A Lorentzian Equivariant Index Theorem

Imagine you are trying to count the number of "special" solutions to a complex puzzle. In mathematics, this puzzle is often a Dirac operator—a fancy machine that takes a shape (like a sphere or a twisted tube) and tells you about its hidden geometry and topology.

Usually, mathematicians work in "Riemannian" space, which is like a calm, static landscape (think of a rubber sheet stretched out). But this paper tackles a much wilder environment: Lorentzian spacetime. This is the kind of space Einstein described, where time flows, things move at the speed of light, and the geometry is dynamic (think of a movie playing, not a still photo).

Here is the story of what the authors, Islam and Ronge, have discovered, broken down into simple analogies.

1. The Problem: Counting in a Moving World

In the static world (Riemannian), mathematicians have a famous rule called the Atiyah-Singer Index Theorem. It says: "If you want to know how many special solutions a puzzle has, you don't need to solve the whole puzzle. You just need to look at the 'fixed points' (places that don't move) and do a little bit of math on the edges."

But in the Lorentzian world (our universe with time), things are messy. Time moves forward, and the rules of the game change. For a long time, no one knew if the "counting rule" still worked here.

Recently, two other mathematicians (Bär and Strohmaier) proved that the rule does work for Lorentzian space, but only for the total count. They didn't account for symmetry.

Symmetry is like a dance. Imagine a group of people (a "group") dancing on a stage (the spacetime). If the dancers move in a coordinated way (isometrically), some parts of the stage might stay still (fixed points), while others spin around. The authors wanted to know: If we count the solutions while respecting this dance, does the rule still hold?

2. The Big Idea: The "Time-Slice" Trick

The authors' main breakthrough is a clever way to turn the "moving movie" (Lorentzian) back into a "still picture" (Riemannian) so they can use the old, trusted rules.

Imagine the spacetime is a loaf of bread.

The Lorentzian loaf: The slices are moving and changing shape as you move through time.
The Riemannian loaf: The slices are static and frozen.

The authors realized that even though the "movie" is moving, you can mathematically "slice" it up in a very specific way. They showed that if you look at the "slices" of time (the Cauchy surfaces), the problem of counting solutions in the whole 4D spacetime is actually the same as counting how the solutions flow from the bottom slice (the past) to the top slice (the future).

They call this "Spectral Flow."

Analogy: Imagine a river (time) with fish (solutions) swimming in it. Some fish swim upstream, some downstream. The "Index" is just the net number of fish that crossed a specific line.
The authors proved that counting the fish in the Lorentzian river is exactly the same as counting the fish in a frozen, Riemannian lake, provided you look at the "fixed points" where the group dance leaves things still.

3. The "Group" Twist: The Magic of Eigenspaces

This is the most creative part of their work. They had to deal with the "Group" (the dancers).

Usually, when you have a group acting on a space, it's like having a choir singing. The math gets complicated because everyone is singing different notes at once.

The Authors' Trick: They realized they could break the choir down into individual singers (eigenspaces).
The Metaphor: Imagine the group action is a prism. When white light (the complex problem) hits the prism, it splits into a rainbow of pure colors (eigenspaces).
On each pure color (each specific "note" the group is singing), the group acts like a simple multiplier (just turning the volume up or down).
This allows them to say: "The total answer is just the sum of the answers for each individual color."

This is a "surprisingly simple technique" that lets them take a known result for a single number (the non-equivariant index) and turn it into a complex function that describes the whole group's dance.

4. The Result: The Formula

The final formula they derived looks scary, but the concept is beautiful. It says:

The Number of Solutions = (The Geometry of the Fixed Points) + (The Boundary Terms)

The Fixed Points: Just like in the static world, the answer depends heavily on the places where the group's dance leaves the spacetime unchanged. You integrate a special "shape formula" (involving curvature and twisting) over these spots.
The Boundary Terms: Because the spacetime has an edge (a past boundary and a future boundary), you have to add a "correction" based on what happens at the edges. This is like paying a toll when you enter or leave a city.

5. Why Does This Matter?

Physics: This helps us understand the quantum behavior of particles in curved spacetime (like near black holes or in the early universe) when symmetries are involved.
Mathematics: It bridges a huge gap. It proves that the elegant rules of static geometry (Riemannian) can be successfully translated into the chaotic, time-flowing rules of our universe (Lorentzian), even when symmetries are involved.

Summary in One Sentence

The authors built a mathematical "time-machine" that lets us freeze a moving, time-dependent universe into a static snapshot, allowing us to use old, trusted counting rules to solve new, complex problems involving symmetry and time.

The "Aha!" Moment:
They didn't invent a new, complicated machine to solve the Lorentzian problem. Instead, they found a simple "adapter" (the spectral flow and eigenspace decomposition) that lets the old, simple machine work perfectly on the new, complex problem. It's like realizing you can use a standard screwdriver to fix a high-tech gadget if you just understand the right angle to hold it.

1. Problem Statement

The Atiyah-Singer Index Theorem is a cornerstone of geometric analysis, relating the analytical index of an elliptic operator (the difference between the dimensions of its kernel and cokernel) to topological invariants of the underlying manifold. When a compact Lie group $\Gamma$ acts on the manifold, the equivariant index (a function on the group) can be computed via a fixed-point formula (Atiyah-Segal-Singer).

However, extending these results to Lorentzian manifolds (spacetimes) presents significant challenges:

Hyperbolicity: Dirac operators on Lorentzian manifolds are hyperbolic, not elliptic, meaning they do not naturally possess a Fredholm index without specific boundary conditions.
Boundary Conditions: Bär and Strohmaier (2019) established that on a compact, globally hyperbolic spacetime with a spacelike boundary, the Dirac operator becomes Fredholm under Atiyah-Patodi-Singer (APS) boundary conditions. They proved a non-equivariant index theorem where the index equals the spectral flow of the boundary Dirac operators.
The Gap: Prior to this work, there was no equivariant version of the Lorentzian index theorem. Existing generalizations (e.g., by Damaschke) focused on $L^2$ -indices for non-compact manifolds where the group compensates for non-compactness, rather than computing a refined equivariant index on a compact manifold.

Goal: To derive a formula for the equivariant index of a twisted Dirac operator on a compact, globally hyperbolic spacetime with a timelike boundary, subject to APS boundary conditions, under the action of a compact Lie group.

2. Methodology

The authors employ a multi-step strategy combining geometric decomposition, functional analysis, and spectral theory to reduce the equivariant Lorentzian problem to known Riemannian results.

A. Spacetime Splitting and Equivariant Geometry

Global Hyperbolicity: They utilize the fact that a compact globally hyperbolic spacetime $M$ with spacelike boundary $\partial M = \Sigma_0 \sqcup \Sigma_1$ is diffeomorphic to $[0, 1] \times \Sigma$ .
Equivariant Splitting: A key technical contribution (Section 2) is proving that for a compact group $\Gamma$ acting by isometries, one can choose a time function and a splitting such that the group action is constant in time. Specifically, $\gamma(t, x) = (t, \tilde{\gamma}(x))$ .
Metric Structure: This allows the metric to be written as $g = -N^2 dt^2 + g_t$ , where the lapse function $N$ is identically 1 near the boundaries, and the group action preserves each Cauchy slice $\Sigma_t$ . This ensures the family of boundary Dirac operators $A(t)$ is $\Gamma$ -equivariant.

B. Reduction from Equivariant to Non-Equivariant

The core technical innovation (Section 3) is a method to reduce equivariant indices and spectral flows to non-equivariant ones using spectral decomposition.

Eigenspace Decomposition: For a group element $\gamma$ , the Hilbert space decomposes into eigenspaces $E_\lambda(\gamma)$ .
Linearity: Since $\gamma$ acts as a scalar $\lambda$ on each eigenspace, the equivariant index and spectral flow are simply weighted sums of the non-equivariant quantities restricted to these subspaces:
$\text{ind}_\gamma(D) = \sum_{\lambda} \lambda \cdot \text{ind}(D|_{E_\lambda(\gamma)})$
$\text{sf}_\gamma(A) = \sum_{\lambda} \lambda \cdot \text{sf}(A|_{E_\lambda(\gamma)})$
Challenge: Restricting a Dirac operator to an eigenspace does not yield a Dirac operator. To overcome this, the authors relate the geometric Dirac operator to an abstract model operator of the form $\partial_t - iB$ (or $\partial_t + B$ in the Riemannian case) via unitary equivalence (Section 5). This model operator behaves well under restriction to eigenspaces.

C. Bridging Lorentzian and Riemannian

Model Operator: The authors construct a unitary map $W$ that transforms the Lorentzian Dirac operator $D$ into a form involving the abstract operator $\partial_t - iB$ , where $B$ is unitarily equivalent to the family of hypersurface Dirac operators $A(t)$ .
Index = Spectral Flow: Using the reduction technique, they prove that the equivariant index of the Lorentzian operator $D_{APS}$ equals the equivariant spectral flow of the family $A(t)$ (Theorem 6.6).
Riemannian Comparison: They construct an auxiliary Riemannian Dirac operator $\hat{D}$ on the same manifold (using the metric $\hat{g} = N^2 dt^2 + g_t$ ). They show that $\text{ind}_\gamma(D_{APS}) = \text{ind}_\gamma(\hat{D}_{APS})$ .

D. Application of Riemannian Equivariant Index Theorem

Since $\hat{D}$ is a Riemannian operator, the authors apply the Atiyah-Segal-Singer-Donnelly index theorem (combining results by Berline-Vergne and Donnelly).
They compute the necessary characteristic forms ( $\hat{A}$ -genus, Chern character) and transgression forms to handle the boundary terms arising from the difference between the original metric and a product metric near the boundary.

3. Key Contributions

First Equivariant Lorentzian Index Theorem: The paper provides the first formula for the equivariant index of a Dirac operator on a compact Lorentzian manifold with boundary.
Reduction Technique: The development of a "surprisingly simple" technique to reduce equivariant spectral flow and index problems to non-equivariant ones via eigenspace decomposition, applicable to abstract operator families.
Unitary Equivalence to Model Operators: Generalizing previous results (van den Dungen) to show that the Lorentzian Dirac operator is unitarily equivalent to an abstract operator $\partial_t - iB$ even when the lapse function $N$ is not constant, provided $N=1$ near the boundary.
Explicit Fixed-Point Formula: Deriving a concrete formula that includes bulk integrals over the fixed-point set and boundary terms involving $\eta$ -invariants and traces of the group action on kernels.

4. Main Results

Theorem 1.1 (Main Theorem):
Let $(M, g)$ be an even-dimensional compact globally hyperbolic spin spacetime with spacelike boundary $\partial M = \Sigma_0 \sqcup \Sigma_1$ . Let $D_{APS}$ be the twisted Dirac operator with APS boundary conditions, equivariant under a group $\Gamma$ . For any $\gamma \in \Gamma$ , the equivariant index is:

$\text{ind}_\gamma(D_{APS}) = \int_{M_\gamma} (2\pi i)^{-\ell - \ell^\perp} a + \int_{\partial M_\gamma} (2\pi i)^{-\ell - \ell^\perp} T_a + b$

Where:

$M_\gamma$ is the fixed-point set of $\gamma$ .
$a = \frac{\hat{A}(M_\gamma) \wedge \text{ch}_\gamma(E)}{\det^{1/2}(1 - \gamma^\perp \exp(-R^\perp))}$ is the bulk integrand (involving the $\hat{A}$ -form and localized Chern character).
$T_a$ is the transgression form of $a$ .
$b = -\frac{1}{2} \left( \text{tr}(\gamma|_{\ker A(0)}) + \text{tr}(\gamma|_{\ker A(1)}) + \eta_\gamma(A(0)) - \eta_\gamma(A(1)) \right)$ represents the boundary contribution involving the $\eta$ -invariant and kernel traces of the boundary Dirac operators.
$\ell$ and $\ell^\perp$ are half the dimension and codimension of the fixed-point set.

Key Insight: The formula is structurally identical to the Riemannian equivariant index theorem, but the proof relies on the Lorentzian "index = spectral flow" relationship.

5. Significance

Unification: It unifies the geometric analysis of Lorentzian and Riemannian manifolds in the context of equivariant index theory, showing that the topological nature of the index persists even in hyperbolic settings.
Physical Relevance: The setting (globally hyperbolic spacetimes with boundaries) is relevant to mathematical physics, particularly in the study of quantum field theory on curved spacetimes and the analysis of fermions in cosmological models with boundaries.
Methodological Advancement: The reduction technique (decomposing into eigenspaces to handle equivariance) offers a versatile tool that can potentially be applied to other equivariant problems in global analysis where direct equivariant proofs are difficult.
Foundation for Future Work: The authors note that this framework can likely be extended to non-compact settings (with appropriate potentials), $Spin^c$ structures, and general Dirac-type operators, opening new avenues for research in Lorentzian geometry.

In summary, the paper successfully adapts the machinery of equivariant index theory to the Lorentzian setting, proving that the index is determined by local geometric data on the fixed-point set and boundary spectral data, exactly as in the Riemannian case, but achieved through a novel reduction to spectral flow.