Recently I've had some questions about cardinality and the real numbers:

For any infinite subset $S \subseteq \mathbb R$ , can we find a set $H$ such that $S \subseteq H \subseteq \mathbb R$ , such that $H$ is closed under addition, and $|H| = |S|$ ? Can we require $H$ to be closed under multiplication as well and still have this result?

I have no experience in these sorts of problems, but I have a feeling these statements are true. Could results like this be generalized to apply to infinite subsets of an arbitrary set closed under some (possibly infinite) set of binary operations? Thanks a lot!


The answers to all your questions are yes. The general construction proceeds by setting $S_0 = S$ and inductively: $$S_n = \{s+t : s , t \in S_{n-1}\}$$ and letting $H = \bigcup_{i\geq 0} S_i$. It is easy to check that $H$ is closed under addition. Further, we have$$|S_n| \leq |S_{n-1}\times S_{n-1}| \leq |S_{n-1}| + \aleph_0.$$ In the case that $S_{n-1}$ is infinite, we do not need the $\aleph_0$ but you can do this construction in the finite case too. Either way, we have $|H| = |S| + \aleph_0$. This probably requires the axiom of choice or some variant.

You can generalize this to countably many $n$-ary operations where $n < \omega$ in the obvious way.

EDIT: Note that $S$ does not need to lie in some set where the binary operation exists to carry out this construction, we can carry out something like this by defining $S_n' = \{(s,t) : s,t \in S_{n-1}$ and define $s+t = (s,t)$ and quotient out by the axioms of the binary relation at each stage.

  • $\begingroup$ By $\mathcal N_0$ I think you mean $\aleph_0$, which is \aleph_0 in TeX. But I'm not entirely clear on what it does in that formula. $\endgroup$ – Henning Makholm Oct 25 '15 at 1:35
  • $\begingroup$ That was quick, thanks! Could you clarify on $|S_{n-1} * S_{n-1}| = |S_{n-1}| + \aleph_0$, as couldn't $S$ be larger than $\aleph_0$? Or am I misreading your symbol? I don't want to assume the continuum hypothesis. $\endgroup$ – B Gunsolus Oct 25 '15 at 1:36
  • $\begingroup$ For a finite $S_{n-1}$ you get $\le$ instead of $=$ in the second displayed equation, though. $\endgroup$ – Henning Makholm Oct 25 '15 at 1:40
  • $\begingroup$ Thank you. I hope there aren't any more mistakes :( $\endgroup$ – Asvin Oct 25 '15 at 1:41
  • 1
    $\begingroup$ This can be strengthened drastically: en.wikipedia.org/wiki/L%C3%B6wenheim%E2%80%93Skolem_theorem $\endgroup$ – Noah Schweber Oct 25 '15 at 1:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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