Paratransitive algebras of linear operators. II (Q2435462)

The authors continue the analysis of Part I [Linear Algebra Appl. 439, No. 7, 1955--1973 (2013; Zbl 1285.47010)] for algebras \(A\) of linear transformations \(T\) on a finite-dimensional, complex vector space \(V\), that are paratransitive in the sense that, if for any subspaces \(W_1\) and \(W_2\) of \(V\) with dimensions \(1 \leq \dim W_1 =k\), \(\dim W_2= m \leq \dim V\) for some \(k, m\) fixed, the intersection \(A W_1 \cap W_2\) with the image \(A W_1\) under \(A\) is nonzero. They study the algebraic or geometric structure of a paratransitive algebra \(A\) under the assumption that the image \(A W_1\) has codimension \(m-1\) for all subspaces \(W_1\) of \(V\) with \(\dim W_1 =k\), in which case the algebra \(A\) is defined to be exactly \((k, m)\)-transitive. Used as a key ingredient is a spatial version of Wedderburn's Principal Theorem, obtained in [the authors, ``A spatial version of Wedderburn's theorem and applications'', Preprint, 2012], that states that every algebra \(A\) of linear transformations on a complex vector space has an unhinged, block-wise, upper-triangular form with respect to a direct sum decomposition of the space, in the sense as that \(A\) has the decomposition into its semi-simple part of semi-simple block-wise diagonals and its radical of block-wise strictly upper-triangular forms.