Saturday, 24 June 2017

Government as Steering: Cybernetics and the Coming Labour Government

The joy surrounding Jeremy Corbyn's success in the election masks a need to do some very difficult work if a left wing labour government is going to deliver on the promise to transform society. There is muddle-headedness about the practicalities of government, the way events can overtake good intentions (no politician would have wanted a Grenfell on their watch), or the sheer challenge of keeping a political machine together which always seems hell-bent on self-destruction (all political parties seem to have this tendency).

Now is a golden opportunity to do this. Corbyn has the luxury of opposition where his grip on the party has been strengthened, and public expectation of a Corbyn victory (unthinkable before the election) has shifted significantly. These are real achievements.

Labour, and Corbyn, have got here because the Tories don't know how to govern. They see the world in a linear and hierarchical way, where simple "strong and stable" solutions can solve intractable problems. When things don't work out the way they wished (like the deficit coming down), the Tories tend to carry on regardless: strong and stable. This isn't government. It is ideological extremism.

"Government" and "governor" come from the same latin root: Gubernator. The Watt governor is the simplest idea of governing:

The Watt governor 'steers' the engine, by increasing the flow of steam if the engine runs too slow, and decreases it if it runs to fast. The Greek word for governor is kybernetes, from which we get cybernetics. The Kybernetes was the steersman on the ship, so cybernetics is about steering. And so is government.

Stafford Beer is the cybernetic thinker who considers the problems of government (and its related problem, management) in most detail. I have thought about the Viable System Model (see for many years, and the Cybersyn experiment in Chile of 1971-3 (see remains the most significant attempt to rethink government (apart from some promising experiments in the Soviet Union which didn't get off the ground properly - see

There is a fundamental problem that the VSM addresses: the problem of attenuating descriptions of the world. In hierarchical power structures like governments, or bosses of universities, hospitals or any institution for that matter, the "top" relies on filters to give them the most important information from the ground. This is where the pathology starts, because the filter entails removing most of the other descriptions which are not considered important. This is why the election opinion polls got it so wrong - because they didn't listen to the variety of description that was out there. Technology has made the situation worse - it can filter more effectively than anything else - although this is a stupid way to use technology!

The VSM is a set of nested loops within which there is attenuation of description (there has to be), but at the same time the attenuated descriptions are organised into the production of a generative model whose engagements with the organisation (or country) that is being managed is continually monitored. The circular loop continually asks "Are we right?", "In what ways are we wrong?", "What have we learnt about the world that we didn't know before?", "How should the model be changed?". In other words, there is attenuation, and there is amplification of the abstracted model in a continual process of organic adaptation (Beer described his model using the metaphor of the human body). This is steering.

In theory, this is fine, and the VSM is often used in management consultancy to help heal organisational pathology: I'm hosting a conference in November on this very topic:

But apart from Cybersyn, there has been no real-time empirical attempt to exploit this thinking in government or management. We should do it, because our existing models of government cannot deal with the obvious circular causality which is endemic in our world, from overseas wars and local terrorism to austerity and burning tower blocks.  We have to have a practical way of dealing with circular causation, and I worry that Corbyn's labour isn't prepared.

Beer's Cybersyn was a data-driven operation in a world where data was hard to come by (they transmitted it with Telex machines). Today, we have data everywhere - but we don't know how to use it. Most approaches to "big data" seek to amplify automatic "filters" of complexity - this is basically what machine learning does. That's fine up to a point, but whatever filters are produced, are used to create a model which must be tested and improved. The human thinking about the rightness of the models used doesn't appear to happen. All "big data" results are the opportunity for humans to produce new descriptions of the world, and for these new descriptions to feed into higher level steering processes. But it doesn't happen. Consequently, we allow the "big data" to dictate how the world should become without thinking about what we've missed.

One of the critical signs that any government or management should worry about is a decrease in the variety of description about something. This is usually the harbinger of catastrophe. Our Universities are heading straight for this, because they are removing vast chunks of variety in the conversations and descriptions which are made within them as they close departments, sack staff, become fixated on metrics of academic performance which mean nothing, or chase government targets for "teaching excellence" in the hope of getting more money. Nobody is monitoring the richness of conversation in Universities. Yet, the true strength of any university is the richness of the conversations which it maintains.

The same goes for a healthy society. The urgency of thinking about this was impressed upon me a couple of days ago when I received a text message from a bright and brilliant academic and friend in my old institution (one of only a few in that awful place). It's a dismal reminder of how much trouble we are in: "I've just been told I'm being made redundant". So that's another conversation killed.

1 comment:

Paul Hollins said...

Mark , as ever a thoughtful reflective post describing the awful political situation we find ourselves in extremely well. One wonders what Beer would make of it , no I know what Beer would make of it. Government and Universities for that matter seem intent on using attenuation as their only strategy in dealing with the complexity you describe. Meaningless (yet persuasive and powerful) metrics as you say "amplify" the issue. I find comfort in that I think Corbyn and Beer are aligned politically . When, and I do think it is a "when" , Corbyn assumes government it will be interesting to see how things pan out but I do have hope , hope more in the resurgence of community above the self...