Saturday 17 February 2018

The Cybernetics of Competence and Capability: Revisiting Pask

Enid Mumford noted that the difference between competence and capability lay in the difference between attenuation and amplification between complex systems. Competence involved the attenuation of an environment to fit the acceptable parameters of an individual who had learnt a set number of appropriate responses. Capability involved the production of multiple descriptions of understanding of a complex situation such that a solution to a new situation may be creatively generated rather than retrieved from memory.

Gordon Pask noted that the production of multiple redundant descriptions of a single thing was a fundamental part of the learning conversation. When he clarified his "teach-back" process, he goes much further than Laurillard's simple comparator approach to check if the learner has taught back what they were taught. Pask says (in the "Cybernetics of Human Learning and Performance", (1975))

Teachback goes as follows: the teacher says of the student (or ‘subject’) that the student understands a topic to the extent that he can teach it back to the teacher. This is, understanding is inferred if the student can furnish an explanation of the previously discussed topic and can also explain why he gave that explanation of how he constructed it. The crucial point is that the student’s explanation and the teacher’s explanation need not be, and usually are not, identical. The student invents an explanation of his own and justifies it by an explanation of how he arrived at it (in fact an identical explanation is generally rejected unless the student can give a reason why the teacher’s explanation was particularly good). 

The difference between the teacher's utterances and the student's is critical in the teachback process. Pask goes on to say:

the resilience of a memory will depend upon the number of explanations produced in teachback; for example, that a student impelled to give many explanations will fare better at session 2 than a student required to give only one. He has many ways of reconstructing a concept and this  redundancy will combat the effect of interfering and incompatible learning experiences during the intervening week. 
What this suggests is that redundancy is the principal indicator of learning, not information. In Shannon information theory (which underpinned Pask's thinking), redundancy is the inverse of information: it is the context within which messages are formed. Education is not about the message; it is about the context!

How might this work?

Well, compressing Pask's diagram into the exchanges between teacher and learner, it might be drawn like this:
At the stage of "rigidity", the teacher present many alternative descriptions of what they are trying to convey (this is generally what teachers do!). The learner is only able to reproduce one of the descriptions they are given, if that. Gradually, they acquire greater flexibility to utter more descriptions. Having two descriptions of the same thing is very powerful, and builds towards a generative capacity to create (or guess) other explanations.

Then there comes a stage when the learner is able to respond to the different descriptions of the teacher with effective matching alternative descriptions. At this point, we might say that the learner is competent. 

Finally, the true generative power of the learner's understanding is revealed at the stage where from a simple prompt by the teacher, the learner is able to generate all manner of descriptions - some of which may not have been conceived by the teacher. At this stage, we can say that they are capable, confident and adaptable in their knowledge.

What happens from stage 1 to stage 4 is a gradually awakening to the constraint which lie behind the teacher's generation of their own descriptions. By the end, what the learner has learnt are not facts, but the mechanisms of transduction within the teacher whereby the teacher is able to generate the descriptions and skilled performances that they demonstrate.

If only we'd thought of this when people sent their students off with their dreadful e-portfolio systems, we would have done it all very differently!

Friday 9 February 2018

When is a musical note a different note?

I've been talking a lot about transduction recently.  Transduction is the process whereby a distinction gets maintained.  In engineering, it is the process of taking one form of energy and turning it into another - like an electric transformer. But the point is it produces a boundary.

The idea of transduction is useful because it turns what we usually think of as fixed 'categories' into processes. So think of a category or a subject.. 'maths', 'geography', 'chair', 'happy', 'ill', etc... Now think of the process which makes that category. It turns out that any category has two sides: a category is a boundary. The process of maintaining it works from both sides of the category. If you want to change a category, you have to change the process.

So you want to change the culture in an organisation? You need to understand where the transduction is happening and 'tweak' it.

Transductions are recursive. One category depends on many many other categories. More to the point, a category is not necessarily something that can be expressed in language. All perception is transduction: to perceive a difference is to experience a transduction.

That's useful when we think about music. To differentiate one note from another is to experience a transduction. What does that tell us about how transduction actually works? Well, to distinguish one note from another depends partly on there being multiple descriptions of a note. A note is never a simple thing: it is a multiplicity of frequencies to start with, which give it a timbre or colour. It also has a beginning and an end, it has a volume, and so on. A note is different from a silence. So a single note is perhaps a kind of transduction between a silence and the note. Since a silence also has multiple descriptions (silence has many qualities), a note is the difference between one set of multiple descriptions of something and another set of multiple descriptions of something.

If we were to compare one note with another note, then some descriptions between one note and another note might be the same - they might have the same volume for example. But they might have different frequencies, or they might have different timbres. 'Another' note is a change in the arrangement of multiple descriptions. That is what the transduction does: it shifts from one set of descriptions to another.

What about detecting that a note is "the same" note as another? That's an interesting transduction. To say that it is "the same" is to still detect that it is "different", but it is different in a way where the boundary between one and the other produces a new category of "the same".

A new category? Ah! A new transduction!