Algebraic models of Poincaré embeddings (Q2571342)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Algebraic models of Poincaré embeddings
scientific article

    Statements

    Algebraic models of Poincaré embeddings (English)
    0 references
    0 references
    0 references
    1 November 2005
    0 references
    Suppose \(f \colon P^n \to W^m\) is a smooth embedding of smooth, compact manifolds. Then \(f(P)\) admits a regular neighbourhood \(T \subseteq W\). Set \(C = \overline{W\backslash T}\). Then \(C\) and \(T\) are both smooth, compact manifolds of dimension \(n\) with common boundary. From these ingredients, one obtains a pushout diagram \[ \begin{tikzcd} \partial T \ar[r,"-{i}"]\ar[d,"k" '] & P \ar[d,"f"]\\ C \ar[r,"-{l}" '] & W \rlap{\,.}\end{tikzcd} \] These ingredients may be adapted to provide a (homotopy-theoretic) definition of a Poincaré embedding of a finite complex \(P\) in a Poincaré duality space \(W\). Namely, suppose that \(P\) is of dimension \(n\) and \(W\) of dimension \(m\). Then a commutative diagram of the above form, in which \(\partial T\) and \(C\) are now just ordinary spaces, is called a \textit{Poincaré embedding of \(P\) in \(W\)} if it is a homotopy pushout diagram, with \((P,\partial T)\) and \((C,\partial T)\) both Poincaré duality pairs of dimension \(n\), and with the map \(i\) \((n-m-1)\)-connected. The basic question addressed in this paper is that of whether, or to what extent, the homotopy type of the map \(f\) in a Poincaré embedding determines the homotopy type of the space \(C\) (referred to as the complement of \(P\) in \(W\), in analogy with the smooth case) or, more generally, the homotopy pushout diagram, up to homotopy equivalence. In general, the answer is no, as may be seen by considering knots and their complements, for example (see specific references given in the article). In the PL setting (with the smooth set-up suitably adapted into that setting), work of Wall and of Hudson gives positive results along these lines, under suitable codimension and connectivity hypotheses on the map \(f\). The main purpose of this paper is to study the question in the rational homotopy setting. Thus, the authors consider the question of to what extent an algebraic model of the map \(f\) in a Poincaré embedding determines an algebraic model for the rational homotopy type of the space \(C\) or, more generally, the homotopy pushout diagram. Here, algebraic model means in a sense appropriate to rational homotopy theory. The authors give a number of positive results. Under various codimension and connectivity hypotheses on the map \(f\), they establish that an algebraic model for \(C\), or even for the map \(l\colon C \to W\), is determined fairly concretely from an algebraic model of \(f\). Since an algebraic model determines the rational homotopy type, they are able to conclude, for instance, the following: (Corollary 1.5) If \(f \colon P \to W\) is a Poincaré embedding of codimension at least \(3\), with \(P\) and \(W\) simply connected and \(H_*(f;\mathbb{Q})\) \((2m-n+2)\)-connected, then the rational homotopy type of \(C\) depends only on the rational homotopy type of the map \(f\). The rational models for \(C\) or for the pushout diagram constructed by the authors from that of the map \(f\) have applications to other work of the authors on blow-ups and configuration spaces, and also are connected to earlier work of the first-named author on thickenings (see the references for this earlier work).
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    Poincaré embedding
    0 references
    rational homotopy
    0 references
    algebraic model
    0 references
    0 references
    0 references