Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

No. Those aspirations are not possible. Mathematics cannot be reduced to pure logic alone - you have to assume something.

However we are now extremely clear on what you need to assume. And why some assumptions are necessary. And we have collections of assumptions that are very reasonable indeed.

The result is that there is a step of faith. But it is a step that people are very comfortable and confident with.

That said, I would be fascinated to know if this proof can be transformed into a proof that does not need all of the power of Lean. Can it be turned into a proof using the Peano Axioms alone? I think that has a good shot of being true. But it would take a lot of work.

 help



Is it even true that there is a step of faith?

We have many proofs given a set of axioms. But those axioms aren’t even taken on faith as much as they are assumed to be true. In a universe where they are true, the conclusions are true. We could assume they are false and prove other things about universes for which they are false.

I’m thinking along the lines of Euclidean vs. non-Euclidean geometry. Euclid made certain assumptions about geometry that were instrumental in proving various theorems. But you can also find geometries where those axioms are not true! And the existence of other geometries doesn’t invalidate Euclid’s conclusions, it just means that they’re only true when its axioms are taken.


Assuming we're not going down the route of paraconsistent logics (which ZFC etc. are not), consistency isn't really something you can do anything interesting with if it's violated. If it turns out you can also prove the Riemann Hypothesis false under ZFC (which would take out every other theorem along with it), that's just bad for math. So we have to take the leap of faith that won't happen.



Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: