Chesterton's fence[0] could sort of be phrased as "there's information about what makes a system successful just in the state of the system itself". [0]: https

Chesterton's fence[0] could sort of be phrased as "there's information about what makes a system successful just in the state of the system itself". [0]: https://en.wikipedia.org/wiki/G._K._Chesterton#Chesterton's_fence https://en.wikipedia.org/wiki/G._K._Chesterton#Chesterton's_... Lots of systems get to a place where there are load-bearing components and no one is around to explain how they got there, why they're important, or even that they are load-bearing. In systems that evolve via selection, even going back historically, there might never have been a good justification for those components. There's probably systems where we've had almost exclusively _incorrect_ explanations for why those components are important. I suspect a lot of human culture is like this. There's traditions and religious rules and other norms, and we don't really know why or even that certain traditions are important. The justifications for things like sexual modesty and religion aren't great, in fact maybe were never good, but the fact that society adopted them among all possible things it could have adopted _may_ indicate that they in fact serve useful purposes, even if we're not currently totally clear on what those are. So I think as society becomes more rational, we see lots of components with bad justifications, get rid of them, and only then realize that those components were serving important functions that we weren't (perhaps ever) really consciously aware of. Not to say this is bad: there's a lot of genuine cruft that ought to be removed, and it's good to discover what's cruft and what's actually important, but I think this is an explanation for why you see pendulums like this.