Why are conservatives for punitive correction while progressives do not think it works? I think this can be explained by the difference between stable equilibria and saddle points.
If you have a system where people make random “mistakes” an ϵ amount of the time, the stable points are known as trembling-hand equilibria. Or, similarly, if they transition to different policies some H of the time, you get some thermodynamic distribution. In both models, your system is exponentially more likely to end up in states it is hard to transition out of (Ellison’s lemma & the Boltzmann distribution respectively). Societies will usually spend a lot of time at a stable equilibrium, and then rapidly transition to a new one when the temperature increaes, in a way akin to simulated annealing. Note that we’re currently in one of those transition periods, so if you want to shape the next couple decades of policy, now is the time to get into politics.
In stable equilibria, punishment works. It essentially decreases ϵ so it’s less likely for too many people to make a mistake at the same time, conserving the equilibrium. But progressives are climbing a narrow mountain pass, not sitting at the top of a local maximum. It’s much easier for disaffected members to shove society off the pass, so a policy of punishing defectors is not stable—the defectors can just defect more and win. This is why punishment doesn’t work; the only way forward is if everyone goes along with the plan.
Why are conservatives for punitive correction while progressives do not think it works? I think this can be explained by the difference between stable equilibria and saddle points.
If you have a system where people make random “mistakes” an ϵ amount of the time, the stable points are known as trembling-hand equilibria. Or, similarly, if they transition to different policies some H of the time, you get some thermodynamic distribution. In both models, your system is exponentially more likely to end up in states it is hard to transition out of (Ellison’s lemma & the Boltzmann distribution respectively). Societies will usually spend a lot of time at a stable equilibrium, and then rapidly transition to a new one when the temperature increaes, in a way akin to simulated annealing. Note that we’re currently in one of those transition periods, so if you want to shape the next couple decades of policy, now is the time to get into politics.
In stable equilibria, punishment works. It essentially decreases ϵ so it’s less likely for too many people to make a mistake at the same time, conserving the equilibrium. But progressives are climbing a narrow mountain pass, not sitting at the top of a local maximum. It’s much easier for disaffected members to shove society off the pass, so a policy of punishing defectors is not stable—the defectors can just defect more and win. This is why punishment doesn’t work; the only way forward is if everyone goes along with the plan.