A characterization of operators preserving disjointness in terms of their inverse (Q1588011)

A disjointness preserving operator between vector lattices is one where the image of disjoint elements are disjoint. In general, disjointness preserving for an operator \(T\) does not imply disjointness preserving for \(T^{-1}\). The characterization utilizes a kind of monotonicity property. Specifically, \(T\) satisfies condition \((\beta)\) if whenever \(a\) is in the band generated by \(b\) (\(a\) and \(b\) in the domain), then \(Ta\) is the band generated by \(Tb\). This generalizes the condition introduced by Randrianantoanina for spaces of measurable functions; here the statement ``\(x\) is in the band generated by \(y\)'' replaces ``the support of \(x\) is contained in the support of \(y\)''. The authors establish the following characterization: for a class of spaces including those that are Dedekind complete, a bijective operator \(T\) is disjointness preserving if and only if \(T^{-1}\) satisfies the \((\beta)\) condition. Several interesting examples demonstrate that condition \((\beta)\) and disjointness preserving are independent. However, the authors show that for Banach lattices (and other cases) the two conditions are equivalent. Related results and a condition similar to \((\beta)\) are studied.