On H-separable extensions of two sided simple rings. II (Q1821849): Difference between revisions

Let A be a ring with identity 1 and B be a subring of A with \(1\in B\). Then A is called an H-separable extension of B if \({}_ AA\otimes_ BA_ A\) is a direct summand of a finite direct sum of copies of A as a two-sided A-module. Further, let S be a subring between A and B. If the natural mapping of \(S\otimes_ BA\) to A splits, then S is called a left relatively separable extension of B in A. For a subset X of A put \(V_ A(X)=\{a\in A\); \(ax=xa\) for all \(x\in X\}\) (the centralizer of X in A). Now, consider the case where A is a two-sided simple ring and B is a two- sided simple subring of A, \(C:=V_ A(A)\). Denote by \({\mathbb{L}}\) the class of simple C-subalgebras of \(V_ A(B)\), and by \({\mathbb{G}}_{\ell}\) the class of two-sided subrings of A which are left relatively separable extensions of B in A. Then Theorem 2 of the author's earlier paper [Part I, ibid. 11, 246-252 (1982; Zbl 0496.16007)] shows that one can obtain mutually inverse one to one correspondences between \({\mathbb{L}}\) and \({\mathbb{G}}_{\ell}\) by letting each member of \({\mathbb{L}}\) or \({\mathbb{G}}_{\ell}\) correspond with its centralizer in A. In the paper the author improves this result by proving that a two-sided simple subring S of A containing B belongs to \({\mathbb{G}}_{\ell}\) if and only if A is a finitely generated left S-module.