Combinatorial Lefschetz theorems beyond positivity

TL;DR

This paper extends the Lefschetz theorems beyond positivity constraints, applying them to combinatorial and topological problems, leading to solutions of longstanding conjectures in face enumeration and embeddings of simplicial complexes.

Contribution

It introduces a generalized Kähler package that proves the hard Lefschetz theorem without positivity, resolving multiple conjectures in combinatorics and topology.

Findings

Resolved the g-conjecture for simplicial rational homology spheres.

Verified K"uhnel's conjecture on triangulated manifolds.

Generalized the crossing lemma for simplicial complexes in Euclidean space.

Abstract

Consider a simplicial complex that allows for an embedding into $\mathbb{R}^d$. How many faces of dimension $\frac{d}{2}$ or higher can it have? How dense can they be? This basic question goes back to Descartes' "Lost Theorem" and Euler's work on polyhedra. Using it and other fundamental combinatorial problems, we introduce a version of the K\"ahler package beyond positivity, allowing us to prove the hard Lefschetz theorem for toric varieties (and beyond) even when the ample cone is empty. A particular focus lies on replacing the Hodge-Riemann relations by a non-degeneracy relation at torus-invariant subspaces, allowing us to state and prove a generalization of theorems of Hall and Laman in the setting of toric varieties and, more generally, the face rings of Hochster, Reisner and Stanley. This has several applications: - We fully characterize the possible face numbers of simplicial rational homology spheres, resolving the $g$-conjecture of McMullen in full generality and generalizing Stanley's earlier proof for simplicial polytopes. The same methods also verify a conjecture of K\"uhnel: if $M$ is a triangulated closed $(d-1)$-manifold on $n$ vertices, then \[\binom{d+1}{j}\mathrm{b}_{j-1}(M)\ \le \ \binom{n-d+j-2}{j}\ \quad \text{for}\ 1\le j\le \frac{d}{2}.\] - We prove that for a simplicial complex that embeds into $\mathbb{R}^{2d}$, the number of $d$-dimensional simplices exceeds the number of $(d-1)$-dimensional simplices by a factor of at most $d+2$. This generalizes a result going back to Descartes and Euler, and resolves the Gr\"unbaum-Kalai-Sarkaria conjecture. We obtain from this a generalization of the celebrated crossing lemma: For a map of a simplicial complex $\Delta$ into $\mathbb{R}^{2d}$, the number of pairwise intersections of $d$-simplices is at least \[\frac{f_d^{d+2}(\Delta)}{(d+3)^{d+2}f_{d-1}^{d+1}(\Delta)}\] provided $f_d(\Delta)> (d+3)f_{d-1}(\Delta)$.