But, of course, everyone always derides the idea because "we tried this" with New Math in the 70s. No, we didn't. We tried to teach some parts of discrete math (which are formal abstractions) to 5- or 6-year-olds, who weren't yet at the formal operational stage of cognitive development. You can't teach a 6-year-old about set theory any more than you can teach a 6-week-old about object permanence. You have to start when they're older. But "older" doesn't mean "an optional thing to be taught at university!" It's a formative skill, and if you don't pick it up right when you're first able to pick it up, trying to force it into your mind later is somewhat like trying to learn a new spoken language in adulthood when you've been monolingual for your entire life until then.