The quasi-linearity problem for \(C^*\)-algebras (Q1919885): Difference between revisions

Let \(\mathcal A\) be a \(C^*\)-algebra and let \(A\) be the real Banach space of selfadjoint elements of \(\mathcal A\). Let \(A_1\) be the unit ball of \(A\). Let \(\mu\) be a complex-valued function on \(\mathcal A\) which is linear on each maximal Abelian \(^*\)-subalgebra of \(\mathcal A\) and such that the restriction of \(\mu\) to \(A_1\) is bounded. Then \(\mu\) is said to be a quasilinear functional. When \({\mathcal A}= M_2(\mathbb{C})\), the \(C^*\)-algebra of all two-by-two matrices over \(\mathbb{C}\), there exist examples of quasilinear functionals on \(\mathcal A\) which are not linear. The solution of the Mackey-Gleason Problem shows that every quasilinear functional on a von Neumann algebra \(M\), where \(M\) has no direct summand of Type \(I_2\), is linear. This was first established for positive quasilinear functionals by the conjunction of work by Christensen and by Yeadon, and for \(\sigma\)-finite factors by the work of Paschciewicz; the general solution was obtained by Bunce and Wright. All build on the fundamental theorem of Gleason. Although quasilinear functionals on general \(C^*\)-algebras seem much harder to tackle than the von Neumann algebra problem, we can apply the von Neumann results to make progress. In particular, it is proved that: Let \(\mathcal A\) be a \(C^*\)-algebra with no quotient isomorphic to \(M_2(\mathbb{C})\). Let \(\mu\) be a quasilinear functional on \(\mathcal A\). Then \(\mu\) is linear if, and only if, the restriction of \(\mu\) to \(A_1\) is uniformly weakly continuous.