2

Background: The transfer principle in nonstandard analysis implies that any nonstandard model of the reals is a commutative (for additively and multiplicatively). It is also well-known that the set $\beta(\mathbb R)$ of all ultrafilters on $\mathbb R$ is not a nonstandard model of $\mathbb R$, for instance since addition and multiplication do not extend commutatively.

Among the standard results in nonstandard analysis, is that a set is compact iff every point in an enlargement of the set is near standard.

Now, in the context of $\beta(\mathbb R)$, call an ultrafilter on a subset $S\subseteq \mathbb R$ near standard if it contains the neighborhood filter $N_x$ for some $x\in \mathbb R$. The classical result that a set is compact iff every ultrafilter on it converges can now be restated as "$S$ is compact iff every element in $\beta(S)$ is near standard".

Is this analogy just conincidental, or can $\beta(\mathbb R)$ be considered as a non-commutative nonstandrard model of $\mathbb R$ in a useful way?

Ittay Weiss
  • 79,840
  • 7
  • 141
  • 236
  • You don't need the transfer principle to say that any first-order statement true of the reals is true of any nonstandard model; that's what it means to be a nonstandard model. – Qiaochu Yuan Mar 20 '15 at 04:38
  • @QiaochuYuan, I know, this is just an issue of terminology. I'm just curious about the analogy and whether or not it indicates that $\beta(\mathbb R)$ can be though of as a nonstandrad model of $\mathbb R$, but not in the usual technical sense of nonstarndard (i.e., transfer does not apply). – Ittay Weiss Mar 20 '15 at 05:09

0 Answers0