PREreview de Inverse Mobius Spacetime in 1+1D Quantum Gravity: Functional Analytic Structures, Dirac Spectrum, and Pin Geometry

PREreview of “Inverse Möbius Spacetime in 1+1D Quantum Gravity: Functional Analytic Structures, Dirac Spectrum, and Pin Geometry”

27 ott 2026 · Luca Eliseo Pavesi

A. Chakraborty — arXiv:2507.22417v2 [gr-qc] · doi:10.48550/arXiv.2507.22417

Summary

The topological part of the manuscript is sound, but the spectral part, which carries the main physical conclusions, contains errors that must be fixed before those conclusions can be considered established.

The manuscript formulates a 1+1-dimensional Jackiw–Teitelboim (JT) gravity toy model on the Möbius band M and studies its Pin geometry and the spectrum of the Dirac operator. Since M ≃ S¹, one has H²(M; ℤ₂) = 0, so both Pin⁺ and Pin⁻ structures exist. Each parity admits exactly two inequivalent structures, classified by H¹(M; ℤ₂) ≅ ℤ₂ (Theorems 3.1 and 3.6).

The author then imposes the twisted equivariance condition ψ(x+1, w) = γʷ ψ(x, −w) on the universal cover M̃ = ℝ × [−1, 1]. From it the author derives a spectrum with half-integer momentum modes and concludes that ind D = 0, that η(0) = 0, and that the sum over Pin structures doubles the JT partition function without changing the leading Bekenstein–Hawking entropy.

Non-orientable surfaces in JT gravity are a timely topic. An explicit, solvable example of the Dirac spectrum on the Möbius band would be a useful pedagogical contribution.

Major issues

• Question on the compatibility of the twisted condition with the Dirac operator (Sections 4.1–4.2, Theorem 4.4). Take γˣ = σ₁, γʷ = σ₂ and (Rψ)(x, w) = σ₂ ψ(x, −w). If I am not mistaken, a direct computation gives D(Rψ) = σ₁σ₂ (∂ψ/∂x)∘R − σ₂σ₂ (∂ψ/∂w)∘R = −σ₂ (Dψ)∘R. This would mean that D maps sections satisfying ψ(x+1, w) = +γʷ ψ(x, −w) to sections satisfying the condition with the opposite sign. In that case an eigenfunction with λ ≠ 0 would have to satisfy both conditions, and would therefore vanish. Could the author clarify on which space of sections the eigenvalue problem Dψ = λψ is posed? A related question concerns the proof of Theorem 4.4: since γ⁵γʷ = −γʷγ⁵, the spinor γ⁵ψ satisfies the opposite-sign condition, and it is not clear to me how rescaling ψ by −1 can change a linear condition. Would it not be more natural for the lift of the reflection w ↦ −w to commute with γˣ and anticommute with γʷ, for instance A ∝ γˣ, or γʷγ⁵ up to a phase, depending on the Pin⁺/Pin⁻ convention? All subsequent spectral results depend on this choice, so I suggest that the author revisit the equivariance condition and state explicitly which convention is used.

• Momentum quantization (Theorem 4.1). The eigenvalues of σ₂ are ±1, not ±i. The condition exp(2πik) φₖ(w) = σ₂ φₖ(−w) admits non-trivial solutions for every k ∈ ½ℤ, including integer k, with a parity condition on the components of φₖ in the σ₂ eigenbasis. The restriction k ∈ ℤ + ½ does not follow from the argument given.

• Heat-kernel asymptotics (Proposition 4.6, Appendix B). For a rank-2 spinor bundle the leading Weyl term is Tr exp(−tD²) ∼ 2·Area(M)/(4πt) = 1/(πt), with Area(M) = 2, not 1/(2πt). Summing the spectrum of Theorem 4.3 explicitly, with Poisson summation over j and Euler–Maclaurin over n, does give 1/(2πt). This suggests that half of the modes are missing, consistently with the previous point. In addition, on a manifold with boundary the expansion contains a boundary term of order t⁻¹ᐟ², proportional to the length of ∂M, which is absent here.

• Definition of the η-function (Appendix B). The Mellin representation of the η-function of D is η(s) = [1/Γ((s+1)/2)] · ∫ t⁽ˢ⁻¹⁾ᐟ² Tr(D · exp(−tD²)) dt, integrated over t ∈ (0, ∞), with D inside the trace. The expression used in the manuscript, with Tr exp(−tD²), corresponds to a zeta function and cannot detect spectral asymmetry.

• Gluing argument (Section 5.2, Theorem 5.3). Cutting the Möbius band along its central circle {w = 0} does not produce two annuli M₊ and M₋. It produces a single connected annulus of twice the length. The application of the Kirk–Lesch formula, which assumes two components, therefore needs to be reformulated.

• Local trivializations (Section 3.2, Lemma A.1). Because of the identification (x, w) ∼ (x+1, −w), the sets {w > ε} and {w < −ε} coincide in M, so the cover {U₊, U₀, U₋} is not well defined on M. Moreover U₊ ∩ U₀ ∩ U₋ = ∅, so the cocycle condition of Lemma A.1 is vacuous. A cover adapted to the non-orientable direction, along x, would make the construction of the four Pin bundles explicit.

Minor issues

• Fermionic determinant (Section 5.3). In the partition function, a fermionic determinant should enter with a positive power, det D or (det D)¹ᐟ² as a Pfaffian, not (det′D)⁻¹ᐟ², which is the bosonic case. The notation Zᵒʳⁱᵉⁿᵗ(M) is also unclear, since M is not orientable; please define it.

• APS boundary condition (Definition 4.2, Theorem 4.3). The condition should be stated precisely. With outward normal ∂/∂w, the tangential boundary operator is γʷγˣ ∂/∂x (up to conventions), rather than γˣ ∂/∂x. The derivation of μ = (n + ½)π should be shown explicitly.

• Fredholm property and spectral flow (Appendix C). Proposition C.3 states that D: Hˢ⁺¹ → Hˢ (twisted Sobolev spaces) is Fredholm for all s. On a manifold with boundary this requires elliptic boundary conditions, such as APS; without them the kernel is infinite-dimensional. In Theorem C.4, the APS variation formula usually carries a factor ½ in front of η(0; D₁) − η(0; D₀).

• Proposition 2.3. In the proof, the boundary of M is described as “non-orientable”, but it is a circle. The statement H²(M; ℤ₂) = 0 follows directly from M ≃ S¹.

• Cross-references. There are references to non-existent items: “Theorem 6.3” and “Section 5” in Remark 5.4, presumably Theorem 4.3 and Section 3.2.

• Bibliography and keywords. Several references, for example [2], [6], [9], [10], [22] and [29] on quantum reference frames and canonical gravity, are not used in the main text, and neither are some keywords (quantum reference frames, topological vector spaces). The bibliography could focus on Pin geometry, spectral asymmetry and non-orientable JT gravity. A comparison with the treatment of Pin⁻ structures by Stanford and Witten [27] would be valuable.

• Terminology. The term “inverse Möbius band” for the universal cover is non-standard. Since the text itself acknowledges this, using “universal cover” throughout would avoid confusion.

Competing interests

The author declares that they have no competing interests.

Use of Artificial Intelligence (AI)

The author declares that they did not use generative AI to come up with new ideas for their review.