Amenable operators on Hilbert spaces

From MaRDI portal





A major open problem -- commonly attributed to G. Pisier -- in the theory of amenable Banach algebras is to characterize the, not necessarily self-adjoint, amenable operator algebras on Hilbert space. The conjecture is that a closed subalgebra of \({\mathcal B}({\mathcal H})\) is amenable if and only if it is similar to a (necessarily nuclear) \(C^\ast\)-algebra. In [J. Oper. Theory 34, No. 2, 239--249 (1995; Zbl 0855.46029)], \textit{G. A. Willis} corroborated this conjecture for algebras \({\mathcal A}_T\) generated by a single compact operator \(T\): he proved that \({\mathcal A}_T\) is amenable if and only if \(T\) is similar to a normal, compact operator. Subsequently, J.\ Gifford -- in a still unpublished paper -- generalized Willis' result to general operator algebras of compact operators: a closed subalgebra of \({\mathcal K}({\mathcal H})\) is amenable if and only if it is similar to a a \(C^\ast\)-algebra. In the paper under review, the authors focus on singly generated operator algebras where the generating operator need not be compact. It is easy to see that, for a normal operator \(N \in {\mathcal B}({\mathcal H})\), the algebra \({\mathcal A}_N\) generated by \(N\) need not be amenable; in fact, as the authors show, \({\mathcal A}_N\) is amenable -- and, equivalently, weakly amenable -- if and only if \(\sigma(N)\) is Lavrentieff, i.e., has connected complement and no interior. Hence, if \(T \in {\mathcal B}({\mathcal H})\) is such that \({\mathcal A}_T\) is amenable and \(T\) is similar to a normal operator, then \(\sigma(T)\) has to be Lavrentieff. The main result of the paper is that \({\mathcal A}_T\) is amenable if and only if \(\sigma(T)\) is Lavrentieff and \(T\) is similar to a normal operator. Related questions are also discussed. For instance, it is shown that \({\mathcal A}_T\) can never be weakly amenable if \(T \in {\mathcal B}({\mathcal H})\) is quasi-nilpotent. This stands in marked contrast to the Banach space situation where such algebras do indeed exist [\textit{R. J. Loy, C. J. Read, V. Runde} and \textit{G. A. Willis}, J. Funct. Anal. 171, No. 1, 78--114 (2000; Zbl 0946.46041)]. Added by the editor in 2007: In the erratum the authors explain the difficulties with Proposition 3.8 and withdraw their main claims.











This page was built for publication: Amenable operators on Hilbert spaces

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4680393)