Generator coalgebras are not necessarily quasi-coFrobenius. (Q431839): Difference between revisions

If \(C\) is a left quasi-co-Frobenius (QcF for short) coalgebra then \(C\) is a generator in \(^C\mathcal M\), the category of left \(C\)-comodules. It has been asked by \textit{C. Năstăsescu, B. Torrecillas} and \textit{F. Van Oystaeyen} [Algebr. Represent. Theory 11, No. 2, 179-190 (2008; Zbl 1155.16029)] if the converse of this result holds true, that is, whether the fact that \(C\) is a generator for \(^C\mathcal M\) is actually enough to imply that \(C\) is left QcF. In the paper under review the authors prove that the converse does not hold in general, by showing that the tail closure \(T(C)\) of an admissible subcoalgebra \(C\) of a path coalgebra \(kQ\) associated to a (oriented) quiver \(Q\) is a coalgebra that generates its category of left comodules, and that \(T(C)\) is not left semiperfect if \(C\) is so. A concrete example is provided by the coalgebra \(k\mathbb A_\infty\), \(\mathbb A_\infty\) the left bounded infinity quiver \(\bullet\to\bullet\to\cdots\to\bullet\to\cdots\), which is not semiperfect, and so the same is \(T(k\mathbb A_\infty)\), although \(T(k\mathbb A_\infty)\) generates \(^{T(k\mathbb A_\infty)}\mathcal M\). Though, it is proved that the converse holds for the so called monomial pointed coalgebras, this means, for subcoalgebras of path coalgebras \(kQ\) that are pointed, admissible and admit a basis consisting of all paths in \(Q\).