No label defined (Q4934055)

Let \(K(n)\) be the quiver with two vertices \(a,b\) and \(n\) arrows \(a\to b\). It is known that for every algebraically closed field \(k\) the category \(\text{rep}_kK(n)\) of finite dimensional \(k\)-representations of \(K(n)\) is of finite, tame and wild representation type if \(n=1\), \(n=2\) and \(n\geq 3\), respectively. In particular there exist full exact embeddings of \(\text{rep}_kK(n)\) into \(\text{rep}_kK(m)\) for \(m\geq 3\) and \(n\) arbitrary whereas there is no such embedding of \(\text{rep}_kK(3)\) into \(\text{rep}_kK(2)\). The situation changes if we allow infinite dimensional representations. Let us denote by \(\text{Rep}_kQ\) the category of all \(k\)-representations of a given quiver \(Q\). An explicit construction of a full exact embedding of \(\text{Rep}_kK(3)\) into \(\text{Rep}_kK(2)\) is given in the paper. It can be repeated for any base field \(k\). Therefore the tame (with respect to finite dimensional representations) quiver \(K(2)\) is Wild -- here the capital letter refers to arbitrary, not necessarily finite dimensional representations. The fact that tame hereditary algebras are Wild in the above sense was already established by \textit{C. M. Ringel} [in Symp. Math. 23, 321-412 (1979; Zbl 0429.16022)]. Now the author presents a short and direct proof.