From “Gratis and libre open access,” SPARC Open Access Newsletter, August 2, 2008.
In February 2002, the Budapest Open Access Initiative called for a kind of online access to research literature that was free of charge and free of most usage restrictions. It offered a name (“open access”) for the unified concept, but it didn't suggest names for the two component parts.
In a February 2003 article, I distinguished those two parts and called them the “removal of price barriers” and the “removal of permission barriers.” But those were negative terms, and I didn't think to offer matching positive ones describing kinds of access rather than kinds of access barriers.
When the Bethesda and Berlin statements came out (June and October 2003) they followed the Budapest statement in calling for the removal of both price and permission barriers. As a result, all three components of the Budapest-Bethesda-Berlin (BBB) definition of OA now call for both sorts of free online access.
But unfortunately we still don't have widely accepted terms for the two sorts of free online access: (1) the kind which removes price barriers alone and (2) the kind which removes price barriers and at least some permission barriers. This gap in our vocabulary has caused confusion and conflicts, not least because it created pressure to use the term “open access” for each.
In April 2008, Stevan Harnad and I proposed the terms “weak OA” and “strong OA” for these two species. I wrote a blog post to explain what we meant.
But we quickly realized that “weak OA” was needlessly pejorative and started looking for more neutral and descriptive language. Stevan launched a public discussion on the American Scientist Open Access Forum.
Unfortunately, however, the discussion hasn't come up with clear winners. If the terms Stevan and I introduced in April had been better chosen, and widely supported, I'd have used them 100 times already in my blog and newsletter. I need them almost every day.
In the absence of terms that are neutral, accurate, and widely supported, I've decided to make a provisional decision as an individual writer while the larger discussion continues. For now, my choice is to use “gratis” and “libre.” They are accurate, neutral, and descriptive. In the neighboring domain of free and open source software, they exactly express the distinction I have in mind.
The terms may be unfamiliar in the domain of OA or scholarly communication. But as far as I can see at the moment, that's their only drawback, and it's one I may be able to overcome by writing this article. Their relative unfamiliarity is even a kind of advantage. They're not common words with clouds of common connotations. “Weak/strong” were not objectionable because of their new definitions, but because of their preexisting connotations, and “gratis/libre” won't run into that problem.
This choice is personal in the sense that I'm making a decision for my own writing. It's provisional in the sense that I'll continue to look for better terms.
I've already used “gratis/libre” in a few blog posts. But I didn't want to use them frequently until I had time to write up this longer case for them.
Here's the updated heart of my April 2008 blog post using “gratis/libre” in place of “weak/strong.” But while my original post described a decision Stevan Harnad and I reached together, in this version I'll speak for myself.
The term “open access” is now widely used in at least two senses. For some, “OA” literature is digital, online, and free of charge. It removes price barriers but not permission barriers. For others, “OA” literature is digital, online, free of charge, and free of unnecessary copyright and licensing restrictions. It removes both price barriers and permission barriers. It allows reuse rights which exceed fair use.
There are two good reasons why our central term became ambiguous. Most of our success stories deliver OA in the first sense, while the major public statements from Budapest, Bethesda, and Berlin (together, the BBB definition of OA) describe OA in the second sense.
I've decided to use the term “gratis OA” for the removal of price barriers alone and “libre OA” for the removal of price and at least some permission barriers. The new terms allow us to speak unambiguously about these two species of free online access.
On this new terminology, the BBB definition describes one kind of libre OA. A typical funder or university mandate requires gratis OA. Many OA journals provide libre OA, but many others provide only gratis OA.
There is more than one kind of permission barrier to remove. Therefore, there is more than one kind or degree of libre OA.
I've often wanted short, clear terms for what I'm now calling gratis and libre OA. But I've also wanted a third term. In my blog and newsletter I often need a generic term which means “gratis or libre OA, we don't know which yet.” For example, a press release may announce a new free online journal, digital library, or database, without making clear what kind of reuse rights it allows. Or a new journal will make its articles available online without charge but say nothing about its access policy or licensing terms. I will simply call them “OA.” I'll specify that they are gratis or libre OA only when I learn enough to do so.
The two new terms will help us avoid ambiguity without resisting current usage, which would be futile, or revising the BBB definition, which would be undesirable.
I learned a lot from the emails leading up to the “weak/strong” announcement in April, and especially from the emails and blog posts afterwards. Here's a mini FAQ responding the sorts of questions I've heard.
Why introduce new terms at all? Weren't we doing fine before?
We were not doing fine. Our central term was (and is) widely used to cover two non-equivalent sorts of free online access. As long as we don't have narrower terms for the two sorts, then we'll continue to use the broader term “OA” for each, aggravating the ambiguity rather than resolving it.
Are you saying that we should stop using the term “OA” and only use the narrower terms?
Not at all. I'm only introducing terms for sub-species when we need to speak unambiguously about sub-species. When we don't need that level of precision, “OA” is the perfect term, indeed, the only term.
“Gratis OA” and “libre OA” will supplement “OA,” not supplant it—roughly the way “simple carbohydrate” and “complex carbohydrate” supplement “carbohydrate” without supplanting it.
Isn't the green/gold distinction the same as the gratis/libre distinction?
No. The green/gold distinction is about venues (repositories and journals), while the gratis/libre distinction is about user rights or freedoms. Green OA can be gratis or libre, but is usually gratis. Gold OA can be gratis or libre, but is also usually gratis.
It's easier for gold OA to be libre than for green OA to be libre. But both can be libre. It follows that the campaign to go beyond gratis OA to libre OA is not just about journals (gold OA), even if it is primarily about journals.
For more on how these two distinctions differ, see the table I posted to my blog this morning,
Are you trying to legislate usage?
No. I'm in no position to legislate usage. If I were, usage would never have become ambiguous!
I am proposing these terms for others to use as well. But even if I were able to legislate usage, this choice would still be personal and provisional. I need terms to use in my own writing, and I'll use these until I find better ones. I don't want to preempt the search for better terms, by myself or others, and I don't want to tie my hands about using the fruits of that search.
Are you trying to revise the BBB definition of OA?
No. It hasn't changed and I don't want to change it. Nor—slightly different thing—am I retreating from my endorsement of it. (I was the principal drafter of the Budapest statement and stand by it.)
I'm simply trying to clarify communication by introducing terms for different sorts of free online access. It's all about vocabulary, and not at all about policy.
There is a problem to solve. It's not that the BBB definition has changed, or needs to change, but that the term “open access” has changed, and is now widely used in both a BBB and non-BBB sense. As I've argued elsewhere, our term has spread faster and further than the BBB definition. That usage is a fact of life, and support for the BBB definition doesn't make it go away. There are roughly two ways to approach this problem. We could fight the tide of usage and try to make “OA” refer to nothing but BBB OA again. But that's unwinnable. (I deliberately say nothing about the advantages and disadvantages of winning it if it were winnable; that's a pointless exercise.) Or we could cure the ambiguity by using separate names, like “gratis OA” and “libre OA,” for the two important things which have been going under the same name. That's more than winnable. It's easy. It will support unambiguous communication without fighting usage, without modifying the BBB definition, and without giving anyone a reason to diminish their support for it.
Is “libre OA” synonymous with “BBB OA”?
No. Because there is more than one kind of permission barrier to remove, there is more than one kind or degree of libre OA. BBB OA is one kind or subset of libre OA. But there are others, and not all libre OA is BBB OA.
For example, permitting all uses except commercial use (the CC-NC license) and permitting all uses except derivative works (the CC-ND license) are not equivalent to one another and—ignoring certain subtleties—not compatible with the BBB definition. But they all remove price barriers, they all remove at least some permission barriers, and therefore they are all libre OA.
We shouldn't speak as if there were just one kind of libre OA. Gratis OA may be just one thing (freedom from price barriers), but libre OA is a range of things (freedom from price barriers and one or more permission barriers).
What's the best way to refer to a specific type of libre OA?
With a license. We'll never have unambiguous, widely-understood technical terms for every useful variation on the theme. But we're very likely to have clear, named licenses for every useful variation on the theme, and we're already close. Licenses are more precise than single terms and not nearly as susceptible to misunderstanding or divergent usage.
What's the borderline between gratis and libre OA?
Gratis OA removes no permission barriers and libre OA removes one or more permission barriers. (Both of them remove price barriers.)
But what does it mean to remove a permission barrier? If copying a short excerpt is permitted by “fair use” (or “fair dealing” or the local equivalent), then users may do it without asking anyone's permission. Hence, there are no permission barriers in the way. If copying full text and redistributing it to others exceeds fair use, then users must ask permission, take the legal risk of proceeding without it, or err on the side of non-use. In general, when a use requires permission, users face a permission barrier. This doesn't mean that permission is denied, only that permission is not already given and must be sought if one wishes to proceed. When rightsholders grant permission in advance for uses that exceed fair use, then they remove permission barriers. As a practical matter, there are two ways to remove permission barriers: (1) with copyright holder consent, through a license or statement permitting uses that would otherwise be impermissible or doubtful, and (2) with the expiration of copyright and the transition of the work into the public domain.
In short, gratis OA alone allows no uses beyond fair use, and libre OA allows one or more uses beyond fair use.
What's wrong with “full OA” (instead of “libre OA”)?
“Full OA” implies just one degree or kind of OA, in fact a maximum. But libre OA is a range of many positions, corresponding to the many permission barriers which we could remove.
Hence, “full” isn't a good word to contrast libre OA with gratis OA. But it's a perfectly good word to contrast OA journals with hybrid OA journals. For example, I see no problem with saying that “full OA” journals provide OA to all their articles, while “hybrid OA” journals provide OA to some and not others.
What's wrong with “free” access?
The main problem is that some people already use “free” to mean gratis (“free as in beer”) and some people already use it to mean libre (“free as in speech”). It would be very hard to give this widely used and versatile word a narrower meaning for some special purpose, and make it stick. Making “free” a technical term would increase ambiguity, not decrease it.
It's relevant that “gratis” and “libre” emerged to resolve an ambiguity endemic in the “free” software movement.
Why was the original “weak/strong” announcement presented as a joint decision with Stevan Harnad?
Simply because Stevan and I came up with it together. (Not because we are a cabal.) The background is important. Stevan wanted to recognize the movement's many gratis OA success stories. But because gratis OA doesn't meet the terms of the BBB, he felt it necessary to revise the BBB. I wanted to join him in recognizing the many gratis OA success stories, but I didn't want to revise the BBB. Whether the BBB needed revision became a tactical disagreement of growing importance between us, showing up in many of our blog posts, taking more of our time, and perhaps even overshadowing our agreements on most issues of substance and strategy. As we talked it out, however, we realized that our disagreement on the BBB arose from an inadequate vocabulary for the varieties of free online access. Once we had that vocabulary we could agree in our speech as much as we agreed in substance, and we could take revision of the BBB off the table. It was a beautiful resolution—except that we settled too quickly on the wrong pair of words (“weak/strong”).
To show the new terms in action, here's how they help clarify the major points of substance and strategy on which Stevan and I agree. We agree that gratis OA is a necessary but not sufficient condition of libre OA. We agree that gratis OA is often attainable in circumstances when libre OA is not attainable. We agree that gratis OA should not be delayed until we can achieve libre OA. We agree that libre OA is a desirable goal above and beyond gratis OA. We agree that the desirability of libre OA is a reason to keep working after attaining gratis OA, but not a reason to disparage the difficulties or the significance of gratis OA. We agree that the BBB definition of OA does not need to be revised.
Why do we have to recognize this distinction at all?
Because there really is a difference between removing price barriers alone and removing both price and permission barriers, and because this difference really matters to users, strategies, and policies. The distinction by itself isn't new or even controversial. All that's new here is the proposal to use certain terms to name its two parts. But even if you don't like the terms I plan to use, and even if you don't plan to use any special terms yourself, understanding the distinction itself is necessary to understand the day-to-day progress and discussions of the OA movement.