Doing Cybernetics (and talking about it)

Talking about cybernetics is not the same as doing it. Doing cybernetics is a way of speculating about nature whilst actively engaging in a methodical way with nature. Talking about cybernetics is a largely descriptive and analytical exercise. Academia, on the whole, encourages people to talk about things. The kind of speculative activity that cybernetics engages in does not fit academic expectations. There are some areas of academia where doing cybernetics is possible - most notably, the performing arts, and also the practice of education itself. However, cybernetic speculations struggle to establish academic legitimacy - particularly when their codification is divorced from the practice that gave rise to their findings. There seems to be an unseemly rush to codify cybernetic speculations, and then to turn them into objects of discursive inquiry, or (worse) blueprints for interventions or analysis. As a result speculations get reified. The most appalling example of this is in the reification of the various speculations about learning contained within constructivism: speculation becomes dogma.

This has led me to think that maybe a meta-description of what 'doing cybernetics' means might help. I think we need a kind of meta-model of cybernetic activity which defends the highly speculative activities of cybernetics within the context of scientific practice recognised by institutions.

I start by thinking about modelling. Modelling is often associated with prediction and control, and there are many second-rate descriptions of cybernetics which commit this mistake. In fact, modelling is fundamentally about coordinating understanding through mapping a set of idealised constraints. A model, or a machine or software which articulates a model, provides an opportunity for cyberneticians to point at abstract mechanisms and reach shared understanding. Moreover, models are generative of many emergent possibilities. In agreeing about fundamental mechanisms and components, cyberneticians can also agree about the logical emergent potential of a model and its fundamental properties. It is through this logical emergent potential that modelling acquires the connotation of prediction. But this is really to get the wrong end of the stick.

If a model presents logical possibilities, reality (or nature) produces events which may be mapped and measured in the model. The most important process in modelling is the 'indexing' of real events with particular behaviours or emergent processes in the model. The indexing of events is not a trivial exercise: it is at the very heart of the cybernetic process. To index something is to determine similarities and differences between events: it is to reduce perceptions to manageable and measurable components. Although this is a reduction, it is not necessarily reductionist. It would only become reductionist if a particular reduction became dogma, washing over emerging differences which did not fit the model. In actual cybernetic practice, reductions are useful precisely because they reveal their deficiencies in the light of actual experiences.

To measure events is to index them, and then count similar events and their relations. In this process, the key feature is the distinguishing between unsurprising events and surprising ones. In measuring surprise among events, the degree of surprise in nature can be compared to the logical model of how such surprise might be generated. Knowledge is gained by recognising how the constraints of nature are different from the
constraints contained within the abstract model. Usually this leads to the identification of new indexes in nature (new things to count), or critiquing some indexes of things that were thought to be the same, but in fact are distinct. Models consequently have to be adapted to take account of new indexes, and in being adapted generate new sets of possibilities.

The disparity between constraints discovered in nature and constraints within a model is only part of the story. There are also many constraints which bear upon the minds of those who produce models. Among the most serious constraints which appear to bear on all humans is the constraint which encourages individuals to insist on the veracity of particular models even in the face of evidence that they are deficient. Cybernetics requires more than a logical approach to the examination of models and participation with nature. It also requires a high degree of self-examination, reflection and a letting-go of ego in the continual maintenance of a speculative attitude.

In the philosophy of science, the social identification of causal paradigms or explanations has been the central focus. Doing cybernetics brings a different social focus: the coordination of a discourse through the shared participation in nature, and social agreement about indexes within both nature and in models. Agreeing similarities is to determine constraints. To measure the surprisingness of events against a model is to learn about new constraints.

Perhaps the surprising thing is that I think good teachers do this all the time, and always have done!