Embeddings of homology equivalent manifolds with boundary (Q2498030)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Embeddings of homology equivalent manifolds with boundary
scientific article

    Statements

    Embeddings of homology equivalent manifolds with boundary (English)
    0 references
    4 August 2006
    0 references
    For a space \(X\), the deleted product of \(X\) is \(\widetilde{X} = X^2 - \Delta X\). Let \(X\) be of dimension \(n\), and \(f: X \to \mathbb R^m\) an embedding. Then \[ \widetilde{f}(x,y) = (f(x) -f(y))/(| f(x) -f(y)| ) \] defines a map \(\widetilde{f}: \widetilde{X} \to S^{m-1}\). Since \(\mathbb Z_2\) acts on \(\widetilde{X}\) and on \(S^{m-1}\) by exchanging factors and antipodes, respectively, \(\widetilde{f}\) is a \(\mathbb Z_2\)-equivariant map. So the existence of an embedding of \(X\) into \(\mathbb R^m\) implies the existence of a \(\mathbb Z_2\)-equivariant map \(\widetilde{X} \to S^{m-1}\). If \(X\) is an \(n\)-dimensional polyhedron \(N\), one may ask: For which pairs \((m,n)\) does it hold that for each \(n\)-polyhedron \(N\) the existence of an equivariant map \(\widetilde{X} \to S^{m-1}\) implies embeddability of \(N\) into \(\mathbb R^m\)? The answer was known for each pair \((m,n)\) but for (3,3) and (3,2). The authors cover these two cases obtaining the answer to be negative. In order to prove this, they study and use the Haefliger-Wu invariant [see e.g. \textit{A. Skopenkov}, Comment. Math. Helv. 77, 78--124 (2002; Zbl 1012.57035)], and \(\pi^{m-1}_{\text{eq}}(\widetilde{N})\)= the set of equivariant maps \(\widetilde{N} \to S^{m-1}\) up to equivariant homotopy and prove the following theorem, which is of interest in itself and is used to obtain the above mentioned puzzling result: Let \(N\) and \(M\) be compact orientable connected \(n\)-manifolds with non-empty boundaries such that \(M \subset N\) and the inclusion \(M \to N\) induces an isomorphism in cohomology with \(\mathbb Z\)-coefficients. Then the restriction-induced map \(\pi^{m-1}_{\text{eq}}(\widetilde{N}) \to \pi^{m-1}_{\text{eq}}(\widetilde{M})\) is bijective.
    0 references
    embedding
    0 references
    deleted product
    0 references
    Haefliger-Wu invariant
    0 references
    homology equivalence
    0 references
    equivariant homotopy
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references