REVIEW 2 major objections 4 minor 6 references
The Equivariant Fried Conjecture for Suspension Flow of an Equivariant Isometry
T0 review · 2 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash
Pith's one-line read The paper proves the equivariant Fried conjecture for suspension flows of equivariant isometries: the group-equivariant Ruelle zeta function extends to the origin, and its value there equals the square of the equivariant analytic torsion.
desk verdict Serious paper with a real, fixable flaw in its discrete-group example; the main theorem and the compact/compact-centralizer corollaries look credible. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The argument is carried by three pieces. First, the suspension construction: the manifold $M$ is the mapping torus $(Y\times\mathbb{R})/\mathbb{Z}$, where $n\cdot(y,t) = (T^n y, t-n)$, and the suspension flow is $\varphi_s[y,t] = [y,t+s]$; $g$-periodic flow curves correspond exactly to points of the fixed sets $Y^{g^{-1}T^n}$, so the Ruelle zeta function becomes a fixed-point sum (Theorem 5.9). Second, the equivariant analytic torsion is computed by a fibration formula (Theorem 4.2), which expresses $T_g(\nabla E,\sigma)$ for the suspension in terms of equivariant Euler characteristics $\chi_{(g,n)}(\nabla E_1)$, evaluated by a generalisation of the Atiyah-Bott-Lefschetz fixed-point formula to non-compact manifolds (Corollary 3.20). Third, the two fixed-point sums are matched term by term: the sign $\operatorname{sgn}\det(1 - D_y(g^{-1}T^n))$ and the holonomy trace $\operatorname{tr}(gT_{E_1}^{-n})$ appear identically on both sides, and the cutoff compatibility identity (2.22) forces the flow-side weights to equal the torsion-side weights $\psi_g(y)$. The hypothesis that $g$ lies in a compact subgroup enters through Lemma 2.15(II): it is exactly what makes all flow curves periodic, so the fibre over each class in $Y^{g^{-1}T^n}$ has finite size $p(y)$.
What would settle it
A concrete calculation would settle the compact case: on $Y = T^2$ with a flat line bundle whose holonomy is nontrivial in one direction, take commuting rotations $T$ and $g$ such that no $g^{-1}T^n$ has eigenvalue 1 (so the fixed sets are finite and nondegenerate) and the induced map $T^*$ on $H^\bullet(Y,E_1)$ has no eigenvalue 1 (so the suspension bundle is acyclic). Corollary 2.27 then gives both sides as convergent products over eigenvalues of $g^*(T^*)^n$; evaluating $R_g(0)$ and $T_g(\nabla E)^2$ for explicit rotation angles and checking equality tests the theorem directly. For the general statement, the decisive observation is an equivariant isometry meeting every hypothesis of Theorem 2.26 for which identity (2.22) fails for every choice of cutoff functions and Borel section, which would show the condition to be restrictive rather than automatic.
Extended reading notes
Core claim
The central assertion, Theorem 2.26, is that if $g$ lies in a compact subgroup of $G$, the fixed point sets $Y^{g^{-1}T^n}$ are discrete and nondegenerate, and the cutoff compatibility identity (2.22) holds, then the equivariant Ruelle zeta function of the suspension flow satisfies $R_g^{\varphi,\nabla E}(\sigma) = \exp(\sigma\chi_{(g,0)}(\nabla E_1))\,T_g(\nabla E,\sigma^2)^2$ for $\operatorname{Re}\sigma$ large, so that $R_g(0) = T_g(\nabla E)^2$ whenever the equivariant analytic torsion is well-defined. The equality is proved by computing both sides as fixed-point sums over the same set $Y^{g^{-1}T^n}$: the Ruelle side (Theorem 5.9) sums signs and holonomy traces with weights $(1/p(y))\int_{G/Z}\int_0^{p(y)}\psi_M[hy,s]\,ds\,d(hZ)$, while the torsion side (Proposition 6.2, via the fibration formula Theorem 4.2 and the non-compact Atiyah-Bott-Lefschetz fixed-point formula Corollary 3.20) sums identical terms weighted by $\psi_g(y)$; identity (2.22) makes the weights coincide. When $G$ is compact the identity holds with constant cutoffs, the torsion is automatically well-defined, and both sides become explicit cohomological traces, $R_g(\sigma) = \exp(\sum_{n\neq 0} e^{-|n|\sigma}|n|^{-1}\operatorname{Tr}_{H^\bullet(Y,E_1)}((-1)^F g^*(T^*)^n))$ and $T_g(\nabla E) = \exp(\sum_{n\neq 0}(2|n|)^{-1}\operatorname{Tr}_{H^\bullet(Y,E_1)}((-1)^F g^*(T^*)^n))$, from which $R_g(0) = T_g^2$ is immediate.
Load-bearing premise
The general theorem stands on an unproved matching condition, identity (2.22) of the paper: the auxiliary weighting functions (cutoffs) on the base manifold and on the suspension must be choosable so that their fixed-point averages agree at every fixed point of the isometry; if no such choice exists for some equivariant isometry, the general equality between the zeta function and the squared torsion is not established, although the paper does verify the condition in each of its three main corollaries.
Editorial extensions
If this is right
- When $G$ is compact, the equivariant Fried conjecture holds for every suspension flow of an equivariant isometry, with both sides given explicitly by the cohomological traces $\operatorname{Tr}_{H^\bullet(Y,E_1)}((-1)^F g^*(T^*)^n)$; setting $g = e$ recovers the classical Fried conjecture for suspension flow.
- For non-compact $G$ with $g$ in a compact subgroup, the zeta function is determined entirely by the fixed-point data of the isometry $T$: it converges absolutely for $\operatorname{Re}\sigma > c$ and equals the fixed-point sum, so exponential bounds on the sizes of $Y^{g^{-1}T^n}$ translate directly into the domain of convergence.
- The identity-element case for a discrete group is a positive result in a setting where the equivariant Fried conjecture is known to fail in general (through the $L^2$-torsion counterexample), showing the failure requires the group action to interact nontrivially with the flow, not merely that $G$ is non-compact.
- The equality $R_g(\sigma) = \exp(\sigma\chi)T_g(\nabla E,\sigma^2)^2$ holds as an identity of functions in $\sigma$, not just at $\sigma = 0$, so the meromorphic behaviour of the equivariant Ruelle zeta function is governed entirely by the analytic torsion of the suspension.
- In the compact-centraliser case with all fixed sets contained in one uniformly discrete set, the zeta function converges for all $\operatorname{Re}\sigma > 0$, the sharpest possible convergence domain.
Reading between the lines
- The cutoff compatibility identity (2.22) is verified only in the three corollary settings; for the general theorem it is an unproved hypothesis. I read it as a geometric rigidity condition on the $G/Z$-orbits of the fixed points $Y^{g^{-1}T^n}$: it should hold whenever those orbits vary uniformly enough that a product-type cutoff on the suspension averages to the same function as the $G$-cutoff on
- The paper's template, matching a Ruelle fixed-point sum to a torsion fixed-point sum term by term, suggests the equivariant Fried conjecture for any flow whose periodic orbits are parameterised by fixed points of a single diffeomorphism could be attacked the same way; the obstruction for genuinely ergodic flows would be the absence of such a fixed-point parametrisation.
- The identity $R_g(\sigma) = \exp(\sigma\chi)T_g(\nabla E,\sigma^2)^2$ hints at a strengthening of the equivariant Fried conjecture: zeta function and torsion may be related as functions of $\sigma$, with $\sigma\mapsto\sigma^2$ the only substitution needed, making the value at $0$ just the shadow of a functional identity.
- The isometry assumption on $T$ appears to do real work: it trivialises the linearised Poincaré data (Remark 5.4 and Appendix A) and forces fixed sets to be discrete automatically once they are discrete as sets. Dropping it would require controlling linearisations along whole orbits and would likely break the clean term-by-term matching.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proves an equivariant version of the Fried conjecture for the suspension flow of a G-equivariant isometry T on a proper, cocompact Riemannian G-manifold Y, with the flow acting on the mapping torus M=(Y×R)/Z. The central result, Theorem 2.26, states that under hypotheses including g lying in a compact subgroup, discreteness and nondegeneracy of the fixed point sets Y^{g^{-1}T^n}, vanishing of the L² kernel of the relevant Laplacian, and a cutoff compatibility identity (2.22), the equivariant Ruelle zeta function satisfies R_g(σ)=exp(σχ_{(g,0)}(∇E_1)) T_g(∇E,σ²)² for large Re σ, hence R_g(0)=T_g(∇E)². The paper derives unconditional corollaries for compact G, for g with compact centralizer and closed conjugacy class, and for the identity element of a discrete group, and gives explicit formulas for both sides in these cases. The proof uses a fibration formula for equivariant analytic torsion, an Atiyah–Bott type fixed point formula for proper noncompact actions, and a direct computation of the Ruelle zeta function by reducing to fixed point data on Y.
Significance. If the results stand, the paper provides the first general class of equivariant Fried-type theorems beyond the specific examples in the authors' earlier framework, and it explicitly matches the equivariant Ruelle side with the equivariant torsion side for suspension flows. The paper is careful and detailed: the fibration formula in Section 4, the generalized fixed point formula Corollary 3.20, and the explicit formulas (5.15) and (6.5)/(6.6) are nontrivial and appear to be correct in structure. The treatment is honest about the conditional nature of Theorem 2.26 via the cutoff identity (2.22), and the compact and compact-centralizer corollaries genuinely verify that identity in their settings. The main weakness is a concrete advertised example that does not satisfy the paper's own standing assumptions.
major comments (2)
- [Section 2.4 / Example 2.29 and Section 6.5] Example 2.29 does not verify the standing assumption ker(Δ_E)=0, and for the concrete case X=S^n with the trivial flat line bundle that assumption is false. In that case Y=Γ×X and T acts trivially on the Γ factor, so M is the disjoint union over γ∈Γ of copies of the mapping torus M_X. For any nonzero a=(a_γ)∈ℓ²(Γ), the function f([γ,x,t])=a_γ descends to a smooth function on M, is square-integrable because Γ is countable, is locally constant in x and t, and is therefore harmonic. Thus 0≠f∈ker(Δ_E); since Γ is infinite, the kernel is infinite-dimensional. Consequently Example 2.29 is not a valid instance of the equivariant Fried conjecture under the hypotheses of Section 2.4, even if the analytic torsion is well-defined in the sense of Definition 2.3. This does not invalidate Theorem 2.26 or Corollaries 2.27, 2.28, and 2.30, but the advertised noncompact discrete-group application is not established as stated. The example could be repaired by choosing a coefficient system with vanishing L² harmonic forms on X (e.g., an acyclic local system), but that additional hypothesis is absent.
- [Theorem 2.26 / identity (2.22)] The equality in Theorem 2.26 rests exactly on the cutoff compatibility identity (2.22), which matches the Ruelle fixed-point sum from Theorem 5.9 with the torsion fixed-point sum from Proposition 6.2. Since (2.22) is stated as a hypothesis, the theorem is internally consistent; however, the paper should state more prominently that the general theorem is conditional and that, outside Corollaries 2.27, 2.28, and 2.30, no existence criterion for cutoffs satisfying (2.22) is proved. In particular, a reader of the abstract's 'we prove ... in several cases' should be able to see immediately that the general noncompact statement is a conditional result, not an unconditional proof of the equivariant Fried conjecture for arbitrary suspension flows.
minor comments (4)
- [Abstract] The abstract contains a typo: 'maniofold' should be 'manifold'.
- [Corollary 2.30] The phrase 'assumption (II) of Theorem 2.26' is ambiguous: Theorem 2.26 has hypotheses labelled (I) and (II), while Section 2.4 has standing assumptions labelled (I), (II), and (III). Please specify which set of hypotheses is intended.
- [Equation (6.18)] The notation 'Tr(T^{-n}_E |(E1)_y)' in (6.18) is inconsistent with the notation used elsewhere, e.g. 'tr(gT^{-n}_{E1}|_{(E1)_y})' in (5.3) and (5.12). Please unify the notation.
- [Section 2.4, paragraph before Corollary 2.28] The sentence 'This is not a strict assumption, as G=Z_G(e) acts properly on Y^{T^n} which is discrete. Hence there exists a compact subgroup H⊆G such that G/H is discrete' is unclear and the implication is not used later; either justify it or remove it.
Circularity Check
No circular derivation: the main theorem is a conditional comparison of two independently computed fixed-point sums; the main caveat is a correctness gap in Example 2.29, not circularity.
full rationale
The central result, Theorem 2.26, is a conditional comparison theorem. The Ruelle side is computed independently in Theorem 5.9 as a fixed-point sum over Y^{g^{-1}T^n}, while the torsion side is computed independently in Proposition 6.2 via the fibration formula and the Atiyah–Bott fixed-point formula. The equality R_g^{φ,∇E}(σ) = exp(σ χ_{(g,0)}(∇E1)) T_g(∇E, σ^2)^2 is obtained by imposing the cutoff compatibility identity (2.22), which equates the two fixed-point summands. This identity is an explicit hypothesis, not a fitted parameter, and it is verified in the corollaries for the compact, compact-centraliser, and discrete-identity cases. The cited prior work [HS23a], [HS23b], and [HS25] supplies definitions, g-trace criteria, and lemmas on periodicity and discreteness of flow curves; none of these inputs is the target equality, and the key new computations of the Ruelle zeta function and the torsion of the suspension are carried out in this paper. Appendix A explicitly recovers the classical suspension-flow results from [She21], which is presented as a recovery rather than a new prediction. The most definite weakness is Example 2.29: for the concrete sphere/trivial-coefficient case, the standing assumption ker(Δ_E)=0 fails because nonzero L^2 harmonic functions exist on the suspension, as functions constant on each Γ-component of M = Γ × M_X. This is a correctness gap in the example, not circular reasoning, so it does not raise the circularity score.
Assumptions & free parameters
assumptions (6)
- standard math The equivariant Ruelle zeta function and equivariant analytic torsion are well-defined as in [HS23a, HS23b], including the g-trace formalism with cutoff functions.
- domain assumption G is a locally compact unimodular group acting properly, isometrically, cocompactly on Y, with unimodular centralizer Z_G(g), and g lies in a compact subgroup of G.
- domain assumption The L2 kernel of the Hodge Laplacian on the suspension vanishes: ker(Δ_E)=0.
- domain assumption For all n in Z without 0, the fixed point set Y^{g^{-1}T^n} is discrete or empty and consists of nondegenerate fixed points.
- ad hoc to paper The cutoff compatibility identity (2.22) holds for some Borel section and cutoff functions ψ_G, ψ_Y, ψ_M.
- domain assumption Either (I) G/Z is compact and delocalised Novikov-Shubin numbers α_p are positive, or (II) Tr_g(T^n_{E1}P^{E1})=0 for all n and the specific Švarc-Milnor functions with exponential volume growth exist.
Cite this review
Pith. "Pith review of The Equivariant Fried Conjecture for Suspension Flow of an Equivariant Isometry." pith.science (2026). https://pith.science/paper/MRYRR72N
@misc{pith2026250706792,
author = {Pith},
title = {Pith review of: The Equivariant Fried Conjecture for Suspension Flow of an Equivariant Isometry},
year = {2026},
howpublished = {\url{https://pith.science/paper/MRYRR72N}},
note = {Machine review of arXiv:2507.06792}
}
abstract
The Fried conjecture states that the Ruelle dynamical $\zeta$-function of a flow on a compact maniofold has a well-defined value at $0$, whose absolute value equals the Ray-Singer analytic torsion invariant. The first author and Saratchandran proposed an equivariant version of the Fried conjecture for locally compact unimodular groups acting properly, isometrically, and cocompactly on Riemannian manifolds. In this paper we prove the equivariant Fried conjecture for the suspension flow of an equivariant isometry of a Riemannian manifold in several cases. These include the case where the group is compact, the case where the group element in question has compact centraliser and closed conjugacy class, and the case of the identity element of a non-compact discrete group.
Reference graph
Works this paper leans on
-
[349]
issn: 0022-040X. doi: 10.4310/jdg/1453910456 (cit. on p. 29). [Yam22] Yoshikazu Yamaguchi. “Dynamical zeta functions for geodesic flows and the higher-dimensional Reidemeister torsion for Fuchsian groups”. In:Journal für die reine und angewandte Mathematik (Crelles Journal)2022.784 (2022), pp. 155–176. doi: doi:10.1515/crelle-2021-0075 (cit. on p. 3). 65
-
[400]
Heat kernel, large-time behavior, and representation theory
issn: 0003486X, 19398980.doi: 10.2307/1971077 (cit. on p. 3). [She18] Shu Shen. “Analytic torsion, dynamical zeta functions, and the Fried con- jecture”. In:Analysis & PDE11.1 (2018), pp. 1–74.issn: 1948-206X. doi: 10.2140/apde.2018.11.1 (cit. on p. 3). [She21] Shu Shen. “Analytic Torsion and Dynamical Flow: A Survey on the Fried Conjecture”. In:Arithmeti...
work page Pith review arXiv 2018
-
[656]
On the Minimal Sets of Non-Singular Vector Fields
issn: 1944-9992. doi: 10.4310/CAG.2022.v30.n3.a3 (cit. on p. 3). [Wil66] F. Wesley Wilson. “On the Minimal Sets of Non-Singular Vector Fields”. In: Annals of Mathematics84.3 (1966), pp. 529–536.issn: 0003486X, 19398980. doi: 10.2307/1970458 (cit. on p. 3). [Wot08] Artur Wotzke. “Die Ruellesche Zetafunktion und die analytische Torsion hyperbolischer Mannig...
-
[2011]
A Ruelle dynamical zeta function for equivariant flows
isbn: 9783540643241. doi: 10.1007/978-3-662-12494-9 (cit. on p. 9). 62 [Bis11] Jean-Michel Bismut.Hypoelliptic Laplacian and Orbital Integrals. Vol. 177. Annals of Mathematics Studies. Princeton University Press, 2011. isbn: 9780691151298 (cit. on pp. 3, 36). [Car+14] Alan Carey, Victor Gayral, Adam Rennie, and F Sukochev.Index theory for locally compact ...
work page Pith review arXiv 2018
-
[7928]
Heat kernels on covering spaces and topological invariants
doi: 10.1093/imrn/rnab324 (cit. on pp. 7, 22). [Lot92] John Lott. “Heat kernels on covering spaces and topological invariants”. In: Journal of Differential Geometry35.2 (1992), pp. 471–510.doi: 10.4310/ jdg/1214448084 (cit. on pp. 8, 14). 63 [Lot99] John Lott. “Delocalized L2-Invariants”. In:Journal of Functional Analysis 169.1 (1999), pp. 1–31. issn: 002...
arXiv 1992
-
[8708]
R-Torsion and Zeta Functions for Analytic Anosov Flows on 3-Manifolds
doi: 10.1016/0001-8708(71)90045-4 (cit. on p. 8). [Sán96] Héctor Sánchez-Morgado. “R-Torsion and Zeta Functions for Analytic Anosov Flows on 3-Manifolds”. In:Transactions of the American Mathematical Society 348.3 (1996), pp. 963–973.issn: 00029947, 10886850.doi: 10.1090/S0002- 9947-96-01611-X (cit. on p. 3). [Sch74] Paul A. Schweitzer. “Counterexamples t...
Reviewed August 6, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.