In US law, copyright magically comes into existence when someone creates an original work of authorship. The copyright is actually based in the original expression contained in the work, not whether there was effort involved in the work’s creation. So for the specific arrangement of words (+/- images) that make up a particular journal article, there is definitely a copyright in the final (post-review) manuscript – but subsequent changes that do not substantially alter the expression in that manuscript do not themselves create a new copyright.
So when an author transfers the copyright “to an article”, they unquestionably transfer the copyright in the final manuscript -and- the copyedited, typeset version of that manuscript. The original expression in those two versions of the article is -one and the same-, and thus, so is the copyright.
I have personally gotten pushback on this point, and seen public disagreements on this point, from both authors and publishers. The substance of publishers’ disagreement has usually focused on the value that they add to the author’s contribution – but the absence of a copyright in a particular product or activity does not equate to an absence of value. Copyrightability is not a commentary on value, it’s reflective of -original expression-. Publishers add value (varying amounts, depending on publisher, sometimes) through managing peer review, copyediting, typesetting, and other services – but unless they contribute significantly different and new authorial expression, these valuable services don’t have any relation to copyright.
Most of the authors I’ve seen contesting the idea that the copyright in a final manuscript and the typeset published article is the same, on the other hand, have had a different disagreement – they’re pretty sure they’re allowed to post their final manuscript online, but that the typeset version is in the exclusive purview of the publisher. In many cases this is a true statement, but not because those different versions of the article present different copyrights, with the copyright in the manuscript belonging to the author and the copyright in the typeset version belonging to the publisher. Authors who have transferred their full copyright to their publisher are often allowed to post final manuscript versions – or sometimes, pre-peer-review manuscript versions – by publisher policy, or by a specific clause in the publication agreement.
Where neither publisher policy nor any extra publication contract clauses allow posting manuscript versions of an article, authors who have transferred away their copyrights do not have the right to post those versions. (Except to the extent that any other non-owner individual might have the right to copy or post the article, fair use or otherwise.)
This seems to be a controversial statement to many people who read Kevin Smith’s blog post last week (which was itself partially in response to the surprise of authors who’ve received takedown notices about self-posted copies of articles in the last few months.) I said on Twitter, and say here again, that there seems to be an element of wishful thinking in this ‘controversy’ from authors who have not been paying much attention to their publication contracts until recently (or at all.) If the manuscript and the typeset version of the article are different copyrights, then they haven’t been transferring -tons of control over their works- to publishers.
But if that’s true, there’s no reason for SHERPA/RoMEO to exist. There’s no reason several authors I’ve talked with have received quotes for the permissions fees needed to include their own previously published works in (newly typeset) retrospective volumes. There’s no reason why, when an author wants to re-use a figure from an earlier publication, most publishers require authors to get permission from the publisher of the earlier publication.
The details of publication contracts matter.
(While I love the resources that library folk at UMN and elsewhere have developed to help academic authors understand what they own and what publication contracts do, I also encourage authors to consider some of the resources available for/by freelancers who work on other aspects of content production. From certain perspectives, academic authors look a lot like freelancers producing content for publishers. Katie Lane at Work Made For Hire has some eminently readable, and relevant, stuff – check out points 2 and 3 of this post on contracts, for example.)
Obviously, this post is a response to the response to Kevin Smith’s piece
from last week in which he addressed longstanding confusion around copyright in
different versions of a published article, and in what kinds of things
authors can do with different versions of their articles. I had a really interesting exchange on Twitter this morning with Charles Oppenheim, who wrote a detailed response to Kevin’s post, arguing that Kevin is incorrect. I still agree overall with Kevin’s analysis, which he clarified in another post today – but I do agree that at some point, different versions of a work can be
legally separate entities. I’m not sure that point is where Charles thinks it is – that is, I’m not sure that the pre-review draft has a separate copyright than the version that contains post-review edits (in his article, he uses the useful shorthand of “D” for draft, and “F” for final manuscript.) If the pre-review draft has a separate copyright, why do some publisher self-archiving policies distinguish what can be done with “D” and “F” (subsequent to a transfer of “the copyright” in an article)?
I think the disagreements of these thoughtful commentators really lodge in the nuances of rights in works developed
over time, through major revision. If copyright arises at the “moment” of creation, what are the relevant moments? In law school, I had a computer that shut down unpredictably, so I developed a “Ctrl-s” reflex. I’m pretty sure there are not 693 different copyrights in my 3L independent study paper, but I’m also pretty sure that some of my drafts contained entirely unrelated expression (yay starting over partway through) and thus would have different copyrights.
These nuances are something I do want to dig in further on; hopefully in another blog post soon (hence, “part I” above). The way
copyright law deals with issues like the “moment” of
creation, what constitutes a derivative work, and “substantial similarity” is not very
responsive to the actual processes of human creativity. Not that this is
news, really; the relationship between copyright and reality is fraught