Quantum Reference Frames and Correlation Geometry

Abstract: The aim of this paper is to provide a largely self-contained, compact and comprehensible introduction to the basic ideas behind correlation geometry, which underlies the theory of causal fermion system (CFS). A key focus here is on the manner in which the framework deals with gauge transformations, including diffeomorphisms via the principle of unitary equivalence. We will argue that, conceptually, the fundamental description of a physical system in terms of its correlation geometry is much closer to thermodynamics than quantum theory.

• The paper presents correlation geometry as a fundamental framework in which reference fields and their correlations encode the complete effective physical model through a measure over self-adjoint operators.
• The paper shows that local correlation maps are resolution-limited by the reference-system Hilbert-space dimension, while unitary equivalence unifies electromagnetic gauge transformations and diffeomorphisms.
• The paper argues heuristically that mixtures of spacetime geometries generally lack compact effective descriptions beyond the weak-field regime, while acknowledging that nonequilibrium dynamics and key symmetry conjectures remain unproven.

Overview and motivation

This paper by Claudio F. Paganini is a conceptual introduction to correlation geometry, the mathematical structure underlying the theory of causal fermion systems (CFS). Written as a response to the quantum reference frame (QRF) framework of Kabel et al. [(2604.16631)'s motivating reference], the paper's central claim is that correlation geometry represents the QRF idea "taken to its logical conclusion": rather than using reference fields merely as labels for comparing descriptors across different physical models, the framework encodes all aspects of the effective physical model directly in the reference fields and their correlations. The paper advances a strong conceptual thesis: the fundamental description of a physical system via its correlation geometry is much closer to thermodynamics than to quantum theory, with direct consequences for the viability of spacetime superpositions.

The author distinguishes carefully between a physical system (a real object), an effective physical model (a description valid at some emergent scale, characterized by descriptors such as manifold, metric, and fields), and a fundamental physical model (the elementary description giving rise to the effective one). The thermodynamic analogy runs throughout: temperature, volume, and pressure are descriptors of an effective model; the density function on phase space is the fundamental description. The paper argues that conventional spacetime descriptors play exactly the role that $T$, $V$, $p$ play in thermodynamics—labels identifying special correlation geometries among all possible ones.

The structure of correlation geometry

A correlation geometry is defined as a triplet $(\mathcal{H}, \mathcal{F}^{p,q}, \rho)$: a separable complex Hilbert space $\mathcal{H}$; the subset $\mathcal{F}^{p,q}$ of bounded self-adjoint operators with at most $p$ positive and $q$ negative eigenvalues; and a Borel measure $\rho$ on this space. For $p = q = n$, this recovers a causal fermion system of spin dimension $V$0. A key structural fact is that the regular part $V$1 is a smooth manifold of dimension

$V$2

with $V$3; this is what justifies calling the structure a "geometry" of correlations. The triplet itself is taken to be the fundamental physical model.

Encoding effective models via the local correlation map

The bridge from effective to fundamental descriptions is the local correlation map. Given an effective model $V$4 with a fiber bundle $V$5, a choice of sections $V$6 closed into a Hilbert space $V$7, and a section of hermitian forms $V$8 on the fibers, each point $V$9 is identified with a finite-rank self-adjoint operator $p$0 on $p$1 via $p$2. The signature of $p$3 determines which $p$4 the map targets, and a volume form $p$5 pushes forward to the measure $p$6. The choice of sections $p$7 is precisely what plays the role of the quantum reference frame.

Two worked examples illustrate both the power and the fragility of the construction:

• Metric encoding via vector fields: for a compact Riemannian manifold with tangent-bundle sections, the metric at each point becomes a self-adjoint rank-$p$8 operator in $p$9. If $(\mathcal{H}, \mathcal{F}^{p,q}, \rho)$0 consists of exactly $(\mathcal{H}, \mathcal{F}^{p,q}, \rho)$1 tetrad fields on a parallelizable manifold, the entire local correlation map collapses to $(\mathcal{H}, \mathcal{F}^{p,q}, \rho)$2 everywhere, and the measure reduces to $(\mathcal{H}, \mathcal{F}^{p,q}, \rho)$3—all geometric information except total volume is lost.
• Scalar field encoding: using gradient-based hermitian forms on complex line bundle sections, compactness guarantees degeneracies at extrema of real scalar fields, again showing that undersized reference systems destroy essential information.

These examples establish a concrete lesson: the dimension of the reference system controls resolution. With $(\mathcal{H}, \mathcal{F}^{p,q}, \rho)$4, the local correlation map is "bandwidth limited," and full small-scale information is recoverable only in the continuum limit $(\mathcal{H}, \mathcal{F}^{p,q}, \rho)$5. This directly contrasts with Kabel et al.'s use of four scalar coordinate fields; here the Hilbert space should be taken as large as possible. The paper also notes a technical subtlety: well-definedness of the hermitian form requires Sobolev regularity ($(\mathcal{H}, \mathcal{F}^{p,q}, \rho)$6 for infinite-dimensional $(\mathcal{H}, \mathcal{F}^{p,q}, \rho)$7 under the $(\mathcal{H}, \mathcal{F}^{p,q}, \rho)$8 norm), or alternatively an $(\mathcal{H}, \mathcal{F}^{p,q}, \rho)$9-regularization averaging over geodesic balls.

Matter fields and gauge invariance through unitary equivalence

Matter fields are encoded by making the reference system depend on them: one chooses $\mathcal{H}$0 for a suitable operator $\mathcal{H}$1. In CFS the natural choice is the spin bundle with the Dirac operator—the massive Dirac equation for vacuum spacetimes, or the minimally coupled equation $\mathcal{H}$2 when an electromagnetic potential is present. The resulting measure $\mathcal{H}$3 then carries all physical content.

Comparison of fundamental models proceeds via unitary equivalence: two triplets are equivalent if, after isometric embedding of the smaller Hilbert space, there exists a unitary $\mathcal{H}$4 such that $\mathcal{H}$5. Effective models are declared indistinguishable when their fundamental representatives are unitarily equivalent. This single notion subsumes both classical gauge transformations and diffeomorphisms:

• Electromagnetic gauge transformations: if $\mathcal{H}$6 differs from $\mathcal{H}$7 by $\mathcal{H}$8, choosing the transformed reference system $\mathcal{H}$9 yields identical measures, so the models are equivalent—as expected.
• Diffeomorphisms: if $\mathcal{F}^{p,q}$0 is the pullback of $\mathcal{F}^{p,q}$1 along a diffeomorphism $\mathcal{F}^{p,q}$2, and the reference systems, scalar products, and hermitian forms are related by the corresponding pullbacks, the measures agree and the models are equivalent. Notably, equivalence requires the entire relational apparatus—not just the metric and fields—to be transported consistently.

Gauge transformations are then defined abstractly as any redundant mathematical object removable without changing the equivalence class, and symmetries of a correlation geometry are unitaries leaving the measure invariant. The representation is trivially coordinate-invariant by construction. The paper states a conjecture that isometries preserving all fields, combined with reference systems solving isometry-invariant equations, induce symmetries in this sense; this remains unproven.

Against spacetime superpositions: the thermodynamic picture

The paper's most consequential claim concerns superpositions of spacetimes, widely discussed in the QRF literature. The argument is that convex combinations of measures, $\mathcal{F}^{p,q}$3, are perfectly valid correlation geometries—but generically admit no compact effective description in terms of few descriptors, just as a mixture of two equilibrium distributions at different temperatures is a valid statistical distribution but not an equilibrium state describable by a handful of macroscopic parameters. Whereas a quantum superposition evolves under the same dynamical laws as its components, the thermodynamic analogue relaxes toward a new equilibrium with qualitatively different dynamics. The author concludes that spacetime superposition is not a well-defined operation beyond the weak-field/linearized regime accessible to tabletop experiments—a position aligned with the claims in prior CFS work but in tension with a substantial body of literature the paper does not fully engage mathematically.

This claim carries an explicit caveat within the paper itself: whether the mixed measure can be interpreted as an out-of-equilibrium system with well-defined relaxation dynamics is, by the author's own admission, unclear at present. The analogy is heuristic, not a derived result.

Limitations and open questions

Several limitations are conceded explicitly. The paper deliberately sacrifices mathematical rigor—the rigorous framework exists only partially, in prior work on topological fermion systems, with continuation promised elsewhere. The central conjecture linking isometries to measure symmetries is unproven. The thermodynamic analogy, while suggestive, rests on the unresolved question of whether nonequilibrium dynamics exists for correlation geometries at all. Additionally, the diffeomorphism example requires assumptions that the scalar products and hermitian forms themselves transform covariantly, prompting the author to note these structures might need to be appended to the list of effective descriptors—an ambiguity left open. Finally, the claim that subsystem spacetimes at the fundamental level generally do not admit classical effective descriptions is asserted heuristically rather than demonstrated.

Conclusion

The paper reframes quantum reference frames as the foundational encoding mechanism of correlation geometry, in which all physical content lives in the measure $\mathcal{F}^{p,q}$4 over spaces of self-adjoint operators, and conventional descriptors become mere labels for special geometries. Its principal contributions are a unified treatment of gauge transformations—including diffeomorphisms—via unitary equivalence of measures, and a thermodynamically grounded argument that spacetime superpositions lack operational meaning outside the linearized regime. Whether the out-of-equilibrium dynamics suggested by the analogy actually exists for correlation geometries remains the central open question the paper leaves to future work.