FOM: Conservative extensions

[Stephen G Simpson]

Joseph Shoenfield writes:
 >      Steve's proof that ZF plus real choice is conservative over ZF for
 > second order arithmetic seems to be marred by a confusion between two
 > models of ZF.

No, no!  That's not what I proved!  What I proved is that ZFC is
conservative over ZF + real choice, for 2nd order arithmetic.

By the way, more is true: ZFC + GCH is conservative over ZF + real
choice, for 2nd order arithmetic.  This needs an additional argument:
collapse c to aleph_1 by forcing with countable conditions, without
adding new reals.

What you said I proved isn't even true!  ZF + real choice is *not*
conservative over ZF, for 2nd order arithmetic!  (To see this, look at
the Feferman-Levy model of ZF in which aleph_1 = aleph_omega^L.  In
this model, Sigma^1_3 choice fails.)

 > The first, the class of sets constructible from a real, does not
 > necessarily satisfy choice; in fact, a popular axiom is that this
 > model satisfies AD.  The second, the class of sets constructible
 > from R (the set of reals) does not necessarily contain all the
 > reals.

I didn't use either of these models.  What I used was the smallest
transitive model of ZF containing R and W.  Here R is the set of all
reals, and W is any well ordering of R.  This model by definition
contains W, R, and all elements of R.  It is a model of ZFC.

-- Steve