Wednesday, February 20, 2008

Copyright and OA - Response to Stevan Harnad

Prompted by differences of opinion about the Harvard FAS policy, I want to clarify where Stevan Harnad and I agree and disagree about the relationship between copyright and open access.

I understand Stevan's position to be:

1. Open Access policies should conceptually separate a requirement to deposit an electronic copy of a post-peer-review manuscript in a repository from a requirement that the repository make that copy publicly accessible on the Web.

2. Deposit at the time the manuscript is accepted for publication should be unconditionally required.

3. Public access should be allowed any time the publisher's agreement says it may be.

4. If the publication agreement does not permit posting of the manuscript, a repository may still distribute copies by email whenever requested to do so by a user.

5. The combination of 3 and 4 effectively provide open access.

6. Those who argue that open access should also include an explicit public copyright license giving the public more than the right to read (e.g., the rights to republish or to translate or otherwise adapt the work) are mistaken. Either (a) these rights have already been implicitly granted by the public posting of the work; (b) they are not necessary to effective scholarly communication; or (c) even if they would marginally improve scholalry communication, the costs of negotiating copyright with publishers is not worth this benefit.

My response:

Points 1 and 2 are exactly right. Under U.S. copyright law, it is a fair use for an author to send, and for a repository to make, an archival copy of the post-peer-review manuscript. I think it's also a fair use to make an archival copy of the published version of the article. Copyright law in many other parts of the world also would deem this to be legal.

Deposit mandates are highly desirable. Please note that under the Harvard policy, even if a faculty author feels it necessary to seek a waiver of the copyright license to Harvard, there is no reason that author couldn't and shouldn't deposit a copy of the manuscript in the repository.

With respect to public access, I disagree that faculty authors should simply adapt themselves to the arrangements that publishers offer/demand. Moreover, I have have a different view about what those arrangements permit.

So, on point 3, I agree insofar as authors should use all legal rights they have to make their work freely accessible on the Internet. If a publication agreement gives the author the right to post the work in a repository, the author should do so and the repository should make the work freely accessible.

However, I don't think that the starting point for the analysis should be what the publisher's form says. I think authors have an obligation to consider whether signing the publisher's form is ethical behavior.

Copyright is an author's right granted to the author by the public to achieve a public purpose - the promotion of science and useful arts. (I realize that many have a natural rights view of copyright's purpose. I don't share that view.) With rights come responsibilities. Authors of scholarly journal articles do not need the promise of a royalty to have an incentive to perform research or report the results and their analysis of that research. So although the premise of one-size-fits-all copyrights is that authors need exclusive rights to be stimulated to create, that premise is largely false with respect to much scholarly research.

Journal article authors know that they will not receive a royalty nor will those who provide referee services. Instead, the progress of science and useful arts is driven by these authors' desire to achieve broad dissemination of their research. The Internet opens up a (not so) new avenue of scholarly communication. Thus, journal article authors have a duty to consider whether they are making proper use of the copyrights that the public has given them when they agree to the terms of a publisher's agreement that limit how, when or where the author may provide free access to their work on the Internet.

I have a different understanding about the legal consequences of number 4, and therefore I also do not agree with number 5.

As for number 6, clarifying re-use rights through public licensing is desirable. If his view is (a) or (b) I disagree. If his view is (c), however, I agree that the effort necessary to achieve this goal should be subject to cost-benefit analysis. Under current circumstances, where subscription-funded publishers have shown some willingness to permit free access to post-peer-review manuscripts but have not been willing to agree to public licensing, I think an author could responsibly decide to be satisfied with a copyright agreement that permits free access but does not provide for re-use licensing.

1 comment:

Stevan Harnad said...

Upgrade Harvard's Opt-Out Copyright Retention Mandate By Adding a No-Opt-Out Deposit Mandate: No Loss, Only Gain

Michael Carroll, Peter Suber and I are in complete agreement on every point of substance save one: What is the mandate that is the most likely to generate the most OA?

Michael and Peter (and Harvard!) think it is a Copyright Retention Mandate with opt-out (CRM). I think it is a Deposit Mandate without opt-out (DM), which can be trivially added to the Copyright Retention Mandate with opt-out (CRM).

In other words, Harvard can have its (CRM) cake, and eat it (DR) too!

That contingency is completely missing in Michael's analysis of my proposal.

Michael points out that even if a Harvard author opts out of CRM, he can still deposit his article if he wishes to.

But if voluntary deposit -- just for the sake of the benefits of OA, or just because one's university or funder had invited deposit -- had been capable of generating enough OA, then (1) mandates would not be necessary, (2) NIH's invitation policy would not have failed and would not now have had to be upgraded to an immediate deposit mandate, and (3) the hundreds of institutional repositories with invitations instead of mandates worldwide would not be hovering for years at spontaneous deposit rates of 15% while the (still few) mandated repositories approach 100% within two years.

An opt-out mandate is not a mandate. That is why I urge that the Harvard opt-out CRM mandate be upgraded to add a non-opt-out DM clause. Absolutely nothing is lost, and a great deal is gained.

For the papers whose authors can and do opt for CRM, Harvard will have immediate OA. For the papers whose authors opt out of CRM, Harvard will still have 100% immediate deposit, with immediate OA for about two thirds of it and almost-OA for the remaining third.

(Even if it is assumed that the articles for which the authors opt for CRM are identical to the deposits that could have been set to immediate OA anyway, there is still the gain of one third almost-OA and 100% deposit with DM+CRM but not with CRM alone.)

I hope that makes the logic and the contingencies of my proposal still clearer. I might add that exactly the same logic was used in designing the ID/OA (immediate-deposit, optional-access) mandate itself (the one Peter calls the Dual Deposit/Release mandate):

There the logic was that if an institution could not reach agreement on adopting the stronger immediate OA mandate (for copyright reasons, say), then it makes no sense to adopt a delayed-deposit mandate, or, worse, an opt-out "mandate," which allows the publisher's embargo policy to determine that date at which the deposit is made:

It makes far more sense to mandate immediate deposit in every instance, with the publisher's embargo policy applicable only to the date on which the deposit is made OA ("released"), thereby allowing almost-OA to tide over the embargo, thanks to the Button.

Last point: my position would be closest to Michael Carroll's option "(c) even if they would marginally improve scholalry communication, the costs of negotiating copyright with publishers is not worth this benefit" -- except that it is not exactly the benefits of OA I am talking about here but the benefits of a successful OA mandate. (Because every opt-out means no OA.)

Once immediate deposit is safely mandated universally, and it has generated free online access for researchers worldwide to all refereed postprints, we can examine at our leisure what else researchers needed that didn't already come with that free online territory.

But please, let's get there first, and not be held back by pre-emptively over-reaching (and thus inviting opt-out) when 100% free online access is already within our grasp -- if only we manage to mandate the keystrokes!

Stevan Harnad