A theorem of Sanderson on link bordisms in dimension \(4\) (Q5934073): Difference between revisions

\textit{B. J. Sanderson} [J. Lond. Math. Soc., II. Ser. 35, 367-376 (1987; Zbl 0622.57026); Colloq. Math. Soc. János Bolyai 55, 457-471 (1993; Zbl 0815.57016)] studied link bordism in codimension two. A 4-dimensional \(n\)-component link is represented by a locally flat embedding of a disjoint union \(F=K_1\cup\cdots\cup K_n\) of closed oriented surfaces \(K_1,\dots, K_n\) in \(\mathbb{R}^4\). Two such links are said to be bordant if there is a compact oriented 3-manifold \(W\subset \mathbb{R}^4\times [0,1]\) having \(n\) components and such that \(W\cap \mathbb{R}^4\times 0\) is the first link and \(W\cap \mathbb{R}^4\times 1\) is the second. B. Sanderson computed the bordism group \(L_{4,n}\) of 4-dimensional \(n\)-component links. He showed that any bordism class is determined by \(n(n-1)/2\) double linking numbers (with values in \(\mathbb{Z}_2)\) and \(n(n-1)(n-2)/3\) triple linking numbers (with values in \(\mathbb{Z})\). The paper under review suggests an explicit set of geometric representatives (generalized Hopf links) for any given value of Sanderson's invariant. The paper also provides a self-contained geometric proof of Sanderson's classification theorem.