On an isomorphic Banach-Mazur rotation problem and maximal norms in Banach spaces (Q2338954)

The Banach-Mazur rotation problem asks whether a separable Banach space with a transitive norm has to be isometric or isomorphic to a Hilbert space. It is not even known whether a separable Banach space with a transitive group of isometries is isomorphic to a Hilbert space and it was unknown if such a space could be isomorphic to \(\ell_p\) for some \(p\neq 2\). The authors disprove it by showing that the spaces \(\ell_p\), \(1 <p <\infty\), \(p \neq 2\), and all infinite-dimensional subspaces of their quotient spaces do not admit equivalent almost transitive renormings. Recall that a norm on a space \(X\) is called almost transitive if the orbit under the group of isometries of any element \(x\) in the unit sphere of \(X\) is norm dense in the unit sphere of \(X\). They also obtain a version of Krivine's theorem for spaces with asymptotic unconditional structure and a subspace which admits an almost transitive norm. In addition, they prove that the spaces \(\ell_p\), \(1 <p <\infty\), \(p \neq 2\), have continuum many different renormings with \(1\)-unconditional bases each with a different maximal isometry group, and that every symmetric space other than \(\ell_2\) has at least a countable number of such renormings. They also pose the following question: ``Does there exist a separable Banach space \(X\) with a unique, up to conjugacy, maximal bounded subgroup of the group of all isomorphisms from \(X\) onto \(X\)? If yes, does \(X\) have to be isomorphic to a Hilbert space?''