In this blog post, we feature an article by IDRC Senior Program Officer Dr Matthew Smith who considers some of the debates around definitions of openness and the problems of an apparent consensus that privileges legal permissions as necessary and essential to what constitutes openness. This article was first published in the February-March 2016 edition of the ROER4D Newsletter.
As someone who thinks about and funds research on openness in developing country contexts, I’ve often wanted to ditch the word open altogether. It is such a value-laden term, with so many potential interpretations that people attribute whatever meaning they like to it – often with great passion. Then we end up in endless debates regarding effectively arbitrary definitions. Given that any application of “open” to a new social innovation (like open educational resources or open government) is really just a social convention, can we really say that one definition is the right one?
I am pretty sure this debate will continue forever – unless, perhaps, we can begin to think differently about it. As a program officer at Canada’s International Development Research Centre (IDRC), a research for development funder, I have seen interesting emerging findings that call into question standard assumptions about openness – and this has important implications for research. But before we get there, let’s visit one recent debate.
Stephen Laster, the Chief Digital Officer at McGraw-Hill Education, wrote a piece entitled The Future of Education Isn’t Free. It’s Open. He argues that openness is a technical feature whereby “technology or content […] can integrate painlessly with other resources.” David Wiley of OER fame, wrote a rebuttal taking issue with Laster’s definition. While agreeing standards are critical, Wiley argues that a “consensus” definition of openness is “free plus permissions”. In other words, openness is a legal characteristic providing 1) free access, and 2) a formal grant of permissions that “provides users with free and perpetual permission to engage in the 5R [Retain, Reuse, Revise, Remix, and Redistribute] activities”.
At IDRC, we have recently supported research on the adoption and impact of OER in the Global South, alongside other research on Open Science, Open Data, and Open Innovation, which is showing that requiring that narrow, technical or legal definitions are used may actually hinder our ability to really understand the important stuff, i.e., the “open practices” — sharing, reuse (5R’s), and collaboration that these legal and technical characteristics are intended to enable. OER by themselves don’t do anything – they don’t have an impact just sitting in the cloud or on someone’s Raspberry Pi. It is only when they are used in particular ways that change can happen – and it is this change that motivates most people interested in “open” in the first place. Critically, what the research suggests is that open standards and/or legal permissions are neither necessary nor sufficient for some people to treat the material as open in practice (i.e., engage in the 5Rs practices) to make or do something useful or valuable with that technology or content. This is true particularly in developing country contexts without active copyright enforcement or culture.
What the research in the developing world is revealing over and over again is that “free with permissions” can happen through social rather than legal means – it may be based on norms rather than law. This is not a new realisation:
“‘Legal commons,’ are commons established by law, or with support in the legal system. An example is the case of open content licensing. By voluntary action, and aware of the legal implications, content producers license their works under terms that will allow for the building and management of a commons. … A social commons, by definition, is not generated by intellectual property regimes, such as copyright law … In situations where intellectual property enforcement is either impossible or counterproductive, people frequently behave toward protected content as if it were part of a commons, and as if intellectual property regimes did not exist, or simply did not matter.” (Mizukami & Lemos 2010)
So, why should we care?
From a definitional standpoint, the legal element is just one of several potential (alongside social and technical) enablers of the “permissions to engage in the 5Rs” that Wiley mentions that are at the heart of openness. Why should we elevate just one of the enablers to the level of the definition?
Unfortunately, however, the consensus definition can be problematic or even detrimental, for several reasons:
1) It can limit or distort our understanding of reality.
If we are bound to an openness construct that necessarily includes legal openness, we end up missing out on a whole world of open practices around socially “open” (but not necessarily legal) content. From a research and policy perspective, it is useful to understand this whole ecosystem (legal and illegal) of access and use.
When we were setting up the ROER4D project, there were many debates about which definition of OER we would take. The legal (and widely accepted at the time) definition won the day – mainly on the strength of the arguments of the lawyer in the room as well as those that wanted to ensure that the research fit into the larger community of OER research.
Turns out that this perspective meant that we had a blind spot for a lot of really important practices that were having a developmental impact. One of the clear findings emerging from the ROER4D project is that in practice the line between closed, free, and open educational resources is often unclear, if it is even there.
For example, in Karnataka, India, an online community of practice of teacher educators create, share, adapt, reuse and redistribute educational content. However, research undertaken by Gurumurthy Kasinathan of IT for Change found that only ~7% of the educational content that is created and shared has an explicit copyleft license. How do we benefit by excluding the other 93% from the world of open education or OER? Aren’t the sharing, reuse, and other open practices predicated on the content and technology what is really important here? The presence or absence of a copyright license effectively makes no difference to these teacher educators.
This similar pattern has been found in many different contexts – particularly at the level of small classrooms where intellectual property oversight is not feasible at scale. For example, research by Laura Czerniewicz in South Africa indicates that in many contexts students and educators can’t successfully distinguish between ‘open’ and ‘closed’ educational resources. (It should be said, however, that in many cases universities in the developing world are becoming increasingly worried about being sued for copyright infringements. Such enforcement crackdowns do happen on occasion, reducing access to educational materials, particularly for marginalized populations).
The same finding also emerged from the Open Data Research network. For example, a recent synthesis piece that looked at 17 case studies of open data in developing countries found that, “Very few datasets are clearly openly licensed, and there is low understanding of what open licenses entail. There are mixed opinions on the importance of a focus on licensing in different contexts” (Davies, 2014).
Conversely, there are cases where a government might put data online with an open license, but never publicizes it, effectively hiding it from public view, “some of them even hope that the data will not be discovered in this pile” (Janssen et al 2012).
From strict interpretation of open as legal, the latter would be considered open data (although it isn’t open in any real sense) while the former would not be open data (although there may be real use of that data).
In other words, scoping the concept of openness as legal permissions is an arbitrary boundary that doesn’t reflect reality, i.e., the practice of openness. A poorly scoped concept can also lead to arbitrary groupings that distract from the real issue at hand. For example, Hilton’s review of OER impact literature included the Carnegie Mellon University’s Open Learning Initiative (OLI). While from the legal definition perspective the OLI’s materials are “open”, it is misleading at best to include it in a study on the “impact” of OER. Rather than its openness vis-à-vis a copyright designation, the impact results from its interactive learning design. Consider the title of the 2012 paper Hilton cites titled, “Interactive Learning Online at Public Universities: Evidence from Randomized Trials”. If openness is really free + permissions, and the free + permissions have no contribution to the impact, is this really an impact of an “open” resource? (If you argue that the impact is just because of increased access, then the impact results from just being free.) My guess is that this is glossed over due to the desire to show that OER can be as good as non OER – and therefore are a cheaper alternative – which is true. However, I believe this is a disingenuous argument for two reasons: 1) the selection of educational resources for any one comparison is (and always will be) arbitrary. For example, one could always select a low quality OER and compare it to a high-quality ER and find the opposite result; and (more importantly) 2) this approach misses out completely on why open advocates like Wiley tout openness: the 5Rs open practices themselves.
2) The consensus definition takes a fairly black-or-white, in or out approach – and might miss out on the benefits of moving towards open.
Strict adherence to free + permissions sometimes can deter people from engaging in open practices in the first place by raising technical and legal barriers. Evidence from the research on open data in developing countries provides stories of officials, “realising the challenges in moving from closed data, to fully machine-readable and openly licensed data, may be deterred from starting if their early steps towards openness are criticised as ‘not open’, rather than recognized as steps on the way towards openness” (Davies, 2014). This approach can also be seen on the technical side. For example, consider how for some pdfs aren’t open or pdfs are bad for open government.
In the education space this can be most easily seen in the free versus open debate. Consider how MOOCs generate a lot of debate because in many cases “open” in MOOC means “free” but doesn’t include permissions. This makes MOOCs not only a misnomer but also a threat to OER. The argument is based on the assumption that “before long the general public will feel that ‘free’ is good / innovative enough, and no one will care about ‘open,’ permissions, or licensing.” (Wiley 2012)
I’m not sure that assumption has been proven yet – and the opposite might just be true. You could have made a similar argument about America Online (AOL) in the 90s – which tried to create a walled-garden at the start of the Internet. Turned out the when most people got a taste of the broader Internet, they moved on. Indeed, this might also be the case here. Just getting your hands on free stuff might move you to want to reuse, republish, etc., which, it seems, is already happening around the world with stuff that is freely available, although technically open. Perhaps free is a stepping stone towards open. This would make for a really interesting research study.
Furthermore, especially in developing country contexts, free content, like MOOCs, might be enough for many situations, bringing great benefits. Similarly, publishing in pdfs might make life more difficult – but in some contexts something may be better than nothing at all. For example, EngagneNY, a common core curriculum, has had hundreds and thousands of downloads, despite it consisting of pdfs. From a development perspective, perhaps we should appreciate when and where free is sufficient to bring the desired benefits – and where pushing for more openness is necessary.
In conclusion: a more open open?
My argument is the following: 1) we care about openness because of the practices the permissions allow, not because of the permissions themselves, 2) that the ‘consensus’ definition gives primacy to only one of several potential enablers of the expression of these openness practices (missing out on social openness), and as a result 3) doesn’t reflect a lot of the reality of practice on the ground.
Definitions, however, are critical – particularly for research. So what can we do?
One alternative approach would be to take a grounded theory approach to the open definition. In other words, we build up a definition based more on what is happening in practice, rather than pre-conceived theory about open. Given the evidence emerging from IDRC supported research, the conclusion would be to focus on openness in practice, what that looks like, how to do it well, and its benefits – regardless of legal or technical status. I see this as the logical evolution of openness: First we define it (arbitrarily), then we research it, and then based on the new evidence, we redefine it.
The proposed definition above is not to argue that copyleft licenses or technical openness are not highly beneficial or necessary in many contexts. It’s just that it is an empirical question as to the extent the different technical and legal characteristics help encourage open practices in different contexts. We should be asking: for a given context, what are the different configurations of legal, technical, financial and social characteristics that are necessary to enable the types of open practices we are interested to achieve a particular goal? Constraining openness to be dependent upon a legal or technical characteristic risks misses out on this nuance.
One of the potential benefits of this approach is that it could be a more inclusive and welcoming notion that might bring more people into the fold. Perhaps then we can see free offerings and technical interoperability, as stepping stones to increased openness rather than a threat. Then we can really embrace the goal of “a more open, collaborative future [where] we can accomplish our goals by putting students and educators in a better position to achieve theirs” as Wiley wrote.
Originally published in the February-March edition of the ROER4D Newsletter on 7 March 2016