Embeddings of finite-dimensional compacta in Euclidean spaces (Q409661)

Applying the second and third authors' result [Sb. Math. 196, No. 11, 1585--1603 (2005); translation from Mat. Sb. 196, No. 11, 33--52 (2005; Zbl 1141.54014), Theorem 1.5], it is obtained that for a metrizable compactum \(X\) with \(\dim X\leq n\), the set \[ \{g\in C(X,{\mathbb R}^{2n+1}):g^{-1}(\Pi^1) \text{ contains at most 4 points for any line } \Pi^1\subseteq{\mathbb R}^{2n+1}\} \] is a dense \(G_{\delta}\)-subset of the space \(C(X,{\mathbb R}^{2n+1})\) with the uniform convergence topology. In this paper, the above result is complemented as follows: Let \(M_{m,d}\) be the metric space consisting of all affine \(d\)-dimensional subspaces \(\Pi^d\) of \({\mathbb R}^m\), and for a metric space \(X\) and non-negative integers \(q\) and \(k\) let \[ {\mathcal H}(q,d,m,k)=\{g\in C(X,{\mathbb R}^m): \dim\{\Pi^d\in M_{m,d}:|g^{-1}(\Pi^d)|\geq q\}\leq k\}. \] Then, for a metrizable compactum \(X\) of dimension \(\leq n\), (a) the set \({\mathcal H}(3,1,m,3n+1-m)\) is dense \(G_{\delta}\) in \(C(X,{\mathbb R}^m)\) provided \(m\geq 2n+1\); (b) the set \({\mathcal H}(2,1,m,2n)\) is dense \(G_{\delta}\) in \(C(X,{\mathbb R}^m)\) provided \(m\geq2n+1\); (c) the set \({\mathcal H}(1,d,m,n+d(m-d))\) is dense \(G_{\delta}\) in \(C(X,{\mathbb R}^m)\) provided \(m\geq n+d\); (d) the set \({\mathcal H}(4,1,3,0)\) is dense \(G_{\delta}\) in \(C(X,{\mathbb R}^3)\) if \(n=1\). Furthermore, the authors show a parametric version of the above result.