Tuesday, 3 January 2017

Rethinking Computers and Education

There are few who would disagree that it is a bad time in education (or that it is a bad time in world right now). Education is afflicted with deep confusion and a lack of conviction coupled with aggressive and pathological managerialism (with its insincere conviction). Technology has played a key role in this - in the bibliometrics which drive the "status game", or the research assessment exercises, league tables and other manifestations of status pathology, through to assessment regimes (MCQs, automatic essay marking), "plagiarism checking", endless student survey questionnaires, VLEs and MOOCs, and 'analytics'. The pioneers of educational technology hoped for better education; what we have is far worse than anyone imagined. The situation was, however, predictable: Nigel Howard puts it well (in Paradoxes of Rationality, p.xxii)
"There are two main causes of all the evil that exists in the world. One is that humans are very wicked; the other is that they are very stupid. Technocrats tend to underestimate the first factor and revolutionaries the second."
The educational technologists were both proto-technocrats (although they would have hated the description) and revolutionaries (which they would have liked). They were good people (on the whole), but some of their tools were adopted by people who were greedy and status-hungry, seeing opportunities to manipulate the political discourse and control awkward academics for their own nefarious ends. Education became a money making opportunity. The senior figures in education and politics were far more stupid and easily led than anyone expected or had hoped for. And this is where we are.

The difference between pathological and effective use of tools is the extent to which one understands the problem one is trying to solve. The problem of education is not much less complex than the problem of society. For most people, it is too difficult, and so they seek solutions to easier problems which they can define, and which are usually determined by existing technologies. This is why we have e-portfolio and VLEs, and not much has changed since: these technologies were answers not to the question "given that we have technology, what might education be like?" (which is the question we should have asked), but more prosaically, "how can education exploit the internet?".

The problem of education is a problem of highly complex inter-relations of constraint. In education, we are each other's constraints, in a world of material, political, economic, social, psychological and technological constraints. Computers in educational have not merely added to the constraints of education, but by their mutable nature, have transformed the pre-existing constraints within which the education system emerged in the first place. The result is deep confusion. In this way, computer technology was a hand-grenade exploding traditional education.

In many walks of life, however, computers have been very useful. The use of simulation and optimisation in the designs of buildings, public spaces, advanced materials, cars and planes, transport networks, space rockets, drugs and high-tech machinery has transformed much of our public and private environment. Why can't computers be used intelligently in education?

I have two answers to this. Firstly, the use of computers in all the fields mentioned above occurs within clearly codified constraints. The design of a London Underground station must account for the flow of a certain number of people, a building must optimise its use of space, aircraft must reduce their air resistance, and so on. In many cases, techniques like linear programming or genetic algorithms can all be gainfully employed because the 'fitness function' and the constraints which must be negotiated, or the desired criteria which must be met, can be clearly specified.

What is the fitness function of education? For the advocates of AI computer-based instruction, this is a worryingly unproblematic question: it's about getting the "right answer". But any teacher knows that it isn't just about getting the right answer. The right answer might be wrong, to start with: if education was simply about right answers, science would stop. But more importantly, the nature of reasoning, and the nature of the conversation which gives rise to reasoning is much more important.

Some advocates of Learning Design had a different answer to the "fitness function": fitness was about stimulating conversation and action. The richer the conversations, the better the fit. It's a better answer than the AI one, but the criteria for design tend to be based on woolly theories of conversational learning whose inadequacies are never explored. Too many students simply drop out of the conversation, and nobody seems particularly interested in what happened to them, preferring to make grandiose claims for the students who keep talking.

Interestingly, in both these approaches, little use is made of the processing power of computers. The question is whether computers might be used to model the complex dynamics of interacting constraint where we become each other's constraints. The first challenge is to specify the problem in a new way: what, then, do we need in education?

On the one hand, education is about managing complexity. It concerns the complexity provisioning resources, activities and tools so as to support effective learning conversations between learners and teachers. This is fundamentally the concern of cybernetics. But there is a deeper level of thinking which concerns the fact that any "effective relationship" whether between a teacher or a learner, or between learners, involves the dynamics of mutual constraint. Teachers maintain learning conversations by manipulating constraints: sometimes by allocating resources ("read this!"), or by organising activities ("do this!"), or by introducing tools ("do this by using this!"). We currently have little understanding about how the dynamics of constraint in education work. But computers do provide new ways of manipulating constraint with far greater capacity and organisational potential than any teacher could possibly manage. And not only can computers manipulate constraints, they can also provide degrees of meta-analysis about the manipulation of constraint (which itself is a constraint).

Perhaps the answer to the question, "what do we need in education?" is simple: better theory and better evaluation. 

No comments: