systems thinking, continued…


Richard Moore

Someone responded on the linkedin forum:
I find quite welcoming the notion of locality versus globalization. …While it complicates the picture, I think that a model consisting of a system of systems might be closer to a working solution.

The standard way of analyzing a system is to break it down into subsystems, and to identify the interactions of those subsystems. We can see that in the various diagrams people have shared here. Each subsystem can then be studied on its own, taking into consideration its inputs and outputs to the larger system. And each subsystem can then be broken down into smaller subsystems, until clarity is achieved. 

We then get several diagrams, at different levels. At the top level, we’ve got the big system, represented as a collection of subsystem boxes, connected by various kinds of arrows. The other diagrams, at each level, have the same kind of general structure. It is easier to understand the overall system by looking at these various diagrams, than by exploding it all into one super-diagram, which is what they did with the ‘Obesity’ diagram.
Similarly, when designing a system, the standard procedure is to create the same kinds of diagrams, at the various levels. The art of systems design has to do with correctly identifying which functions belong together in the same subsystem, so as to make the overall system robust and comprehensible. 
In designing very large dynamic systems, such as the Internet, the principles of decentralization and parallelism turn out to be very important. Decentralization makes the system more robust, more resistant to disruption. Taking down any one node on the Internet does not disturb the overall system; it only causes localized disruption. Parallelism enables different parts of the system to function, to get work done, without depending on or waiting for other parts of the system. Thousands of independent servers are more powerful than one super-mainframe, and a parallel system can grow and evolve much more easily than a centralized system.
When we think about achieving sustainability, there are two approaches. One approach, the one everyone seems to be talking about here, is to ‘fix’ the current system, by first analyzing it, and then twiddling it here and there to make it more sustainable. The other approach, the one adopted by the leading researchers into sustainability, is to envision a whole new system, based on decentralization and parallelism. That is what bioregionalism, and localization, are all about. 
The first approach, twiddling the existing globalized system, is fraught with difficulties, both ecological and political. Ecologically, this approach fails to take full advantage of decentralization and parallelism. Politically, there are no effective levers available to us to do the necessary twiddling.

subscribe mailto:

2012: Crossroads for Humanity:

Climate science: observations vs. models

related websites: