Start Submission Become a Reviewer

Reading: Author rights and the Harvard open access policies: a response to Patrick Alexander


A- A+
Alt. Display

Opinion Pieces

Author rights and the Harvard open access policies: a response to Patrick Alexander


Peter Suber

Director Harvard Library Office for Scholarly Communication Harvard University, US, US
X close


In his opinion piece criticizing the open access (OA) policies at Harvard University, Patrick Alexander makes several factual errors about the policies themselves and Harvard’s experience under them. In response, I discuss several relevant facts about Harvard OA policies, among them that the policies were adopted by faculty votes, not imposed by administrators; that under the policies, faculty only grant Harvard nonexclusive rights to new faculty articles, not exclusive rights or full copyright; that the policy-created Harvard OA license is merely a default that authors can easily waive for any given article; that the policies do not hinder Harvard faculty in publishing and do not limit their freedom to publish in the venues of their choice; and that the policies give Harvard faculty more rights, not fewer rights, over their own work than they typically get from their publishing contracts.
How to Cite: Suber, Peter. 2021. “Author Rights and the Harvard Open Access Policies: A Response to Patrick Alexander”. Insights 34 (1): 8. DOI:
  Published on 21 Apr 2021
 Accepted on 19 Feb 2021            Submitted on 02 Dec 2020


In his opinion piece criticizing the open access (OA) policies at Harvard University,1 Patrick Alexander makes several factual errors about the policies themselves and Harvard’s experience under them. In response, I discuss several relevant facts about Harvard OA policies, among them that the policies were adopted by faculty votes, not imposed by administrators; that under the policies, faculty only grant Harvard nonexclusive rights to new faculty articles, not exclusive rights or full copyright; that the policy-created Harvard OA license is merely a default that authors can easily waive for any given article; that the policies do not hinder Harvard faculty in publishing and do not limit their freedom to publish in the venues of their choice; and that the policies give Harvard faculty more rights, not fewer rights, over their own work than they typically get from their publishing contracts.

Harvard had already answered his questions and objections in public documents that Alexander could have consulted, including a book-length guide to Harvard-style OA policies. (Note 1) He did not quote or cite them. If he found them inadequate to answer his questions and objections, he could have criticized their inadequacy. But he does not mention them.

I readily acknowledge that Harvard is a large and complicated institution, with many public documents explaining its policies and practices. If our public documents on the OA policies are hard to understand, we will try to improve them. In the meantime, I hope this article helps clarify how the policies work, how they benefit authors, and why Alexander’s criticisms miss the target.

Background on the Harvard OA policies

Harvard was not the first university with an OA policy. It was about the twelfth. But it was the first in the U.S., and the first anywhere to adopt what we now call a ‘rights-retention’ OA policy.

What makes rights-retention policies distinctive is the way the university and the faculty members retain key rights, including the right to make future scholarly articles OA.

One short version of Alexander’s criticism is that rights-retention OA policies reduce author rights and perhaps even promote infringement. One short version of my response is, on the contrary, that rights-retention OA policies increase author rights and guard against infringement. (For details, see the following sections.)

Before Harvard adopted its first OA policy, in February 2008, university OA policies focused on repository deposits and ignored rights retention. At their heart, they encouraged or required the deposit of new works in the institution’s OA repository. But when authors did deposit new work, university staffers had to look up the publisher’s self-archiving policy or seek special publisher permission to make the work OA. This took staff time and resources, and even with that investment the answer was often ‘no’.

Under a rights-retention policy, faculty members (Note 2) grant the university a certain set of nonexclusive rights to their future scholarly articles (Note 3), and the university grants the same set of rights back to the authors (Note 4).

Under U.S. copyright law, the authors originally hold these rights, which is why they can grant them to their institution. However, when publishing they typically transfer many or all their rights to the publisher. The purpose of Harvard’s grant-back to authors is to ensure that authors still hold these nonexclusive rights after they sign their publishing contracts.

The grant of nonexclusive rights to Harvard includes the right to make the work OA. But it includes many other nonexclusive rights as well. The reason for enlarging the set of rights granted to Harvard, beyond those needed to make the work OA, is to enlarge the set of rights granted back to authors and thereby give authors as much freedom as possible to use and reuse their own work after publication. Among other things, this frees authors to release revised or expanded editions, to reuse articles in anthologies or monographs, and to authorize translations, text mining, and the copying needed for long-term preservation.

Under the Harvard policies, faculty members also commit to deposit their future articles in the institutional repository. When they do deposit new articles in the repository, Harvard already has the nonexclusive right to make them OA and need not look up the publisher’s policy or seek special permission. This saves time, labor, and money. It also solves the primary copyright problem in implementing a university OA policy by allowing the institution to make policy-covered works OA through the repository.

Before the Harvard policies, university OA policies were adopted by administrators and imposed on faculty. Harvard was the first university to adopt an OA policy by faculty vote.

If we call the faculty grant of nonexclusive rights the ‘license’, then the heart of a rights-retention policy is the faculty grant of the license to the university. In addition to the license, most rights-retention OA policies, including Harvard’s, contain a provision letting authors opt out of the license for a given article, case by case. We call this the waiver option.

There are several ways for policy-covered authors to avoid making their policy-covered work OA, if that is their desire. The waiver option is the most conspicuous because it is explicit in the policy. The primary purpose of the waiver option is to allow authors to submit new work to the journals of their choice, including journals that do not want their published articles to co-exist with OA copies elsewhere.

The Harvard policies are opt-out policies. Once adopted, faculty can opt out of the license, or obtain a waiver, for any given article. The policy shifts the default to OA (or more precisely, permission for OA). Obtaining a waiver for a given article changes the default for that article but leaves the background default standing.

The Harvard OA policies apply to the accepted author manuscript (AAM), not the version of record (VoR). The AAM is the version of the text approved by peer review, without any subsequent publisher edits or enhancements. However, if the journal publishes the article under an open license, such as CC BY, then the open license allows Harvard – and any other person or institution – to redistribute the (VoR) if they wish.

Rights-retention OA policies do not make it easier or harder for universities to gather policy-covered articles for deposit in the repository. But they make everyone’s life easier once articles arrive at the repository. Universities with rights-retention OA policies only face one problem (the gathering problem), not two (the gathering problem and the copyright problem).

If we imagine different universities giving roughly equal effort to gather articles for the repository, with roughly equal success, then rights-retention policies will increase the number of articles the institution can make OA and decrease the implementation work of staffers. The reason is that the policies solve the permission or copyright problem. When the repository receives articles for deposit, it can make them OA without much further ado. (Note 5)

Finally, the Harvard OA policies are implemented by the Office for Scholarly Communication (OSC), which I have directed since 2013, in consultation with the OSC Faculty Advisory Committee. (Note 6)

Responding to Alexander’s objections

Here is my attempt to distill Alexander’s major objections to the Harvard OA policies, and respond to them.

With one exception, I do not address his objections to the SPARC author addendum. However, correcting misunderstandings about rights-retention OA policies can also correct misunderstandings about rights-retention author addenda.

1. Objection: that Harvard took faculty rights without their consent

Alexander claims to expose ‘an intellectual “landgrab” by universities’. Since he focuses on Harvard, he seems to mean that Harvard grabbed faculty intellectual property rights. Later he refers to ‘the potential hijacking of author rights that is endemic in OA policies like Harvard’s’. Despite the word ‘potential’ in the second quotation, both passages suggest that Harvard acquired faculty rights without faculty consent. He leaves the unmistakable impression that he believes the OA policies are based on coercion, seizure, or expropriation. That is the objection I will address in this section.

Alexander does acknowledge that Harvard’s Faculty of Arts and Sciences adopted its OA policy by a unanimous vote. (Note 7) However, after that acknowledgement, he seems to forget this important fact and argues as if the policies were imposed on faculty against their will.

First, the ‘landgrab’ and ‘hijack’ language suggests that the OA policies were adopted by administrators and imposed on faculty. But administrators played no role in initiating, drafting, or adopting the Harvard policies.

To avoid this very misunderstanding, our good-practices guide has long recommended (Note 8) that similar policies at other universities ‘should be adopted by the faculty, not the administration. … This will show that faculty wanted the policy, and that it was not forced on them by the administration. … [The policy] should be a faculty initiative and be perceived to be a faculty initiative’.

Because Harvard is highly decentralized, it could not adopt a single university-wide OA policy. Instead, its nine schools adopted nine separate policies in nine separate faculty votes between 2008 and 2014. (Note 9) However, the nine school-level policies are identical in substance. In particular, they are all rights-retention policies with waiver options adopted by faculty votes.

Alexander seems to object to opt-out policies as such, and to regard them as coercive. He writes, ‘Harvard’s policy and its permutations elsewhere look an awful lot like negative option marketing. That is, you are ‘in’, unless you let university name know that you are out and secure a waiver. The burden falls to you to extricate yourself’.

That is true but incomplete. It is true that the Harvard OA policies are opt-out policies. But because they are adopted by faculty votes, they are self-imposed opt-out policies. After the policy is adopted, faculty are indeed ‘in’ until or unless they choose to get a waiver for a given paper (or until they leave the university (Note 10)). But this is the policy drafted, discussed, approved, and adopted by faculty.

I cannot tell what Alexander would say about voluntary or self-imposed opt-out policies, because he does not discuss them. But think about other voluntary or self-imposed opt-out rules, such as picking the default font in your word processor or choosing to auto-deposit a certain fraction of every paycheck into a retirement account. We can admit that these acts create a subtle force or nudge (Note 11) to take one path rather than another, provided we remember that this nudge is self-imposed, changeable, and easy to change. There is no sense in which it is a ‘landgrab’ or ‘hijacking’.

Alexander seems to think that the waiver process is slow and difficult. For example, he writes, ‘I suspect my provost neither wants to hire someone as his designate nor handle the phone calls, e-mails and paper requests from faculty (and in some cases staff) to waive university name’s irrevocable control of all rights under copyright of faculty intellectual property’.

But at Harvard the process is fast and easy. As we put it in our good-practices guide (Note 12), ‘Faculty who object to opt-out OA policies sometimes believe that the default will be difficult to shift, or that their request to shift it might be denied. But this depends on the policy. We recommend that the policy make clear that the institution “will” grant opt-outs or waivers, whenever a faculty member asks it to do so, not merely that it “may” grant waivers’.

We follow our own advice on this point. Harvard grants requested waivers, no questions asked. It takes an author 30 seconds on a web form.

One important consequence of the waiver option is that Harvard faculty retain the freedom to decide for or against OA for any given article. This is the rationale of the waiver option, and it ensures that Harvard faculty retain the freedom to submit new work to any journal and any publisher.

If waivers are easy to get, does the waiver option make the policy toothless? This is roughly the inverse of Alexander’s objection. Here’s how we address it in our guide (Note 13): ‘Faculty who worry that a waiver option might entail a high waiver rate should not underestimate the power of changing the default. Shifting the default can and does change behavior on a large scale’.

When our kind of policy was new, we did not know whether the waiver rate would be high (say, 40%, 60%, 80%) or low. But now, after 13 years of experience, we can say that it is low, below 5%. The same is true at MIT and the University of California, the only other institutions with rights-retention policies for which we have data on the waiver rate.

Because making the policy toothless is the opposite of Alexander’s objection, the low waiver rate might be the opposite of consolation. I would put the situation this way. We do not think policies that merely change a default are coercive, especially when the new default is self-imposed and easily changed. If the Harvard opt-out OA policies were coercive in an objectionable sense, then either faculty would not have adopted them, or the waiver rate would be much higher than it is.

We can clarify the situation from another direction. To cover Harvard researchers other than faculty, we launched a voluntary Individual Open-Access License (IOAL). (Note 14) If librarians, administrators, research fellows, students, or other non-faculty scholars want the same benefits that faculty get from faculty-level policies, then they may sign the IOAL. If not, not. It is entirely voluntary. But it too, once signed, is a self-imposed opt-out policy. If the opt-out character made it coercive in an objectionable sense, then no one would sign it. Yet many Harvard authors choose to sign it – on average, 500 Harvard authors each year since it launched two years ago.

What benefits? Why would faculty vote for rights-retention OA policies? Why would individuals sign the individualized equivalent? Here is how we enumerate the benefits to authors in our FAQ: (Note 15)

‘The Harvard open access nonexclusive license:

  • Gives authors permission to make their work open access without the difficulty or uncertainty of negotiating with publishers;
  • Enables the university to help authors make their works open access;
  • Preserves author freedom to publish in the journals of their choice;
  • Preserves author freedom to decide for or against open access for each publication; and
  • Enhances author rights to reuse their work, and gives authors more rights over their own work than standard, or even progressive, publishing contracts.’

Faculty discussed the perceived advantages and disadvantages of the policies at length before voting. For example, among the questions discussed in these meetings were the following. Will faculty be required to publish in OA journals? Will some publishers reject policy-covered submissions? Will faculty have to pay to publish? Will Harvard ‘own’ or acquire the full copyright in these works? Will faculty who want waivers have to provide reasons that administrators might question or reject? Will there be a penalty for non-deposit in the repository? Will the versions in the OA repository violate copyright?

Far from being gulled or coerced, faculty were persuaded after lengthy discussion. Some of the Harvard schools scheduled not just one, but a series of Q&A meetings on their draft policies and stopped holding new meetings only when faculty stopped showing up. Only then did they hold the vote.

The same response (at least from Harvard) applies to Alexander’s patronizing and unargued assertion that ‘[Faculty at schools with Harvard-style policies] have embraced OA without being fully informed about what may or may not be in their best interests’.

Alexander points out that under U.S. law, authors own the copyright in their works until or unless they transfer those rights to others. Then he asks, ‘Is there any reason to disagree with this?’. He leaves the impression that Harvard does disagree with it, or does not know this elementary rule of law, or somehow violates it. On the contrary, the Harvard OA policies are built on this rule of law. The premise of these OA policies is that Harvard will not have these rights unless authors grant them to it. The institution does not already have them, cannot simply take them, and would not want to.

Some universities hold that faculty scholarship falls under the work-for-hire doctrine. But Harvard is not one of them. Under Harvard rules, faculty are the copyright holders in their scholarly works. (Note 16) That is why our OA policies depend on the faculty grant of rights to the institution, not on an administrative ‘landgrab’ or ‘hijacking’.

If we only cared to answer the ‘coercion’ objection from the OA policies, we could stop here. But Harvard follows certain principles and practices that make its full answer more complete.

A background implementation principle at Harvard is that there is no penalty for non-deposit in DASH,2 our OA repository. Faculty who are dubious about the policy, too busy to deposit new work, or simply forgetful, can ignore it without penalty. This freedom from penalty brings latitude above and beyond the waiver option.

If there is no penalty for non-deposit, then how does Harvard gather works for deposit? For many authors, deposit is voluntary and spontaneous. They see the benefits to themselves and potential readers. For other authors, as I have pointed out in many places over the years, we depend on ‘expectations, education, incentives, and assistance, not coercion’. (Note 17)

Another background implementation principle at Harvard is that if authors want one of their articles removed from the repository, we will remove it. We follow author wishes on this, and do not wait for the ordinary grounds for retraction. Only half a dozen authors have asked us to remove a work from the repository (averaging one request every two years), and none requested the removal on the ground that they felt deceived, coerced, or that their rights had been violated.

Similarly, while some Harvard authors have received takedown notices from publishers for posting their articles to their personal websites, DASH has not received even one takedown notice.

2. Objection: that Harvard ends up with full copyright over faculty articles

Alexander quotes the Harvard policy language showing that faculty only grant the university nonexclusive rights in their future articles. But in key parts of his discussion, he forgets that point and ignores the distinction between exclusive and nonexclusive rights.

I will re-quote the critical sentence from the Harvard model policy here and number three parts of it for discussion.

‘More specifically, each Faculty member grants to (university name) a [1] nonexclusive, irrevocable, worldwide license to exercise [2] any and all rights under copyright relating to each of his or her scholarly articles, in any medium, [3] provided that the articles are not sold for a profit, and to authorize others to do the same.’

In key parts of his discussion Alexander focuses on clause 2 (‘any and all rights under copyright’) and ignores clause 1 (‘nonexclusive’) and clause 3 (the proviso that faculty do not grant Harvard the right to sell these works for a profit).

For example, he writes, ‘The Harvard policy says that you retain the copyright but that you licensed all rights under copyright to university name’. The policy explicitly qualifies the ‘all rights’ language and Alexander ignores the qualifications. There are two qualifications worth noting. First, faculty only license nonexclusive rights. Second, faculty do not even license all nonexclusive rights. They do not grant Harvard the exclusive or nonexclusive right to sell policy-covered works for a profit.

Alexander writes, ‘So that means that university name (or another entity to whom all rights under copyright were licensed) could take your articles, mash them up with other articles and create a textbook, sell it to students, recover all manufacturing and overhead costs, and not pay the author a dime. I can see why faculty are lining up to do this’.

This is another place where he uses the ‘all rights’ language without qualification. Beyond that, faculty have ‘lined up’ for this outcome in the sense that they voted up rights-retention OA policies at every Harvard school. If Alexander believes that faculty should not be allowed to exercise their rights this way, he needs to offer an argument. Sarcasm is not an argument.

The mash-up Alexander describes is allowed by the policies, just as it is allowed by OA journals using open licenses. (Some open licenses also allow selling for profit.) However, it is not allowed by the DASH terms of use,3 which state that third parties may ‘not sell or charge for any [policy-covered] article (whether or not any profit is involved)’.

Another background implementation principle bears on what Harvard may do with works in the repository, beyond what third parties may do. Although the OA policies give the university the right to license its nonexclusive rights to others, in practice the university only licenses them in full back to the authors. Beyond the permissions granted to the public through the DASH terms of use, Harvard does not license these works to third parties without the consent of the authors.

Alexander asks a related question: ‘Could [the university] give away [my work], license it under a Creative Commons CC BY-NC license, or sell it for less than I could? That would seem to be the case’. The answer is ‘yes’. The university does give away the work through the OA repository, except when authors have obtained waivers or chosen not to deposit the work at all. Faculty voted for the OA policies precisely to facilitate this kind of giveaway.

Finally, Alexander seems to want the right to sell his journal articles. (This is one way to read his question about whether Harvard could sell his article for less than he could.) If so, he is in a small minority. If scholars wanted to sell their journal articles, they would not follow the centuries-old custom of giving their articles and rights (including this right to sell their articles) to publishers without charge. Nor would they vote up university OA policies. But, of course, he has the right to try to sell his journal articles and, if he were to try, he would find that Harvard-style OA policies give him less trouble than journal publishers. At Harvard, he would still have both the waiver option and the non-deposit option. By contrast, journal publishers would not pay him for his articles. On top of that, either they would make the articles OA themselves (sometimes with and sometimes without a right to sell them elsewhere) or insist on rights that block him from selling the articles elsewhere.

3. Objection: that some Harvard faculty may not know they are covered by such a policy

Alexander writes, ‘Under U.S. law, non-exclusive licenses need not be in writing, but there needs to be some mutual understanding. Does the Harvard policy create that understanding, even if faculty disagree or are not aware? But what if I do not agree or do not remember agreeing?’.

Faculty who were at Harvard when their schools adopted a policy were aware of the policy, at least if they attended faculty meetings, followed the meeting agendas, read the associated handouts, or participated in any of the Q&A sessions.

But, of course, many faculty now at Harvard were not at Harvard when their schools adopted their OA policy. All new Harvard faculty are asked to affirm the university’s intellectual property policies by signing the Participation Agreement. (Note 18) By its nature, this expectation extends to faculty who had not been hired when the various parts of the intellectual property policies were adopted. The policies cover patents as well as copyrights, and many aspects of copyright other than OA. After the Harvard schools adopted their rights-retention OA policies, the broader Participation Agreement covered them too. Faculty who read the Participation Agreement know about the OA policies. Faculty who sign it affirm the Harvard OA license in writing.

Similarly, non-faculty researchers who sign our voluntary IOAL know they are signing it. (See Section 1 above.)

Finally, Harvard does not make work OA through its repository until the Harvard authors sign an Assistance Authorization (AA) form. Alexander is right that a nonexclusive license need not be in writing. However, if a prior grant of nonexclusive rights is to prevail over a later grant of exclusive rights, then the safest course is to ensure that the prior grant is affirmed in writing. (Note 19) That is one purpose of our AA form. Hence, regardless of the author’s presence at Harvard at the time of the faculty vote, the author’s reservations about the policy, or the author’s memory, we ask for written affirmation of the Harvard license before we act on the Harvard license. (Note 20)

4. Objection: that the Harvard policies interfere with faculty ability to publish

Alexander predicts that Harvard-style OA policies will limit the ability of authors to publish their work. For example, he writes:

‘The fact that university name can sell the intellectual property (without profit but recovering costs) and distribute it for free means that few publishers would be in a position to contract with you to publish your content. … Some publishers may refuse to publish if an author insists on adhering to a university’s OA policy. As a result, authors are handcuffed by a university policy that stifles possible publishing opportunities for their research. … Giving something away is not always in the best interest of the researcher, particularly if it restricts her ability to publish her research. … What effect will university name’s control of all rights under copyright have on your chances of publishing in a top-flight journal?’

We do not know any examples in which Harvard authors have had trouble publishing articles because of our OA policies. Nor do we know examples from any of the dozens of institutions worldwide – in North America, Europe, Africa, Asia, and Australia – that have adopted rights-retention OA policies like ours. Nor does Alexander cite any examples.

Our faculty have no trouble publishing in top-flight journals. His prediction has been tested for 13 years at Harvard, and for nearly as long at many other universities around the world. There is no supporting evidence for it, and abundant evidence against it.

Finally, what are the odds that Harvard faculty would acquiesce in silence if a Harvard policy, especially a self-imposed policy, limited their freedom to publish? Wouldn’t their protests be notorious enough to quote in an article critical of the policies?

Alexander predicts that the SPARC author addendum, like the Harvard rights-retention policies, will limit author freedom to publish. Suppose an author uses the addendum when submitting a manuscript to ‘a small HSS [humanities and social sciences] society publisher, [which] relies on subscription income to support its membership. It can ill afford to forego revenue, so, in order to meet its financial obligations, it rejects your article and decides to select another piece without the OA restrictions. You are left with an unpublished article that must undergo the entire review process elsewhere, where it will be likely to face the same precarity again unless you are able to find an OA publisher and the funds to publish your work’.

Just as we do not know any cases in which rights-retention OA policies limited faculty freedom to publish, we do not know (Note 21) any cases in which author addenda caused journals to reject submissions. Author addenda do not modify publishing contracts unilaterally or without the consent of publishers. They are proposed contract modifications, which publishers remain free to take or leave. Publishers can reject the proposed addendum without rejecting the article. It is easier to do and preserves the journal’s option to publish good work. (Note 22)

5. Objection: that rights-retention OA policies work against other interests of authors

In this section I will look at alleged harms to authors beyond the supposed obstacle to publishing already discussed in Section 4.

Alexander writes that ‘author rights as a foundational tenet of OA are being scuttled. University-led OA policies [exemplified by the Harvard policies] are silently and slowly undermining that foundational notion that, in Peter Suber’s words, “OA depends on author decisions and requires authors to exercise more rights or control over their work than they are allowed to exercise under traditional publishing contracts.”’.

Alexander seems to endorse the sentence he quotes from me, and believes it cuts against rights-retention OA policies. I stand by the quoted sentence and argue that it does not cut against rights-retention policies. On the contrary, rights-retention policies increase author rights over their own work, which is one reason I support them.

Alexander claims ‘to explore the effects of OA mandates at universities that not only undermine but violate the rights of authors, at times in ways that, ironically, traditional publishing never would’. However, he never makes good on the two strong claims that Harvard-style OA policies ‘violate the rights of authors’ and do so in a way that ‘traditional publishing never would’. If he has details or supporting arguments, I have been unable to find them. (Note 23)

He later writes, ‘What is more, if under these OA policies you are still the copyright holder, it will fall to you financially and practically, not to an institution or a publisher, to register the copyright in your name so that you are entitled to the critical benefits of recordation. It would fall to the copyright holder – you – to defend against a claim of copyright infringement. Universities are claiming all the benefits of having all rights under copyright but share in absolutely none of the risks.’.

This is another passage in which Alexander forgets that Harvard only acquires a nonexclusive license, not full copyright (see Section 2 above). (Note 24) When Harvard grants back the same nonexclusive rights to authors, the authors too only acquire nonexclusive rights, not full copyright. Harvard is careful to use the language of nonexclusive rights and does not claim to have ‘all rights under copyright’ without notable qualifications.

Alexander writes, ‘The policy states, “The Faculty of (university name) is committed to disseminating the fruits of its research and scholarship as widely as possible.” Do we have evidence that licensing rights to a university leads to broader dissemination?’.

We do. One kind of evidence is the number of downloads from DASH, the institutional OA repository. Nearly all the articles in DASH have already been published elsewhere. (That is, most are postprints, not preprints.) If publication in a journal met all demand from would-be readers, there would be little or no demand for the OA copies in DASH. But on the contrary, DASH articles are in high demand. The average article in DASH was downloaded 731 times last year, a rise from 635 the year before. In the period September 2019 to September 2020, DASH had more than 9.5 million downloads. In June 2020 alone, DASH had more than one million downloads, its best month ever.

One cause of the June 2020 spike in DASH downloads was our recent (March 2020) launch of a program to fast-track Harvard’s coronavirus research into DASH. (Note 25)

Another kind of evidence arises from our collection of user stories.4 Every DASH download includes a cover letter inviting readers to drop us a line to say how OA to this work benefits them. For nearly a decade, we have received about five stories every day. (We have thousands of stories still offline that we will add to the online collection as soon as we can.) Some of the stories are from academics at institutions that cannot afford subscriptions to the journals where the articles originally appeared. The rest are from non-academics without privileges at academic libraries, such as physicians, journalists, policymakers, clergy, and nonprofit organizations. Their stories are expressions of gratitude from readers who did not have access to the versions of the same articles published in journals, and constitute strong, direct ‘evidence that licensing rights to a university leads to broader dissemination’.

Many of our faculty initially believed that publishing their articles in well-known, high-impact journals made their articles available to everyone who wanted to read them. However, when they see these user stories, and the download numbers from DASH, they realize that publication in paywalled journals leaves much demand unmet. They often tell us in moving terms how surprised and delighted they are that the OA copies in DASH help meet that demand. (Note 26)

Alexander writes, ‘Open access journals can be top-flight, as long as they are peer reviewed, but what if you want to write a book based on one or two of those articles that are OA and controlled by university name?’.

Harvard’s rights-retention policies positively help authors in this situation. As noted, Harvard has a wide range of nonexclusive rights in these articles and grants them back to authors. That ensures that authors have all the rights they need to write a book based on articles they previously published. They need not seek permission from the original publisher, or from Harvard, unless the new publisher demands exclusive rights.

Alexander writes, ‘But authors and researchers, especially those in HSS [humanities and social sciences] disciplines, seem oblivious to the potential hijacking of author rights that is endemic in OA policies like Harvard’s’.

Unfortunately, it is Alexander who is oblivious to the relevant facts here. There are four: first, that faculty granted these rights to Harvard, which did not ‘hijack’ them from authors; second, that faculty only granted nonexclusive rights, not exclusive rights or full copyright; third, that Harvard grants back to authors all the nonexclusive rights they granted to Harvard under the OA policies; and fourth, that this grant-back increases author rights, and gives authors far more rights to reuse their own work than they receive under standard or even progressive publishing contracts.

6. Objection: that the Harvard license must expire or become revocable

Alexander argues that irrevocable licenses are incompatible with 17 USC 203(a)(3), which says, ‘Termination of the grant may be effected at any time during a period of five years beginning at the end of thirty-five years from the date of execution of the grant; or, if the grant covers the right of publication of the work, the period begins at the end of thirty-five years from the date of publication of the work under the grant or at the end of forty years from the date of execution of the grant, whichever term ends earlier’.

Of course, we acknowledge the right of termination under U.S. copyright law. However, it is largely moot in the case of the Harvard nonexclusive license because Harvard grants back the same license to the authors. The ordinary reason for authors to terminate a prior license is to gain or regain rights previously granted to another party. Because Harvard faculty continue to hold the nonexclusive rights they once granted to Harvard, they do not need to terminate the license in order to acquire or recapture them. (Note 27)

However, if Harvard authors wish to exercise their right of termination anyway, not to gain or regain rights but merely to block Harvard from having or exercising these rights, then of course Harvard will accede. But even in that case, obtaining a waiver (and if desired, removing works from DASH) is faster, easier, and less expensive than exercising the statutory right of termination. Moreover, if they wanted, authors could take these steps right now, decades before they become eligible to exercise the statutory right of termination.

Authors who do choose to exercise their right of termination would most likely terminate against their publisher, not Harvard. Harvard faculty already have all the rights Harvard has, but may not already have all the rights their publisher has.

Finally, as noted earlier, the Harvard OA license terminates on its own, with no legal action needed, when authors leave Harvard or die.


Nothing in Alexander’s piece shakes my support for Harvard-style rights-retention OA policies or my confidence in their legal soundness. I stand by our methods of implementing them and our recommendations to the many other institutions who have consulted with us about them. I stand by the benefits they bring to Harvard authors and their many readers worldwide. I invite anyone curious about our policies, or rights-retention policies in general, to consult our online documents (cited in the notes) or contact us directly.


1Peter Suber and Stuart Shieber, “Good practices for university open-access policies,”. First released in September 2012 and regularly updated. Last updated November 13, 2020. Although the guide is frequently updated, all the quotations from it in the present article were in the guide at the time Alexander’s article appeared, and for several years before that. Accessed November 20, 2020. 

2The Harvard school-level OA policies only cover faculty, but some other Harvard OA policies cover non-faculty researchers. For simplicity and unless noted otherwise, I will focus on the school-level policies for faculty. 

3The Harvard school-level policies only cover scholarly articles, but some other Harvard OA policies cover genres other than articles, such as electronic theses and dissertations. Again, for simplicity, I will focus on the school-level policies for articles. 

4See the entry in our FAQ on “What will Harvard do with the articles covered by its license?” Note this point in particular, ‘When Harvard receives the grant of nonexclusive rights from faculty, it grants the same rights back to the faculty. The result is that faculty receive more rights from the OA policy, to use and reuse their own work, than they received under their publishing contracts.’ Accessed November 20, 2020. 

Also see the longer discussion in the section of our guide on “Transferring rights back to the author.” Accessed November 20, 2020.

5There is a little further ado, much of it shared by institutions without rights-retention policies. Staffers must check to see whether works submitted to the repository are covered by the OA policy, for example, that the author has the right kind of institutional affiliation, that the work was published after the author’s school adopted its OA policy, that the author has not obtained a waiver for that work, and that the deposited copy is a version the institution is allowed to share. Staffers also check and usually improve the article metadata. 

6See the current membership of the OSC Faculty Advisory Committee, (accessed 1 March 2021). 

7Like the OA policy at the Faculty of Arts and Sciences, all the other Harvard school OA policies were adopted by faculty votes. See Harvard Office for Scholarly Communication, “Open Access Policies.” Accessed November 20, 2020. ‘In 2008, Harvard’s Faculty of Arts & Sciences voted unanimously to give the Harvard a nonexclusive, irrevocable right to distribute their scholarly articles … In the years since, the remaining eight Harvard schools voted to establish similar open-access (OA) policies.’ 

Several of these votes were also unanimous. See the Open Access Directory list of university OA policies adopted by unanimous faculty votes. Accessed November 20, 2020.

8See the section of the guide on the “Adopting authority.” Accessed November 20, 2020. 

9Here are the Harvard OA policies, Accessed November 20, 2020. There are nine school-level policies and five center-level policies, plus the Individual Open-Access License, discussed further below. 

10See the entry in our FAQ, “Does the policy apply to articles I write after leaving Harvard?” ‘Once you are no longer affiliated with Harvard, any articles you write will not be subject to the open-access policies at your school and will not be licensed to Harvard.’ Accessed November 20, 2020. 

11See Richard H. Thaler and Cass R. Sunstein, Nudge, (Yale University Press, 2008). Thaler and Sunstein use the term ‘nudge’ in almost exactly this sense, the gentle force to follow a default, or (importantly different) the reasons to follow the default. (accessed 2 March 2021). 

12See our discussion of the terms ‘opt out’ and ‘opt in’ in the section of our guide on “Talking about [an OA] policy.” Accessed November 20, 2020. 

In another section of our guide, we say, ‘The policy should make clear that the institution will always grant waivers (or opt-outs), no questions asked. Faculty needn’t offer a justification or meet a burden of proof. To prevent fear or confusion on this point, the policy should refer to “obtaining” a waiver … rather than “requesting” a waiver’. Accessed November 20, 2020.

13In the section of our guide on the “Waiver option”. Accessed November 20, 2020, 

14See our FAQ entry, “What is the Individual Open-Access License?” Accessed November 20, 2020 

Also see the IOAL launch announcement from October 2018. (accessed 2 March 2021).

15See our FAQ entry, “What are the advantages for authors?” Accessed November 20, 2020. 

Also see Section 5, presenting evidence for the additional benefit that making articles OA leads to wider dissemination.

16This is the rule for faculty. But non-teaching employees are covered by the work-for-hire doctrine. Hence, when we wanted to extend the benefits of the OA policies to librarians, we first had to exempt librarian-scholars from work-for-hire, which we did in 2018. See “Harvard Exempts Its Librarian-Scholars From the Work-For-Hire Doctrine,” October 31, 2018. Accessed November 20, 2020. 

Also see Peter Suber and Martha Whitehead, “A brief history of open access at Harvard,” July 28, 2020. ‘This enabled Harvard librarians to sign the [Individual Open-Access License] and make their scholarship OA by choice rather than by an institutional exercise of the work-for-hire doctrine.’ Accessed November 20, 2020.

17See Peter Suber, Open Access, (MIT Press, 2012), p. 87. (accessed 2 March 2021). 

Here is a deep link to the passage in question. (accessed 2 March 2021).

Also see Peter Suber, “Harvard’s Open-Access Policies.” First released March 17, 2017 and periodically updated. Last updated January 27, 2020. Accessed November 20, 2020. To implement our policies, we depend not on coercion but on ‘expectations, education, incentives, assistance, requests, and persuasion’.

18The text faculty are asked to sign is only available to Harvard affiliates. But here is a public-facing page on the Participation Agreement. Accessed November 20, 2020. The key provision reads, ‘If I am a Faculty member of a Faculty or School of the University that has adopted an Open Access Policy, I hereby confirm my grant to Harvard of a nonexclusive license with respect to my scholarly articles, as set forth in that policy’. 

19See 17 USC §205(e) (2010). (accessed 2 March 2021). Note that §205(e) makes the written affirmation sufficient without making it necessary. We are playing it safe by seeking written affirmation of the Harvard license even if it might not be legally necessary. 

20See our discussion of this in the section of our guide on the need for an “Individualized writing.” Accessed November 20, 2020. 

21Heather Joseph, Executive Director of SPARC, joins me in this assessment. (Offline communication.) 

22Toward the end of the quoted passage, Alexander repeats the long-refuted false assumption that all or most OA journals charge APCs (‘…unless you are able to find an OA publisher and the funds to publish your work’). According to the Directory of Open Access Journals, only about 30% of peer-reviewed OA journals charge APCs. Accessed November 20, 2020. 

23If I could find an argument that Harvard-style rights-retention policies violate author rights, I would examine it in detail. But in the absence of such an examination, I can point out that the Harvard policies were reviewed and approved by the Harvard Office of the General Counsel. The policy at Harvard Law School was one of those adopted by a unanimous faculty vote. And in our 13 years of experience under the policies, our OA repository has received no takedown notices. The university has not been sued or threatened with suits for implementing the policies. Our authors have not been sued or threatened with suits for complying with these policies. 

24Here are three more passages in which he forgets that Harvard (and other institutions with rights-retention OA policies) only acquire nonexclusive rights, not exclusive rights or full copyright. (1) ‘According to U.S. copyright law … transfer of copyright or exclusive rights must be conveyed in writing. … But what about in the case of Harvard’s open access ‘template’, which is being used by so many universities and colleges?’ (2) ‘What effect will university name’s control of all rights under copyright have on your chances of publishing in a top-flight journal?’ (3) ‘I suspect my provost [does not] want[] to hire someone as his designate … to waive university name’s irrevocable control of all rights under copyright of faculty intellectual property.’ 

25Colin Lukens, “Reacting to the unprecedented urgency of COVID-19 research at Harvard: DASH’s fast-tracking deposit program,” OSC (blog), October 19, 2020. Accessed November 20, 2020. 

26For example, see James Russell, Mashtots Professor Emeritus of Armenian Studies at Department of Near Eastern Languages and Civilizations at Harvard. When he received a routine email from our repository reporting one month of downloads for a recent deposit, he wrote, ‘Wow! 193 readers in 44 countries. I am, without any irony at all, humbled and delighted. Thank you for the work you are doing.’ (Reprinted with permission.) 

27For what it’s worth, we can find no case law on the use of 17 USC §203(a)(3) by authors to terminate nonexclusive rights in third parties when the authors themselves already possess the same nonexclusive rights. Also note that the Authors Alliance, a leading expert on termination rights for authors, has recommended that ‘reversionary rights do not apply to non-exclusive licenses’. See “Authors Alliance Submits Brief Supporting Reversionary Rights in Canada,” December 10, 2018. Accessed November 20, 2020. 

Abbreviations and Acronyms

A list of the abbreviations and acronyms used in this and other Insights articles can be accessed here – click on the URL below and then select the ‘full list of industry A&As’ link:


The author would like to thank Kyle Courtney, Heather Joseph, Colin Lukens, and Stuart Shieber for their helpful comments on earlier drafts of this article.

Competing interests

None. I’ve worked full-time for two decades to foster open access. I have done so at Harvard since 2009 and have led the Harvard Office for Scholarly Communication since 2013. Also see


  1. Patrick H. Alexander, “Open access and author rights: questioning Harvard’s open access policy,” Insights 33 (1): 23, DOI: (accessed 2 March 2021). 

  2. Digital Access to Scholarship at Harvard (DASH), (accessed 2 March 2021). 



comments powered by Disqus