This is the story of a surprisingly difficult journey: my attempt to understand* whether it’s possible for a copyright holder, once having made her work available under a Creative Commons (CC) license, to revoke the license and then distribute the work under different terms (or halt its distribution altogether).
At first blush, the answer to this question would seem simple. CC licenses are irrevocable – it says so right in the legal text of each license, as well as in various FAQ and summary documents provided by the Creative Commons Organization (CCO). This means, obviously, that once a copyright holder has applied a CC license to her work, she can’t later remove the license and restrict access to the work.
However, it turns out not to be quite that simple, because that’s not all the license text says. It also indicates that “the Licensor may… offer the Licensed Material under separate terms or conditions or stop distributing the Licensed Material at any time.” And according to the CCO, it’s possible for a copyright holder to “(change) his or her mind and stop distributing under the CC license terms.”
This brings up some fairly mind-bending questions. For example: given that the CC license represents an irrevocable contract between the copyright holder and the public, and given that there is only one public, to whom could the copyright holder subsequently “offer the Licensed Material under separate terms”? (Maybe to someone who isn’t a member of the public?) And given that free distribution is a core right provided by all CC licenses, what could it mean for the copyright holder to “stop distributing” the licensed work? (Does this mean only that the copyright holder is not required to participate actively in distribution? Could anyone reasonably believe that the license would entail that requirement, absent this language?)
So the language in both the licenses and the CCO’s explanatory documents suggests two apparently contradictory realities:
- Once applied, a CC license cannot be revoked;
- After a CC license has been applied to her work, the copyright holder may stop distributing her work under that license and start distributing it in some other way (or not at all).
How to resolve this discrepancy?
Before outlining what I learned, let me take a moment to share what the experience of trying to answer that question was like. It was, in a word, fraught. I should acknowledge that at the beginning of this process, I wasn’t posing a question as clear and straightforward as the one above – because while I could see that there was an issue to be resolved, I wasn’t yet sure exactly what the issue was. Initially I was under the impression that license irrevocability was a function of Creative Commons policy – which, of course, was incorrect, as was quickly (and gleefully) pointed out to me. Irrevocability is, in fact, baked right into the terms of the licenses themselves. But since the opposite of irrevocability also seems to be baked into the license terms, it became clear that the reality is more complicated and that the implications of this complexity could be significant for authors and other copyright holders.
Looking for some expert help, I posed questions on both the Read2.0 and the LIBLICENSE listservs. The response to my queries was swift and, shall we say, mixed. I won’t go through the tick-tock of the whole discussion because it was long and dense and multidirectional. But I will say that many people were hugely helpful to me as I tried to think/write my way through the thickets of apparently (and sometimes actually) competing truths to a place where the contradictions were finally, I think, resolved. A handful of my interlocutors were less helpful – I did get sneered at, condescended to, yelled at (“WRONG WRONG WRONG” was my favorite response), and accused of trying to undermine the whole open licensing regime. But such reactions are par for the course when one raises questions related to open access and open scholarship, and I’ve gotten more or less used to them – and though the process was occasionally a bit painful, it was in the end tremendously useful.
Anyway, after days of what was sometimes labyrinthine discussion and debate (AND YELLING), I was finally able to boil the issues down to two:
Issue #1: What Exactly Is Licensed under CC?
Early in the conversation it became clear that the resolution to the apparent contradiction in license terms likely hinged on a very important question: does the license apply to the intellectual work itself, or does it apply only to copies of the intellectual work?
If the latter, then there is no real contradiction in the license terms: each copy of the work has one license or another attached to it, and that license is irrevocable, whatever it may be for each copy. If I publish a textbook under a CC BY license in 2019, copies of the work that are created under that license are tagged as CC BY copies and are forever available under that license. However, in 2021 I could decide not to continue distributing the textbook under CC BY, and could switch, say, to CC-BY-NC. Copies made subsequently would be subject to that license rather than CC BY (unless, of course, they were copies of the original CC BY-licensed manifestation).
Alert readers will already be seeing problems with this interpretation, though. For one thing, in this scenario, old CC BY copies would be competing in the marketplace with newer CC-BY-NC ones. (This might represent a logistical annoyance, but if it’s the copies that are licensed rather than the underlying work, and if the licenses are irrevocable once applied, then it’s nevertheless a legal possibility.) A more significant objection, though, arises from the language of the licenses themselves, which all refer to the licensed entity as “the work,” not copies of the work. I raised this issue with one expert informant, whose position is, in fact, that it’s the copies and not the work itself that are licensed. Her view was that the word “work” in the formal license text may be intended more colloquially than one would normally assume.
But what if the license means literally what it says (as contracts tend to do), and what’s licensed is the underlying work itself? In that case, the resolution would seem to be that since CC licenses are irrevocable, the copyright holder is permanently stuck with whatever CC license is applied in the first instance: if I apply a CC BY license to my textbook, the terms of the license require that all future manifestations of my textbook be distributed with that license. Which, if true, brings us to the second issue.
Issue #2: How Does Dual Licensing Work?
A couple of people (including several current and former higher-ups at the Creative Commons organization) expressed the view that while it’s not possible to revoke the CC license originally applied to a work, it is possible to apply a different license to the work in addition to the license previously applied. What this creates is a situation in which the work is available under both licenses, and members of the public can simply pick the terms under which they want to use the work.
But this struck me as fundamentally incoherent, legally speaking. So I responded by asking this:
Suppose I write a textbook and make it available under a CC BY license. Two years later, I decide to make it available under a CC-BY-NC license, which is more restrictive than CC BY. In that case, does the application of the CC-BY-NC license create for the public a new restriction on commercial reuse of the work?
- If it does, then in what sense is the CC BY license still in force?
- If it doesn’t, then in what sense have I actually applied a CC-BY-NC license? (In other words, if the public isn’t required to abide by the restrictions of the second license, but has the option of simply ignoring it, then how is it true that the license has been applied, as a matter of legal reality?)
(Please note that simply saying “the work is now available under two licenses, and the public can use the work under either license” doesn’t answer my question; it’s just a restatement of the proposition that led to my question.)
After further back-and-forth, it became clear that the CCO (or at least these CCO executives) see it this way:
- The CC license applies to the work itself, and therefore to all copies of the work made after the license is applied;
- The copyright holder can apply a second license, and it will affect the public’s rights only if it expands them;
- Technically, the copyright holder can apply a second license that is more restrictive, but it will have no legal effect, since the original license is already irrevocably attached to the work.
So in other words: if I write a textbook and publish it under a CC BY license in 2019, and then change my mind and apply a CC-BY-NC license to it in 2021, the second license has no effect on the public’s rights with regard to my work. However, if I apply a CC-BY-NC license in 2019 and then change my mind and apply a CC BY license in 2021, the second license does affect the public’s rights – because in this case, I haven’t tried to revoke any rights previously granted. I’ve only expanded them.
This strikes me as the most obviously reasonable interpretation of the license terms. As is so often the case after I’ve spent a lot of time puzzling through a problem, the solution has ended up seeming pretty obvious – but given how many differing interpretations were offered to me over the course of my many conversations, I feel fairly confident that the question itself wasn’t simply silly. And other readers may have entirely different interpretations, of course, which would be interesting to see in the comments.
The biggest takeaway? I’d say it’s the message communicated by the Creative Commons Organization itself on its FAQ page: “You should think carefully before choosing a Creative Commons license.” (Assuming, of course, you actually get to choose.)
* This essay is an expression of the author’s personal opinion; nothing in it should be construed as legal advice.
16 Thoughts on "Q: Can You Revoke a Creative Commons License? A: No. Er… Sort Of? Maybe?"
The way I always saw it is that the license is between the licensor (copyright holder) and individual licensees – ie not with the entire public. I understand it as public in the sense that anyone can become a party to the license.
So let’s say you license your work CC BY in 2019. Alice downloads it from your website. She can use it CC BY. In 2020, you take your work down. Like the textbook editions analogy, in 2021 you re-release it CC BY-NC. Bob downloads it from your website and can only use it under the NC license. However this doesn’t stop Alice from using it CC BY.
But let’s say Bob can get a copy from Alice who has posted it on her website. Now Bob automatically gets an offer as a downstream recipient to use it under CC BY. So you can sortof revoke a license by ceasing distribution, but your previous licensees have the right to pick up where you left off. Anyone who has contact with those original terms can then become a party to that license. In this way the original license is irrevocable and not automatically attached to the work as it’s being distributed.
Right, this seems like another expression of the idea that the license applies to copies of the work, not to the underlying work itself. If that’s true, then I think it makes the “revocation” scenario pretty straightforward: if the license is only enacted when a member of the public encounters and creates a copy, and each enacted license is irrevocable, then you can have different irrevocable licenses attached to different copies of the work. But if the folks at Creative Commons are right, then the license actually attaches to the work itself, and that means that once I’ve applied a CC license to my work, that license is permanently attached to the work itself. I can grant additional rights if the original license was more restrictive than CC BY, but I can’t take away public rights that I originally granted if the original license was CC BY — which is what would happen under the scenario you’ve outlined.
Right, but in either case, how you distribute matters. So in 2021, if Bob downloads the work and it says CC BY-NC, then how does he know that there is CC BY on it? Unless he finds out from seeing Alice’s 2019 copy with CC BY terms he has no way to know there is actually a more permissive license underneath. So even if it’s attached to the work rather a copy, knowledge of the distribution terms is necessary for practical reasons.
Thanks for a great post and stimulating discussion!
Thanks for this – an interesting read. I have a related question (if this has been answered elsewhere, apologies, just point me in the right direction and I will be on my way). Preprints are generally (usually, often?) posted under a CC BY licence. If you subsequently submit the work based on your preprint for publication in a journal that doesn’t offer open access (or only offers more restrictive CC licences), what happens to the original licence? You’ve modified the work most likely, which is permitted under the CC BY licence, but have you created a new work that can be published under a different licence?
That’s a good question. I think the answer is that if you have made the work available to the public under a CC license, you do not have the option of later making the work available under more restrictive terms than are provided under that original license — because the license applies to the work itself. So in your scenario, I think the question would be whether the formally published version represents the same work as the preprint version. In most cases, I think it would, even though there may be some differences between them. So that’s a good reason to be very aware of the terms under which you make a preprint available — you may be limiting your options for later publication. (Of course, as I noted in the piece, you may not always be fully free to choose how you make your work available, depending on what terms you agreed to as a condition of funding, employment, etc.)
I am not sure this (“in most cases it would”) is correct; the license is a copyright license, and the reasoning of publishers is that the typeset/formatted published version is subject to the license the authors agreed to with that publisher.
A more lenient and earlier preprint license does therefore not apply to that final (typeset/formatted) published…?
First, most preprints are not released under a CC BY license. bioRxiv and medRxiv give authors a choice of license with varying results (CC BY usually in the minority), and most arXiv preprints are under the arXiv license which is very similar to “all rights reserved”.
I am not a lawyer, but my understanding is that for versions of an article, as Rick notes here, creating a new version of an article under a less restrictive license is not a problem (and those terms supersede the previous version’s terms). Adding new, more restrictive terms don’t override the original’s. Arguments are made that anything different between the two, new text, copyedits, changes, new figures, etc. is copyrighted and restricted in its distribution, ditto the typesetting and branding. How any of this would hold up in court is still an unanswered question, just as is what constitutes a “commercial” use of material under a CC BY-NC license.
After reading this discussion I wonder just who wrote the CC license. Was it read by a competent lawyer before being issued. If yes, the competency of the lawyer needs to be questioned it not what we have are amateurs placing creators in peril! What will eventually result is some kind of suit and the courts will decide just what the clause means.
The Creative Commons organization has provided both an account of how the license suite came about, and a database history of litigation involving their licenses.
I think the big takeaway here is that CCC needs to better word its licenses.
CC, not CCC. There are way too many “Cs” in the copyright space. Thanks.
Thank you for sharing your findings and analysis. We publish three online-only journals that are now OA, and authors can choose between CC BY-NC-ND and CC BY. We’ve been asked a few times post-publication to change the license from CC BY-NC-ND to CC BY, presumably because the authors didn’t know their funder’s license requirements when they selected the license at submission. In these cases, we have changed the license in the article online and published an erratum to note the change. So far, the change requests have always been in the direction of expanding access. We’ve wondered what we would do if an author wanted to go in the opposite direction, so this post is very useful.
One question I have about this is history. Lets say an article is initially licensed under a CC-BY-NC license, implying the commercial rights are retained. Anyone using that work commercially (at that time) would theoretically be violating the assigned license terms. If at some later date, that work’s license is “updated” to a more expansive license like CC-BY, does that new license become retro-active? Would it nullify any concerns about prior commercial uses that were against the NC terms before to the change to a more expansive license?
Good question. My guess is that the prior use would still be a license violation and that technically, the copyright holder could pursue some kind of redress — but that the subsequent change in license terms would make prosecution harder.
But an attorney might be able to provide a more authoritative answer.
It has always struck me that the CC license’s need for simplicity leads to several “issues” like this one. As a corporate librarian, the NC license type leaves a lot up for interpretation. I think many authors use it to ensure their content is not republished by another entity for profit, but publisher try to use it to limit my constituents use of OA content, even for legitimate R&D. As COVID proved, access to scientific information is imperative for all scientists, even those working in the corporate sector.
Thank you for the interesting read. I fully agree about the key takeaway: “You should think carefully before choosing a Creative Commons license.”
It seems to me that licencing one’s work as CC0 or CC-BY (if they got to choose) is a big decision to make, and one should understand the consequences before they agree to do so. Personally, I don’t think it makes sense to attach a more restrictive licence to the original, or there’s a need to do so at all (except for rare cases of technical errors?). In a word, it’s a matter of respecting a promise you signed up for. If unsure and if a choice is available, starting with a more restrictive license seems to make obvious sense.
For what it is worth, I would like to share a case where we allow the copyright holder of a CC-BY-NC work to allow commercial use based on a specific request for such use. We based our decision on the information provided by CCO on “non-commercial interpretation” (available here https://wiki.creativecommons.org/wiki/NonCommercial_interpretation):
> NonCommercial licenses are non-exclusive.
> Like all CC licenses, the NC licenses are non-exclusive. This means that an NC licensor is free to offer the material under other terms, including on commercial terms. A frequently discussed use case for the NC licenses is a creator who wishes to allow NonCommercial use but also authorizes commercial uses in exchange for payment. (Additional permissions such as this may always be offered; licensors may also use our CC+ protocol to offer these in a standardized manner.) Also, licensees are always free to contact licensors to ask permission to use the work for commercial purposes.