Connective Knowledge and Open Resources

The following is a set of responses to interview questions I answered over the weekend.

1.      Chinese readers firstly know you and George Siemens by the theory “connectivism”. How did you start your cooperation initially? What is the background (or reason) you create this theory? Can you introduce the meaning of connectivism and illustrate its characteristics in details?

That’s a good question, and I don’t really know the answer to it. George and I have been exchanging email since the late 1990s – we each had our own online newsletter (mine was OLDaily, his was called elearnspace). We read each other’s’ work, talked about the same subjects and encountered each other at conferences. And we had very similar perspectives on the nature of online learning.

To me, the theory is and always has been ‘learning networks’. This is a play on words: it is in part a theory about using networks to support learning, and in part a theory about how networks learn. For me, this is based on an associationist philosophy of knowledge, and rooted in connectionist theories of artificial intelligence. I wrote a long PhD proposal on it in 1990 –

http://halfanhour.blogspot.ca/2009/03/tnp-1-introduction.html

When I began working on online learning, I brought this theory with me. It led me to propose content syndication networks in education, and to embrace social media in education. I outlined my theory in the paper ‘Learning Networks’ in 2004

http://www.downes.ca/post/20 and George wrote his important paper also in 2004.

http://www.elearnspace.org/Articles/connectivism.htm

He came up with the term ‘connectivism’. I think it is derived from the term ‘constructivism’. His version of connectivism is much more socially oriented than mine, much more about social networks. Mine is rooted in artificial intelligence and neural networks.

Either way, the underlying propositions are the same. A person’s knowledge is the state of connectivity in their mind – the connections between their neurons – and to learn is to form or reshape these connections by strengthening or weakening them through interaction and use. It’s a theory where we grow knowledge, rather than acquiring it, where we learn by immersion in a field of study rather than by being told about it.

2.      The “connectivism” is extended to the whole world recent years, how did you combined it with practice in a subsequent study ? Can we think OER is an effective way to implement the theory?

The practical implementation of connectivism is the massive open online course, or MOOC. George and I developed the first MOOC in 2008, called ‘Connectivism and Connective Knowledge’. It was intended very deliberately to be an example of the theory we had formed over the previous four years, and not simply a place to talk about it.

The key aspect of the CCK08 MOOC was that it was distributed – that it, it was not based on a centralized core of content, and it was not located on a single website. Rather, we encouraged people to create course content and to contribute to the course through their own websites, through discussion groups, communities, and and other forum they chose. We had about 170 blogs created to support the course, and we aggregated these using RSS – content syndication supporting online learning.

The CCK08 course and all the other courses we have developed since have been based on open educational resources (OERs). We use OERs as a way to talk about the concepts and ideas in the course. They may be papers, lessons, diagrams and images, videos, or anything else. What was most important what that they were sharable – we needed to be able to move them from one location to another in our distributed course.

The course as a whole was formed in the structure of a network. The OERs were the signals being sent back and forth between individuals in this network. The knowledge created in the course was the result of this interaction, and not the content of the interaction.

3.      From your web & blog, we found you were a passionate advocate of OER, can you share with us your first experience about open online educational resource? and the purpose and meaning of the action at that time?

I knew about open educational resources before anyone knew what the term meant. By no means am I alone in this. It has always been common for teachers and instructors to create and share learning materials. It is only with the arrival of the internet that we could do so globally.

It’s really hard for me to find a ‘first’ instance of experience with OERs, but if I was forced to do so I would probably say it was my ‘Guide to the Logical Fallacies’. You can still see this guide today – http://www.fallacies.ca – and it was created in 1995. I created the text for it when I was teaching at Grande Prairie Regional College in 1994, because my students needed a good way to think about logical fallacies. I was trying to get them to spot fallacies by recognizing them, by seeing signs that would remind them of common patterns of argument.

When I moved to Brandon and began working at Assiniboine Community College, I converted the guide into HTML and posted it on my brand new web site. It was partially to share the content with the (small) internet community, and mostly to show people at the college what could be done with the web. It led to our first web-based course a year later.

This was before Creative Commons or OERs, and only a few years after Richard Stallman created GPL. But there was still a sense in the community that there should be a type of licensing that allowed people to use content for free. I attached such a license to my Guide to the Logical Fallacies – it was modeled on George Reese’s license for the Nightmare Mudlib software package.

4.      You have your own definition of OER which is apart from the official definition, can you explain it? And talk further about the role of OER in personal learning?

When I’m asked to define OERs I use the UNESCO definition, which is, “teaching, learning or research materials that are in the public domain or released with an intellectual property license that allows for free use, adaptation, and distribution.”

Generally, when I think of OERs, I think of digital resources, simply because non-digital resources cannot be shared in the manner envisioned in the definition.

In addition, my definition of ‘teaching, learning or research materials’ is a functionalist definition, rather than an essentialist definition. That is to say, I do not believe it is the nature of a resource that makes it a ‘teaching, learning or research material’, but rather, its use.

This is significant because some people say that OERs must have a certain format or construction – for example, they must have learning objectives, or an instructional focus, or a mechanism for assessment, or some such thing. I don’t agree with this. If a resource, whatever it is, is used to support teaching, learning or research, it is by that fact an educational resource.

Finally, when I employ the term ‘free use, adaption, and distribution’, my sense of the word ‘free’ includes the sense of ‘without cost’. Many supporters of OERs argue that the license must allow commercial reuse, including sale or subscription fees. I don’t agree with this. If someone wants to license their material to permit commercial use, I have no objection, but for me, ‘free learning’ does not have a price tag attached.

There has been in the past some efforts made to define a type of open license that would apply specifically to educational use. I have opposed this because such a definition would define educational use as use by a college or school. But to me, the great strength of OERs is that they enable a person to learn without relying on an educational institution. OERs become, as I said above, the vocabulary learners can use to talk with each other about a certain domain. In this sense, OERs ought to be freely usable, and sharable, not only by institutions, but more importantly, by students and learners.

5.      We all know you are the originator of the Massive Open Online Courses (MOOCs), what’s your perspective on OER’s potential impact on MOOCs? Can you summarize the major elements of MOOCs?

The major types of MOOC are the cMOOC, which wqe created, and the xMOOC, which is the sort of MOOC offered by Coursera, Udacity, and the like. The difference is that a cMOOC is designed as a network, as I described above, while an xMOOC is based on a central course site and content that will be followed by all students.

The xMOOC, in other words, is like a traditional college or university course. It typically requires that custom content be created, and is therefore very expensive to build. Costs can run into the tens of thousands of dollars. Sometimes this is necessary, if for example you are creating a MOOC in a new subject area or new language. But often, the content already exists on the internet as OERs.

In a cMOOC, we begin with the presumption that the content exists out there somewhere. So when the MOOC is set up, we begin with what is a bare outline, usually just a list of topics and ideas. Course participants begin by reconsidering this outline – not everything will suit everyone, of course, but nobody has to do everything. Then, as the course progresses, participants (including the instructor) find OERs from around the internet and link them to the course.

Note that this use of OERs is very different from the traditional online course. In a traditional course, or an xMOOC, all the content is brought into the central website and incorporated as a part of the course. Not only does this create a lot of work, it also raises licensing issues, as you are now copying or republishing the content. In a cMOOC, though, we leave the content where it is, and simply link to the content. So a cMOOC is not a package, like a traditional course, it is a network of related and connected content.

Sometimes this content will be found on the internet, but just as often it will be created by course participants. That’s why it’s important to have a mechanism – like RSS – to link different websites together. So in any given week of the MOOC, participants will be able to read text and watch videos that were either found on the internet, or to read text and watch video that was created by participants in the course.

What, then, are the major elements of the course? In one sense, they are whatever the course participants want them to be. In another sense, the major elements are:

-          The course participants, including the instructor

-          Resources. Either found on the internet or created by participants

-          A mechanism for linking participants and resources with each other

It’s this this third part that traditional courses and xMOOCs lack. Because they are centralized, everything is brought into the course site. But in a cMOOC, we use a system that connects people.

In our courses, we used the gRSShopper application that I created. http://grsshopper.downes.ca  It can be used as an RSS aggregator to harvest resources found or suggested by participants. It also harvests Twitter messages and discussion board posts. It then organizes these resources, most commonly as a daily newsletter, which is then distributed to participants. They can either subscribe by email, reads the RSS version, reads the website, or follow the course on Twitter or Facebook.

6.      Could you analyze the differences between cMOOCs and xMOOCs? As the commercially driven, xMOOCs explode in popularity very quickly, how do you think about the limitation of its development in the next level? And what’s the development space of cMOOCs?

Well as I mentioned already, the xMOOC is centralized and the cMOOC is distributed. The xMOOC has a core content that everyone follows, while the cMOOC has a wide ranging network of content which people browse through and sample according to their own interests.

But more than that, the pedagogy of an xMOOC and a cMOOC is different. The xMOOC is typically based on the transmission model of instruction. That is to say, the instructor will attempt to take some knowledge he or she has, and through a process of communications, transfer it to participants and students. Success in this kind of course occurs if the student’s resulting knowledge resembles or (ideally) is identical to the instructor’s knowledge.

In a cMOOC, by contrast, an immersion model of instruction occurs. That si to say, the instructor creates or sets up an environment similar to one in which the students will want to be successful. The instructor will model correct or successful behavior in the environment. Because the environment is dynamic and continually changing, the students will not be able to replicate the instructor’s knowledge, but instead will have to generate his or her own strategies. Success in such a course occurs if the student is able to successfully navigate the environment without scaffolding or support from the instructor.

The different is like, on the one hand, telling a person how to swim, and having them repeat that information back to you, and putting a person in the water and showing them how to swim, eventually letting them go to swim on their own.

There is room for both types of instruction. But in any field, students will have to graduate from the simpler transmission mode of instruction to the more complex immersion mode of instruction.

7.      In Your article you mentioned MOOCs provides a fair opportunity to learn, and it eliminates one of the great advantages the wealthy have always enjoyed over the poor. Can we think it bring great impact on developing countries’ education?

In the recent Times Higher education article George Siemens talked about students in India telling him their lives had been changed by Coursera. I can easily believe this. The traditional form of learning requires schools and professors and facilities and resources. The cost is enormous compared to what the provision of internet access costs. Whatever people can learn on their own is that much more countries do not have to hire professors to teach them.

A lot of people argue, correctly, that you cannot have a full education without in-person instruction, hands-on experience, and even the social contact you get with other students in a school, college or university setting. They are quite right. But the point here is, we do not need to set up all education in this way. And there is a great deal people can learn on their own.

We’ve seen this in India, even before Coursera. Many major software companies are built using Indian programmers. These people taught themselves on the internet. They accessed open source learning materials, they participated in online discussions and in online communities, they created software working on open source development projects, and through interaction and immersion, they developed their own software development skills. On their own, without college or university. Now they compete worldwide – I just took part in an online WizIQ session, through a very nice collaboration tool built and run by Indian software engineers.

Not every nation will focus on software, but every nation will focus on maximizing the amount of learning a person can do on his or her own. This creates tremendous opportunities for developing nations, because if they can enable their citizens access to these online resources, in their own language, then they will be able to develop skills equal to those in more industrialized countries.

Yes, this by itself will not be enough. But it significantly improves the capacity and opportunity for developing nations.

I should add, it’s not MOOCs specifically that create this capacity. It’s open online learning generally. MOOCs are just one way to structure and deliver educational experiences. Development communities – such as open source software communities – are another way. The main and underlying element here is open access to learning resources and community. Open access to learning networks, in other words.

8.      Accompanied by the MOOCs, a network-based pedagogy have been proposed, can you introduce the pedagogy? What’s the difference between a MOOC and a traditional course?

I think I have addressed that sufficiently above.

9.      We know you develop the gRSShopper application, can you introduce and explain how that affects MOOCs?

I have already described the functionality of gRSShopper. The way to think of it, and similar applications, is as an open content distribution system.

This is important. It is not enough simply to have open educational resources. If nobody can find them, if nobody can access them, they serve no useful function at all. If a person creates an OER and has no way to distribute it to students who are interested, the OER is not very useful.

gRSShopper employs open syndication standards – specifically, RSS and Atom, along with open APIs and other mechanisms, to create this distribution network. In a MOOC, the network is created specifically for a given course, but gRSShopper is designed to support open content distribution and syndication generally.

People who used Google Reader or who use Feedly or Evernote or similar tools are accessing a similar capacity. When we use channels like RSS, we do not need to depend on publishers or distributors to find and package information for us.

In the past, I used gRSShopper mostly to distribute content – to make my own RSS feeds and my own email newsletters. When I first started work on it, there weren’t any good ways to do this – people were building RSS feeds by hand. Over time we saw Blogger and WordPress and Livejournal developed, and RSS feed generation was automated, but I still needed the registration and email newsletter function. And none of these supported aggregation – the RSS readers and blogging engines were always separate products. So gradually it became my integrated tool that did everything.

It was this all-in-one functionality that made the first MOOC possible In this one tool I had a mechanism that would collect student postings and make links to them available to everybody else, in whatever format they liked. Without this tool, our course would have looked like any other course, and would have been nothing special. And when – inevitably – someone mass-marketed an open course (like Stanford AI) there would have been no alternative model to compare it against.

10.  You mentioned a paradox that the more autonomous, diverse and open the course, and the more connected the learners, the more the potential for their learning to be limited by the lack of structure, support and moderation normally associated with an online course. Do you find the appropriate solution or approach for this problem later? Whether the learning analytics(LA) can be think as problem-solving techniques?

Well the reason why their learning is limited by the lack of structure is that they have never learned to be able to learn for themselves in an environment that lacks structure. This makes them unable to function in most environments, not just MOOCs, with the result that even after completing their education they still have a great deal of learning to do in the workplace or the community.

Most educational institutions – and most corporations, and governments – respond by creating more structure, so people are more easily able to cope. And there is certainly a merit to this approach; when we want large numbers of people to be able to do something, the more clearly defined the structure, the more likely they will be successful. But this works only so long as the structure can be clearly defined. And in most of what happens in the workplace or the community, this becomes impossible.

Take transportation, for example. The most structured system of transportation employs buses or trams moving at predefined times. This works for as long as transportation needs are limited. But if people have more complex needs – they need to deliver milk, for example, or they need to go into the country – then the bus and tram system is inadequate. So we have to adopt a less structured system, such as the automobile, and teach people to manage their own transportation for themselves. Even with cars and trucks, there’s still structure – there’s still roads and traffic regulations and signals. But the denser and more difficult the traffic, the more ineffective such structures are, and people have to learn how to self-organize in traffic. Exactly the same is true of learning.

Right now, we have a basic learning system. We put people into schools and colleges and take them through predefined routes. But if we say we do this because people cannot learn for themselves, we are fooling ourselves. People will have to learn how to learn for themselves, because our learning needs are becoming as complex as our transportation needs.

So how do we approach this. Rita Kop and I looked at one way of doing it, using a MOOC to help people learn the basic tools of learning – the course was called Critical Literacies and it was a mixed success. I do intend to try it again, because I think that the basic tools for learning – the critical literacies – are just like the basic tools for traditional learning, like reading and writing, and can be (must be) taught and learned at a basic level of education, with children, as a platform to support future learning.

11.  With respect to actual assessment and credentialing, there are two basic approaches, except learning analytics, your mentioned your own approach-a network clustering approach, can you explain it for us?

I can sketch the idea, but it’s important to understand that it’s an idea that needs a lot of development, and that other people probably have a greater grasp on this idea than I do.

It’s a two-step process. In the first step, we identify what the competencies are that constitute expertise in a given area of study. In the second step, we compare an individual’s performance against that expertise.

What makes the method I propose different is that I propose to use network analytics for both steps. This is importantly different because the competencies won’t be typical competencies, which can be expressed as a proposition (like: ‘staple two pieces of paper together’) but will rather be a complex collection of network behaviours that can be described in words. So people won’t be able to just imitate the right words, they will have to be able to master the actual performance.

12.  Finally as the end of this interview, can you talk about technologies that will impact the development of MOOCs in the future? And make some recommendations for the development of China’s MOOCs? 

I have never even been to China, so I am very hesitant to make recommendations for the development of China’s MOOCs, except to offer wide platitudes: don’t close your options, embrace a diversity of approaches, focus on helping people learn for themselves, and develop a culture of sharing and cooperation. But you know all that.

Every technology will impact the development of MOOCs in the future, but we can probably reduce them to three major types of impact: carbon, carbon, and carbon.

First, carbon will play a key role in the development of super-strength materials for construction, etc. Carbon-fibre already has a wide variety of applications (my laptop, a Lenovo, made in China, is made from it). Carbon fibre redefines where and how we can build computing power into things.

Second, carbon will impact power and information storage. Carbon nanotubes, graphene supercapacitors, etc., will enable powerful computers to hang on the wall, be read like books or worn like clothes.

Third, carbon-based life (such as genetics, organic computing, crops, seeds and medicines) will inform the nature of computing in the future. Connectivism is only the first iteration of this. A next-generation version of connectivism will envision the develop of learning from the creation of network structures to the organization of knowledge in forms resembling DNA.