Cyclic splittings of pro-\( \mathcal{C}\) groups (Q7228780)

A cyclic splitting of a pro-\(\mathcal C\) group is defined as a splitting as an amalgamated free pro-\(\mathcal C\) product or as a pro-\(\mathcal C\) HNN-extension over a procyclic subgroup; if this subgroup is torsion free and isomorphic to \(\mathbb Z_\pi\), it is called a \(\mathbb Z_\pi\)-splitting. The class \(\mathcal C\) is assumed to be closed for subgroups, quotients, finite direct products and extensions with abelian kernel.\N\NOne of the main results shows that, if a pro-\(\mathcal C\) group \(G\) does not act on a pro-\(\mathcal C\) tree with trivial edge stabilizers and without a global fixed point, then \(G\) does not admit a pair of hyperbolic-elliptic cyclic splittings. The paper also analyzes pairs of hyperbolic-hyperbolic cyclic splittings and determines the possible structures of the normalizers of non-trivial subgroups of the edge groups.\N\NA further main result establishes that, under the absence of an action on a pro-\(\mathcal C\) tree with finite cyclic edge stabilizers and without a global fixed point, if \(G\) admits two hyperbolic-hyperbolic cyclic splittings and \(N_G(C_1)\) is not virtually cyclic, then \(G=N_G(C_1)=N_G(C_2)\) is virtually abelian. If \(\mathcal C\) consists of groups of odd order, then \(G\cong C_i\times \widehat{\mathbb Z}_{\mathcal C}\). The arguments rely on pro-\(\mathcal C\) Bass-Serre theory and include the introduction of pro-\(\mathcal C\) analogues of blowing up a vertex and standard refinement.