On locally constructible spheres and balls (Q642118)

The discovery of non-polytopal and non-shellable simplicial spheres was an important step in the process of understanding the nature of simplicial spheres and simplicial balls in higher dimensions \(d\geq 3\). The number of distinct combinatorial types of simplicial spheres with or without additional properties is another important question. In this remarkable article the hierarchy is further explained in a qualitative and in a quantitative way. In Main Theorem 1 it is shown that the class of locally constructible spheres (LC-spheres) is a proper subclass of all spheres. On the other hand the shellable spheres are a proper subclass of the LC-spheres. The concept of LC-spheres was introduced by \textit{B. Durhuus} and \textit{T. Jonsson} in [``Remarks on the entropy of 3-manifolds'', Nucl. Phys., B 445, No. 1, 182--192 (1995; Zbl 0990.57500)] as spheres with \(N\) facets that are obtainable from a tree of \(N\) tetrahedra by identification of adjacent triangles only. Main Theorem 1 disproves a conjecture made in that paper. Main Theorem 2 states that the number of LC \(d\)-spheres with \(N\) facets is not larger than \(2^{d^2N}\). It has been known for longer that a 3-sphere with a knotted triangle is not shellable if the knot is at least 3-complicated. In Main Theorem 3 it is shown that such a sphere is not even LC. Main Theorem 4 gives the corresponding hierarchy for subclasses of simplicial \(d\)-balls. There are the following proper inclusions: vertex decomposable \(\to\) shellable \(\to\) constructible \(\to\) LC \(\to\) collapsible onto a \((d-2)\)-complex \(\to\) all \(d\)-balls. In addition, by Corollary 3.13 all LC 3-balls are collapsible, and by Theorem 3.23 not all collapsible 3-balls are LC. The proofs involve a clever combination of various methods from the theory of polytopes with those from knot theory and PL topology. It seems to be hopeless to sketch that in more detail here.