1

I want to verify whether the statement in the title is indeed true, i.e, if $\vert X\vert=\kappa \geq \aleph_0$ then the collection of subsets of cardinality strictly less than $\kappa$ is also $\kappa$?

It relies in the following propositions:

  1. For all $0<\lambda< \kappa$, $\kappa^\lambda< 2^\kappa$ and $\kappa\cdot \kappa=\kappa$.
  2. There are $\kappa$ cardinalities lesser than $\kappa$. I think this derives from the definition of an ordinal.
  3. $\vert\{ A\subseteq X: \vert A \vert< \vert X\vert \} \vert= \Big\vert \sqcup_{\lambda<\kappa} \{ A\subseteq X: \vert A \vert=\lambda \}\Big\vert = \sum_{\lambda<\kappa} \kappa ^ \lambda < \big( 2^\kappa \big)^{\kappa}=2^{\kappa\cdot \kappa}= \vert 2^X\vert. $

I am unsure about $(1)$, but it seems natural that $2^\lambda< 2^\kappa$ if $\lambda<\kappa$, but I know that in the world of set theory this might be the case. I have a suspicion that the claim is true even if (1) isn't, but I would welcome any counter-example, whether it be for (1) or the whole claim.

Keen-ameteur
  • 7,663

1 Answers1

1

First, your restatement of the title is not the same as the title since less than $2^\kappa$ is not the same as at most $\kappa$.

Second, for any infinite $\lambda<\kappa$, we have that $\kappa^\lambda$ is the cardinality of the subsets of $\kappa$ of cardinality $\lambda$ (see: Proving that for infinite $\kappa$, $|[\kappa]^\lambda|=\kappa^\lambda$). Therefore, any failure of (1) (i.e., $\lambda<\kappa$ and $\kappa^\lambda=2^\kappa$) will constitute a counterexample to the question. Now, it is certainly possible for this to happen. For example, by Easton's Theorem it is consistent that there are regular $\lambda<\kappa$ with $2^\lambda=2^\kappa$. In a different direction, if you assume GCH then, for any singular $\kappa$, if $\lambda=\text{cof}(\kappa)$ then $\lambda<\kappa$ and $\kappa<\kappa^\lambda$, hence $\kappa^\lambda=2^\kappa$ by GCH at $\kappa$.

halrankard2
  • 1,592
  • 2
    You don't need Easton's theorem there. This is true in Cohen's original model of $\lnot\sf CH$. And it follows from Martin's axiom, and therefore from all the stronger forcing axioms. – Asaf Karagila Sep 28 '20 at 13:01
  • Thanks, I had forgotten. Indeed, for any infinite $\kappa$ there is a model of ZFC in which $\kappa^{\aleph_0}=2^\kappa$. So the number of countable subsets of $\kappa$ is $2^\kappa$ in this model. – halrankard2 Sep 28 '20 at 17:53
  • I am not that familiar with the notion of co-finality and forcing, but you're saying that in many cases, $\lambda<\kappa$ and $2^\lambda =2^\kappa$? – Keen-ameteur Sep 28 '20 at 18:02
  • Yes, but the existence of such cases is independent of ZFC. However, the more relevant condition for you is $\lambda<\kappa$ and $\kappa^{\lambda}=2^\kappa$. That can still fail even if you assume exponentiation $\kappa\mapsto 2^\kappa$ is strictly increasing. For example, suppose $\kappa=\aleph_{\omega}$. Then $\kappa^{\aleph_0}>\kappa$ (in ZFC). So if you assume GCH (which implies exponentiation is strictly increasing), you get $\kappa^{\aleph_0}=2^\kappa$. – halrankard2 Sep 28 '20 at 18:34