Pure subnormal operators have cyclic adjoints (Q1283033)

This is the solution of the problem stated in the title, which was outstanding since 1976. From that time on partial results have been established by Sarason, Wagen, Clancey, Rogers, Chan, Bourdon, Shapiro. The essentially new feature of the procedure is the employment of a conceptually simple comparison method: For two bounded linear operators \(S\) and \(T\) on Hilbert spaces \({\mathcal H}\) and \({\mathcal K}\) it is written \(S\succ T\) if there exists a one-to-one linear map \(A:{\mathcal H}\to{\mathcal K}\) such that \(AS=TA\), and then cyclicity of \(T^*\) implies cyclicity of \(S^*\). Beyond the main theorem it is shown that a subnormal operator has a cyclic adjoint if and only if its normal part is cyclic or equivalently if it is *-cyclic. Proofs of already known facts can be simplified by the comparison method. This nicely written paper with a great result ends with open questions arising from the solved ones.