On transitivity-like properties for torsion-free abelian groups (Q2155622)

This paper studies relationships between the classes of transitive, fully transitive and Krylov transitive torsion-free abelian groups. Recall that a group \(G\) is called \textit{transitive} if for any two elements \(x,y\) with \(\chi _{G}(x)=\chi _{G}(y)\) (same characteristics) there exists an automorphism of \(G\) mapping \(x\) to \(y\), \textit{fully transitive} if for any two elements \(x,y\) with \(\chi _{G}(x)\leq \chi _{G}(y)\) there exists an endomorphism \(\phi \) of \(G\) with \(\phi (x)=y\) and \textit{Krylov transitive} if for any elements \(x,y\in G\) with \(\chi _{G}(x)=\chi _{G}(y)\) there exists an endomorphism of \(G\) mapping \(x\) to \(y\). In this paper, a new definition is added: a group \(G\) is called \textit{weakly transitive} if, for any pair of elements \(x,y\in G\) and endomorphisms \(\phi \), \(\psi \) of \(G\) such that \(\phi (x)=y\), \(\psi (y)=x\), there exists an automorphism \(\eta \) of \(G\) with \(\eta (x)=y\). Some old-standing problems posed by \textit{P. A. Krylov} et al. [Endomorphism rings of baelian groups. Boston MA: Kluwer Academic Publishers (2003; Zbl 1044.20037)] are solved. Specifically, Problem 44 from there is answered in the affirmative by constructing a Krylov transitive torsion-free abelian group which is neither fully transitive nor transitive. Some of the results (this is just a personal selection) obtained, follow. Proposition 2.4. If \(G=\prod\limits_{i\in I}G_{i}\) is a vector torsion-free group, where all direct factors \(G_{i}\) are of rank 1 with equal type, then \(G\) is both a fully transitive and transitive group. Theorem 2.6. The vector torsion-free group \(G\) is transitive if and only if \(G=\prod\limits_{i\in I}G_{i}\), where \(\pi (G_{i})\cap \pi (G_{j})=0\) for any \(i\neq j\), and every component \(G_{i}\) is a direct product of rank-1 groups of equal type. Corollary 2.15. The vector torsion-free Krylov transitive group is always fully transitive. Theorem 2.17. If \(A\) is a group such that the direct sum \(A\oplus A\) is Krylov transitive (in particular, is both transitive and fully transitive), then \(A\) is fully transitive. Conversely, there are groups \(G\) which are simultaneously transitive and fully transitive, but such that their square \(G\oplus G\) is not longer a Krylov transitive group. In addition, there is a proper Krylov transitive group \(G\) (which is neither transitive nor fully transitive) such that its square \(G\oplus G\) is also not longer a Krylov transitive group. In particular, in both cases, the direct sum \(G(\lambda )\) is not Krylov transitive for any ordinal \( \lambda >1\), too. Providing a Venn diagram which illustrates the relationships between our four types of transitivity, Section 3 yields constructions and non-trivial delimit examples. On the way, here is one more result. Theorem 3.7. Suppose that \(A\) and \(B\) are weakly transitive groups such that \(\mathrm{Hom}(A,B)=0=\mathrm{Hom}(B,A)\) and for which there is a prime \(p\) such that neither \(A\) nor \(B\) is \(p\)-divisible. Then \(G=A\oplus B\) is weakly transitive, but not transitive nor fully transitive. The paper ends with a ``Concluding discussion and open problems'' section.