This is quite dense to unpack but it does remind me of a main theme from those Adam Curtis documentaries: That overarching master plans, especially social ones, are overconfident by nature and generally lead to large unexpected consequences — many of which are catastrophically bad (ahem Nazism).

My napkin explanation is that governments and organizations (and so on) 'bite off more than they can chew' by implementing theories into policies before really understanding them. They see "this will optimize X (prosperity, civil rights, public health)" without fully understanding the larger system.

However, that's just about the implementation details. In the end, it all comes down to what a society wants and its values. The hard part isn't the intellectualism and engineering of how to 'fix' a society, it's deciding what a 'fixed' society looks like, short term and long term. I think we as technical people need to be reminded often that the social world isn't a machine with an easily identifiable telos.

Nazi-ism is a bit of a bad example, because it's a master plan that, once adopted, went mostly as planned. It was evil, very obviously, but if anything it speaks against the idea that High Modernism is always doomed to fail in implementing its abstract designs.

To your main point: I would go a step further, and say that governments and organizations "bite off more than they can chew" because it's fundamentally impossible for individual leadership to fully comprehend a larger system to enforce a policy on. Abstractions always obliterate local knowledge; worse off, it's usually impossible to even know the unknowns that are being blown away. Rumsfeld's unknown unknowns.

Hayek had it on the mark, but if anything he (and, to a larger extent, his loudest followers) didn't apply his observations thoroughly enough, and instead think you can blindly apply governmental regimentation through universalistic property rights that are somehow immune to the weaknesses of global rules. Smart, technical-minded people are especially prone to this fallacy: see the everlasting temptation to shove waterfall processes into Agile rhetoric.

I'm taking some time to digest this article, but I like it a lot, and I think it might fit in with some thoughts I've been trying to solidify for awhile. Going back to Rumsfeld, he had a trio of the known knowns, the known unknowns, and the unknown unknowns. I want to argue there's a forth category, however: the unknown knowns. In other words, institutions and networks of individuals may be isomorphic to real data and the algorithms that work on it, even though no individual needs to know about the data or algorithms actually being implemented, any more than a register would need to know what bits lies inside it. We thus embed knowledge into society: we've created an emergent social computation through unplanned collective activity. The parallels to Elinor Ostrom and James Scott should be easy to build to from there.

The unknown knowns -> qualia. I am doing some work on the application of these concepts in organisational design and control.

Anything you've written or read on this topic you'd recommend?

There's an Errol Morris documentary from last year about/featuring Rumsfeld called The Unknown Known. https://en.wikipedia.org/wiki/The_Unknown_Known

In regards to your comment about 'this will optimize X,' it reminded me of article [0] I read recently. The main point of it is to 'beware the local maxima.' Governments seem to focus on one goal at a time, the X, and seem to disregard effect to other variables: they care only about the local maxima.

[0]: http://www.iafrikan.com/2014/07/07/beware-the-local-maxima-p...

>overarching master plans, especially social ones, are overconfident by nature and generally lead to large unexpected consequences

in application to reality, a formal system and expressions built in it are always approximations. The more complex an expression the farther it usually is from reality. It is true for linearized equations describing physical processes and even more true for formal systems describing society/economy/etc...

"Rather than design master plans, we designers of systems should instead design federations of little systems linked together by extensible protocols and generic interfaces. This way we can create complex systems without having to explicitly define them. These protocols and interfaces themselves are systems which, in order to work, must constrain what can be said."

I'd be very interested to see a companion piece that investigates specific working examples of this approach. Most importantly, I'd like to see examples of systems of this kind that can actually be attributed to a single person or small group.

This is basically 'systems theory', which was all the rage in the 60s and 70s, and Christopher Alexander's work was linked to. There's a reason it didn't take off as much as expected - not because the overarching premise is wrong (that its possible to describe all things, even design, in a systematic way), but just that its incredibly difficult to the point of impossible for people to do so.

Notes on the Synthesis of Form is a lot more systems-theoretic. The algorithm is more or less recursive min-cut partitioning which is now well-researched.

There is, however, a real logistical problem is getting the input requirements, massaging them to identical conceptual scope, and then connecting them together so you can perform said decomposition. That is a massive, massive bottleneck.

That's why it's so telling that Alexander abandoned the technique in his thesis in favour of using the building site itself, in effect, as a computational medium.

Right: that's what I was trying to imply. It's very appealing to think this way, but the only missing piece is how you actually do it.

The Web from its inception until about ~1996-97 was like this.

If you like this stuff, check out Thriving Systems Theory and Metaphor-Driven Modeling.

It takes the 15 properties and integrates them across system choice centers. Some interesting things fall out of it:

1. Levels of Scale -> Stepwise Refinement 2. Strong Centers -> Cohesion 3. Boundaries -> Encapsulation 4. Alternating Repetition -> Extensibility 5. Positive Space -> Modularization 6. Good Shape -> Correctness 7. Local Symmetries -> Transparency 8. Deep Interlock and Ambiguity -> Composition of Function 9. Contrast -> Identity 10. Gradients -> Scale 11. Roughness -> User Friendliness 12. Echos -> Patterns 13. The Void -> Programmability 14. Simplicity and Inner Calm -> Reliability 15. Not Separateness -> Elegance

http://www.amazon.com/Thriving-Systems-Theory-Metaphor-Drive...

Are those mappings in the book you linked? Or something you just came up with?

They are from the book.

Well, after clicking on the link, I discovered that he had basically copy-and-pasted the chapter titles for Part 3 of the book.

I think this article was written by one of those general-thinking people whom I can never understand.

I think it's hard to understand because he's generalizing a set of experiences or insights he's had, but he's only writing about the generalization. You don't know all of the insights and experiences and so the generalization seems too... general. To that end, did you take a deeper look at any of the things he linked?

Alexander's stuff has changed the world already and I hope it will change the world much more in the future. Hope it is sooner than later.

If you're a fan of Taleb's Antifragile, this stuff is the Antifragile.