Having saddled myself with the agile learning term, one of the hazards I can’t complain about is having to explain it: What does it really mean? What’s different about it? What’s agile about it. There’s a working definition of the key elements on the agile learning wiki, which I continue to develop slowly and sporadically. Recently I’ve been reflecting on some more nuanced, but still half-formed, ideas, which feel more like blog-conversations than wiki-definitions. These are partly prompted by reading Douglas Rushkoff’s excellent Program or Be Programmed (which deserves a blog post of its own), and also by the Learning Analytics course, devised by George Siemens and colleagues, which I’m currently participating in (and blogging about in detail over here).
What I’m toying with at the moment is a distinction between “weak” agile learning and “strong” agile learning. This is after John Searle’s distinction between “weak” and “strong” artificial intelligence, but I suspect this kinship may be tenuous and, certainly, vainglorious. They might equally well be called, say, pragmatic agile learning and principled agile learning — or something else.
- Weak agile learning is based on whatever works, as long as it’s within the definition. Analytics, automated processes, traditional tutor-led power relationships are all fine if they’re open (as in open educational resources), collective and flexible.
- Strong agile learning is committed to the values of making everything — process, resources, algorithms and the context in which the learning is framed — visible, transparent and manipulable to the learner. So it’s taking a more radical definition of openness.
The weak version allows for things like intelligent curriculums, gamification and personalisation by the provider. The strong version wants to trust in learners’ intelligence and give them the information and the data to personalise their own experiences.
Pragmatically, I’m drawn to the weak version. I distrust purism, believing every oyster needs some grit (for most of my three decades as a vegetarian, I’ve eaten meat a few times a year). But ethically and aesthetically I feel the strong version needs shouting about, because gung-ho enthusiasm for the Big Data/Scientific Management seems to be leading down a dangerous path. Let me explain.
Why do people insist on using big data and AI techniques to add more intelligence and power to the provision of learning? Why not focus more on the receiving and acting-on end? Well, we know why, don’t we? It’s because the receiving and acting end is unruly and unpredictable. But what’s the end game of trying to build all the intelligence into the provision? If it were successful, then the education system would have succeeded in deskilling the act of learning. Yes, deskilling learning so that anyone can do it without thinking or stressing too much* — what kind of oxymoron is that? It’s not sustainable, in that it doesn’t lead to robust, resilient learners who can deal with genuinely messy, unpredictable circumstances.
I suspect it’s not achievable, either, because learners too are messy and unpredictable. What they take in is affected by all sorts of environmental contingencies — the state of their blood sugar, hormones, preoccupations with tonight’s football match, or what they learned in another subject yesterday — and a million other factors that are beyond the modelling capabilities of any Intelligent Tutoring System. Sometimes they learn against the grain, wilfully or tacitly detecting that the tutorial agenda is at variance with their own, and thus taking away the opposite lesson to the one intended. The strong version of agile learning doesn’t fight this, but celebrates it, encouraging learners to adapt and own their own learning paths — by all means in dialogue with, or under the guidance of, a tutor. Of course, they will make mistakes. Sometimes they freewheel down the path of least resistance; others they bite off more than they can chew and get well and truly stuck. In the long run, they learn from these mistakes. They not follow a steady, incremental path, but they emerge better able to learn in world that hasn’t been tailored to their needs.
Weak agile learning says that if building the world’s smartest tutor provides quick, accessible and flexible learning opportunities, let’s do it. Strong agile learning worries that this is a Pyrrhic victory: learners win the battle of achieving short-term objectives, but lose the war of becoming resourceful and independent lifelong learners. The drive towards intelligent tutors and curriculums falls into the trap that Nicholas Negroponte identifies in his foreword to Beyond the Hole in the Wall: Discover the Power of Self-Organized Learning (click on “Look inside” to read the foreword):
Learning and teaching are not symmetrical. They are not the flip sides of the same coin, in spite of the fact that almost all papers and conversations on education assume they are.
The working assumption is this: solve teaching and you will get learning. Even if true in part, it addresses only some kinds of learning and never really attempts to understand the learning of learning itself.
I could go on. There are many links my mind wants to make. But I’m trying to take a leaf out of Clive Shepherd’s blogging book, and keep my posts to an hour and a few hundred words, rather than half a day and a few thousand words.
* I know that intelligent curriculums and personalised tutorials would deliver challenges, but they’d always be challenges with just the right amount of stress to motivate without intimidating — a kind of stress-free stress.