On inner actions of Hopf algebras and stabilizers of representations (Q1327024)

This paper begins with a very general treatment of inner actions by coalgebras. Let \(C\) be a coalgebra and \(Q\) an algebra over a field \(k\); let \(R\) be a set with an action \(C \times R \to Q\), \((c,r) \to c \cdot r\); and let \(\rho\) be a mapping of \(R\) into \(Q\) and \(Z\) be the centralizer of the image of \(\rho\) in \(Q\). A subcoalgebra \(D\) of \(C\) is called inner if there exists a convolution invertible element \(u\) of \(\text{Hom}(D,Q)\) such that \(c \cdot r = \sum u(c_ 1) \rho(r) u^{-1} (c_ 2)\) for all \(c\) in \(D\) and \(r\) in \(R\). The author shows that, when \(0 \cdot r = 0\) for all \(r\) in \(R\), there is a largest inner subcoalgebra \(C_{\text{inn}}\) in \(C\). Now let \(H\) be a Hopf algebra over \(k\), let \(R\) be a twisted left \(H\)-module algebra, and assume \(\rho\) is a homomorphism of algebras. Then \(H_{\text{inn}}\) is a Hopf subalgebra of \(H\) if either \(H\) is pointed or the coradical of \(H\) is cocommutative and \(Z\) is a semi-primary ring. Moreover, if \(H\) is cocommutative, both \(Q\) and \(R\) are twisted left \(H\)-module algebras, and \(\rho\) is a homomorphism of twisted left \(H\)-module algebras then \(H_{\text{inn}}\) is a normal Hopf subalgebra of \(H\) under the same conditions. Of particular interest is the author's construction of a Hopf algebra \(H\) and a normal Hopf subalgebra \(R\) such that, for \(Q = R\), the adjoint action of \(H\) on \(R\), and \(\rho =\) identity map, there is no largest inner Hopf subalgebra of \(H\). Next, assume \(H\) is a Hopf algebra over \(k\) with bijective antipode, let \(A\) be a right faithfully flat \(H\)-Galois extension of an algebra \(R\) over \(k\), and let \(M\) be a right \(R\)-module with \(\rho\) the associated map of \(R\) into \(Q = \text{Hom}(M,M)^{\text{op}}\). If \(A\) has a normal basis, i.e., \(A\) is cleft \(H\)-comodule algebra; then \(R\) is a twisted left \(H\)-module algebra, and a subcoalgebra \(D\) of \(H\) is inner if, and only if, \(D\) stabilizes \(M\). In this way a connection is made with the author's earlier study of induced modules and stable modules [Isr. J. Math. 72, 196-231 (1990; Zbl 0751.16015)], and some results from that paper are refined. Finally, the author generalizes and examines the concept of \(X\)-inner actions and \(X\)-outer Galois extensions.