1
$\begingroup$

If B is a Boolean ring is of uncountable cardinality c, does B have 2^c distinct maximal ideals ? Can you please give me a reference where this question is answered (hopefully) positively ? Thanks

$\endgroup$

3 Answers 3

3
$\begingroup$

This particular question is easy to answer. Many related questions about the relationships between various cardinals associated with Boolean algebra (or Boolean rings), such as: cardinality, number of ultrafilters, density, cellularity, distributivity etc, can be found in

  • Monk, J. Donald: Cardinal invariants on Boolean algebras. Progress in Mathematics, 142. Birkhäuser Verlag, Basel, 1996. 3-7643-5402-X. MR1393943
$\endgroup$
1
  • $\begingroup$ Since this is the accepted answer, it may be useful to say that the answer is: no. $\endgroup$
    – YCor
    Nov 9, 2019 at 9:32
6
$\begingroup$

This is false. We might as well count the number of ultrafilters, since an ultrafilter is the complement of a maximal ideal.

Let $X$ be a set of cardinality $c$. If $B$ is the set of subsets $S \subseteq X$ such that $S$ or its complement is finite, then $B$ is a Boolean algebra under the usual set-theoretic operations. If an ultrafilter $U \subset B$ contains a finite set $S$, then it is easy to see it must be principal. If it contains no finite set $S$, then it must be the ultrafilter of cofinite sets. So here the cardinality of the set of ultrafilters or of maximal ideals is $c$ again.

$\endgroup$
3
  • $\begingroup$ Thanks for your answer and the example. I understand that the statement is true for atomless countable Boolean rings. $\endgroup$
    – Rangaswamy
    Dec 24, 2012 at 20:55
  • $\begingroup$ Yes, there is only one such Boolean algebra up to isomorphism, and it is the free Boolean algebra on countably many generators. It is immediate that this has $c$ many ultrafilters in it. $\endgroup$
    – Todd Trimble
    Jan 1, 2013 at 21:19
  • $\begingroup$ Equivalent topological description: $Y$ is the 1-point compactification of a discrete set of cardinal $c$ and $B$ is the Boolean algebra of continuous functions $Y\to\mathbf{Z}/2\mathbf{Z}$. $\endgroup$
    – YCor
    Nov 25, 2016 at 8:05
-1
$\begingroup$

I think this is false—is it supposed to be the other way around?

The Stone representation theorem embeds $B$ in $\prod_{\mathfrak{p} \in \operatorname{Spec} B} \mathbb{F}_2$, where $\operatorname{Spec} B$ is the set of prime (or maximal) ideals of $B$. In particular, $c = \lvert B \rvert \le 2^{\lvert \operatorname{Spec} B\rvert}$.

$\endgroup$
1
  • $\begingroup$ For infinite Boolean algebras you have $|B|\leq|Spec(B)|$, so the other inequality is not sharp except for finite Boolean algebras. $\endgroup$ Dec 22, 2012 at 21:06

Your Answer

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

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