Local limits of uniform triangulations with boundaries in high genus

Imagine you are a tiny ant living on a giant, crumpled piece of paper. This paper isn't just flat; it's a complex, multi-layered surface with holes, twists, and loops, like a pretzel made of paper. In mathematics, this is called a triangulation of high genus.

For a long time, mathematicians knew what happened if you looked at a flat piece of paper (a sphere) as it got infinitely big: the ant would see a specific, infinite, random pattern called the UIPT (Uniform Infinite Planar Triangulation). It's like the ant is standing on an endless, flat, snowy plain.

But what if the paper is a giant pretzel with thousands of holes (high genus)? And what if the ant is standing right on the edge of a hole (a boundary)?

This paper by Tanguy Lions answers that question. Here is the story of the discovery, explained simply.

1. The Setting: The Giant Pretzel

Imagine you have a massive sheet of paper made of tiny triangles.

The Size ( $n$ ): The paper is huge, with millions of triangles.
The Holes ( $g$ ): The paper is twisted into a shape with many holes (high genus). The number of holes is proportional to the size of the paper.
The Edge ( $p$ ): The paper has a long, winding edge (a boundary).

The author asks: If I stand on this edge and look around, what does the world look like?

2. The Two Perspectives

The paper explores two different ways the ant (the "root") can be placed on this giant pretzel.

Perspective A: Standing in the Middle

If the ant is standing randomly somewhere in the middle of the paper (far away from the edges), the paper looks like a Planar Stochastic Hyperbolic Triangulation (PSHT).

The Analogy: Imagine standing in the middle of a vast, hyperbolic forest. The trees (triangles) grow so fast that the space feels "squeezed." It looks flat locally, but the geometry is weird and curved. This is the "standard" view for the middle of these high-genus maps.

Perspective B: Standing on the Edge (The Big Discovery)

This is the main breakthrough of the paper. If the ant is standing on the edge of the paper, looking out, the view is different.

The Result: The local limit is a Half-Plane Hyperbolic Triangulation.
The Analogy: Imagine you are standing on the edge of a cliff. You look out, and instead of seeing the whole forest, you see an infinite, flat wall of triangles stretching out to your left and right, with a "void" (the edge) in front of you.
Why it matters: Before this paper, we knew these "Half-Plane" shapes existed mathematically, but we didn't know how to build them from real, large, physical objects. This paper proves that if you take a giant, hole-filled paper and stand on its edge, you naturally "see" this Half-Plane world. It's like discovering that the edge of a giant, crumpled map naturally unfolds into a specific, infinite shape.

3. The "Peeling" Adventure

How did the author prove this? He used a technique called Peeling.

The Metaphor: Imagine the giant paper is a cake. You start at your edge and eat one triangle at a time. As you eat, you reveal what's underneath.
The Challenge: On a normal flat cake, eating a piece is simple. But on a giant, twisted pretzel with holes, eating a piece might:
1. Reveal a new hole.
2. Connect two distant parts of the edge (folding the paper onto itself).
3. Connect to a completely different hole.

The author had to prove that if the paper is huge and the edge is long, weird folding events are extremely rare.

The "Folding" Problem: Imagine the edge of the paper folding back and touching itself nearby. If this happened often, the "Half-Plane" view would be broken.
The Solution: The author used a clever counting trick. He showed that if the edge did fold back on itself often, the math wouldn't add up. The number of ways to arrange the paper would be too small. Therefore, the edge must stay "straight" and open, creating that clean Half-Plane view.

4. The "Surgery" Trick

To prove that the edge doesn't fold back, the author used a method called Surgery.

The Metaphor: Imagine you have a tangled ball of yarn (the paper). You want to prove it's not a specific kind of knot. So, you take a pair of scissors, cut a specific part of the yarn, and rearrange it.
The Logic: He showed that if the paper did have a "bad" folding pattern, you could cut and rearrange it to create more valid papers than actually exist. Since you can't create more valid papers than exist, the "bad" folding pattern must be impossible. This is a classic "proof by contradiction" using combinatorial magic.

5. Why Should You Care?

This isn't just about paper maps.

Physics: These shapes model the fabric of space-time in quantum gravity. Understanding how the "edge" of the universe behaves is crucial.
Networks: It helps us understand how large, complex networks (like the internet or social networks) behave when they have a lot of "loops" or cycles.
The "First Time" Achievement: This is the first time mathematicians have successfully built these "Half-Plane" shapes out of large, real-world-like objects. It connects the abstract math of infinite shapes with the concrete math of huge, finite objects.

Summary

Think of the paper as a giant, crumpled, hole-filled blanket.

If you stand in the middle, you see a hyperbolic forest.
If you stand on the edge, the blanket naturally unfolds into an infinite, flat wall (the Half-Plane).
The author proved this by showing that the blanket is too big to fold back on itself in a way that would ruin the view.

It's a beautiful piece of work that connects the geometry of the very small (local limits) with the geometry of the very large (high genus), revealing that even in a chaotic, hole-filled world, the edges have a very orderly, predictable nature.

Here is a detailed technical summary of the paper "Local limits of uniform triangulations with boundaries in high genus" by Tanguy Lions.

1. Problem Statement and Context

The Regime:
The paper investigates the asymptotic behavior of uniform random triangulations of high genus. Specifically, it considers a sequence of triangulations with:

Genus ( $g_n$ ): Proportional to the number of faces ( $n$ ), such that $g_n/n \to \theta \in [0, 1/2)$ .
Boundaries: The triangulations possess $\ell_n$ boundaries with perimeters $p_n = (p_n^1, \dots, p_n^{\ell_n})$ .
Scaling: The total boundary length satisfies $|p_n| = o(n)$ , and the average boundary length tends to infinity ( $|p_n|/\ell_n \to \infty$ ).

The Gap:
Previous work by Budzinski and Louf [18] established that uniform random triangulations without boundaries in this high-genus regime converge locally to the Planar Stochastic Hyperbolic Triangulation (PSHT), denoted $T_\lambda$ . However, the behavior of these maps when rooted on a boundary edge was an open problem. Intuitively, a boundary in a hyperbolic surface should look like the edge of a half-plane. The paper aims to prove that the local limit around a typical boundary edge is the Half-Plane Hyperbolic Triangulation ( $H_\lambda$ ), previously defined by Angel and Ray [5].

2. Methodology

The proof relies on coarse combinatorial estimates and probabilistic arguments rather than heavy algebraic machinery (like the Goulden-Jackson recurrence used in [18]). The methodology is divided into two main parts corresponding to the two main theorems.

A. Rooting in the Bulk (Interior)

To prove that rooting a random edge in the bulk yields the PSHT ( $T_\lambda$ ), the author adapts the strategy of [18] but introduces a crucial new argument to establish planarity of the limit without relying on specific recurrence relations.

Tightness: Using the fact that boundaries are small ( $o(n)$ ), it is shown that a random edge is typically far from any boundary.
Planarity via "Copy" Argument: To prove the limit is planar (i.e., contains no genus), the author assumes the contrary (that a genus-1 subgraph appears with positive probability). By the exchangeability of edges, this implies the existence of many disjoint copies of this subgraph. The author then performs a combinatorial counting argument: the number of ways to embed $\eta n$ disjoint copies of a genus-1 map is exponentially suppressed compared to the reduction in the total number of triangulations caused by removing the genus. This forces the probability of non-planarity to zero.

B. Rooting on the Boundary

Proving the limit is the Half-Plane Triangulation ( $H_\lambda$ ) is significantly more technical. The main challenge is ruling out "pathological" geometric configurations where the local neighborhood of the boundary edge does not resemble a half-plane.

Peeling Exploration: The author uses a peeling process (revealing triangles one by one) to explore the neighborhood of the root edge $e_n$ .
Excluding Pathological Cases: The proof rigorously excludes four specific pathological scenarios:
1. The third vertex of the revealed triangle lies on the same boundary but far away (folding).
2. The revealed triangle creates a hole that is filled by a non-planar or large component.
3. The boundary folds onto itself (creating a large hole).
4. The revealed triangle connects to a different boundary component.
Combinatorial Surgery: To rule out the folding and self-intersection cases, the author employs a surgery argument on the peeling diagrams. By constructing an injective map between triangulations that satisfy these pathological conditions and a larger set of triangulations, they show that the probability of these events decays to zero as $n \to \infty$ .
Identification of the Limit: Once tightness and the "half-plane" structure are established, the limit is identified as a mixture of subcritical and supercritical half-plane triangulations. Using the known asymptotic ratio of triangulation counts (derived from the bulk limit), the author proves the limit must be the specific supercritical model $H_{\lambda(\theta)}$ .

3. Key Contributions

First Construction of $H_\lambda$ as a Limit: This is the first rigorous construction of the Half-Plane Hyperbolic Triangulations ( $H_\lambda$ ) as local limits of large random maps. Previously, they were defined axiomatically via spatial Markov properties.
Robust Proof Technique: The proof of the bulk limit (Theorem 1.2) avoids the Goulden-Jackson recurrence relation, relying instead on coarse combinatorial estimates. This suggests the result is robust and applicable to other models of large genus maps (e.g., with decorations like the Ising model).
Resolution of Open Conjectures: The paper confirms conjectures from [5] and [23] that hyperbolic half-plane triangulations arise naturally as limits of large genus maps with boundaries.
Combinatorial Estimates: The paper provides new quantitative estimates on the number of triangulations with boundaries, specifically showing that $\tau_{p}(n, g-1) / \tau_p(n, g) \to 0$ and deriving asymptotic ratios for adding/removing boundaries.

4. Main Results

Theorem 1.1 (Boundary Rooting):
Let $T_{n, g_n, p_n}$ be a uniform random triangulation of genus $g_n$ with boundaries $p_n$ . Let $e_n$ be a uniformly chosen oriented edge on the union of the boundaries. If $g_n/n \to \theta$ and $|p_n| = o(n)$ with $|p_n|/\ell_n \to \infty$ , then:
$(T_{n, g_n, p_n}, e_n) \xrightarrow{d} H_{\lambda(\theta)}$
where $H_{\lambda(\theta)}$ is the supercritical half-plane hyperbolic triangulation with parameter $\lambda(\theta)$ determined by $d(\lambda) = (1-2\theta)/6$ .

Theorem 1.2 (Bulk Rooting):
Let $e_n$ be a uniformly chosen oriented edge in the bulk of $T_{n, g_n, p_n}$ . Under the same conditions:
$(T_{n, g_n, p_n}, e_n) \xrightarrow{d} T_{\lambda(\theta)}$
where $T_{\lambda(\theta)}$ is the Planar Stochastic Hyperbolic Triangulation (PSHT).
Note: This generalizes the result of [18] to the case with boundaries.

Corollary (Boundary with fixed perimeter):
If one fixes a specific boundary of perimeter $p$ and roots on it, the limit is $T^{(p)}_{\lambda(\theta)}$ , the PSHT with a boundary of perimeter $p$ .

5. Significance and Future Directions

Geometric Understanding: The results provide a deeper understanding of the geometry of random surfaces in the high-genus regime. They confirm that while the global topology is hyperbolic (large genus), the local geometry around a "typical" point is the PSHT, and the local geometry around a "typical" boundary is the Half-Plane Hyperbolic Triangulation.
Universality: The proof strategy's independence from specific recurrence relations suggests a universality class for high-genus random maps.
Future Work: The author notes that the regime where the boundary length is proportional to $n$ ( $|p_n| \sim \alpha n$ ) remains open, with the conjecture that the limit would be a triangulation with infinitely many infinite faces. Additionally, the techniques are expected to be applied to study typical distances and the peeling process in these models, potentially refining results on the logarithmic diameter of high-genus maps.

In summary, this paper bridges the gap between the combinatorial enumeration of high-genus maps and their probabilistic geometric limits, successfully identifying the local structure of boundaries in this regime as the hyperbolic half-plane.