A primary obstruction to topological embeddings and its applications (Q5929704)

Two closed topological \(m\)-manifolds \(M_i\), \(i=1,2\), are R-bordant (R comes from retraction) if there exist a compact \((m+1)\)-dimensional topological manifold \(W\) such that the boundary of \(W\) is the disjoint union of \(M_1\) and \(M_2\) and retractions \(r_i:W\to M_i\). For maps \(f_i:M_i\to N\), \(N\) a topological manifold, the authors say that \(f_1,f_2\) are R-bordant if \(M_1\) and \(M_2\) are R-bordant and there is a map \(F:W\to N\) such that \(F|M_i=f_i\). \(F\) is called an R-bordism between \(f_1\) and \(f_2\). \textit{C. Biasi} and \textit{O. Saeki} [Math. Scand. 80, No. 1, 5-24 (1997; Zbl 0888.57023)] have previously introduced a homology class \(\theta(f)\in H_{m-k} (M;\mathbb{Z}_2)\) for a map \(f:M\to N\), where \(M\) is a closed smooth \(m\)-manifold and \(N\) is a smooth \(n\)-manifold, \(k=n-m\). This homology class coincides with the fundamental class of a generic map homotopic to \(f\). They first have shown that \(\theta(f)\) is an obstruction to the existence of a homotopy between \(f\) and a topological embedding, and later its bordism invariance. In the first paper under review [Osaka J. Math. 37, No. 3, 527-535 (2000)] the present authors continue in such a direction and show that if \(f_1\) and \(f_2\) are R-bordant, then \(\theta(f_1)\) and \(\theta(f_2)\) correspond to each other under the canonical isomorphism between \(M_1\) and \(M_2\) induced by R-bordism. In the second paper under review, the authors continue the study of this obstruction using the technique of generic differentiable maps due to \textit{F. Ronga} [Compositio Math. 27, 223-232 (1973; Zbl 0271.57009)]. Besides, by this technique, there are new proofs of some previous results, also generalizations of some known results and new results. As a sample we quote part of the theorem on characterization of embeddings among generic differentiable maps: Let \(f:M\to N\) be a \(C^2\) generic map with \(m=\dim M<\dim N=n+k\), \(M\) closed and denote \(f_0=f: M\to f(M)\). Then the following statements are equivalent: (a) \(f\) is a differentiable embedding. (b) The induced homomorphism \((f_0)*:H* (M;\mathbb{Z}_2) \to\widehat H*(f(M); \mathbb{Z}_2)\) is an isomorphism.