It’s been too long since I’ve blown off some ID-related steam. The claim of pedagogy-neutrality in standards an interesting issue.
As I have said many times, I believe that the term “pedagogy-neutral” is not adequately descriptive, and instead the term “pedagogy-agnostic” should always be used. I purposely choose “agnostic” because of its religious implications: a pedagogy-agnostic standard “doesn’t know if there’s a pedagogy or not.” In other words, it is impossible to design a language which allows the expression of *any* pedagogy without simultaneously allowing the expression of *no* pedagogy.
When using a standard which is “truly pedagogically neutral,” a person must be able to express an instance of any particular pedagogy. If this is the case, then a person must of necessity also be able to express no pedagogy whatsoever. If the language of expression places no constraints on the expressor (let’s pretend that its possible for such a standard of expression to be agreed upon), so that they can express any pedagogy they desire, the language is incapable of keeping them from expressing anything else – because there are no constraints on expression! Expressions of great pedagogy and no pedagogy may both be well-formed and validate.
One might say, “judging whether or not the pedagogy is effective,” (as if this phrase is meaningful without specific reference to specific individuals and specific content) “is not our goal. Our goal is to enable the machine interpretable expression of the pedagogy itself.” This leads us to abstract coneptual languages like IMS-LD. Using this language we can model any one or set of individuals acting in specific roles, in specific environments, with specific resources, as they perform certain activities.
Note that we can model not only teaching in this way. As the IMS-LD spec points out, we could model all of drama using these contructs (actors play parts amongst scenery advancing a storyline). But in addition to all the teaching ever done and all the drama ever performed, one quickly comes to see Wertch’s analytic construct of “person acting with mediational means in a sociocultural context” or Engestrom’s notion of activity theory – people using socioculturally grounded tools to work toward goals. In other words, we see that high-level, abstract conceptual languages like IMS-LD are, in fact, capable of modeling the whole of human activity. What of it does not involve people, resources, environments, and activities?
And so IMS-LD is not only pedagogically-neutral, it is pedagogically agnostic – capable of modeling in machine interpretable format the wide range of human activities, any specific one of which may be neatly within or far beyond the tiny realm of teaching.
Is this a good thing or bad? Realizing that the primary focus of the standard is modeling human activity does raise certainly questions. For example, why would we model human activity? In order to better understand the roles humans play, so that we can neatly replace them with “intelligent” machine subsititutes? Is it because people are so excited to use the automated support option when they call Dell after their Windows machine blows up? In asking over 2,000 people now if, when they need support, they choose the autoamted system or a real-live human support engineer, guess how many hands I have seen go up for the autoamted system? About 8. Ooh! What a demand! Let’s replace other humans who provide support too! Like learning support!
Why would we remove the communications from telecommunications? “It is still communications, surely” some will argue. Excellent. As long as your machine teacher sending me bits counts as communication, then surely my machine student sending back the “right” bits when prompted by the machine teacher will still count? We could learn something from the communications from machine to machine. If the teachers were to broadcast, error-check, and continue rebroadcasting and error-checking until they could confirm that the student accurately received the message, then perhaps something akin to communications would be going on. But we are so obsessed with the technology… determining the XML binding, making sure our XML validates, making sure it looks beautiful after being XSLT’ed, etc.
Why would we turn the greatest enabler of social interaction into a simple data download service? It’s like the great sequence of scenes in Real Genius. At a top school (I always thought it was supposed to be CMU, for some reason) a lecture room is full of students and a professor. But before long, the professor has been replaced by a reel-to-reel, and students simply scrawl down notes from the recorder. By the end, the reel-to-reel plays in the front to a room empty of all but miniature tape recorders. That is the direction all this automation madness is going, you know.
Now, I’m not against automation where it makes sense. I love an exciting computer game as much as the next person – I form hypotheses, test them out in the game environment, receive immediate feedback, tune my hypothesis, test them again, receive feedback again, and so on. But this and other automated systems
are not complete learning environments. They’re a temporary activity I engage in before going back to my primarily social learning interactions with real human beings.
This doesn’t mean I think distance education is a bust. On the contrary, I believe massively multilearner environments hold great promise (for some people learning some content), modeled on things like World of Warcraft. Why? They’re “the best of all possible worlds” – they place opportunities for automated, immediate feedback in a highly social context. But don’t get me wrong. There is no “best approach” for every individual and all content. But I think these MMLEs hold great promise for many people and a sizable family of content.
Likewise I don’t believe that IMS-LD is a bust. I do think, however, it could be put to more fruitful purposes. For example, when people carry out “educational research” of the horse race vareity (you know, my pedagogy is better than yours), IMS-LD could play a vital role. A key to good empircal work is replicability, which we almost never get in educational research reporting. You get garbage like “our constructivist approach worked better than traditional classroom instruction” without enough detail to ever dream of replicating either approach, let alone the study. If all reporters of educational research used IMS-LD to describe their methods, it would go a long way toward bringing needed rigor to our field.
Someone will, of course, rightly point out that, because it is a human activity, individuals engagements in MMLEs can be modeled in IMS-LD, too. Well, good. I hope someone does it.
Not sure what the focus of the ‘rant’ here is Dave. IMS Learning Design can (and should) be used to model human interventions in the edcuational process. I’ve been reviewing a paper recently where extensions to IMSLD are proposed to expand the types and scopes of the humantutor/learner interactions that can and do take place in most forms of campus and distance education. Even without these extensions, IMS can model all or no pedgagogical activities as you seuggest- even those with a great deal of human interactions.
As for automatuion, you could ask how many people would perfer to use a human teller at a bank – especially one with a long linup in front of it, compared to using an ATM. I like using automated systems when they are faster and more effective then human helpers – challenges not really beyond the scope of much AI work today.
Terry Anderson
See my comment in my blog.
… “forget about the notion of “pedagogical free” learning objects and framework. Let try the other way around. Identify powerful and effective pedagogical approaches (learning structure) and implement solutions to these approaches.” …