Geometric Trace Formulas: A Modern Perspective

• Geometric trace formulas are techniques that relate analytic traces to sums over fixed points and orbits in classical, derived, and noncommutative spaces.
• They employ methods like supersymmetric localization, loop space analysis, and derived intersections to compute orbital integrals and categorical traces.
• Applications span automorphic forms, index theory, and the Langlands program, offering a unified framework for spectral and geometric insights.

A geometric perspective using trace formulas interprets spectral and representation-theoretic identities as statements about the geometry of spaces—classical, noncommutative, or derived—by explicitly relating analytic traces to sums over geometric fixed points, closed orbits, conjugacy classes, or intersections in moduli. This viewpoint, central to modern developments in automorphic forms, index theory, and categorical representation theory, underscores fixed-point localization, periodic-geodesic analysis, stack-theoretic intersections, and the structure of orbits, often within the organizing frameworks of field theory or higher-categorical algebra.

1. Geometric Foundations of Trace Formulas

Geometric trace formulas fundamentally equate the trace of a natural operator (e.g., a Laplacian, Hecke correspondence, or group convolution) with a sum over geometric data attached to the underlying space. In classical settings, this takes the form

$$\mathrm{Tr} K_f = \sum_{[\gamma]} a_\gamma O_\gamma(f)$$

where $O_\gamma(f)$ is an orbital integral—averaging $f$ along the conjugacy class of $\gamma$—and $a_\gamma$ captures the volume of centralizers or stabilizers. The geometric side arises by unfolding the kernel $K_f(x,y)$ and integrating over the diagonal, localizing the trace to fixed points and periodic orbits (Frenkel, 2012, Frenkel et al., 2010). In higher-categorical and derived contexts, this is generalized: the trace is the self-intersection of the diagonal in a correspondence category, and fixed-point loci become derived intersections or loop spaces (Ben-Zvi et al., 2013).

Crucially, geometric trace formulas organize automorphic, representation-theoretic, and index-theoretic identities in a manner that tightly binds spectrum (eigenvalues, representations) to explicit geometric orbits (conjugacy classes, fixed point loci, closed geodesics), setting the stage for powerful localization mechanisms.

2. Supersymmetric Localization and Fixed-Point Loci

Recent advances provide path-integral and supersymmetric approaches to geometric trace formulas, exploiting the structure of the loop space of a manifold or coset. In particular, the path-integral formalism for supersymmetric quantum mechanics on compact symmetric spaces produces exact trace formulas by localizing the integral onto the quantum fixed-point set: periodic solutions of a derived localization equation corresponding to closed geodesics, constant loops, or flat connections (Choi et al., 2023, Choi et al., 14 Feb 2025, Choi et al., 2021). The general scheme is:

• Construct a supersymmetric gauged sigma model whose configuration space is the relevant loop space $L(\Gamma\backslash G)$ or $L(G)$.
• Introduce a nonstandard $\delta$-exact deformation of the action whose critical locus consists of those loops with constant momenta (e.g., closed geodesics or constant gauge fields).
• Evaluate the path-integral in the localization limit $\lambda\to\infty$ by summing Gaussian contributions at fixed points, weighted by one-loop determinants encoding normal bundle data (Euler classes, root denominators).

Explicitly, for the Selberg trace formula on a compact Riemann surface $X=\Gamma\backslash SL(2,\mathbb R)/SO(2)$,

$$\sum_j h(r_j) = \frac{\mathrm{Vol}(X)}{4\pi}\int_{-\infty}^\infty r h(r)\tanh(\pi r)dr + \sum_{[\gamma]\neq[I]}\sum_{n=1}^\infty \frac{\ell(\gamma_0)\chi(\gamma^n)}{2\sinh(n\ell(\gamma)/2)} \widetilde h(n\ell(\gamma))$$

where each term corresponds, respectively, to contractible loops (volume, continuous spectrum) and nontrivial closed geodesics (primitive conjugacy classes) (Choi et al., 2023). In higher rank, fixed-point loci for the localization correspond to conjugacy classes of all semisimple elements and their associated geometric cycles.

This approach unifies geometric and spectral data and forges a direct analytic bridge between representation theory, periodic pseudodifferential calculus, and topological field theory.

3. Loop Spaces, Derived Correspondences, and Nonlinear Traces

In derived algebraic geometry and higher categories, the geometric perspective on traces is further formalized through the theory of correspondences, dualizability, and higher categorical traces. The central insight is that in a category of correspondences (spans) of derived stacks, the trace of the identity endofunctor is the derived loop space $\mathcal L X = X \times_{X\times X} X$, while the trace of a general (possibly nonlinear) correspondence $Z$ is its self-intersection,

$\mathrm{Tr}(Z) = Z \times_{X\times X} X,$

identifying the categorical trace with generalized fixed-point loci (Ben-Zvi et al., 2013).

When a sheaf theory $S$ (e.g., coherent sheaves, D-modules) is applied:

• $HH_*(S(X)) \cong$ volume forms or distributions on $\mathcal L X$
• The categorical trace corresponds to functions on $\pi_0(\mathcal L X)$, often interpreted as classes in Hochschild or cyclic homology.

This machinery yields categorical, universal versions of the Atiyah-Bott-Lefschetz, Grothendieck-Riemann-Roch, and Frobenius/Weyl character formulas, all interpretable as geometric traces over self-intersection or fixed-point data. A table summarizing these geometric identifications:

Classical Notion	Geometric Stack Perspective	Categorical/Sheaf-theoretic Trace
Euler characteristic	Trace of identity functor	Sum over fixed points of identity
Lefschetz trace	Trace of correspondence	Sheaf cohomology over fixed-point locus
Weyl character formula	Class function on $G/G$	Trace over category of $G$-modules
Riemann-Roch	Pushforward along $f:X\to Y$	Trace over loop space maps $\mathcal L f$

4. Geometric Decomposition of Orbital Integrals

A key structural element of geometric trace formulas is the explicit decomposition into contributions labeled by geometric orbits—conjugacy, double cosets, closed geodesics, or flat bundles—each interpreted as a fixed-point component or cycle associated to the symmetry being traced. For example:

• For $\mathrm{GL}(3)$, the geometric side splits into sums over unramified and ramified orbits, with the latter arising as limits (regularizations) of unramified orbital integrals as eigenvalues coalesce (Xinghua et al., 29 Aug 2025).
• The Selberg and generalized Selberg trace formulas decompose the trace into elliptic, hyperbolic, mixed, and parabolic classes, with each associated to geometric features: isolated fixed points, closed geodesics, flat orbits at cusps, and more intricate higher-rank centralizers (Biró et al., 2023).
• In the context of relative trace formulas, orbits correspond to symmetric spaces or double cosets (e.g., $H\backslash G / H$), with the geometric side organizing as weighted sums over these orbits, sometimes requiring regularization via truncation or periodization (Delorme et al., 2015, Delorme et al., 2016, Lekkas et al., 16 Sep 2025).

Weighted orbital integrals encode normalization by volumes of stabilizers, determinants of normal bundles (Weyl discriminants), and symmetry group factors, enforcing functoriality and invariance under geometric symmetries. In higher-categorical settings, orbital integrals become traces over derived stacks of fixed loci, and can be computed via cohomological tools.

5. Geometric Trace Formulas in Noncommutative and Categorical Settings

Noncommutative geometry offers a further extension, interpreting trace identities for spectral triples, pseudodifferential operators, and C*-algebras as geometric formulas for noncommutative spaces:

• Heat-trace asymptotics for abstract pseudodifferential operators yield expansions of the form

$$\mathrm{Tr}(e^{-tA}) \sim \sum_k a_k t^{\frac{k-\dim(A)}{r}}$$

generalizing the Minakshisundaram-Pleijel result, with coefficients expressed via Dixmier or Wodzicki residues—these are geometric invariants of the underlying operator-algebraic structure (Hekkelman, 27 Jun 2025).

• The Dixmier trace formula generalizes the classical integral to noncommutative spaces: $\mathrm{Tr}_\omega(a|D|^{-d})$ serves as a noncommutative measure of volume, recoverable as a limit of traces over spectral projections or operator integrals.
• Noncommutative Taylor expansions and Szegő-type limits connect operator perturbation theory to quantum ergodicity, with the limiting behavior controlled by the geometry of spectral subspaces and dynamical flow on noncommutative cospheres.

This demonstrates that geometric trace formulas are not confined to finite-dimensional stacks or classical spaces, but extend, via operator-theoretic and categorical formalism, to broad classes of quantum and noncommutative spaces.

6. Geometrization in the Langlands Program and Categorical Functoriality

A geometric perspective on trace formulas is essential in the geometric Langlands program and modern approaches to functoriality. Here, the classical trace is interpreted as a cohomological or categorical trace: moduli spaces of bundles (Bun$_G$), stacks of local systems (Loc$_G$), and Hecke correspondences are organized such that the (orbital) geometric side is the trace of a sheaf or complex over the stack, while the (spectral) side is described via derived or categorical Langlands correspondences (Frenkel, 2012, Frenkel et al., 2010).

The geometric trace formula equates sums over Frobenius traces on the cohomology of moduli spaces with spectral-counting formulas, with the identification mediated by categorical equivalences (e.g., $D^b(\mathrm{Loc}_{^LG}) \simeq D^b(\mathrm{Bun}_G)$). Derived fixed-point formulas (Atiyah-Bott-Lefschetz for stacks) and cohomological pushforwards underpin the formal matching, and relative geometric trace formulas (e.g., via Whittaker sheaves or period integral functors) encode automorphic periods and $L$-function residues in derived terms.

The geometric viewpoint clarifies stabilization and endoscopic transfer, interprets transfer of functors as derived push-pull along L-group embeddings, and underlies the proof of foundational conjectures, such as the Fundamental Lemma, by tracing identities to symmetry and topology in moduli spaces.

7. Geometric Meaning and Implications

The geometric perspective reframes trace formulas as localization principles: analytic traces of (pseudo)differential, group, or categorical operators localize on fixed loci or periodic orbits dictated by the symmetry in question. For classical manifolds, this is reflected in the sum over closed geodesics or periodic flows; for stacks and categories, in terms of self-intersections and derived fixed points; for C*-algebras and noncommutative spaces, via residues and dynamical invariants.

Each term in a geometric trace formula corresponds to a concrete geometric structure:

• Volume or density terms reflect the measure of contractible or trivial loci.
• Sums over geodesics/conjugacy classes correspond to nontrivial cycles or orbits, with combinatorics informed by the topology of the space and symmetries of the problem.
• Orbital integrals and local measures record the geometry of stabilizers and the nature of singularities or degeneracies.

This perspective not only organizes spectral and arithmetic data but also provides robust tools for calculations and generalizations in representation theory, quantum field theory, noncommutative geometry, and categorical representation theory, extending the power of the trace formula into fundamentally geometric domains (Choi et al., 2023, Ben-Zvi et al., 2013, Hekkelman, 27 Jun 2025, Frenkel, 2012).