Valuated vector spaces, Kurepa's hypothesis and Abelian \(p\)-groups (Q1328235)

A few years ago \textit{P. Keef} [Abelian group theory, Proc. 4th Conf., Perth/Aust. 1987, Contemp. Math. 87, 31-41 (1989; Zbl 0673.20031)] observed that the existence of certain abelian \(p\)-groups depends on the existence of Kurepa trees. Recall that a tree \(T\) is a \(\kappa\)-Kurepa tree for some infinite cardinal \(\kappa\) if \(T\) has height \(\kappa\) with levels of cardinality \(< \kappa\) and at least \(\kappa^ +\) branches of ordinal type \(\kappa\). In order to gain more insight into Keef's important observation, it is interesting to isolate the Kurepa condition. This can be done by a now classical method of abelian \(p\)-groups, namely by investigating valuated vector spaces first, followed by existence results on abelian \(p\)-groups which depend on particular socles. The authors follow this concept and show a characterization of the existence of Kurepa trees in terms of valuated vector spaces. Say that a valuated vector space \(V\) with values in a cardinal \(\kappa\) or \(\infty\) is a \(\kappa\)-Kurepa vector space if the ground field has cardinality \(< \kappa\), \(| V| = \kappa\), \(| V/V(\alpha)| < \kappa\) (where \(V(\alpha) = \{x \in V : v(x) \geq \alpha\}\) \((\alpha < \kappa)\) induces a Hausdorff topology on \(V\)) and the completion \(\widehat V\) has cardinality \(> \kappa\). The existence of \(\kappa\)-Kurepa trees turns out (Theorem 1) to be equivalent to the existence of such \(\kappa\)-Kurepa vector spaces. Then, following Keef, the authors prove the existence of certain abelian \(p\)-groups under the Kurepa hypothesis on the existence of \(\aleph_ 1\)-Kurepa trees; see details of the paper for the results related to abelian \(p\)-groups.