The generalized continuum hypothesis revisited (Q1976618)

For cardinals \(\lambda\) and \(\theta\), the author defines \(\lambda^{[\theta]}\), \(\lambda\) to the revised power \(\theta\), to be the minimal cardinality of any family \(\mathcal P\) of \(\theta\)-size subsets of \(\lambda\) such that every \(\theta\)-size subset of \(\lambda\) is contained in the union of strictly fewer than \(\theta\) members of the family \(\mathcal P\). The main theorem proved in the paper is the following. Let \(\mu\) be a singular strong limit cardinal, and take any cardinal \(\lambda > \mu\). For some cardinal \(\kappa < \mu\), there is a family \(\mathcal P\) of \(<\mu\)-size subsets of \(\lambda\) with \(|{\mathcal P}|=\lambda\), such that every \(<\mu\)-size subset of \(\lambda\) is equal to the union of \(<\kappa\) members of \(\mathcal P\). The proof uses generic ultrapowers. Then a related but stronger theorem is proved using pcf theory. The main theorem has the consequence that, for \(\lambda\) and \(\mu\) as in the main theorem, there is some \(\kappa < \mu\) such that for all regular cardinals \(\theta\) with \(\kappa \leq \theta < \lambda\), we have \(\lambda^{[\theta]}=\lambda\). (Thus for each \(\lambda\), for most \(\theta\) sufficiently smaller than \(\lambda\), \(\lambda^{[\theta]}=\lambda\).) Somewhat tongue in cheek, the author argues that this is a reasonable way to rephrase the generalized continuum hypothesis so as to obtain a replacement for Hilbert's first problem which is provable in ordinary set theory. The paper concludes with a number of further applications of the main theorem: to diamonds, to a topological problem, to the existence of small models, etc.