# Defining $\mathbb{Z}$ in $\prod_p \mathbb{F}_p(t)/\mathcal{U}$

Is it possible to show that there is a simple formula, preferably existential, that characterizes a nonstandard model of the ring of integers among the elements of $\prod_p \mathbb{F}_p(t)/\mathcal{U}$ where $\mathcal{U}$ is a nonprincipal ultrafilter over the prime numbers?

Thank you

-
This is a duplicate of mathoverflow.net/questions/86905 . – Emil Jeřábek Feb 2 '12 at 13:59
No this is not a duplicate because, this time, the ultraproduct is over $\mathbb{F}_p(t)$ and not over $\mathbb{F}_p$ – user16974 Feb 2 '12 at 14:06
Oh, I see. But the answer is still the same. By Joel’s argument, any subset of your field containing 1 and closed under + contains $\prod_p\mathbb F_p/\cal U$ (and possibly more stuff), hence it is not elementarily equivalent to the integers: e.g., it makes $-1$ (or any other integer for that matter) a sum of two squares. – Emil Jeřábek Feb 2 '12 at 14:08
I meant: any definable subset etc. – Emil Jeřábek Feb 2 '12 at 14:11
Emil, why not post that as an answer? – Joel David Hamkins Feb 2 '12 at 14:12

No, such a formula does not exist. If $R$ is a definable subset of $\prod_p\mathbb F_p(t)/\cal U$ which contains $1$ and is closed under addition, then by the argument Joel David Hamkins gave in his answer to your previous question, we must have $R\supseteq S:=\prod_p\mathbb F_p/\cal U$. In $S$, every element is a sum of two squares (because this holds in all finite fields). It follows that any integer (e.g., $-1$) is a sum of two squares in $R$, hence $R$ does not satisfy the universal theory of $\mathbb Z$.