Universal Coefficients and Mayer-Vietoris Sequence for Groupoid Homology

Imagine you are trying to understand the shape of a very complex, shifting city. This city isn't made of bricks and mortar, but of relationships and movements. Some parts are static, some are chaotic, and some are perfectly organized. In mathematics, this "city" is called a groupoid. It's a tool used to describe everything from how particles move in physics to how patterns repeat in music.

Luciano Melodia's thesis is like a new surveying manual for this city. He wants to measure its "holes," "loops," and "twists" (mathematical properties called homology). But because the city is so complex and constantly changing, standard measuring tools don't work well. He invents a new, specialized toolkit called Moore Homology.

Here is a breakdown of his three main discoveries, explained with everyday analogies:

1. The "Moore" Method: Counting with a Flashlight

Imagine you are in a dark, infinite maze (the groupoid). You want to know its structure.

The Old Way: You try to map the entire maze at once, including every shadow and dead end. This is messy and often impossible.
Melodia's Way (Moore Homology): You use a flashlight with a limited battery. You only care about the parts of the maze you can illuminate right now (compact support). You walk through the maze, taking snapshots of small, manageable sections (chains) and counting how they connect.
The Result: By stitching these small, illuminated snapshots together, you can figure out the global shape of the maze without needing to see the whole thing at once. This method is perfect for "ample" groupoids, which are like cities built from tiny, disconnected blocks (totally disconnected spaces) that fit together in specific ways.

2. The "Universal Coefficient" Rule: Translating Languages

Imagine you have a map of the city drawn in English (Integers, $\mathbb{Z}$ ). You want to translate this map into French (a different set of numbers, $A$ ) to see if the shape changes.

The Problem: Usually, when you translate a complex text, you might lose some nuance or gain some new meaning. In math, this is the "Universal Coefficient Theorem" (UCT). It tells you how to convert your map from one language to another.
Melodia's Discovery: He found that for his specific "flashlight" method, the translation works perfectly only if the target language is simple and discrete (like counting numbers: 1, 2, 3).
The Catch: If you try to translate into a "continuous" language (like real numbers with decimals, where you can have infinite precision), the translation breaks. The "flashlight" method simply cannot capture the infinite smoothness of continuous numbers in this specific setup. It's like trying to describe a smooth curve using only Lego bricks; you can get close, but you can't get the smoothness. This proves that his method is inherently designed for "discrete" (blocky) data.

3. The "Mayer-Vietoris" Puzzle: Gluing the Pieces

Imagine you have a giant jigsaw puzzle of the city, but it's too big to solve in one go.

The Strategy: You cut the puzzle into two overlapping pieces (say, the North and South halves). You solve the North, you solve the South, and you solve the overlapping middle part.
The Magic: Melodia provides a mathematical formula (the Mayer-Vietoris sequence) that acts like a glue recipe. It tells you exactly how to combine the solutions of the North, South, and Middle pieces to reconstruct the solution for the whole city.
Why it matters: This allows mathematicians to break down incredibly complex groupoids into tiny, manageable chunks, solve the chunks, and then snap them back together to understand the whole.

The Big Picture: Why This Matters

Melodia's work is like giving engineers a better way to build bridges over a chaotic river.

Before: Engineers (mathematicians) were struggling to measure the river's flow because the water was too turbulent and the tools were too heavy.
Now: Melodia gave them a lightweight, modular toolkit.
1. It works by looking at small, manageable chunks (Moore chains).
2. It has a strict rulebook for how to change the "units of measurement" (Universal Coefficient Theorem), but warns you not to use it for overly smooth, continuous measurements.
3. It provides a step-by-step guide for solving the problem by breaking it into pieces and gluing them back together (Mayer-Vietoris).

In short: This thesis provides the "instruction manual" for measuring the shape of complex, shifting mathematical worlds using a method that is robust, modular, and perfectly suited for the "blocky" nature of modern dynamical systems. It tells us exactly when our tools work, when they fail, and how to use them to solve the hardest puzzles in the field.

Title: Universal Coefficients and Mayer–Vietoris Sequence for Groupoid Homology
Author: Luciano Melodia
Institution: Friedrich-Alexander Universität Erlangen-Nürnberg
Degree: Master of Science in Mathematics

1. Problem Statement

The thesis addresses the computation and structural understanding of the homology of ample étale groupoids. While the homology of groupoids (specifically via the nerve construction) is a well-established tool in operator algebras and dynamical systems (e.g., Matui's work), several foundational gaps remain regarding:

Coefficient Dependence: The behavior of homology when coefficients are not discrete. Specifically, does the classical Universal Coefficient Theorem (UCT) hold for topological abelian groups?
Computational Tools: A robust, chain-level framework for computing homology by decomposing the unit space of a groupoid into simpler pieces (Mayer–Vietoris principle) without relying solely on classifying space homotopy types.
Invariance: A rigorous proof that Moore homology is invariant under Kakutani equivalence (a form of Morita equivalence for ample groupoids) at the chain level, rather than just via homotopy type arguments.

The central challenge is that standard algebraic tools (like tensor products) often fail to capture the topology of compactly supported continuous functions when coefficients are non-discrete, creating a disconnect between the "Moore" model (compactly supported chains) and classical singular homology.

2. Methodology

The author employs a Moore chain complex approach, defined on the nerve of the groupoid $\mathcal{G}_\bullet$ .

Chain Groups: $C_c(\mathcal{G}_n, A)$ , the group of compactly supported continuous functions from the space of composable $n$ -tuples to a topological abelian group $A$ .
Boundary Operator: Defined via pushforwards along face maps ( $d_i$ ). Since $\mathcal{G}$ is étale, face maps are local homeomorphisms, allowing for well-defined pushforwards (finite fiber sums) for compactly supported functions.
Key Assumptions: The groupoids are second countable, locally compact, Hausdorff, and totally disconnected (ample). Coefficients are primarily treated as discrete abelian groups, with specific analysis of the failure of these methods for non-discrete coefficients.

The methodology combines:

Simplicial Topology: Utilizing the nerve $\mathcal{G}_\bullet$ and its geometric realization $B\mathcal{G}$ .
Algebraic Topology: Applying classical homological algebra (exact sequences, UCT) to the specific chain complexes of groupoids.
Geometric Groupoid Theory: Using reductions to clopen saturated subsets, Kakutani equivalence, and the construction of specific functors (similarities) to establish invariance.

3. Key Contributions

A. Universal Coefficient Theorem (UCT) for Discrete Coefficients

The thesis proves a natural short exact sequence for ample groupoids with discrete coefficients $A$ :
$0 \to H_n(\mathcal{G}) \otimes_{\mathbb{Z}} A \xrightarrow{\iota} H_n(\mathcal{G}; A) \xrightarrow{\kappa} \text{Tor}^{\mathbb{Z}}_1(H_{n-1}(\mathcal{G}), A) \to 0$

Mechanism: The proof relies on the identification $C_c(\mathcal{G}_n, \mathbb{Z}) \otimes_{\mathbb{Z}} A \cong C_c(\mathcal{G}_n, A)$ . This isomorphism holds because, for ample groupoids and discrete $A$ , compactly supported continuous functions are locally constant and have finite image, allowing them to be generated by characteristic functions of compact open sets.
Obstruction to Non-Discrete Coefficients: The author rigorously demonstrates that this isomorphism fails for non-discrete coefficients (e.g., $A=\mathbb{R}$ ) on non-discrete spaces (e.g., the Cantor set). In such cases, $C_c(\mathcal{G}_n, A)$ contains functions with infinite images, which cannot be represented as finite sums of tensors from $C_c(\mathcal{G}_n, \mathbb{Z}) \otimes A$ . Consequently, the classical UCT does not hold for general topological coefficients in this framework.

B. Mayer–Vietoris Sequence for Ample Groupoids

A Mayer–Vietoris long exact sequence is constructed for homology with discrete coefficients.

Setup: Given a clopen saturated cover of the unit space $\mathcal{G}_0 = U_1 \cup U_2$ , the groupoid decomposes into reductions $\mathcal{G}|_{U_1}$ , $\mathcal{G}|_{U_2}$ , and $\mathcal{G}|_{U_1 \cap U_2}$ .
Result: A short exact sequence of chain complexes is established:
$0 \to C_\bullet(\mathcal{G}|_{U_1 \cap U_2}; A) \to C_\bullet(\mathcal{G}|_{U_1}; A) \oplus C_\bullet(\mathcal{G}|_{U_2}; A) \to C_\bullet(\mathcal{G}; A) \to 0$
This yields a long exact sequence in homology, allowing the computation of $H_*(\mathcal{G}; A)$ from the homology of the pieces. This is tailored specifically for the "cut-and-paste" nature of ample groupoids.

C. Kakutani Invariance at the Chain Level

The thesis provides a chain-level proof that Moore homology is invariant under Kakutani equivalence (equivalence via full clopen reductions).

Method: Instead of relying on the weak homotopy equivalence of classifying spaces (which does not automatically imply isomorphism of Moore homology), the author constructs explicit étale functors and similarity maps (natural transformations).
Chain Homotopy: It is shown that similar functors induce chain-homotopic maps on the Moore complexes. This establishes that $H_n(\mathcal{G}; A) \cong H_n(\mathcal{H}; A)$ for Kakutani equivalent groupoids $\mathcal{G}$ and $\mathcal{H}$ , validating the use of reduced models for computation.

D. Long Exact Sequences for Subgroupoids and Quotients

The paper extends Matui's framework to establish long exact sequences for:

Regular Subgroupoids: Where the complement behaves well under the range map.
Quotient Groupoids: Using "regular proper surjections" to define reduction maps that handle the fibers of the quotient map, allowing for a Mayer–Vietoris-style sequence for quotients.

4. Results and Examples

Computational Framework: The combination of the UCT and Mayer–Vietoris sequences provides a complete calculus for computing homology of ample groupoids. One can reduce a groupoid to a simpler model, decompose the unit space, compute integral homology, and then derive homology with arbitrary discrete coefficients.
SFT Groupoid Example: The author computes the homology of a disjoint union of Deaconu–Renault groupoids associated with Shifts of Finite Type (SFT).
- By calculating the integral homology $H_*(\mathcal{G})$ using the matrix $I - A^T$ , they demonstrate how torsion in $H_0$ (e.g., $\mathbb{Z}/2\mathbb{Z}$ ) contributes to $H_1$ when using finite field coefficients via the $\text{Tor}$ term in the UCT.
- Specifically, for a prime $p$ , $H_1(\mathcal{G}; \mathbb{Z}/p\mathbb{Z})$ is shown to depend on the torsion of $H_0(\mathcal{G})$ and the prime $p$ , illustrating the "torsion contribution" phenomenon explicitly.

5. Significance

Bridging Algebra and Topology: The work clarifies the precise boundary where algebraic tools (tensor products) fail in the context of topological groupoids. It establishes that the "Moore" model is inherently a discrete coefficient phenomenon for ample groupoids.
Computational Utility: By providing explicit chain-level constructions for Mayer–Vietoris and Kakutani invariance, the thesis offers practical tools for researchers in $C^*$ -algebras and symbolic dynamics who need to compute invariants for complex groupoids without relying on abstract homotopy theory.
Foundational Rigor: The proof of Kakutani invariance via chain homotopies (rather than just classifying space homotopy) resolves a subtle gap in the literature, ensuring that the specific "compactly supported" nature of the homology is preserved under equivalence.

In summary, Melodia's thesis establishes a robust, computationally tractable homology theory for ample groupoids, proving that while it aligns with classical algebraic topology for discrete coefficients, it possesses unique structural constraints regarding non-discrete coefficients that must be carefully navigated.