The weakest \(m\)-convex topology stronger than an \(A\)-convex topology need not exist (Q862012)

Let \(E\) be an associative algebra over the real or complex field. A subset \(V\) of \(E\) is called \(m\)-convex (resp., \(A\)-convex) if it is absolutely convex and idempotent (i.e., \(V^2\subset V\)) (resp., for every \(x\in V\) there exists a number \(\beta=\beta(V,x)>0\) such that \(xV\cup Vx\subset \beta V\)). A topological algebra is called locally \(m\)-convex (resp., \(A\)-convex) if the origin has a fundamental system of neighborhoods consisting of \(m\)-convex (resp., \(A\)-convex) sets. Locally \(m\)-convex algebras were introduced by \textit{E. A. Michael} [Mem. Am. Math. Soc. 11 (1952; Zbl 0047.35502)], where locally \(A\)-convex algebras were also defined implicitly. Continuing his investigations on locally \(A\)-convex algebras, the author constructs several examples of locally \(A\)-convex algebras which solve in the negative several questions posed in some previous papers, such as, for instance, that the collection of all locally \(m\)-convex topologies in general does not have a weakest element.