Design Patterns for Decentralization
Nature (evolution) has shown us beautiful design patterns for intelligent systems. One such intelligent system is the physical body of any species in the animal kingdom. An animal is centralized in that it has a central nervous system and central blood pump (heart) that supplies blood throughout the system.
On the other hand, we have recently understood that there is no central point of control in the animal brain. We now know that emergent intelligence is pretty much a decentralized thing: there are no central neurons in the same sense as a Central Processing Unit (CPU) in the digital computer. The brain is a fully distributed, decentralized system in which not one neuron (or even group of neurons) controls the system. Its intelligence is an emergent property of the system as a whole, in which each neuron has limited but wide-area input tentacles and both far and wide output coverage.
In other words, nature has shown us beautiful and highly effective designs that have both centralized and decentralized characteristics. I think what nature is showing us is that there are parts of the system in which decentralization make sense and there are parts in which centralization make more sense. The decentralization/centralization question is not a system-wide issue; it is rather like a design pattern issue. You cannot have a system that is wholly decentralized; some parts of it must necessarily be centralized.
In the animal body, there are only two sub-systems that are centralized: command and feedback signals (nerves and the central nervous system) and blood supply. The brain itself is highly decentralized, as we have already noted. Throughout the animal body, outside of the input and output signal connections and cardiac sub-system, everything else is decentralized. We can go down to the cell level and see that each cell is a complex system that functions independently except as regards to energy distribution and waste disposal. When an invading foreign bacteria enters the system, the response is local (even though the effect can be global, as in a viral infection). Some groups of cells receive signals from the central nervous system (muscles). Sensor cells stimulate the input pathways to the brain.
We can think of an independent country or human society as a “system” with lots of sub-systems. We have determined that there are sub-systems that are better centralized (the military, for example) and sub-systems that are better decentralized (like mass media and business). We have tried different system designs for society. In the communist design we have centralized control of both mass media and business; in the capitalist system we have decentralized mass media and business. In either design, government must necessarily be both centralized and rigid.
The whole point here is this: a system cannot be wholly decentralized; otherwise, it’s not a system but total chaos. This is as true in blockchain system design as it is in societal system design. Therefore, the ideal is not a totally decentralized society; the ideal, rather, is a design that affords the most freedom and yet allows us all to advance to a better place.
References:
Niall Ferguson, “In Praise of Hierarchy”;
Carlos C. Tapang, “Stablecoins: The Free Market as Ultimate Decentralizing Force”.
