[FOM] Relevance of reverse mathematics to constructivity

[Daniel Méhkeri]

> If you don't mind, please do fill in a bit. Specifically, if one can 
> safely lose the witness, why isn't AC_{00} valid?

When I said that the difference between primitive N and set-theoretic
omega wasn't important, I really meant to say countable choice (even dependent choices) would be _valid_ in both contexts.

In fact dependent choices might even be the easier one to see, because the
hypothesis would basically be the same. Use dependent choices over N,
then find a way to encode the result as a countable subset of omega times x.

Regards,
Daniel

-- 




      Offrez un compte Flickr Pro à vos amis et à votre famille.
http://www.flickr.com/gift/