Pure state transformations induced by linear operators

From MaRDI portal



Abstract: We generalise Wigner's theorem to its most general form possible for B(h) in the sense of completely characterising those vector state transformations of B(h) that appear as restrictions of duals of linear operators on B(h). We then use this result to similarly characterise the pure state transformations of general C*-algebras that appear as restrictions of duals of linear operators on the underlying algebras. This result may be interpreted as a noncommutative Banach-Stone theorem.


The main goal of the paper is the study of those transformations between the pure state spaces of two C*-algebras which in a canonical way correspond to linear mappings on the underlying algebras. Following the notation on the paper, for each pure state \(\omega\), \([\omega]\) will stand for the class of all other pure states which are unitarily equivalent to \(\omega\). A transformation \(\varphi^{\sharp}\) between the pure state spaces of two C*-algebras \(A\) and \(B\) is called \textit{fibre-preserving} if for each pure state \(\omega\) we have \(\varphi^{\sharp} ([\omega]) \subset [\varphi^{\sharp} (\omega)]\). We shall say that a property holds locally for \(\varphi^{\sharp}\) if it holds for the restriction of \(\varphi^{\sharp}\) to each equivalence class \([\omega]\). Finally, \(\varphi^{\sharp}\) \textit{has a locally solid range} if \(\varphi^{\sharp} ([\omega]) = \varphi^{\sharp} ([\omega])^{\perp \perp}\), where for each subset \(V\) of the pure state space, \(V^{\perp}\) denotes the set of all pure states which are orthogonal to each element in \(V\). The author establishes the following characterisation which could be regarded as a non-commu\-tative version of the Banach--Stone theorem: Let \(\varphi^{\sharp}: \mathcal{P}_{B} \to\mathcal{P}_{A}\) be a transformation between the pure state spaces of two C*-algebras \(B\) and \(A\). Suppose that no irreducible representation of \(B\) is of the form \(M_2 (\mathbb{C})\), then there is a linear map \(\varphi : A\to B\) whose adjoint is \(\varphi^{\sharp}\) (i.e., \(\varphi^{\sharp} (\omega) = \omega\varphi)\) if and only if \(\varphi^{\sharp}\) is uniformly \(\sigma(B^*,B)-\sigma(A^*,A)\) continuous, locally bi-orthogonal and has a locally solid range. The case in which \(\varphi\) is a Jordan *-isomorphism is also discussed. The paper generalises the work of \textit{F.~W.\ Shultz} [Commun.\ Math.\ Phys.\ 82, 497--509 (1982; Zbl 0488.46050)] where only the case of bijections is studied. One of the main tools developed in the paper is a generalisation of Wigner's theorem in Theorem 2.13.











This page was built for publication: Pure state transformations induced by linear operators

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q813948)