Designing Freedom

Stafford Beer

• Publisher: House of Anansi Press
• Year: 1993
• ISBN:
  9780887848551
• Topic: Philosophy
• Subtopic: Philosophy & Ethics in Science

Distinguished cyberneticist Stafford Beer states the case for a new science of systems theory and cybernetics. His essays examine such issues as The Real Threat to All We Hold Most Dear, The Discarded Tools of Modern Man, A Liberty Machine in Prototype, Science in the Service of Man, The Future That Can Be Demanded Now, The Free Man in a Cybernetic World.

Designing Freedom ponders the possibilities of liberty in a cybernetic world.

THE FUTURE THAT
CAN BE DEMANDED NOW V

A famous summit conference, not to say confrontation, about freedom once occurred about twenty miles from my home in England. It lasted from 15th to 23rd June, in the year 1215. During those negotiations between King John and his barons, the Magna Carta was signed: a document that spoke for all time of the decentralization of power and of the rights of individuals, and is still much quoted 750 years later. I remember from my boyhood a humourous monologue explaining these events, which ended something like this:

So it’s thanks to that Magna Carta
that was signed with the barons of old
that in England today we can do as we like
—so long as we do as we’re told.

The blatant contradiction embodied in this joke remains the desperate problem that it always has been. How do we sustain individual liberty and societary cohesion at the same time? It is right that this problem should be incessantly discussed, and it is discussed. But the discussion always seems to lead straight into the same disastrous trap; a false dichotomy between the notions of centralization and decentralization. The vehemence with which this matter is debated is extraordinary, because the most cursory consideration of what constitutes a viable system reveals how false the dichotomy must be.

For example: if you personally were a fully centralized system, you would need to remember to tell your heart to beat. If you paid too much attention to what I am saying, and forgot, you would collapse on the floor. That would be dramatic, but it is not what I am trying to achieve with you. But if you were a fully decentralized system, you would trot off from this broadcast to investigate any sound. Neither solution would leave you a viable system for long. Let us analyse the blend of central and peripheral command and see what happens.

We discovered earlier that viable systems are bombarded continuously with high-variety stimuli, the variety of which has to be attenuated if the system is not to be overloaded. The attenuation must be done according to a pattern, if it is not merely an arbitrary discard. If that pattern is to have survival-value (which is a necessity for a viable system) then it must be a regulatory model of whatever is regulated. Then it follows that this has to be a central function for the system, because only the system as a whole can have a model of its own relationship with its own environment.

Nonetheless, when the central function of variety attenuation is operating, it is by definition not appropriating to itself the discarded variety. But by Ashby’s Law we know that variety must be absorbed somewhere. Therefore whatever variety is not appropriated by the central function must needs be handled by a decentralized function. This variety handling is by definition autonomous. That is to say that some part of any viable system does what it likes. But of course the autonomous part of the system remains part of the system, and to do that it must take notice of the central regulatory model. To that extent, then, it does what it is told.

If we make a terrible mess of interpreting these simple cybernetic discoveries in our society, and I feel that we do, it is because there is no agreed machinery for settling clearly which parts of the system are which. To do so is indeed virtually impossible unless the models in question are made explicit for each level of recursion. To remind ourselves: a level of recursion is a level at which a viable system is in operation, as an autonomous part of a higher-level viable system, and containing within itself parts which are themselves autonomous viable systems. We spoke of this set-up before as being like a set of Chinese boxes.

We find the process of settling autonomy going on continuously within any viable family, for example. As the children grow older, they exert more and more their personal freedom of action. But this has to fit into the family’s general regulatory model, at a higher level of recursion. Thus a great deal of time will be spent in discussing the notion of autonomy for the younger members, and the time required has to be a great deal because requisite variety must be attained. In the upshot, families often manage to preserve their model of the family, which is centralizing, and the liberty of the younger people, which is decentralizing—and forget altogether to consider the right of freedom for the parents, who then become identified with the centralizing authority. This, you will note, is not at all to the parents’ advantage (since they have lost freedom in the process) although it may appear to be so to the children.

This homely example is repeated, with great force, in the yet higher levels of societary recursion. A well-intentioned corporation or national service tries to hang on to its systemic policy—because this is what makes it itself, this is what embodies its aims, this structures its regulatory model. But that policy is centralizing. Then the corporation and the national service, being well-intentioned, embark on high-variety negotiations with the parts of the system, in order to delineate autonomy—which they really wish to be maximal. But the tools they are using are not cybernetic tools, that undertake variety engineering, but administrative tools which do not. As we have several times noticed, bureaucracies install amplifiers in one loop of the homeostat when they should be installing attenuators in the other—and vice versa.

The upshot now is really quite strange, but nonetheless extremely common. The parts of the system—subsidiary companies, sectors of the national service—who have been in good conscience given maximal autonomy, believe that they have been totally centralized. This is because their variety is attenuated by wrongly installed central amplifiers. But, on the other hand, the higher management at the centre, in conceding in good conscience maximal autonomy, believes that it has somehow been robbed of any role at all. This is because its variety is over-attenuated by wrongly installed peripheral attenuators. In the family, given quite a lot of hard work, everyone can just about hang on to happiness. But in big institutions, where—we have to remember—the brains of all the men and women involved are still the same size as they are in family roles, disenchantment spreads.

I often reflect that our organizations are so constructed in their typically pyramidal shapes, so that they could work only if the people in them grew bigger heads as they became more senior. In that case, of course, there would at least be a chance that they could maintain requisite variety. However, as we know, everyone’s head is roughly the same size—except, perhaps, metaphorically. I also reflect upon the device whereby “decentralization” is often advocated as the solution for an institution’s problems, when it is held to be over-centralized, while “centralization” is simultaneously prescribed for another institution on the grounds that it is too diffuse. And I have seen the two policies advocated alternately, and what is more alternately adopted for the same institution by successive groups of consultants. It is a kind of managerial madness. It leads, as it can only lead, to exacerbated oscillations in the system’s search for stability.

The solutions lie, they can only lie, in good variety engineering—and here is the key point. We must not confuse the pattern of the regulatory model with its specific content. It is enough to attain requisite variety by specifying the pattern. To specify the content is too much. Yet this is what endlessly happens, and I have noted that it usually happens—in those well-intentioned institutions, in good conscience—for one fundamental reason. This is called “fairness”. But I believe this kind of fairness to be an excuse for avoiding responsibility.

Take the example of a big institution that has a salary policy, or an employee automobile policy, or an inventory policy. We need a way of saying what the policy—which is to say the variety-attenuating regulator model, or pattern-really is without specifying its specific content. We fail. We ought to say: this much can be spent on salaries, on automobiles, on inventories, and leave it to managers at lower levels of recursion to apportion the money. Instead we do the variety engineering in the wrong place. “This is the salary scale”; “You are entitled by your job to this range of automobiles”; “All inventories must be cut by ten per cent”. It is done, as I said, in the name of fairness, but that is delusory.

It is nonsense to say that two men of the same age, with identical qualifications, with identical commitments, are necessarily worth the same wage. Obviously not. One may be useless, and the other a paragon. It is nonsense to say that my job ought somehow to determine my needs where automobiles are concerned. How and where I live, how many children I must push into the automobile—these are my own affair. It is nonsense to penalize a good manager who works on a scientifically calculated minimal inventory because his colleague managers are inefficiently tying up the firm’s capital. This is a recipe to encourage inefficient managers. Why do we blandly accept so much nonsense?

The variety attenuators to use here are not policy documents from the centre, but the managers themselves. That is what managers are for. As to the criterion of fairness, the manager—or any individual, in whatever he does—ought to be ready to take responsibility for his own decisions. Our society militates against that morality (for that is what it is) with all its force—in the name of an efficiency which is thoroughly bad cybernetics, in the name of a fairness that is manifestly unfair. But please remember the precept: each of us should take responsibility for his own acts. The practice is precisely the contrary.

As usual, then, we have the amplifiers and attenuators on the wrong side of the equation. But all of this is now written into our culture; all of this is now underwritten by our bureaucratic formulae. That is why I have repeatedly argued that fundamental change in our modes of organization is essential. Merely to juggle with existing forms simply increases the swing of the oscillating pendulum that never can find its stable state. And, as I have mentioned before, this means that the system is robbed of the crucial reference point without which it cannot learn, cannot adapt, cannot evolve.

How do we set about making so fundamental a change as would bring our ways of working into line with the scientific rules of this game? To answer this question it is necessary to understand the nature of resistance to change. Here is a phrase that is on everybody’s lips—“there is a resistance to change”—but it is a phrase which is not analysed according to the principles of effective organization. People seem to imagine that they are confronted with a psychological hang-up, insofar as men and women are not supposed to like change.

But pause one moment, is that true? People, considered as individuals, it seems to me, like change rather a lot. Don’t you get bored when nothing changes? I know I do. Then just why do we go around saying that there is a resistance to change? Of course, the answer is simple. It is not the living, breathing human who resists change in his very soul. The problem is that the institutions in which we humans have our stake resist change. Therefore we feel as individuals that we cannot afford to embrace it. And this is an extremely sound argument.

If you have spent a lifetime working your way up a ladder, you literally cannot afford to be robbed of the prizes when your turn comes to collect them. One of my earliest experiences in industry was to listen to managers explaining to senior operatives that they were to be deprived of their life-times’ ambitions, because the whole technology of the process was to be changed. That was in the steel industry, in which the skills of a first-hand melter—a job that it took a lifetime to learn—were replaced in a year by clever instrumentation. But ten years later those same managers were themselves confronted by the computer, which would have made many of their skills redundant in turn. However managers have power, and the computer none. It was easier to misuse the computer than to accept the institutional change—because the consequences would have been quite personal.

Now I have come to what I consider to be the explanation of the abuse of science and technology in our society. The power has remained where it resided. The tools of modern men have been disregarded at this level of recursion. And there is none left to say a loud NO to that—until the people themselves say No. So this is why I contend that we are considering a future that can be demanded now.

Every time we hear that a possible solution simply cannot be done, we may be sure on general scientific grounds that it can. Every time we hear that a solution is not economic, we ought to ask: “for whom?”—since it is people, just people, who will have to pay. Every time we hear that a proposal will destroy society as we know it, we should have the courage to say: “Thank God; at last.” And whenever we hear that it will destroy our freedom, we should be very cautious indeed. For such freedom as we have is our most treasured possession, and we know how to be vigilant. Yet for that very reason, this is the simplest method that the powerful have to cling to power: to convince people that any other concession of that power would be unsafe.

But I would like to stop philosophizing to you about all this, at the expense of introducing another technical term. We have not had a new one for a couple of lectures, and I hope that the very few that I have introduced make a useful vocabulary. We are by now used to the notion that institutions are not just entities, with certain characteristics. They are instead dynamic viable systems, and their characteristics are in fact outputs of their organizational behaviour. The variety that is pumped into them is absorbed by regulating variety, through an arrangement of amplifiers and attenuators. A system that, through this kind of exercise in requisite variety, achieves stability against all perturbations, is called a homeostat.

A homeostat can resist perturbation, not only against expected disruption, but against unexpected disruption too. For this reason it is not only stable, but ultra stable. Whatever happens to it—provided that its relaxation time is sufficiently short—it will not go into oscillation, and still less will it explode in catastrophic instability. The sign of this homeostasis, now so deficient in our major institutions (and perhaps, as I said earlier, even in ourselves) is that critical outputs of the system are held steady.

Why produce this extra terminology at this late stage? It is because I want to answer the question about resistance to change in a very precise way. All homeostatic systems hold a critical output at a steady level. But some of them have a very special extra feature. It is that the output they hold steady is their own organization. Hence every response that they make, every adaptation that th...

Table of contents

• Cover Page
• Title Page
• Copyright Page
• Contents
• The Real Threat to “All We Hold Most Dear”
• The Disregarded Tools of Modern Man
• A Liberty Machine in Prototype
• Science in the Service of Man
• The Future That Can Be Demanded Now
• The Free Man in a Cybernetic World
• The CBC Massey Lectures Series