of the Creative Commons, has said in an analysis of whether authors can put preprints of articles published using standard author agreements under Creative Commons licenses:
Although technically distinct, the copyrights in the pre-print and the post-print overlap. The important point to understand is that copyright grants the owner the right to control exact duplicates and versions that are "substantially similar" to the copyrighted work. (This is under U.S. law, but most other jurisdictions similarly define the scope of copyright).
A pre-print will normally be substantially similar to the post-print. Therefore, when an author transfers the exclusive rights in the work to a publisher, the author precludes herself from making copies or distributing copies of any substantially similar versions of the work as well.
Much progress has been made in the area of author agreements, but authors must still pay careful attention to the details of agreements, which vary considerably by publisher. The SHERPA/RoMEO—Publisher Copyright Policies & Self-Archiving database is a very useful and important tool and users should actively participate in refining this database; however, authors are well advised not to stop at the summary information presented here and to go to the agreement itself (if available). It would be very helpful if a set of standard author agreements that covered the major variations could be developed and put into use by the publishing industry.