A II\(_{1}\) factor approach to the Kadison-Singer problem (Q461423): Difference between revisions

Let \(B(\ell_2)\) be the space of all bounded linear operators on the Hilbert space \(\ell_2\) and let \(\mathbb D\) be the subalgebra of diagonal operators on \(\ell_2\). \textit{R. V. Kadison} and \textit{I. M. Singer} [Am. J. Math. 81, 383--400 (1959; Zbl 0086.09704)] asked the question if every pure state on \(\mathbb D\) has a unique state extension to \(B(\ell_2)\). Several equivalent re-formulations were obtained over the years until \textit{A. Marcus}, \textit{D. A. Spielman} and \textit{N. Srivastava} [``Interlacing families. II: mixed characteristic polynomials and the Kadison-Singer problem'', to appear in: Ann. Math. (2) 182, No. 1, 1--24 (2015), \url{http://annals.math.princeton.edu/wp-content/uploads/Marcus_Spielman_Srivastava_IF2.pdf}, \url{arXiv:1306.3969}] solved the problem in the affirmative; see [\textit{P. G. Casazza} and \textit{J. C. Tremain}, Proc. Natl. Acad. Sci. USA 103, No. 7, 2032--2039 (2006; Zbl 1160.46333)] for more information about the history of the problem. The author provides a new equivalent re-formulation of the Kadison-Singer problem in the setting of \(\text{II}_1\) factors. Recall that a \(\text{II}_1\) factor is a von Neumann algebra that is infinite dimensional, has trivial center and a completely additive trace state. The most ``basic'' example of a \(\text{II}_1\) factor is the hyperfinite \(\text{II}_1\) factor \(R\) of F. Murray and J. von Neumann, defined as the infinite tensor product \((R, \tau) = \overline \bigotimes_k (M_{2 \times 2}(\mathbb C), \text{tr})_k\), where \(M_{2 \times 2}(\mathbb C)\) denotes the algebra of \(2 \times 2\) matrices with complex entries and \(\text{tr}\) the normalized trace operator on \(M_{2 \times 2}(\mathbb C)\). Let us also consider the infinite tensor product \((D, \tau) =\overline \bigotimes_k (D_2, \text{tr})_k\), where \(D_2\) denotes the subalgebra of diagonal \(2 \times 2\) matrices. \(D\) is a so-called Cartan subalgebra of \(R\), i.e., a maximal abelian self-adjoint subalgebra (MASA) of \(R\) whose normalizer generates \(R\). The author now proves that the Kadison-Singer problem has an affirmative answer if and only if, for every free ultrafilter \(\omega\) on \(\mathbb N\), every pure state on \(D^\omega\) has a unique state extension to \(R^\omega\). The hyperfinite \(\text{II}_1\) factor \(R\) also contains a MASA whose normalizer is trivial, a so-called singular MASA. The author shows that, if \(A\) is a singular MASA of \(R\) and \(\omega\) a free ultrafilter on \(\mathbb N\), then every pure state on \(A^\omega\) has a unique state extension to \(R^\omega\).