Finiteness properties and quasi-isometry of group pairs

This paper establishes that both geometric and homological finiteness properties of group pairs remain invariant under a suitably defined notion of quasi-isometry for group pairs.

The Gist

Imagine you are an architect trying to describe the "shape" and "complexity" of a city. In mathematics, groups are like these cities, and mathematicians have developed a way to measure how "finite" or "manageable" a city is. Is it built with a finite number of blocks? Can you navigate it without getting lost in an infinite maze?

This paper, "Finiteness Properties and Quasi-Isometry of Group Pairs," by Kevin Li and Luis Jorge Sánchez Saldaña, tackles a specific, tricky version of this problem. It asks: If two cities look roughly the same from a distance, do they share the same fundamental rules of construction?

Here is the breakdown in simple terms, using some creative metaphors.

1. The Cast of Characters

• The Group ($G$): Think of this as a City. It has streets (generators) and intersections.
• The Subgroups ($P$): These are like Special Districts within the city (e.g., the Financial District, the University Campus).
• The Group Pair ($G, P$): This is the City plus its Special Districts. You can't just look at the city; you have to look at how the city interacts with these specific districts.
• Finiteness Properties ($F_n$ and $FP_n$): These are the Building Codes.
  • Type $F_n$: Can you build a model of the city using only a finite number of bricks up to a certain height? (Geometric finiteness).
  • Type $FP_n$: Can you describe the city's rules using a finite list of instructions? (Algebraic finiteness).
• Quasi-Isometry: This is a Blurred Photo. If you take a photo of two cities from a very far away, blurry helicopter, and they look identical (same number of districts, same general layout), they are "quasi-isometric." The paper asks: If the blurry photos look the same, are the actual building codes the same?

2. The Problem: The "Coned-Off" City

In the past, mathematicians knew that if two simple cities (groups without special districts) looked the same from a distance, they shared the same building codes.

But when you add Special Districts (subgroups), things get messy.

• The Old Way: To study these cities, mathematicians used a tool called the "Cayley Graph" (a map of the city).
• The New Way: For group pairs, they use a "Coned-Off" Map. Imagine taking every Special District and collapsing it into a single giant "Cone Vertex" (a lighthouse). You connect every street in that district to the lighthouse.
• The Problem: This new map is weird. It has infinite loops, and the "lighthouses" make the map non-finite in ways that break the old rules. You can't just apply the old "blurred photo" logic because the cones mess up the geometry.

3. The Solution: The "Unicone" Trick

The authors invented a clever workaround called the "Unicone Rips Complex."

• The Metaphor: Imagine you are exploring the city, but you are only allowed to carry one lighthouse in your backpack at a time.
• The Rule: You can look at a neighborhood of streets, but if you see a second lighthouse (a second Special District) in your view, you ignore it or treat it as "too far away." You only build your map based on clusters that contain at most one lighthouse.
• Why it works: By restricting the view to "one lighthouse at a time," the map becomes manageable again. It behaves nicely, allowing the mathematicians to apply their tools.

4. The Main Discovery

The paper proves two massive things using this "One Lighthouse" rule:

1. The "Quasi-Retract" Rule: If City A is a "quasi-retract" of City B (meaning City A is a simplified, distorted version of City B that you can map back and forth without losing too much information), and City B has good building codes (is "finite" in the right way), then City A must also have good building codes.
2. The "Blurred Photo" Rule: If two Group Pairs (City + Districts) look the same in a blurry photo (are "strongly quasi-isometric"), then they share the exact same finiteness properties.

5. Why This Matters

This is a big deal for several reasons:

• It Unifies Math: It connects the geometry of shapes (topology) with the algebra of rules (group theory) in a relative setting.
• It Solves a Mystery: For a long time, people wondered if these properties held true when you added "Special Districts." This paper says, "Yes, they do, provided you look at them the right way."
• It Helps with "Relatively Hyperbolic" Groups: These are groups that act like hyperbolic spaces (like a saddle shape) but have some "flat" or "weird" districts. This theory helps mathematicians understand the structure of these complex shapes, which appear in physics, computer science, and geometry.

Summary Analogy

Imagine you have two massive, complex theme parks (Group Pairs).

• Park A has a few special VIP zones (Subgroups).
• Park B has a few special VIP zones.

You can't walk through the whole park to check the rules. Instead, you take a drone photo from high up.

• If the drone photos look identical (same number of VIP zones, same general layout), the authors prove that Park A and Park B are built on the same fundamental rules.
• Even if Park A is a "distorted" version of Park B (a quasi-retract), as long as the distortion isn't too wild, the rules of construction remain the same.

The authors' secret weapon was realizing that if you focus on the park one VIP zone at a time (the "Unicone" trick), the distortion doesn't matter, and the math works out perfectly.

Technical Summary

Here is a detailed technical summary of the paper "Finiteness Properties and Quasi-Isometry of Group Pairs" by Kevin Li and Luis Jorge Sánchez Saldaña.

1. Problem Statement and Context

The paper addresses the invariance of finiteness properties for group pairs under the geometric equivalence relation of quasi-isometry.

• Background: In geometric group theory, finiteness properties of a discrete group $G$ (such as being of type $F_n$ or $FP_n$) are known to be invariant under quasi-isometry. These properties relate to the existence of finite CW-models for the classifying space $BG$ (geometric) or finite projective resolutions of the trivial module $\mathbb{Z}$ (homological).
• The Gap: While these properties are well-understood for single groups, the theory for group pairs $(G, \mathcal{P})$—consisting of a finitely generated group $G$ and a finite collection of subgroups $\mathcal{P}$—is more complex. Group pairs arise naturally in the study of manifolds with boundary, Poincaré duality groups, and relatively hyperbolic groups.
• The Challenge: Defining a suitable notion of quasi-isometry for group pairs is non-trivial. Unlike standard Cayley graphs, the coned-off Cayley graphs associated with group pairs are not locally finite. This lack of local finiteness causes standard tools (like the Rips complex) to fail to be cocompact, creating infinitely many orbits of loops of a fixed length. This prevents the direct application of classical proofs (e.g., Alonso's proof for single groups) to the relative setting.

2. Methodology and Definitions

The authors develop a rigorous framework to overcome the geometric obstructions of coned-off graphs.

A. Group Pairs and Finiteness Properties

• Definition: A group pair $(G, \mathcal{P})$ consists of a finitely generated group $G$ and a finite collection of subgroups $\mathcal{P}$.
• Type $F_n$: Defined via the existence of a $G$-CW-pair $(X, G/\mathcal{P})$ where $X$ is contractible, cells in $X \setminus G/\mathcal{P}$ have trivial stabilizers, and the $n$-skeleton is cocompact.
• Type $FP_n$: Defined homologically. The pair is of type $FP_n$ if the augmentation kernel $\Delta_{G/\mathcal{P}} = \ker(\mathbb{Z}[G/\mathcal{P}] \to \mathbb{Z})$ admits a projective resolution finitely generated in degrees $\le n-1$.
• Relation: For $n \ge 2$, $(G, \mathcal{P})$ is of type $F_n$ if and only if it is relatively finitely presented and of type $FP_n$.

B. Quasi-Isometry of Group Pairs

The authors define a strong quasi-isometry (and quasi-retraction) of pairs $(G, \mathcal{P}) \to (H, \mathcal{Q})$.

• A map $f = (f_1, f_2)$ consists of a Lipschitz map $f_1: G \to H$ and a map $f_2: G/\mathcal{P} \to H/\mathcal{Q}$ on the coset spaces.
• Condition: The map must "coarsely preserve" cosets, meaning the Hausdorff distance between $f_1(A)$ and $f_2(A)$ is bounded for every coset $A \in G/\mathcal{P}$.
• Strong Quasi-Isometry: Requires the existence of a quasi-inverse that also respects the coset structure strictly ($r_2 \circ f_2 = \text{id}$).

C. The Core Technical Innovation: "Unicone" Structures

To handle the non-locally finite nature of coned-off Cayley graphs, the authors introduce unicone variations of standard geometric objects:

1. Unicone Rips Complex: Instead of the full Rips complex on the coned-off graph (which has infinitely many cone vertices), they consider the subcomplex consisting of simplices containing at most one cone vertex.
   • Significance: This restriction ensures that the resulting complex has a cocompact $G$-action, allowing the use of Brown's criterion.
2. Unicone Loops: In the context of relative finite presentability, they focus on loops in the coned-off Cayley graph that contain at most one cone vertex.
   • Significance: This allows them to characterize relative finite presentability via "coarse unicone simple connectivity," which is preserved under quasi-retractions.

3. Key Contributions and Results

A. Relative Brown's Criterion (Theorem 3.5)

The authors establish a relative version of Brown's criterion. They prove that a group pair $(G, \mathcal{P})$ is of type $FP_n$ if and only if a specific filtration of the unicone Rips complex is essentially $(n-1)$-acyclic.

• Mechanism: They utilize a filtration of the unicone Rips complex by subcomplexes with cocompact quotients and apply homological algebra (specifically properties of directed systems of homology groups) to link the acyclicity of the filtration to the finiteness of the module $\Delta_{G/\mathcal{P}}$.

B. Invariance under Quasi-Retracts (Main Theorem 1.5)

The central result of the paper is that finiteness properties are inherited by quasi-retracts.

• Theorem: If $(G, \mathcal{P})$ is a quasi-retract of $(H, \mathcal{Q})$ and $(H, \mathcal{Q})$ is of type $FP_n$ (resp. $F_n$), then $(G, \mathcal{P})$ is of type $FP_n$ (resp. $F_n$).
• Corollary: Consequently, these properties are invariant under strong quasi-isometry.
• Proof Strategy:
  • For $FP_n$: They show that the essential acyclicity of the unicone Rips complex filtration is preserved under quasi-retractions (Lemma 4.2).
  • For $F_n$: They prove that relative finite presentability is preserved. This relies on showing that if the coned-off graph of $(H, \mathcal{Q})$ is "coarsely unicone simply-connected," then so is that of $(G, \mathcal{P})$ (Lemma 4.9).

C. Connection to Bredon Cohomology (Theorem 1.6)

The paper connects group pair finiteness properties to Bredon finiteness properties relative to a family of subgroups $\mathcal{F}_{\langle \mathcal{P} \rangle}$ (the smallest family containing $\mathcal{P}$).

• Result: If the collection $\mathcal{P}$ is malnormal (a condition ensuring subgroups intersect trivially or are conjugate), then the finiteness properties of the pair $(G, \mathcal{P})$ are equivalent to the Bredon finiteness properties of $G$ relative to $\mathcal{F}_{\langle \mathcal{P} \rangle}$.
• Implication: This implies that Bredon finiteness properties for malnormal collections are also invariant under strong quasi-isometry of group pairs.

4. Significance and Impact

1. Resolution of a Technical Obstacle: The paper successfully navigates the failure of local finiteness in coned-off Cayley graphs. By introducing the "unicone" restriction, the authors recover the cocompactness necessary to apply geometric group theory techniques (like Rips complexes and Brown's criterion) in the relative setting.
2. Unification of Theories: It provides a unified framework linking the geometric properties of group pairs, their homological algebra, and Bredon cohomology.
3. Applications to Relatively Hyperbolic Groups: Since relatively hyperbolic groups are a primary example of group pairs of type $F_n$, these results solidify the geometric invariance of their structural properties. This is crucial for understanding the stability of relative hyperbolicity and related structures under quasi-isometry.
4. Open Questions: The authors note that their results currently rely on strong quasi-isometry (where the map on cosets is a bijection). It remains an open question whether these properties hold for weaker notions of quasi-isometry where the coset map is not necessarily bijective.

In summary, this paper extends the fundamental principle that "finiteness properties are quasi-isometry invariants" from single groups to group pairs, providing the necessary geometric and homological machinery to handle the complexities of relative settings.