Thread by @MattPirkowski

While the ideas of "localism" get a lot right, they also demonstrate an ideological blind-spot in their unwillingness to deal honestly with well-known advantages to centralized structures, such as:

- Superlinear returns to network scale.
- Higher short-term collective agency.

Centralized structures do not emerge arbitrarily. They are not mere products of corruption, dominance, and exploitation.

They emerge because they represent a formidable and robust strategy upon a wide range of adaptive landscapes.

They also minimize in-group transaction costs.

For these reasons, amongst a host of others, it is exceedingly naive to pursue a strategy of localism without fully admitting the likely outcomes:

- Collusive opportunity amongst a sub-net triggering collapse into centrality.
- Invasion by outside force willing to act centrally.

Heterogenous emergent structures, far from being naturally stable, tend to collapse into more efficient structures once lower energy configurations are discovered.

This is as true for complex social mechanisms as it is for a spherical lipid membrane.

This is the process of self-organizing criticality. Typically the evolution of system will eventually lead to a destabilizing cascade, bringing the system to a lower energy configuration.

Now, there are systems in which this isn't true. The brain is an interesting example.

With the brain, we see a form of emergent localism (sub-cortical structures), paired with relatively recent layers of global structure (the dual folded sheets of the cortical hemispheres).

But why not more structural localism in the cortex? Of course, we see some specialization.

But in general, the communicative flows of the cortex are far more global in nature, functionally integrating the more highly localized activity of sub-cortical structures.

Perhaps it is because of some of the advantages of global coherence mentioned above, especially...

...when it comes to navigating increasingly complex state-space, the Coasian costs of coordinating localized interactions start to explode without some form of globally emergent coherence, which then constrains other emergent processes, top-down.

Sound familiar? Perhaps that's because our large-scale emergent social structures seem to demonstrate a similar fractal pattern.

Corporations, states, etc emerge for the same adaptive reasons that our cortex did: superordinate organization possesses deep competitive advantages.

It also possesses limitations, but let's not get carried away and act as if those limitations make this pattern unnatural or entirely undesirable.

Both centralized and decentralized–global and local–systems will play a part in the next phase of stable human growth / adaptation.