Zulip Chat Archive

Stream: Is there code for X?

Topic: Hartog's theorem


Lucas Polymeris (Aug 23 2021 at 21:23):

Hi. Anyone knows if Hartog's theorem ( for every set XX there exists an ordinal α\alpha such that αX|\alpha| \nleq |X|) has been formalized in Lean? I haven't found anything on mathlib.

Mario Carneiro (Aug 23 2021 at 21:25):

It is trivialized in lean because of the axiom of choice

Mario Carneiro (Aug 23 2021 at 21:26):

you can use (cardinal.mk X).succ as such a cardinal

Lucas Polymeris (Aug 23 2021 at 21:33):

Wouldn't have any interest to have it without choice? I'm new to Lean, so I don't know if there is (as there is in set theory) a tendency to avoid choice when not strictly necesary

Mario Carneiro (Aug 23 2021 at 21:34):

the short answer is that there is not

Mario Carneiro (Aug 23 2021 at 21:35):

there is not really any space for developments that depend on LEM but not AC, as with most AC-free set theory

Mario Carneiro (Aug 23 2021 at 21:36):

Incidentally, metamath is a much friendlier environment for that kind of pure set theory, and there is a proof of Hartogs's theorem there

Lucas Polymeris (Aug 23 2021 at 21:36):

Right, yes. I'm starting to understand that here LEM kind of takes the role of choice in other places

Reid Barton (Aug 23 2021 at 23:07):

In Lean, choice is an axiom and used to prove LEM. So there's no way to track (using #print axioms) that a proof uses only LEM and not choice.


Last updated: Dec 20 2023 at 11:08 UTC