You can't work well for a "mental model" based on frequency because in reality you have two different mental models. Sure, if one is more prevalent you might re

You can't work well for a "mental model" based on frequency because in reality you have two different mental models. Sure, if one is more prevalent you might reduce the frequency of a mismatch, but that's it. > doing the expected thing for most people most of the time citation needed.