O-minimal Hauptvermutung for polyhedra. I (Q2449842)

The author proves a version of the famous Hauptvermutung from topology, and this employs methods from the model theory of ordered structures. A \textit{polyhedron} in \(\mathbb R^d\) is the union of finitely many simplices. The Polyhedral Hauptvermutung is the conjecture that if two polyhedra are homeomorphic then there even exists a homeomorphism which is piecewise linear. This goes back to \textit{E. Steinitz} [Sitzungsber. Berl. Math. Ges. 7, 29--49 (1908; JFM 39.0545.12)] and \textit{H. Tietze} [Monatsh. Math. Phys. 19, 1--118 (1908; JFM 39.0171.01)], and it was refuted by \textit{J. W. Milnor} [Ann. Math. (2) 74, 575--590 (1961; Zbl 0102.38103)]. See the book edited by \textit{A. A. Ranicki} (ed.) et al. [The Hauptvermutung book. A collection of papers of the topology of manifolds. Dordrecht: Kluwer Academic Publishers (1996; Zbl 0853.00012)] for further details on the subject. The reason why the Polyhedral Hauptvermutung fails in general is that there are surprisingly wild ways to construct homeomorphisms among polyhedra. The theory of o-minimal models provides tools to precisely distinguish between tame maps and those which are not. This comes at the price of a somewhat elaborate language which allows to make such careful distinctions. The key notion here is \textit{definability} (with respect to the natural language of ordered fields). For instance, the definable subsets of \(\mathbb R\) are precisely the finite unions of open, closed or half-open intervals (which may be unbounded). The main result of the paper under review (Theorem 1.1) is a proof of the Polyhedral Hauptvermutung, over an arbitrary ordered field with a fixed o-minimal structure, for definably homeomorphic polyhedra which are bounded and closed. Notice that boundedness and closedness is not enough to force compactness in the topological sense over a general ordered field. Beware, however, that it is common practice in this area to call such bounded and closed sets \textit{compact} nonetheless. The proof is rather long and technical. Yet this cannot be avoided since definability is such a subtle notion. Along the way the author establishes a number of results which are of independent interest. For instance, Lemma~3.1 is a version of the Simplicial Approximation Theorem in this setting. The main result is very fundamental. It explains quite well to what extent the intuition of the founding fathers of topology was correct when they came up with their conjecture.