Some mildly wild circles in \(S^ n\) arising from algebraic K-theory (Q810931)

Among wild circles in \(S^ n\) there are some which are ``less wild'', called mildly wild here, and those agree with tame embeddings in a sense introduced by \textit{F. Quinn} [Ann. Math., II. Ser. 110, 275-331 (1979; Zbl 0394.57022)], which are also called end-tame embeddings. In the paper this means the following: Let X and Y be compact ANR's and let r: \(N\to Y\) be a retraction from a closed neighborhood N of Y in X to Y. Y is end-tame in X if for every neighborhood U of Y and \(\epsilon >0\) there exists a neighborhood V of the end and a homotopy h: (X-Y)\(\times I\to X-Y\) such that: (1) \(h=id\) on (X-Y)\(\times \{0\}\cap (X-U)\times I\); (2) h takes (U-Y)\(\times I\) into U-Y; (3) h((X-Y)\(\times \{1\})\subset X-V\); (4) The tracks of \(r\circ h\) have diameter less than \(\epsilon\). If i: \(X\to Y\) is an embedding, then i is end-tame (or mildly wild) if i(Y) is end-tame in X. For Y being end-tame in X, the complement X-Y is an ANR which is homotopy dominated by a finite simplicial complex. So if one thinks of the homotopy type of X-Y to be that of a finite simplicial complex, then the Wall obstruction (finiteness obstruction) \(\sigma =\sigma (X-Y)\in \tilde K_ 0(\DJ\pi_ 1(X-Y))\) comes into play. The authors consider \(Y=S^ 1\), \(X=S^ n\) and end-tame embeddings i: \(S^ 1\to S^ n\), so that under appropriate conditions on \(\pi_ 1\) they prove that all corresponding finiteness obstructions can be realized on \(S^ n-i(S^ 1)\). Here is the main theorem on realization of the finiteness obstruction: for each \(n\geq 7\) and each finitely presented perfect group \(\pi\) with \(H_ 1(\pi)=H_ 2(\pi)=0\), and \(\sigma\in \tilde K_ 0({\mathbb{Z}}\pi)\), there is an end-tame embedding i: \(S^ 1\to S^ n\) such that \(\pi_ 1(S^ n-i(S^ 1))=\pi\), and such that \(S^ n-i(S^ 1)\) is finitely dominated with \(\sigma (S^ n-i(S^ 1))=\sigma.\) Beside this result there are some further interesting and useful results on realization of finiteness obstruction like: If \(\pi\) is as above and \(n\geq 5\), then there exists a finitely dominated integral homology n- sphere (i.e. a complex which has the integral homology of an n- dimensional sphere) with \(\pi_ 1(X)=\pi\) and \(\sigma (X)=\sigma\), or for \(n\geq 7\) there exist embeddings i: \(S^ 1\to S^ n\) such that \(S^ n-i(S^ 1)\) is not a Poincaré duality space over \({\mathbb{Z}}\pi_ 1(S^ n-i(S^ 1))\), even though \(\pi_ 1(S^ n-i(S^ 1))\) is finite and \(S^ n-i(S^ 1)\) is finitely dominated. The paper also brings a theorem on realization of controlled end obstruction. So part of Theorem 4 of the paper says: Under the same assumptions on \(\pi\) and \(n\geq 7\) it holds: For each \(\tau\in \tilde K_{-1}({\mathbb{Z}}\pi)\) there is an embedding i: \(S^ 1\to S^ n\) such that \(\pi_ 1(S^ n-i(S^ 1))=\pi\), \(S^ n-i(S^ 1)\) has the homotopy type of a finite complex, and such that the controlled end obstruction of \(S^ n-i(S^ 1)\) over \(S^ 1\) is \(\tau +(-1)^ n\tau^*\).