Describing learning – should we?

This is an issue that come up frequently for me – an indicator that it’s time to distil my thoughts.

A lot of effort, from a lot of people, has been devoted to finding adequate (let along accurate) ways to describe learning. There’s a problem, though – learning doesn’t lend itself well to description. In the realm of education and learning, language falters.

An example of where description works – food

Consider the domain of food. Here are some Things that we find in food:

Ingredient
Preparation technique
Recipe

Now we can agree that these are Things, and we can agree that these Things are related to Food. We also have something very precious: user consensus on what is what.

If I show you a carrot, you’ll say ‘that’s an ingredient’.
If I show you a page of a Jamie Oliver book with the heading ‘Jamie’s yummo roast beef’, you’ll say ‘that’s a recipe’.
If I show you carrots mentioned in ‘Jamie’s yummo roast beef’, you will understand the carrot-to-recipe relationship immediately and unambiguously.

…and so will everyone else who encounters instances of our Things, with very few exceptions. Yay! Consensus!

Now let me ask a different question: what type of Thing would you look at and say ‘that is a Competency’?

Would your definition match – or, would you have consensus with – the person next to you?

 

An example of where description didn’t work – LOM

Fortunately, we seem to now be leaving behind the Dark Days of the Learning Object Metadata (IEEE LOM) profile. We’re leaving it behind because it didn’t work very well. But what was it about it (aside from being un-indexably and un-editably stuck in the middle of a bunch of sequencing information in an imsmanifest.xml file, which is itself tucked away deep in a compressed file) that made it not work?

I’d suggest that LOM tried to describe things that can’t be described in schemas, and that was the seed of it’s downfall.

Here it is, in all it’s glory:

Let’s take an example (it’s page 25, for those of you playing along at home):

IEEE LOM 5.3: Interactivity Level

Explanation:

The degree of interactivity characterizing this learning object. Interactivity in this context refers to the degree to which the learner can influence the aspect or behavior of the learning object. NOTE 1:–Inherently, this scale is meaningful within the context of a community of practice.

Value space: 

very low
low
medium
high
very high

Example:

NOTE 2:–Learning objects with 5.1:Educational.InteractivityType=”active” may have a high interactivity level (e.g., a simulation environment endowed with many controls) or a low interactivity level (e.g., a written set of  instructions that solicit an activity). Learning objects with 5.1:Educational.InteractivityType=”expositive” may have a low interactivity level (e.g., a piece of linear, narrative text produced with a standard word processor) or a medium to high interactivity level (e.g., a sophisticated hyperdocument, with many internal links and views).

 

Now my questions are – and I mean this literally, in the literal sense of literally:

Who cares?

What value does this description have to a user?

What do the values actually mean? (And no, I don’t see much value in the stated CoE proviso.)

Who uses the term ‘hyperdocument’?

 

(I want special credit for not using the even less meaningful field 5.4, ‘semantic density’, but I will award five points and a star to anyone who can guess what that means.)

I’m leery of any descriptive metadata that is not user-meaningful on it’s face. My issue with this, however, is different – by providing definitions in the first place, this schema tries to introduce the idea that level of interactivity is an attribute that a learning object can meaningfully have, and by implication, that learning can have. I disagree – I don’t think that that is an accurate way to describe learning content, let alone a useful one.

In terms of usefulness – can you imagine a teacher poring over a database of content muttering ‘well this is grand, but I really need something with an interactivity level of at least ‘high’? I can’t. That’s why I say it’s not useful.

 

In bite-size:

Learning is probably the most sophisticated/complex thing that human beings do.

The more sophisticated/complex the learning is, the less it can be described using schemas and controlled vocabularies.

Sophisticated/complex learning is the best kind – and therefore the kind we want to support with learning content the most.

 

LOM is fading away, but the idea behind it – that the often intangible learning outcomes supported by learning content both can (I disagree) and should (I disagree more) be described in structured metadata. This idea pops up in the projects like CELEBRATE and MELT and Ariadne and ReSIST  (and ReSIST, if I’m honest, I had never heard of before today).

Educational concepts aren’t like self-raising flour and basting and baking. They defy attempts to encapsulate them. The various curricula of the world can’t even decide if ‘social science’ and ‘social studies’ are equivalent*, so there’s little hope for ‘interactivity level’, ‘difficulty’, or even ‘learning resource type’.

So when it comes to creating stable definitions for educational concepts, well – a lot of the time, I don’t think we should.

 

Coda:

I have a lot of hope for the LRMI specification, especially with its use of DC terms and (more importantly) the backing of schema.org. Where Things are included, they are included at a high level of abstraction: ‘alignmentType’ has pretty sensible values like assesses, teaches, requires, and the apparently-meaningful-but-I’ve-never-heard-of-it textComplexity. 

But I note that even here, we have fields where no fields should really go  - does ‘interactivityType:mixed’ mean anything to you?

I didn’t think so.

 

*As it turns out, they’re not.