Zero product determined some nest algebras (Q1932605): Difference between revisions

Let \({\mathcal A}\) be an algebra over a commutative unital ring \({\mathcal C}\). Denote by \({\mathcal A}^2\) the \({\mathcal C}\)-linear span of all products \(AB\), where \(A, B\in {\mathcal A}\) are arbitrary. A \({\mathcal C}\)-bilinear map \(\phi:{\mathcal A}\times {\mathcal A} \to {\mathcal V}\), where \({\mathcal V}\) is a \({\mathcal C}\)-module, is said to be zero product preserving if \(\phi(A,B)=0\) whenever \(AB=0\). Typical examples of maps satisfying the above condition are those of the form \(\phi(A,B)=T(AB)\), where \(T:{\mathcal A}^2 \to {\mathcal V}\) is a \({\mathcal C}\)-linear operator. If, for a given algebra \({\mathcal A}\), these are the only possible examples, then \({\mathcal A}\) is said to be a zero product determined algebra. In the above definition, one can replace the ordinary multiplication by a Jordan or Lie product. In the paper under review, it is studied which nest algebras are zero product determined. It is proved, for instance, that the algebra of upper triangular complex (respectively, real) \(n\)-by-\(n\) matrices has this property. The same holds for a nest algebra which is associated to a finite nest of subspaces in an infinite-dimensional complex Hilbert space. Similar results are also obtained for zero Jordan (respectively, Lie) determined nest algebras.