Given a nonprincipal ultrafilter $\mu$ on $\mathbb{N}$ and a sequence of groups $G_i$, one can define its ultraproduct as:

$$ ^*\prod_{i\in \mathbb{N}}G_i:=\{(x_i)_{i \in \mathbb{N}}| x_i\in G_i\}/\sim$$, where $(x_i)_i\sim (y_i)_i$, iff $x_i=y_i$ $\mu$-almost everywhere.

Suppose you are given also group homomorphisms $f_{i+1}:G_{i+1}\rightarrow G_i$, then one could also consider something like an ultra- inverse limit:

$$\{[x_i]_i|f_i(x_i)=x_{i-1} \quad \mu-\mbox{almost everywhere}\}$$

Has this been studied before? Is there a good source, where I could learn about this?

  • $\begingroup$ I've never seen this before, but it looks like a fascinating idea. Are you taking this as a subgroup of the Cartesian product, or a subgroup of the ultraproduct? One guess I would make is that it won't be possible to put a nice topology on this construction, in the way that e.g. profinite groups have a nice topology. $\endgroup$
    – Colin Reid
    Oct 21, 2010 at 12:39
  • $\begingroup$ I'll bet he means a subgroup of the ultraproduct; the square brackets in the definition surely indicate an equivalence class? $\endgroup$ Oct 21, 2010 at 13:41

1 Answer 1


Either the even numbers or the odd numbers are $\mu$-large, so the condition degenerates to ultraproduct.

  • $\begingroup$ OK this explains why I didn't find anything about it $\endgroup$ Oct 21, 2010 at 13:43
  • 1
    $\begingroup$ But does the question still make sense for more general ultra-inverse limits (on large index sets, etc.)? $\endgroup$ Mar 16, 2012 at 4:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.