Sapient Governance II — Coordination Level, Part A

Logistical Coordination

Natural Coordination in Living Systems

In Sapient Governance I, Part A, I introduced homeostasis as a form of operational level self-control for processes. Homeostasis is a form of closed-loop control in cybernetics. As long as individual processes are cooperating and are tightly coupled (see below) systems can operate smoothly. The origin of life itself probably went through a quasi-Darwinian evolution in which groups of autocatalytic but interacting processes aggregated in protocell arrangements. Since evolution involves copy with error along with selection by external conditions, these protocells were fragile and subject to internal disruptions as well as external ones. Life had to develop a way to deal with the external environment and make sure internal processes were in balance and coordinated when cooperation wasn't enough.

And life did evolve such a facility. In order to deal with the exigencies of existence in a changing world life quickly learned the trick of tactical and logistical control as a layer above homeostasis and cooperation. One of the more compelling models of this is due to Humberto Maturana and Francisco Varela, called autopoiesis, meaning self-creating. For life there are the production processes, like protein synthesis or ATP production. Then there are additional functions that monitor these productions in order to match demand with supply. If supply exceeds demand, these functions provide feedback signals to lower production rates, thus matching supply to demand. Very little is wasted in a cell!

At first glance this appears to be nothing more than feedback control covering more than one process, and it does incorporate that. But it is qualitatively different in that it invariably involves more complex signaling (information processing) and operates over many interacting processes and, very importantly, over longer time scales than operational control. It thus must have a more complex model of processes and flows in order to optimize overall functions. Generally, this form of control also involves feedforward signals to a greater extent. And, perhaps most importantly, it provides anticipatory signals to processes prior to their getting out of balance.

A simple sequential coordination (logistical) control mechanism

This diagram shows a very simple sequential flow controller. It gets feedforward error information from sensors in the flow stream. Shown also is the feedback that each process (P1, P2, and P3) gets from its own output. The flow controller uses feedforward and feedback information (open arrows) as input to an optimization model and uses the output of that model to send signals back to the processes to change their rates of processing. The model, in the case of a living cell, is biochemical analog computer. There is no real separation between the controller and the model.

In truth most process interactions are in complex networks and/or have side inputs/outputs. But processes tend to cluster around a larger general function. See this article in Wikipedia about the the Citric Acid Cycle (Kreb cycle) for an example of a complex process, that is a process composed of smaller processes, that uses cooperation and autopoietic logistic control (the little arrows that swing into the main cycle at various points). Biochemists have worked out many details of the control structures in the living cell. Of course the grand control system involves the genetic DNA and its own autopoietic control system of epigenetics. This is obviously a subject far broader then can be covered here, so I will leave the reader to follow the pointers and discover more about how hierarchical control plays out in the living cell.

Autopoiesis also provides services to produce new component processes when something is degraded (the entropy law is forever being fought by the autopoietic processes). When stresses occur, such as temperature or lack of food, autopoietic controls re-budget internal energy flows to make sure critical functions are supplied with enough, sometimes even shutting down non-critical functions until the stress passes. Finally, the autopoietic network of controls will direct the cell to excrete substances (depending on cell type) in response to external conditions and signals from other cells. Thus autopoiesis is a model for coordination control, both logistical and tactical, for cell behavior.

Now, how can we use these concepts in the governance of human society, and in particular the economy?

Coordination Control in Social Governance

Until very recently, in the United States, the conventional wisdom (or I should say the prevailing wisdom) has been that the market is the ultimate solution to all problems in the economy. Since the Reagan Revolution we have been making steady progress on deregulating everything imaginable. The government has tried, especially in the last seven years, to get out of the regulation business under the ideological belief that the economy would take care of itself if we just let the magic of the market (a la the modern interpretation of Adam Smith) do its thing. Of course what no one in Washington was willing to admit was that, that philosophy only applied to regulatory actions, but not to subsidies. Only now we see this was completely fallacious reasoning. We are currently in the throes of major market failures in housing, finance, and the stock market. Oil prices have skyrocketed causing severe economic shock, but in this case, in spite of all the witch hunting going on to find the evil people behind this, it looks like the so-called fundamentals (supply and demand) are to blame. Hence the market is working in this case [sadly many of the voices that promoted free market solutions previously are the same ones positing evil-doers manipulating the market now.] But massive market failures are coming to light everywhere else.

But the idea that a market economy without some form of coordination control could be sufficient flies in the face of what we know from hierarchical control theory and the natural systems that have been studied. The market economy, as I argued in my prior two pieces, SG I, Part A and SG I, Part B, is part of an overall operational level control mechanism that would work best for cooperative processes and then only if information is freely available in a timely manner. In natural systems, like living systems, there are analogous processes and they cooperate, for the most part, to produce a smooth functioning that benefits all components. Competition comes about when resources become constrained and some kind of priority scheme has to be enforced to assure that the living system continues to function. So too in the economy, competition creates a situation where priorities have to be set. And there is no inherent mechanism within the market to do so.

In this installment I want to explore the nature of coordination control as it might apply to wise governance of society. I am focusing mostly on economic governance since it is the most pervasive aspect of our daily lives. Other issues such as public protection functions (e.g. police, fire departments, etc.) or private protection functions like healthcare, are certainly important. And, at least in the case of the latter, are intertwined with the economic issues. But I think I can make my point focusing on the economics of society and, besides, without economics nothing else would probably matter much.

Coordination control consists of two major functions. One is logistical control, which is what I will describe in this installment. The other is tactical control, to be discussed in Part B. In truth these two functions must actually operate in coordinated fashion themselves. They are partners in coordinating the operations level and in coordinating the whole system with its environment. In the next installment I will describe the coordination of the system — human society — with the rest of the Ecos.

Cooperation and Competition in Operational Subsystems

If all processes in the operational level cooperated all the time there would be no need for any additional levels of control. Cooperation here means that processes are perfectly coordinated in terms of their receiving and supplying roles. An example of high internal coordination would be the packet communications protocols used in the Internet. Such protocols are possible due to the tight coupling between processors. Another example might be the modern computer-mediated supply chain used to stock retail inventory. But most real world processes are not tightly coupled.

Coupling strength depends on time and number of 'hops' involved in a flow.

Each process in this diagram receives flows of some stuff from other processes. The diagram follows one stream from process A through to process D. Along the way processes receive other influences that can introduce various kinds of disturbances in the flow of interest. Additionally time delays or inherent latencies in the flows means that disturbances will not be felt immediately by downstream processes. Processes that are further away from A (C and D) have a lower coupling strength with regard to the certainty of the flow path. Many things can interfere with the flow before it arrives at D. The dashed lines represent the relative strength of coupling. Clearly in a highly complex, dynamic system where many couplings intercede between point A and point Z and many interferences may affect the flows, coupling strengths can be highly variable. Such systems are subject to disruptions and uncontrolled oscillations in flow levels.

Typically natural systems use buffers at each process to help mitigate the exigencies of flows. Inventories are a natural example of such buffers in economic processes. But inventories carry a maintenance cost that might exceed its value to the whole system at times. An alternative to expensive buffers is to use fast information channels to communicate disturbances before they reach distal points.

Establishing an information channel between points can smooth the flow and strengthen the coupling.

This, by itself, does not ensure the system as a whole will perform optimally, or even satisfactorily. But it does help coordination between the two end points. Also, this solution does not work in terms of coordinating all of the players. In the diagram we provide communications between distal points. What about the in-between points. A full solution to coordination requires informational links between all point pairs. Since the number of connections between all points in a network goes up much faster than the number of points, this is really not feasible in very complex systems. In a long chain such as shown in the diagram, an alternative partial solution is to have communications links just between the pairs with a direct flow link (input/output relation) and set up computational coordination along the path. This, of course, introduces time lags and so has inherent problems as well. We will see how a coordination level of control can help with this problem.

The problem with competition is more bothersome. Competition occurs when two or more processes use the same resource from a single rate limited source. In the primary flows of materials and energy this can be seen to create a problem when the total resource being produced is less than the aggregate of demand. Processes then compete for the scarce resource based on their own self interest. In the economic system this is also seen in terms of customers being scarce resources. Or rather their money is. Multiple processes (products and services) compete for scarce customer dollars. Traditional economics teaches us this is a good thing because it causes companies to keep their prices low to attract customers. But, this has its own logical flaw when companies can fudge on quality aspects to lower their production costs. Price competition is a double-edged sword. Without some higher level coordination based on global optima we stand in danger of misallocating resources — having lower priority (in the larger scheme of things) processes out compete higher priority processes to the detriment of the whole system.

I don't want to leave an impression that I think competition is a bad thing. As with all aspects of systems dynamics, it has its place. Some level of competition is good for 'jostling' the system as a whole so that it tries out new connections from time to time. Introducing a little noise into a system allows it to find a new, possibly more efficient, configuration. In biological cells Brownian motion keeps molecules jostling about as the system searches for an energy minimum. Competition in economic systems can fulfill this same principle to keep the system from getting stuck in a rut. It is when competition becomes cutthroat, or becomes the raison d'être of the system that things get out of control and it no longer serves a useful purpose.

Coordination control, properly understood, is a service to operational subsystems that help ensure adequate communications and prevent excessive competition. It is true that when a system is put under certain kinds of stress, that some form of so-called 'command and control' is necessary to intervene for the sake of coordinating priorities. We'll look at that as well.

Logistical Control

The primary purpose of logistical control is to assure sufficient smooth operations and prevent undue waste of resources. In a fully specified system (one in which all of the relevant variables are known and all functions on those variables are known), engineers will talk of optimization. That is, the system is coordinated in a fashion that guarantees 'perfect' performance and perfect balance of all of the subsystems. Unfortunately real complex, dynamic systems, like life or societies, are nowhere near fully specified. We wouldn't even know where to begin. This fact, by the way, is why the Soviets failed in their attempts at a planned economy. They treated the problem as if the system were fully specified and they simply needed to tweak the control signals. What they simply failed to comprehend is that operational control is the first line of system performance and that logistic control has to operate as a service to the former. One example of a nearly fully specified operational system is the military. Here the notion of command and control for coordination is more like what the Soviets imagined they were dealing with. So naturally they ran their system like a military operation. But the truth is a complex economy is greatly under specified and needs an evolutionary model of service-oriented coordination.

This is not to say that there aren't some important aspects of the economy that are not specified. Indeed there are several (below and in Part B). But, as will be seen, the trick of an evolutionary coordination is to be able to command when required and simply provide service in every other area. This will allow the underlying operational level a degree of freedom in exploring and discovering new ways of doing things.

There are the problems of time scales, delays, and inertia, in inter-component flows that create problems which mere cooperation between components cannot solve. On top of it all there is noise in the system. Uncertainty is the rule of the day in real life.

Information Management Service

There are several areas where logistical coordination can serve to ensure adequate information flow in the operational level. The most important of these is in the markets themselves. The markets are mediated in a common currency used in transactions. Thus the currency carries, or should carry, the best information available for purposes of pricing. Another area that is crucial to operations processes is having some global measures of availability and quality of resources.

Price Signals

Markets are mediated by the flow of currency, counter to the flow of goods and services, and those flows are determined by prices. Everything in the market has a price. Customers are told what price they should pay and producers learn what price customers (in the aggregate) are willing to pay. It is a continuous dance. Under a system where profit motive is king (our current system) each party is trying to get the best deal and would love to take advantage of the other party to do so. Not too much advantage, mind you, but just enough so that they come out a winner and get to pocket the difference.

The reason this dance is choreographed this way is that nobody really knows how much things cost to make and move. Money, today, has no standard basis that gives a sense of value. Dollars are worth what you can buy with them and economists are set about trying to establish some kind of average purchases basket of goods as an indicator of what a dollar is worth. It is a good idea to tie the dollar to actual assets but the problem is that we then turn around and cost the assets in dollars. It is a horribly circular argument. The result is prices do not truly reflect actual costs (in some real currency) and the rule caveat emptor reigns. Its a lousy way to encourage a cooperative environment.

But if money were valued by a standard that actually made physical sense then costs would be easy to assess and prices would be transparent. In other words, the exchange of currency for goods and services would convey the real information such that buyers could judge value of purchases much better (assuming of course that no one cheats).

Since all economic work (useful work to humanity) requires the expenditure of energy I propose that money be based on energy. In my Question Everything blog I address this proposal in "What is money, really?" So I won't push the idea any further here than to say that an economy in which money has a standard value based on a real physical quantity would go a long way (maybe 80% of the way, but don't quote me) to increasing the transparency of the market. While the current vogue in economics is to bash the rational agent, and not without reason, I think people would be able to make better judgments about their purchases if full costs were worked into the asking price. I'm betting that many items we currently take for granted as being affordable would suddenly, if the full cost in energy were taken into account, become prohibitively high relative to what people would be paid for their labor/mind energies. In the USA we have a long history of government subsidies for things that are inherently energy expensive, indeed for energy itself. So we buy food from Central America or China or who knows where, but not from our neighborhood farm. Next time you buy a mango, ask yourself what was the cost of this item in total BTUs expended? If you bother to do the math, you will either be thankful that energy isn't the basis of money, or think twice when going for the next one.

In the blog post "What's wrong with this picture," I provide a set of cost accounts from a series of businesses, from parts manufacture to final product (reverse order) to show how energy, while a seemingly small contribution to products actually, and ultimately, accounts for all costs. Labor is energy, materials have an inherent energy component (called embodied energy*). And if you go all the way back to mining the ores and farming the food, you find it is energy all the way down! Supplying the energy to these basement industries also takes energy. That, in fact, is the real reason oil costs so much now. The net energy gain is diminishing the harder it is to get the oil out of the ground. Just like 'it takes money to make money', it takes energy to get energy and that is the true baseline factor.

Coordination control comes into play in several ways in this scheme. One way that is really a tactical issue, is the measurement of the amount of energy that is available to the economy to do useful work. This doesn't mean just measuring the number of barrels of oil, or the tonnage of coal, but the actual energy measure such as British Thermal Units (BTUs) or joules. Some agency would monitor, in semi-real-time, the influx from the environment of all net energy ready to do work (oil needs to be refined into directly useful products). I will address in Part B, in the tactical coordination treatment, how this information is used to regulate the inflow of energy. But here the important point is how the information is used logistically within the economy. What we need to know, globally, is how much representational currency (i.e. money) is to be available in circulation. If the value of the currency is fixed to a given amount of energy, and if the aggregate of net energy into the economy is changing, then the number of currency units circulating has to be adjusted accordingly. Reducing the amount of money in the whole system would tend to slow things down since there would be less cash around. But that makes perfect sense since there would also be less actual energy to do work. This situation is where priority setting might be necessary if the tactical arm assigned to boost energy availability is slow on the job!

An energy standard for money, with the appropriate monetary policy, would assure that a phenomenon like inflation could not happen. The economy might contract, as mentioned, but prices would still reflect true costs. Moreover, market bubbles, such as we have seen recently, could not even get started since prices could only vary with the actual amount of energy put into the product. In fact, under this scheme, and where efficiency standards reduce the amount of energy needed for a given product, the prices of things should tend downward over time. That is, within physical limits.

Detecting and Exposing Cheaters

Sapient governance might still be achievable if we had better information about the unsapient among us who can't help themselves when it comes to lack of moral/ethical behavior. Well all of us are at risk under the seeming right circumstances. So anomaly or cheater detection is one of the prime forms of logistical control service. We do this now (or did before Reagan/Bush I/Bush II and maybe a little blame goes to Clinton too) in the form of regulatory agencies of various kinds. So it isn't a new idea. The efficacy of this function, however, needs work. We need clearer ideas about what constitutes appropriate production functions (including, of course, waste processing) and set up monitoring systems that can actually do the job.

At this point let me introduce a claim that will not be welcomed in certain quarters, but here goes anyhow. The problem with our current systems of regulation are that they are decided and run by lawyers. OK, the law is an important element in enforcement and ownership and other areas. But the fact is that in most of these cases we have got good specifications of what the process should be doing, or we know how to develop good specifications. There is an inherent truth in the matter, for example of what constitutes reasonable discharges of waste matter into the environment. This is the ken of scientists and engineers. Lawyers are trained to win contests. Scientists are trained to discover truth to within an acceptable margin of error. Big difference in motivations. Currently our agencies are run by lawyers who occasionally consult with the scientists and engineers. I would have it that scientists and engineers, with some cross training in legalities, ran the agencies and consulted with lawyers when necessary. The main issue is that many of the regulatory agencies are established because there is a veridical position taken for the benefit of society (and the environment) as a whole. Scientist and engineers are motivated to cut through the complexities and get to the central issue. Lawyers too often regale in complexity and nuance because it makes their form of reasoning (and the challenges) seem extra important. But science is not decided in a court of law. Laws, on the other hand, can be decided by the product of science.

Command and Control (C&C)

Under Stress

What is going to happen after the peak of energy is reached and we enter a long emergency, as James Howard Kunstler calls it? Our economic system, currently dependent on mindless consumption, is going to go into a tail spin. Any system that experiences a restriction on a major resource input has to marshal its internal resources and go on a crash budget. Cells do this when deprived of some necessary foodstuff or subjected to temperature stress. The system has to have a command and control backup that takes over to handle the emergency. Think about the way we and the Brits operated during WWII. There was rationing and sacrifices so as to deploy resources to the fighting machine. This is a natural response to emergencies.

In our economy today, and I would suppose long into whatever future we have, there are micro-emergencies, or local conditions that arise all the time. Think of floods in the Midwest, or Hurricane Katrina. Under these situations the coordination level has to respond with emergency actions. Local processes cannot know what their relative importance is to the whole system. So it falls to a higher level of control to coordinate and budget resources accordingly.

Under Strategic Decisions to Change

One form of long-term coordination involves transformation of the underlying operations level due to strategic decisions. From time to time things need to change and new processes created, old ones disassembled. In living systems evolution is usually the strategic decision maker (although in a hierarchical organization like a multi-cellular organism some cells may change functions in response to a signal from some other part of the body). In society strategic-level decisions are thrown out at us all the time without any real strategic thinking behind them. For example, when a technology radically changes or is newly introduced it causes substantial long-term changes in the economy. Automobiles put buggy whip manufacturers out of business. Computers and telecommunications have so altered our society that we still have no real idea what the significance is.

But once a strategic-level decision is made, the command and control mechanism takes over to direct the changes needed in the operations level in natural systems. This is easily observed in human organizations like companies and even governments. It is generally not the case in the economy where no one has a strategic view of what the change means. Automobiles were introduced to solve a seeming operational problem - faster (supposedly cheaper) personal transportation. No one could foresee the extensive social changes that would ensue or the calamitous consequences of CO2 and contaminants emissions. No one could have predicted (and this time that statement is truth Condelezza) suburban sprawl and all that that entails. Thus there are few, if any, examples of C&C controls working with a market-based economic system. And where there have been explicit C&C

Life doesn't have a strategic view, in general. That is why evolution works by trial and error. Our economy, indeed our whole social system is currently without a strategic view, even though bits and pieces, such as organizations, have shown an ability to operate strategically within their own sphere. I also have suggested that sapience is part of humankind transcending the boundaries of Darwinian evolution by the evolution of a brain that is capable of strategic thinking. That means life has produced an autonomous unit for which some strategic level control is not just a matter of Darwinian evolution. I will return to that in the last installment in this series.


This has been a world wind tour of a small part of the concept of coordination control in complex, dynamic systems, both living and social. It has only really covered the logistic aspects of coordination and I've limited it to economic governance for the most part. Logistical controls act to coordinate between internal processes at all scales in a complex system. Coordination may be somewhat rigid if all of the variables were known. But in a complex system like the economy this is completely unrealistic. Therefore coordination has to be of an adaptive sort. Overall goals need to be specified, but detailed specifications of how they will be accomplished are not possible. This is why strict command and control economies, such as that tried in the Soviet Union fail. Most of coordination comes from ensuring a free flow of necessary information, so comes as a service provided to operations level processes, rather than dictates. This actually can work when processes themselves are adaptive but responsive to the larger picture goals (for the entire system). Laissez-faire economics, mediated only by the market and trusting completely in Adam Smith's invisible hand metaphor (the 'as-if' part) cannot work unless every process were strongly coupled with every other process, and that is impossible.

In the next installment I will cover tactical coordination, which involves positioning the system favorably with respect to its environment. Tactical coordination can be viewed as trying to find the best cooperation with other systems that does not leave you worse off.