[FOM] Question for Nik Weaver

Nik Weaver nweaver at math.wustl.edu
Thu Feb 23 05:23:07 EST 2006

Bill Taylor asked (quoting me):

> Nik, you made this comment in an article yesterday...
> > it is possible to treat things like the real line and
> > the power set of N essentially as proper classes
> Presumably this apposition implies that you do NOT think
> of P(N) and R as being essentially isomorphic?

I do think there's a straightforward bijection between them.
Sorry for misleading you.

> Incidentally, I think this identification, accepted by many
> (most?) mathematicians, and especially by almost all "pure
> mathematicians", is a clear-cut example of what Friedman
> called "coding" just the other day.  He commented as that
> mathematicians loathe coding whenever they see it, and try
> to avoid it at all costs.  This seems to be a counterexample.

Steve Simpson made the point somewhere that there is plenty of
coding in classical mathematics.  (I don't have the reference at
hand.)  I think what he meant by this was that, for example, any
classical "construction" of the real line (as Cauchy sequences of
rationals, Dedekind cuts, decimal expansions) is really a "coding".

Probably most mathematicians, to the extent they are interested
in foundations at all, would prefer a foundational scheme that
involves as little coding as possible.  I don't see that as a
really fundamental issue though.

Friedman's particular beef was that in my J_2 model, some coding
is involved in forming the dual of a separable Banach space.  That
could be a legitimate esthetic criticism.  I don't insist that
J_2 is the "best" predicative setting in which to develop core
mathematics, though I do think it is rather nice.


More information about the FOM mailing list