Operator hyperreflexivity of subspace lattices (Q607771)

Given a (complex) Hilbert space \({\mathcal H}\), a subspace lattice \({\mathcal L}\) is a collection of orthogonal projections onto subspaces of \({\mathcal H}\) such that \({\mathcal L}\) contains \(0\) and the identity projection \(I\) and is closed in the strong operator topology. For a given subspace lattice \({\mathcal L}\), we denote by \(\text{Alg}{\mathcal L}\) the algebra of all bounded linear operators \(A\) on \({\mathcal H}\) for which \(PAP=AP\) for all \(P\in{\mathcal L}\), and by \(\text{LatAlg}{\mathcal L}\) the subspace lattice containing all projections \(P\) for which \(PAP=AP\) for all \(A\in\text{Alg}{\mathcal L}\). The lattice \({\mathcal L}\) is called reflexive if \(\text{LatAlg}{\mathcal L}={\mathcal L}\). The ordinary metric distance between a given projection \(P\) and a given subspace lattice \({\mathcal L}\) is \[ d(P,{\mathcal L}) := \inf \{||P-Q||:Q\in{\mathcal L}\}. \] Motivated by Arveson's distance formula for nest algebras, define an alternate distance by \[ \beta(P,{\mathcal L}) := \sup \{||P^\perp AP||:A\in(\text{Alg}{\mathcal L})_1\}, \] where \((\text{Alg}{\mathcal L})_1\) denotes the set of all contractions in \(\text{Alg}{\mathcal L}\). It is always the case that \(\beta(P,{\mathcal L})\leq 2 d(P,{\mathcal L})\) and \({\mathcal L}\) is called hyperreflexive if there is a positive constant \(\kappa\) such that \(d(P,{\mathcal L})\leq \kappa \beta(P,{\mathcal L})\) for all projections \(P\). In [J.~Lond.\ Math.\ Soc., II.\ Ser.\ 39, No.\,2, 309--323 (1989; Zbl 0723.47003)], \textit{K.\,R.\thinspace Davidson} and \textit{K.\,J.\thinspace Harrison} showed that every hyperreflexive lattice is also reflexive, but that the converse is not true. They also showed that certain subspace lattices, including nests, commutative subspace lattices, and the projection lattices of hyperreflexive von Neumann operator algebras, are hyperreflexive, and they gave bounds on the value of \(\kappa\) in each case. In the present paper, the authors examine two analogous concepts, called operator reflexivity and operator hyperreflexivity, which they compare to each other and to the concepts of reflexivity and hyperreflexivity just discussed. Specifically, the reflexive hull of a subspace lattice \({\mathcal L}\) is defined by \(\text{Ref}{\mathcal L} :=\{P:Px\in\overline{{\mathcal L}x}\;\forall x\in {\mathcal H}\}\), and \({\mathcal L}\) is operator reflexive if \(\text{Ref}{\mathcal L}={\mathcal L}\). An alternative measure of the distance from a projection \(P\) to a subspace lattice \({\mathcal L}\) is given by \(\alpha(P,{\mathcal L}):= \sup\{d(Px,{\mathcal L}x):x\in{\mathcal H},\;||x||\leq 1\}\). It is always true that \(\alpha(P,{\mathcal L})\leq d(P,{\mathcal L})\), and \({\mathcal L}\) is operator hyperreflexive if there is a constant \(c>0\) such that \(d(P,{\mathcal L})\leq c\alpha(P,{\mathcal L})\) for all projections \(P\). Every operator hyperreflexive lattice is also operator reflexive, and the authors show here, culminating in Corollary 4.3, that there is an operator reflexive subspace lattice that is not operator hyperreflexive. Theorem 2.1 shows that every hyperreflexive subspace lattice is also operator hyperreflexive with \(c\leq 4\kappa\). (That is, the operator hyperreflexivity constant is at most 4 times the hyperreflexivity constant.) Theorem 3.2 shows that every finite subspace lattice is operator hyperreflexive, and an example is presented of a finite lattice that is not hyperreflexive. It has been shown by \textit{V.\,S.\thinspace Shulman} and \textit{I.\,G.\thinspace Todorov} [Integral Equations Oper.\ Theory 52, No.\,4, 561--579 (2005; Zbl 1081.47070)] that reflexivity implies, but is not equivalent to, operator reflexivity. Thus, the present paper completes the lattice of implications that hyperreflexivity implies both reflexivity and operator hyperreflexivity, while each of these implies operator reflexivity. All of these implications are strict.