On minimal norms on \(M_{n}\) (Q2472096)

Let \(M_{n}\) be the space of all \(n\times n\) matrices over \(\mathbb{C}\). There is a partial ordering on the (vector) norms on \(M_{n}\) given by \(\| ~\| _{1}\leq\| ~\| _{2}\) if and only if \(\| A\| _{1}\leq\| A\| _{2}\) for all \(A\in M_{n}\). A norm \(\| ~\| \) on \(M_{n}\) is called an algebra norm if it satisfies the additional property \(\| AB\| \leq\| A\| \| B\| \) for all \(A,B\). It is known that an algebra norm is minimal in the set of all algebra norms on \(M_{n}\) if and only if it can be induced from a vector norm on \(\mathbb{C}^{n}\) [see Theorem 3.2.3 of \textit{G. R. Belitskii} and \textit{Yu. I. Lyubich}, Matrix norms and their applications, Birkhäuser, Basel (1988; Zbl 0645.15019)]. The authors object in this paper is to generalize this result to minimal norms in the set of all norms on \(M_{n}\). The reviewer believes that a difficulty arises because, if there is no restriction on the set of norms, no (nonzero) norm is minimal. The authors have informed the reviewer that the use of ``norm'' in the main theorem should be understood as ``algebra norm''. In that case the theorem is the known result quoted above.