BRIAN RUSHTON
5 min
This paper investigates the properties of combinatorial modulus in subdivision rules where the valence of vertices is unbounded. Specifically, it examines how different rates of vertex growth—linear versus exponential—affect the conformality of these rules, building upon the foundational work of Cannon, Floyd, and Parry regarding finite subdivision rules.
To analyze these subdivision rules, the author employs the concept of combinatorial modulus, which serves as a discrete analog to the modulus of topological annuli in complex analysis. The study utilizes barycentric subdivision and a subdivision rule derived from Borromean rings as primary examples. The author evaluates whether these rules satisfy the two axioms of conformality: Axiom 1 (equicontinuity of approximate moduli) and Axiom 2 (non-degeneracy of points). The analysis is supported by circle packing simulations generated using Ken Stephenson’s Circlepack software.
The author finds that the 1,2,3-tile criterion, originally developed for bounded-valence rules, is sufficient to prove conformality for rules with linear growth at every vertex. In contrast, for rules with exponential growth, such as barycentric subdivision, the criterion is insufficient to satisfy both axioms of conformality. In these exponential cases, the modulus shrinks too rapidly to allow for the construction of annuli with unbounded modulus, leading to a weaker form of conformality. The author also demonstrates that the subdivision rule associated with the Borromean rings is conformal and conjectures that this property extends to all alternating links.
Understanding the conformality of subdivision rules is essential for the broader program of proving that hyperbolic groups with a 2-sphere at infinity are hyperbolic 3-manifold groups. By extending the study of modulus to unbounded valence rules, this research provides deeper insight into the geometric behavior of these systems and their relationship to rational maps and circle packings, offering a framework for analyzing more complex topological structures.
Cannon, Floyd and Parry have studied the modulus of finite subdivision rules extensively. We investigate the properties of the modulus of subdivision rules with linear and exponential growth at every vertex, using barycentric subdivision and a subdivision rule for the Borromean rings as examples. We show that the subdivision rule arising from the Borromean rings is conformal, and conjecture that the subdivision rules for all alternating links are conformal. We show that the 1,2,3-tile criterion of Cannon, Floyd, and Parry is sufficient to prove conformality for linear growth, but not exponential growth. We show that the criterion gives a weaker form of conformality for subdivision rules of exponential growth at each vertex. We contrast this with the known, bounded-valence case, and illustrate our results with circle packings using Ken Stephenson's Circlepack.
Sam: And that's enough to recover a usable modulus, even if not full conformality?
Alex: It recovers symmetry in the modulus definition, which is the prerequisite for the broader classification program. Whether it recovers full conformality depends on the specific rule.
Sam: Which brings in the Borromean rings case?
Alex: Yes, and that's the most instructive example in the paper. The Borromean rings subdivision rule has unbounded valence — so by the naive argument it should fail — but Rushton shows it remains conformal. That's not just a curiosity; it's evidence that unbounded valence doesn't automatically preclude conformality. The question is which structural properties of the rule determine which side of the line you're on.
Sam: So the paper is really mapping the boundary between rules that survive the exponential growth and rules that don't.
Alex: Precisely. And the main technical tool for doing that is the Layering Theorem — the result that lets you stack disjoint annuli to get a lower bound on the modulus of a larger region. That's the supporting scaffolding. It doesn't establish conformality on its own, but it guarantees the modulus doesn't collapse entirely, which is what you need before you can say anything useful about the limit space.
Sam: Where does the argument have genuine gaps?
Alex: Two places. First, the upper bounds. The paper establishes lower bounds on modulus in the exponential growth regime, but the corresponding upper bounds — which you'd need for a complete characterization — aren't proved in general. Second, the treatment of alternating links relies on conjectures that haven't been resolved. So the classification of subdivision rules associated with hyperbolic 3-manifold groups is conditional on those conjectures holding. A careful referee would flag both of those as constraints on how far the conclusions generalize.
Sam: What would closing those gaps actually enable?
Alex: A complete classification of subdivision rules for all hyperbolic 3-manifold groups — which is the long-term target of this research program. Right now you can handle the bounded-valence cases cleanly, and you have partial results for the unbounded ones. Filling in the upper bounds and resolving the link conjectures would let you extend the classification to the full unbounded regime, which is where most of the geometrically interesting cases live.
Sam: So the paper is a meaningful advance on a hard problem, with clearly marked open questions.
Alex: That's a fair read. It moves the conformality criterion into territory where it previously had no traction, identifies exactly where it still falls short, and gives a constructive mechanism — the vertex-weighting — that future work can build on. The open questions aren't weaknesses so much as an honest account of what remains.
Sam: A rigorous piece of work with a well-defined frontier.
Alex: Exactly the kind of result that makes the next paper possible. Thanks for listening to ResearchPod.