Abelian subalgebras of von Neumann algebras from flat tori in locally symmetric spaces (Q813931)

For a semisimple Lie group \(G\) with no center and no compact factors, take \(K\) to be a maximal compact subgroup of \(G\). The associated Riemannian symmetric orbit space \(G/K\), denoted by \(X\), is contractible and of non-positive curvature. Denote the dimension of a maximal Euclidean space isometrically embedded in \(X\) by \(r\). Let \(\Gamma\) be a torsion-free co-compact lattice in \(G\) and let \(M\) denote the compact quotient manifold \(\Gamma \backslash X\). Let \(T^r\) be a totally geodesic embedding of a flat \(r\)-torus in \(M\). Then, for \(\xi \in T^r\), the fundamental group \(\pi(T^r,\xi)\) can be identified with a subgroup \(\Gamma_0\) of \(\pi(M,\xi)\) and is an Abelian subgroup of \(\Gamma\) which acts freely, as \(\mathbb{Z}^r\), on \(X\). Let \(\mathcal{N}\) be a von Neumann subalgebra of a type II\(_1\) von Neumann algebra \(\mathcal{M}\). Denote the (unique) trace-preserving conditional expectation onto the von Neumann subalgebra \(\mathcal{N}\) of \(\mathcal{M}\) by \(\mathbb{E}_{\mathcal{N}}\). The operator norm of a linear mapping \(T:\mathcal{N} \to \mathcal{N}\), where it is assumed that the range has the \(L^{2}\)-norm associated with the trace on \(\mathcal{M}\), is denoted by \(\|T\|_{\infty,2}\). A maximal Abelian self-adjoint subalgebra (masa) \(\mathcal{A}\) of a von Neumann algebra \(\mathcal{M}\) is called singular if every automorphism leaving \(\mathcal{M}\) invariant is implementable by unitaries of \(\mathcal{A}\). For von Neumann algebras \(\mathcal{A} \subset \mathcal{C} \subset \mathcal{M}\), \(\mathcal{A}\) and \(\mathcal{C}\) are called a strongly singular pair if for all von Neumann subalgebras \(\mathcal{B}\) with \(\mathcal{A} \subset \mathcal{B} \subset \mathcal{C}\), for every unitary \(u\) in \(\mathcal{M}\), \(|E_u|_{\mathcal {B}_u*}{ }- E_{\mathcal {B}}|_{\infty,2} \geq |u - E_{\mathcal{C}}(u)|_2\). When \(\mathcal{A} = \mathcal{C}\), this reduces to the definition of strong singularity of \textit{A.~M.\ Sinclair } and \textit{R.~R.\ Smith} [Geom.\ Funct.\ Anal.\ 12, No.~1, 199--216 (2002; Zbl 1027.46078)]. Let VN\((\Gamma)\) denote the left von Neumann algebra of \(\Gamma\) on \(\ell^{2}(\Gamma)\). Let \(\Gamma_1\), identified with \(\mathbb{Z}\), denote the subgroup of \(\Gamma_0\) generated by the class of a regular closed geodesic in \(T^r\). The author's Theorem 1.1, a.k.a. Theorem 3.1, shows that for a unitary \(u \in \mathcal{M}\) to be in VN (\(\Gamma_0)\) it is sufficient that \(u\)VN(\(\Gamma_1)u^{-1} \subset VN (\Gamma_0)\). The author uses exceedingly technical geometric proofs involving Weyl chambers for semisimple Lie groups [cf.\ \textit{A.~Borel}, ``Linear algebraic groups'' (W.~A.\ Benjamin, New York etc.) (1969; Zbl 0186.33201), Chapter IV] and geodesics in \(T^r\), though the proofs are similar to those used in \textit{G.~Robertson, A.~M.\ Sinclair } and \textit{R.~R.\ Smith} [Int.\ J.\ Math.\ 14, No.~3, 235--258 (2003; Zbl 1050.46043)]. The author shows that if \(x_0\) is in the support of a unitary \(u\), there are only a finite number of cosets \(\Gamma_0y\) with \(y \in \Gamma\) such that \(|u|_{\Gamma_0 y_2} \geq |u(x_0)|\). Calling these cosets \(\Gamma_0y_1,\dots,\Gamma_0y_n\), \(x_0\Gamma_1\) is seen to be in the union of these cosets and so \(x_0 \in \Gamma_0\); hence, using the above sufficient condition, \(u \in VN(\Gamma_0)\). A similar result is his Theorem 3.5 that VN(\(\Gamma_1) \subset VN(\Gamma_0)\) is a strongly singular pair, lessening the conditions set by Robertson et al., for a strong singularity of VN(\(\Gamma_0)\). It is not clear whether this \(\Gamma_0\) is the same as that in Theorem 3.1 or whether the author has assumed that \(\Gamma\) is ICC. \textit{L.~Pukanszki} [Can.\ J.\ Math.\ 12, 289--296 (1960; Zbl 0095.10002)] and \textit{S.~Popa} [Math.\ Scand.\ 57, 171--188 (1985; Zbl 0641.46034)] developed invariants, relative to algebraic isomorphism, for masas to show that there are infinitely many inequivalent singular masas of a type II\(_1\) von Neumann factor. These invariants are sets of numbers in \(\mathbb{N} \cup \{ +\infty \}\). In his last section, for \(X\) of constant negative curvature, \(x_0\) the class of a simple closed geodesic and \(\Gamma_0= \langle x_0\rangle\), the number of elements in \(\Gamma_0/\Gamma \backslash \Gamma_0\) is shown to be infinite so that the Pukanski invariant for VN(\(\Gamma_0)\) is \(+\infty\). The different sections of the article appear to be somewhat hurriedly pieced together and as such should not have gone past the referee unchecked.