Link & Learn



The Memex concept

One of the highlights of this year’s Hypertext conference (which I missed) was Andrew Dillon‘s opening keynote. He is a great speaker—the Irish accent doesn’t hurt—and it would have been great to see  it. Perhaps a recording will materialize eventually. In the meantime, there is the written version that reviews the state of Hypertext research 65 years after some of its tenets were articulated by Vannevar Bush in the famous “As We May Think” article in the Atlantic.

Dillon observed that over these past 65 years, we’ve made some progress toward Bush’s vision, but all the hypertext rhetoric notwithstanding, we still read pretty much the same way as before. He reluctantly admitted that a considerable amount of reading is happening through the mediation of a computer these days, but pointed out that designers of online reading tools often strive to imitate the paper versions. He concludes that

Part of what we have learned over the last two decades is that genre forms are highly embedded in the cognition of communities. It is plausible to consider the emergence of new digital genres but we are witnessing to a greater degree is the shaping of communicative practices across media to the demands of our own architecture, with its preference for patterns and order.

Thus, although associative linking has its place, printed text has its own logic that we have designed to support our cognitive processes. There are, of course, echoes of the “Nature vs. Nurture” debate here; I think the Nature side of the argument has a stronger case.

Dillon also raised an interesting issue for the Hypertext research community, one related to evaluation. While hypertext has been touted as a natural technology for supporting learning, much of the evaluation of hypertext effectiveness has focused on navigation. Reflecting on the history of the discipline, he pointed out that

…the argument seemed to be that research would be better spent solving the problem of being lost in hyperspace than in addressing instructional impact.

Vannevar Bush imagined a technological solution to empower reading, writing, and learning. In this review, Dillon points out that while we have some mastery of the reading and writing aspects of hypertext, we’ve still got a lot to learn about  learning.

Share on: 


  1. Interesting — I’ve also read some stuff that the more hyper links the less learning.

  2. My sense is that learning is a complex enough phenomenon that no single factor such as the presence or absence of links can be shown to have reliable effects. But the burden of proof should fall on advocates of HT’s effectiveness in supporting learning. At the moment, the jury is out.

  3. Jim:

    Though there are a few studies that suggest hyperlinks may detract from learning, these tend to define “learning” in a very old-fashioned and problematic way. Because links allow the reader to explore more broadly and in greater depth, it may be difficult to measure exactly what the student has learned in the past minutes. Links make research harder, but there is no evidence at all that they are deleterious.

    Gene: Why should the burden of proof fall on advocates of hypertext? In any case, the base proof is trivial: hypertext links provide better access to a wider range of information than print. If you want to know something, you’re obviously better off exploring the docuverse and using links to gather information and weigh sources and approaches than you’d be with any single text.

  4. This is beginning to sound a bit like an argument around gun control: links don’t detract from learning, bad links detract from learning. I think Andrew Dillon’s point was that while we as a culture have developed pretty effective patterns for creating text that is useful for learning, we don’t have an equivalent store of experience with linking. It’s certainly possible to have effective links and ineffective linear text; what’s not clear is how easy it is to tell whether the link structure you’re creating (manually or automatically) will support the kinds of exploration that you would like it to support.

  5. There is a difficulty, Gene, in your meanings of “we” and “equivalent”.

    If one accepts “we” to mean “human knowledge”, then the picture is quite good.

    – People understand Wikipedia style declarative links well.
    – People understand comedic links well (xkcd, homestarrunner, rickroll)
    – There are generally understood conventions for advertising and cross referencing.
    – Many online forums have nuanced conventions for ironic links, political google bombing links, and attribution links.
    – The indexing systems of legal and procedural documents have worked well with links from day one.
    – Writers of electronic literature understand patterns of hypertext.
    – Designers such as Tufte have set out the basic principles for data-rich linked visuals.
    Semantic links are acceptably read by machines and humans
    – Social media has helped us understand the social and ethical implications of things like attribution and politeness in linked writing
    – Educators have rather complete frameworks by which to teach and evaluate linked writings

    We live in good times for the link. How we read and write is fundamentally different from how it was before, and we’re able to be intentional about how we do it.

  6. You raise good points about the potential for links to be useful; I think Dillon’s point was that this potential is not always realized and that much scholarship around hypertext concerns itself with the mechanics of link creation and link following, rather than considering the overall effect that the hypertext is intended to impart as an effective education tool.

    There undoubtedly have been effective uses of HT for education and other purposes; but are these the rule or the exception?

  7. Agreed. For the broader picture, you may wish to take a look at Mark Bernstein’s paper on Criticism at HT2010.

    And you’re right about reach: hypertext is by far the exception in schools. Most schools don’t even have books.

  8. […] Nathan Matias pointed me to Mark Bernstein’s paper (‘paper’ is an inadequate label for the work) on […]

  9. Oh, fyi, I now see that my very last statement could look like a troll. I apologize. I should explain that I’m very interested in education in the developing world, and that I spend time looking at things like:

    The Consequences of Literacy
    Literature in poor countries

    Having visited schools at all levels of quality and wealth, from the very best in the world to some of the most shoestring, it’s unfair for me to make a throwaway comment like that which doesn’t add to informed discussion. Sorry.

  10. I wonder if there is a Godwin’s Law equivalent for health and education in the developing world :-)

Comments are closed.