Hypertext & Pedagogy

Tags: | |
POSTED: June 14, 2006

This paper addresses the question: is there a politics of representation concerning digital knowledge schemes?

I will look at the history of hypertext, and locate it within a specific trend in cultural politics. I will then show how some of the ambitions of the hypertext visionaries, specifically Ted Nelson, have been realised in general terms in the world wide web, and in educational terms in CmapTools. I will briefly compare CmapTools and study.log in order to demonstrate the different political positions they occupy with regard to the representation of knowledge. Finally I will suggest how a politics of representation concerning digital knowledge schemes relates to the development of pedagogy and specifically e-pedagogy.

A very short history of hypertext, Xanadu and the web

Vannevar Bush is usually credited with the first description of hypertext-like capability in an article he published in 1945 entitled As We May Think. He named the system proposed in the article Memex, and it was based on the use of microfilm. Ted Nelson was the first to conceive of a computer-based hypertext, in a paper written in 1960 while he was a graduate student at Harvard University. He first used the term hypertext publicly when he presented a paper on “zippered lists,” at a national conference of the Association for Computing Machinery in 1965.

In 1974 Nelson founded the Xanadu project, and “zippered lists” initially played a central role in the Xanadu system. Xanadu was intended to provide a worldwide storage system in which all information would be available socially, and all information would be stored only once. In Dream Machines (1974) Nelson claimed that this

reflects a fundamental philosophy. For those to whom the world is hierarchical, cut and dried, there is no reason to want to keep reorganising ideas, writings, indexes, directories. But for anyone who must constantly work with ideas, writings, indexes, directories, the problem is their ever-swirling change. Basically, then, the Xanadu system is just one thing: a new form of interconnection for computer files – corresponding to the true interconnection of ideas – which can be refined and elaborated into a shared network.

Nelson has worked on Xanadu continually since 1974, although it has still to be completely implemented and publicly released. It has been through many incarnations, and a number of influential people, including Eric Drexler, have worked on it. Xanadu has served as an inspiration to many people during its thirty year history, although it has also been derided as an unfinished and unfinishable fantasy. In 1999 all the existing code for the Xanadu project was released as open source.

It is interesting to note that Tim Berniers-Lee spent time in discussion with Ted Nelson in the period when he was developing the ideas that became HTML, and thus led to the development of the world wide web. Nelson sees the web as

a brilliant simplification. As I understand it, and maybe I have this wrong, but Tim Berners-Lee came and we had lunch, in, oh I guess it was 1989, 90, something like that, in Sausalito, and I really liked the guy, and he’d done this very simple thing, and it sounded too trivial to me {laughs} but he certainly was a nice fellow and I expected to keep in touch with him, although I am a very bad correspondent, and the next thing I knew suddenly the thing had caught on.

The world wide web has many of the features Nelson envisaged for Xanadu, but by no means all of them. To cite just two examples of the differences, Nelson envisaged links that went in many directions, as well as a system of automatic payment for citation, quotation, or copying that would replace copyright payments. The mechanisms he envisaged thirty years ago would have enabled the recording industries to make a smooth transition to digital downloading instead of embarking upon unpopular and restrictive measures such as so-called “digital rights management”.

The politics of hypertext

Nelson conceived hypertext and Xanadu as computer counterculture projects that arose as much from cultural politics as from technology. In Nelson’s words: “not the nature of the machine, but the nature of ideas, is what matters”. His pursuits were part of a broader academic movement that included the pioneering educational work of John Holt, author of How Children Fail and How Children Learn, as well as the sociological work of Paul Goodman. They were informed by the thinking of Thomas Kuhn, whose book The Structure of Scientific Revolutions had demonstrated and then popularised the idea that creativity and invention were necessarily social activities.

Nelson was concerned with the use of computing in schools, and especially with then-current practices of “computer aided instruction” which he regarded with abhorrence. He wrote that :

1. The human mind is born free, yet everywhere it is in chains. The educational system serves mainly to destroy for most people, in varying degrees, intelligence, curiosity, enthusiasm, and intellectual initiative and self-confidence. We are born with these. They are gone or severely diminished when we leave school.

2. Everything is interesting until ruined for us. Nothing in the universe is intrinsically uninteresting…

3. There are no ‘subjects’. The division of the universe into ‘subjects’ for teaching is a matter of tradition and administrative convenience.

4. There is no natural or necessary order of learning…

His ideas of “responding media” and hypermedia were aimed primarily at education and literature, and were necessarily collaborative, in the same way as the telephone is. They were an attempt to create tools to implement what Kuhn had uncovered and Holt was promoting: the idea that all learning and discovery was collective. At this period Nelson wrote many polemics attempting to establish a social, politically aware computing at the service of progressive thinking and autonomous living and learning.

Hypertext, then, originated as part of a liberational cultural and political movement with aims that were as political as they were technical.

Hypertext, constructivism and concept mapping

In How Children Learn, John Holt wrote that it is “a serious mistake to say that, in order to learn, children must first be able to “delay gratification,” i.e., must be willing to learn useless and meaningless things on the faint chance that later they may be able to make use of some of them. It is their desire and determination to do real things, not in the future but right now, that gives children the curiosity, energy, determination, and patience to learn all they learn”. He also pointed out that “when we feel powerful and competent, we leap at difficult tasks. The difficulty does not discourage us”.

This critique was a statement that paralleled the emerging constructivist epistemology that was becoming influential in both in education and psychology; a thinking reflected in more academic publications such as DP Ausubel’s Educational Psychology: a cognitive view , in which Ausubel stated that ” If I had to reduce all of educational psychology to just one principle, I would say this: The most important single factor influencing learning is what the learner already knows. Ascertain this and teach him accordingly.”

The pedagogy that arose from constructivism implied a new political position with regard to learning and learners. Not only did it argue that learning, and the construction of knowledge, was a necessarily social activity, but they also argued that the learner must be trusted to learn at her own speed in her own way. In fact, many including Ted Nelson and John Holt, went as far as to say that the current institutions, and institutional arrangements, for education were actually counterproductive and merely serving to enslave people. This position, and the political dimensions that it housed, was voiced most eloquently by Ivan Illich in Deschooling Society.

This movement was not merely philosophical. There were many practical aspects to it, including the birth of the “free school” and home schooling movement. There were also attempts to create learning tools that would reflect constructivist thinking.

One such attempt began to bear fruit in 1972, when Joseph D Novak and his research team at Cornell University created concept mapping. This began as a way to map the results of interviews with children, but was soon developed as a powerful tool in its own right.

Novak has described concept maps as showing “the specific label (usually a word or two) for one concept in a node or box, with lines showing linking words that create a meaningful statement or proposition. We define concepts as perceived regularities or patterns in events or objects, or records of events or objects, designated by a label. Concepts are arranged hierarchically with the most general, most inclusive concept at the top, and the most specific, least general concepts toward the bottom. Propositions are statements about some event or object that shows a relationship between two or more concepts. There may also be cross-links showing relationships between concepts in two different areas of the concept map. Identifying a new crosslink may sometimes lead to a creative insight. Concept maps are also based on an explicit cognitive psychology of learning, and constructivist epistemology.”

In 1987, before the advent of the world wide web, Novak began working with the Institute of Human & Machine Cognition to further develop concept mapping as a collaborative, network-based tool, suitable for learning and other creative purposes. This led to the release of the free CmapToolkit, consisting of CmapTools and CmapServer. Used together these provide a collaborative means, a social software, for sharing concepts; and Novak and his co-workers have explicitly encouraged this by establishing open CmapServers where anybody can store their concept maps.

Concept maps link directly to the rest of the world wide web because each concept and proposition can contain embedded links to related resources on the server or anywhere else on the web. These embedded links can also be to other concept maps, and Novak has been encouraging the use of such links to tie different areas of public knowledge into what he terms “knowledge soups”.

Politics, hypertext and concept mapping

Currently, for most people think, hypertext is simply another name for the web. From this perspective, it is useful to see the world wide web itself in the way that Nelson has suggested: as a very clever simplification of some of the core ideas of the Xanadu project. It is these ideas that make the web interesting, and have made it important culturally and politically.

Ted Nelson’s vision of hypertext and Joseph Novak’s vision of concept mapping have several important features in common. They both propose systems that are explicitly linked to calls for the primacy of the learner and learner’s right to learning autonomy. They both distinguish between meaningful learning and rote, or mechanical, learning. They both propose schemes which are essentially socialist in the nineteenth century sense that Raymond Williams (1983) describes as “an emphatic and distinguishing term, explicitly contrasted with individual and especially individualistic theories of society”.

Both Nelson and Novak are much more concerned with ideas, with concepts, than with information. Both, in different ways, are concerned to establish a digital commons in which ideas are pooled and shared. Both have deliberately produced open-ended systems in which documents or maps have no fixed edges and no natural stopping points.

Web pages are social objects that have no natural borders. A link on a page can direct the user to almost any other page on the web. A reference to a book can take the user directly to the relevant page on Amazon. A passing reference to Snap, Crackle & Pop can take the user directly to a page deep inside the Kellogg’s website.

Concept maps can be created and edited collaboratively, synchronously and asynchronously. They can be blended together into different knowledge soups.

This is in sharp contrast to individualist ways of representing knowledge and information digitally. The University of Hamburg’s study.log, for example, is an entirely individual tool, that privatises knowledge and makes what I know available only to me and to those people to whom I choose to give a copy of a “context”. If CmapTools gives us access to a shared, social digital commons, study.log provides a door into a closed and private library. Concept maps offer a method to communicate ideas visually, in a non-sequential, non-linear way. As a communication tool CMapTools is necessarily social in its biases. and the model it offers is open and democratic. Study.log offers a way to store material that may or may not be useful in constructing concepts and propositions, and is geared towards private contemplation in both its structures, metaphors and interface.

This can clearly be seen by examining how information flows through a group of users. One person, perhaps a teacher, uploads a study.log context, containing embedded information, and the others download a personal copy of this into their own study.log. If one person subsequently adds to the material in the context, or rearranges the context into a more useful form, nobody else in the group know about this. Nobody has any way of finding out, unless the new arrangement is again uploaded and downloaded by all members of the group. If two people make different changes to the context simultaneously then, from a group perspective, the situation is hopeless, because there is no way to socialise the group’s changes into a single new context (or no easy way).

At a deeper level study.log privileges the collection of information, as opposed to the construction of concepts. Epistemologically it is merely a method for bundling material that lies somewhere between a name-based system such as desktop folders and a keyword system like GMail. Information is necessarily individual, in terms of meaning, for information only becomes social when it is attached to a concept and thus becomes explicable in terms of its importance and context. Knowledge, however, is necessarily social.

Hypertext, concept mapping and pedagogy

For a tool to be useful as a learning object it needs to represent knowledge in an open and social way. It needs to allow the user to navigate it, and to understand why they are navigating it. It can only do this if its structure is clearly visible; if it is designed to privilege the conceptual stream rather than the individual droplets of information.

Such a tool also needs to be open in the sense that the learning objects it produces are easily publishable, reproducible and reconfigurable. CmapTools has at least three different levels of publication, allowing for different levels of interaction.

Firstly any Cmap may be published as an image. This provides a read-only outline, which may have utility as an illustration within a learning object.

Secondly Cmaps can be published as web sites. This creates a web page with the Cmap outline as an image mapped in HTML with automatically generated Javascript actions. In this form of publication any links on the map (to other maps, to locally stored resources, or to links on the web) are live and may be edited from the HTML on the web page. The outline itself cannot be edited.

Thirdly Cmaps may be published on a CmapServer. In this case editing permissions can be granted to users, as in most content management systems, and users with sufficient permissions can edit any aspect of the Cmpa including the outline itself. Using CmapServer users may work together from a distance on a map, with several users working on the same map simultaneously.

Using these methods truly social learning objects can be constructed in which project-based (constructivist) learning occurs, and knowledge increases, collaboratively. This represents a distinct cultural and political position in the field of distance learning, very different from the positions taken by the creators of WebCT or study.log.


There is necessarily a point of view about any scheme for representing knowledge. As Ted Nelson has pointed out; “all simulation is political. Every simulation program, and thus every simulation, has a point of view. Just like a statement in words about the world, it is a model of how things are, with its own implicit emphases: it highlights some things, omits others and always simplifies”.

Hypertext, concept maps and study.log contexts are simulations of thought processes, models of “how things are”. As I have attempted to demonstrate they contain political and cultural points of views embedded into the ways in which they represent knowledge: views about who owns knowledge, how knowledge is derived and constructed, and how it is shared. The move to embrace social software in many areas represents a political and a cultural shift, and one that is firmly on the side of cultural democracy.

Hypertext, and more specifically CmapTools, can be important tools through which this shift can be applied to democratic learning and pedagogy, and to distance learning and epedagogy.


DP Ausubel, Educational Psychology: a cognitive view, 1968
Eric Drexler Engines of Creation, 1986
John Holt, How Children Learn, 1966
Ted Nelson, Computer Lib, 1974
Ted Nelson, Dream Machines, 1974
Joseph D Novak & Alberto J Canas, The Origins of the Concept Mapping Tool and the Continuing Evolution of the Tool, 2001
Raymond Williams, Keywords, 1983