Excellent comment, and just to clarify the state is the top level of authority, lets call state 'social construct -A-'. An anarchist going in to dismantle the social construct, lets call him 'agent -B-'. We can call the executive order, or referendum or whatever legal document(s) that would be used to dismantle the state, 'social construct -C-'.

Now 'social construct -A-' is not the same as 'agent -B-' so asking the question :

How can the top levels of authority chisel away at themselves?
Is similar to saying social construct -A- and agent -B- is the same entity. That issue from what I can see is not logically correct.

There can be a case made that social construct -C- may not be able to destroy social construct -A-. I think what you are trying to say, and correct me if I'm wrong, but that there is near a zero probability of an event that social construct -C- will destroy social construct -A-.

I completely understand that position. I don't buy in to the idea the probability will be high, but also don't immediately assume it will be zero. On the grounds of probability is where you and I would likely start disagreement.

Historically, it has typically taken a social construct to destroy another social construct, so the tactic or strategy of using -C- to destroy -A- is sound. The biggest barrier I see is the degree of complexity of the constructs.

I completely agree that people are afraid to suffer, and agree to an extent that is why they suffer. The question I ask is if social construct -A- is creating more perceived discomfort than comfort? If significant discomfort then social construct -C- has support and agent -B- is just an agent.