The bow-tie, the law of requisite variety, and the global economy

John Robb has written an essay viewing the consolidation of the global economy as an instance of the bow-tie formalism from the theory of control systems. Here’s a diagram of the idea as applied to adversarial situations in human society, usefully melded with concepts from John Boyd’s OODA theory of adversarial interaction:

There are a few very significant implications in viewing the global economy in this way that bear explicit emphasis:

  • The center of the bow-tie has an adversarial relationship with the rest of the environment;
  • The heirarchies on either side of the bow-tie that place the resources of the environment at the disposal of the center and impose the decisions from the center on the environment serve to buffer or isolate the center from the environment’s complexity by pre-emptively reducing the complexity of the inputs and translating the simple goals of the center into more complex actions on the environment to attempt to enforce them;
  • The actions of the center, if its spectrum of purpose is total, will have an inherent tendency to reduce the total complexity of the environment to make it more compatible with the capacities of the center (to narrow the wide ends of the bow-tie down to the width of the center);
  • The less complex the environment is, (or the narrower the ends of the bow-tie,) the less complex (narrower) the center may become while preserving the same level of viability with respect to its environment.

Another concept from the theory of control systems, the law of requisite variety, has an important implication: Since the variety of perturbations a system can potentially be confronted with by its environment is unforeseeable and large, (typically due to combinatorial growth of possibilities with system size,) the system should always try to maximize its internal variety (or diversity), so as to maximize its chances of being prepared to benefit from any foreseeable or unforeseeable contigency. (Closely related to Taleb’s antifragility concept.)

Combining this with the previous observations, we can reach much farther. The more the center of the bow-tie with total spectrum of purpose consolidates itself and grows homogeneous, eliminating its internal diversity, the more it must rely on passively buffering itself from its immediate environment by mutilating away its environment’s complexity instead of engaging with it, and the more fragile and less viable it becomes. In turn, the more the center succeeds in mutilating away the environment’s complexity, the more room it has to consolidate itself despite the greater homogeneity consolidation imposes. This dynamic is a positive feedback loop of complexity reduction, the net effect of which is that the broader system in which the control system is embedded loses diversity on the whole, due to the globally homogenizing influence of the center’s attempts to retain control while simultaneously exploiting that control through greater consolidation, and the more fragile and less viable the broader system becomes with respect to external perturbations.

However, when growth of possibilities with system size is uniform and faster than linear, (which is quite true in the case of combinatorial growth,) there is another interesting consequence: the more consolidated the center becomes, the smaller is the size of the environment it can successfully control. So the center is increasingly the center of a narrower and narrower portion of its overall environment, while that portion grows more and more sensitive to perturbations from a growing exterior.

The application of these ideas to the case of the economy is straightforward. Consider Kevin Carson’s recent independent characterization of the historical trajectory of capitalism in this light:

Throughout history, propertied classes have relied primarily on artificial scarcities of material resources to extract a surplus from labor.  With the help of state-enforced artificial property rights, a ruling class can control great concentrations of land and capital. These monopolies prevent competition from driving down the price of capital and land to their natural values. Thus the means of production are artificially scarce and expensive, and labor is forced to pay tribute for access to them.

Today, however, the imploding cost of production means that concentrated ownership of land and capital is becoming less and less effective as a means of rent extraction. The desktop revolution has reduced the cost of setting up a “publishing house” or “music studio” a hundredfold. Micromanufacturing with open source desktop CNC tools will soon do likewise to the cost of a factory. Intensive raised-bed horticulture grows many times more food per acre than mechanized agribusiness. In fact most “farming” is a real estate investment in which the government pays rent for the “farmer” to hold land out of use!

In this age of abundance, when the falling cost of machinery and exploding efficiencies of extracting value from inputs threaten to make control of physical resources worthless as a source of rent, rents accrue mainly to “property rights” like the right to do certain things, or criminalizing competition from more efficient ways of doing things.

Under old-style capitalism, rents were extracted by using artificial property rights to restrict access to physical opportunities for production. Now that the cheapening of physical means of production has made this strategy untenable, the ruling classes must instead charge rents on the right to produce with one’s own physical resources.

Now, what is the way around this self-destructive dynamic? Referring to the bow-tie diagram, we see that a control system with a limited spectrum of purpose and tacit ways of engaging with the environment will be able to deal with inputs in a focused way that doesn’t put stress on its internal capacity for complexity, and will be able to further its purposes through actions that influence the environment by engaging with its complexity rather than violently mutilating the environment and destroying its complexity. Totality of purpose implies an inherently self-destructive dynamic, which casts doubt on the usefulness of such a sweeping ontological category as an economy as a focus for system design, and gives a new perspective on the troubles faced by totalitarian political systems and the tragedies that surround them. Also, smaller control systems interacting with smaller portions of the environment will face less uncertainty in establishing a balance between their internal complexity and their environments’ complexity, due to the rapid growth of complexity with size. The observation that “the most severe fragilities created by bow-tie architectures involve hijacking or manipulating the universally used central protocol and carriers, rather than simple destruction,” made by Marie Csete and John Doyle in their 2004 paper entitled “Bow ties, metabolism, and disease” indicates multiple smaller, simpler systems over fewer, larger, more complex ones to reduce vulnerability to adversarial attacks and worst-case vulnerability to chance failures. And so do the successes of systems designed as collections of individually minimalist, loosely-coupled parts, for example those based on the Unix philosophy, which speak strongly in favor of a society based on principles analogous to Eric Raymond’s design rules for software systems. Let me call your attention to these three:

  • Rule of Robustness: Robustness is the child of transparency and simplicity.
  • Rule of Diversity: Distrust all claims for “one true way”.
  • Rule of Extensibility: Design for the future, because it will be here sooner than you think.

Comments Off

Filed under Uncategorized

Comments are closed.