Neighbourhoods in root systems of infinite Coxeter groups

Yuhan Cai, Xiang Fu, Lawrence ReevesarXiv 2018 · DOI · arXiv

TL;DR

Infinite Coxeter groups possess root systems that, when projected onto a hyperplane, form complex sets with intricate accumulation points. This paper provides a rigorous topological description of these "limit roots" by constructing a fundamental system of neighborhoods based on the algebraic dominance order of the roots.

What Problem Does This Paper Address?

Coxeter groups are mathematical abstractions of symmetry groups generated by reflections (like the symmetries of a kaleidoscope or a crystal lattice). While finite Coxeter groups are well-understood, infinite ones are far more mysterious.

A key tool for studying these groups is the root system—a collection of vectors perpendicular to the group's reflecting hyperplanes. In an infinite Coxeter group, these roots are discrete and spread out to infinity. However, if you "normalize" them—essentially projecting them onto a fixed affine hyperplane—they cluster together into a compact set. This creates a "limit set" of roots, similar to the fractal-like limit sets seen in Kleinian groups or dynamical systems.

Before this paper, researchers knew these limit sets existed and were contained within a specific geometric region (the isotropic cone), but the fine-grained topology of these sets—how points are clustered and what their neighborhoods look like—was not well-defined.

Methods

The authors employ a mix of geometric group theory and linear algebra. Their approach relies on several key pillars:

  1. Normalization: They project the infinite root system $\Phi$ onto a "transverse hyperplane" $V_1$. This turns an unbounded set of vectors into a bounded set of "normalized roots" $\widehat{\Phi}$ that can have accumulation points.
  2. Dominance Order: They utilize a partial ordering on roots called "dominance." A root $x$ dominates $y$ if, roughly speaking, any reflection that flips the sign of $x$ must also flip the sign of $y$.
  3. Dihedral Subgroup Analysis: Much of the heavy lifting is done by analyzing rank-2 (dihedral) reflection subgroups. Because these subgroups are simpler, the authors can use explicit matrix calculations (using hyperbolic functions like $\cosh$ and $\sinh$) to track exactly how sequences of roots converge to specific limit points.
  4. Neighborhood Construction: They define a sequence of sets $N_i$ based on the accumulation points of roots that dominate a specific sequence. They then use geometric arguments to prove these sets shrink down to a single limit point.

Datasets

As a theoretical mathematics paper, this work does not use empirical datasets. It relies on the internal logic and axioms of Coxeter system theory.

Key Insights and Ideas

  • Algebraic Topology Bridge: The most striking idea is using the dominance order—an algebraic property—to define a topological neighborhood. This suggests a deep link between the group's internal structure and the visual geometry of its limit set.
  • The Role of Dihedral Subgroups: The authors show that limit points arising from infinite dihedral (rank-2) subgroups are the "building blocks" of the larger limit set. Since these points are dense in the total limit set $E$, understanding their neighborhoods is the key to understanding the whole space.
  • Geometric Constraints: Through Theorem 4.8, the authors prove that their constructed neighborhoods $N_i$ are "natural." They show that the intersection of all such neighborhoods around a point $\hat{a}_\infty$ contains only that point, satisfying the requirements for a fundamental system of neighborhoods.

Key Results

  • Theorem 4.8: This is the central result. It proves that the intersection of the constructed sets $N_i$ is exactly the limit point $\hat{a}_\infty$.
  • Corollary 4.9: Formally establishes that the collection $\{N_i\}$ forms a fundamental system of neighborhoods for limit points arising from non-affine infinite dihedral reflection subgroups.
  • Convergence Proofs: The paper provides explicit limits for sequences of roots in dihedral subgroups, showing they converge to the intersection of the span of the roots and the isotropic cone $Q$ (where the bilinear form $B(v,v) = 0$).

Limitations and Future Work

The primary limitation acknowledged by the authors is that their specific neighborhood construction is demonstrated for limit points arising from non-affine infinite dihedral reflection subgroups.

While these points are known to be dense in the overall limit set $E$, the authors have not yet formally proven that the collection of all such neighborhoods forms a complete basis for the topology of $E$. They describe this final step as "work in progress," aiming to eventually provide a universal topological foundation for the limit sets of all finitely generated infinite Coxeter groups.

Summary updated

Abstract

Let $W$ be a finitely generated infinite Coxeter group, with $Φ$ and $Π$ being the corresponding root system and set of simple roots respectively. It has been observed by Hohlweg et la that the projections of elements of $Φ$ onto suitably chosen hyperplanes, called \emph{normalized roots}, are contained in the convex hull of $Π$ (which is a compact set), and hence the set of all normalized roots may exhibit interesting asymptotical behaviours. In this paper we investigate the topology of the limit set of the normalized roots and demonstrate a natural system of neighbourhoods around each limit point arising from a non-affine infinite dihedral reflection subgroup of $W$.
BibTeX
@misc{cai2018neighbourhoods,
  author = {Yuhan Cai and Xiang Fu and Lawrence Reeves},
  title = {{Neighbourhoods in root systems of infinite Coxeter groups}},
  year = {2018},
  eprint = {1811.05628},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1811.05628},
}

Loading related content...

Figure