{"id":"79bddac3-92f0-4266-af1e-6eb291387ec0","arxiv_id":"2602.12802","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"For the most general parity-symmetric four-dimensional scalar-tensor EFT with four-derivative terms, stationary (including rotating) cone-defined black hole horizons are shown, under explicit technical assumptions, to be metric Killing horizons where all three polarisation cones touch.","lead":"Two new theorems show that stationary, including rotating, black hole horizons in a broad class of higher-derivative scalar-tensor gravity theories — defined by the fastest wave cone — are still Killing horizons of the spacetime metric, where all propagation cones touch. This matters because wave polarizations that travel on different cones seemed to threaten the usual laws of black hole thermodynamics.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The physical reading depends on the unproven identification of the nonlinear event horizon with a Q-characteristic surface (Sec. II.B); this supports the CONDITIONAL verdict but does not sink the proof.","rationale":"Read in good faith: the paper proves nontrivial conditional theorems with a detailed order-by-order induction and explicitly discloses its main limitations. The reader's weakest assumption is exactly right: the physical interpretation of both theorems requires that the domain-of-dependence boundary of the full nonlinear EFT be a characteristic surface of the linearized Q-cone, and this is left open in Sec. II.B. My stress-test found no internal inconsistency or circular step sufficient to reject. The factor-of-two spot-check in (D39a) vs (D48) is not damaging because the sign contradiction forces both L_{k(1)}h0 and L_{k(1)}Φ0 to vanish regardless of the numerical prefactor. Assumption (v) of Theorem A is indeed close in spirit to the conclusion, but the theorem's substantive content is the derivation of nullness/normality and the triple-cone touching; it is not circular. The most important unsupported premise is the nonlinear identification, and it is the authors' own stated open problem. This warrants a conditional verdict: accept the theorems as conditional on (iii)/(iii'), the small-ℓ analyticity framework, and the additional symmetry assumptions, but do not accept the abstract's unconditional phrasing about 'the horizon of a stationary black hole' as a proven statement. No verdict change is needed.","tokens_in":34983,"tokens_out":9052,"duration_ms":86041,"concrete_test":"Compute the full nonlinear principal symbol for the gauge-fixed EFT of [21] and determine whether the boundary of the domain of dependence of the nonlinear system is necessarily ruled by bicharacteristics of Q. Concretely: perform a second-order progressing-wave expansion of (5)-(6) around a weakly coupled stationary solution with |ℓ∇Φ| not small; derive the eikonal equation for the leading discontinuity. If the eikonal equation is Q(ξ)=0 to all orders in ℓ, assumption (iii) is a theorem for small-ℓ solutions; if an explicit family of solutions yields a different front speed (e.g., because of the ℓ²β'R∇²Φ term), the physical reading fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The single most load-bearing concern is the gap between the mathematical objects in Theorems III.1/IV.1 and the physical horizon advertised in the Abstract. Assumption (iii)/(iii') posits that H+(ℓ) is a characteristic surface of the linearized quartic Garding cone defined by Q in Eq. (12a), and the proof shows that such a surface is a Killing horizon. But the EFT equations (5)-(6) are not quasi-linear, and the authors state in Sec. II.B: 'It seems plausible to us that this will remain true even in the full non-linear EFT (4), at least in the weakly coupled case... We leave this as an open problem.' The boundary of the domain of dependence of the nonlinear system need not coincide with the Q-cone unless this is proved; for non-quasi-linear systems, the propagation of discontinuities can be amplitude-dependent, so the linearized principal symbol (10) may not control the causal structure of strong-field solutions. If this identification fails, the theorems establish only that a certain Q-characteristic surface is a Killing horizon, not that the actual event horizon is one, and the thermodynamic conclusions would not follow. This is a genuine open problem, not a contradiction; the authors' disclosure is explicit. A secondary fragility, also noted by the authors, is Theorem A's assumption (v): the bicharacteristics are assumed collinear with a Killing field, which is close to the rigidity content of the conclusion; this narrows the theorem but is not circular. I checked the factor-of-two discrepancy between (D39a) and (D48): it is harmless, since both sides are nonnegative/nonpositive and force L_{k(1)}h0 = L_{k(1)}Φ0 = 0 independent of the factor.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the most general parity-symmetric scalar-tensor EFT in four dimensions with up to four-derivative terms, whose linearized characteristic structure comprises three nested propagation cones (one quadratic, one quartic with two sheets). The authors define the black-hole horizon as the boundary of the causal past with respect to the fastest (quartic Garding) cone. The main results are Theorem III.1 and Theorem IV.1: under analyticity in the EFT parameter ℓ, stationarity, and further technical assumptions, a characteristic horizon H+(ℓ) is in fact a metric-null Killing horizon with respect to g_ab, and Corollary III.2 adds that all propagation cones touch there. The proofs proceed by an order-by-order induction in ℓ using generalized Gaussian null coordinates and explicit curvature decompositions. The stated upshot is that the usual Killing-horizon framework and black-hole thermodynamics survive in this EFT class, and the nested-cone perpetual-motion paradoxes are evaded.","tokens_in":35255,"tokens_out":4458,"duration_ms":46131,"significance":"If the theorems are correct, this is a substantial step: it extends Reall's analysis of the characteristic cone structure of scalar-tensor EFTs and shows that, under stated assumptions, the fastest-cone horizon is compatible with the standard metric Killing-horizon formalism. The paper is technically serious: the proof is a detailed order-by-order induction with explicit gGNC curvature formulas, and the assumptions are stated with unusual clarity, including the authors' own identification of the most restrictive hypotheses. The conclusions, if they apply to actual black-hole horizons, would safeguard zeroth-law and thermodynamic reasoning in a wide class of higher-derivative scalar-tensor theories. However, the advertised physical conclusion is conditional on an unproven identification between the nonlinear event horizon and a characteristic surface of the linearized principal symbol, and Theorem A assumes a nontrivial part of the rigidity content it aims to prove.","major_comments":[{"comment":"The theorems assume that H+(ℓ) is characteristic with respect to the linearized quartic cone defined by Q in Eq. (12a). The paper itself states in Sec. II.B: 'It seems plausible to us that this will remain true even in the full non-linear EFT (4), at least in the weakly coupled case... We leave this as an open problem.' Therefore the Abstract's claim that 'the horizon of a stationary black hole' is metric-null is not established for the genuine nonlinear event horizon; it is established only for Q-characteristic hypersurfaces. This gap is load-bearing for the thermodynamic conclusions and should be addressed by either proving the identification for the nonlinear system or explicitly restricting all physical claims to characteristic horizons.","section":"Abstract; Sec. II.B; Assumptions (iii)/(iii')"},{"comment":"Assumption (v) requires the bicharacteristic generators of H+ to be collinear with a Killing vector field χ^a. In standard black-hole rigidity, this collinearity is a theorem following from stationarity and horizon structure; here it is an input that is very close to the desired conclusion. Thus Theorem III.1 proves that a surface whose generators are already aligned with a Killing field is null with respect to g_ab and hence a Killing horizon. This is a useful result, but it is weaker than the 'vice versa' advertised in the Introduction, and the paper should clearly distinguish the assumed rigidity content from what is derived.","section":"Theorem A, assumption (v) (Sec. III.A)"},{"comment":"In the proof of Theorem B, the rotational Killing field ϕ^a is extended from the ℓ=0 solution to all small ℓ using the arguments of [23]. However, [23] concerns horizons that are null with respect to the metric g_ab, whereas H+(ℓ) is only assumed to be a Garding-cone characteristic surface. The footnote says one may use metric lightcones because the Killing field extends, but the logical steps are not given in detail. As written, this is a potential gap in the derivation of the second symmetry; it should be clarified how the extension is obtained without presupposing the conclusion of the theorem.","section":"Sec. IV.C, extension of ϕ^a"}],"minor_comments":[{"comment":"Typo: 'asstationary' should be 'a stationary'; also 'goverened' in the opening sentence of the full text.","section":"Abstract"},{"comment":"Footnote 7 has 'corollary7' with a missing space; please fix the typesetting.","section":"Sec. II.E"},{"comment":"In assumption (v), the phrase 'and either c=0 or c≠0 on all of H+(ℓ)' is slightly awkward; since c is a function, consider stating 'c is either identically zero or nowhere zero on H+(ℓ)', which is what the subsequent argument uses.","section":"Sec. III.A"},{"comment":"The discussion of the 't−φ' reflection symmetry in Theorem B would benefit from a brief explicit example or a pointer to the appendix where the coordinate form (E1) is derived, since the covariant condition (iii') is somewhat hard to parse on first reading.","section":"Sec. V"}],"recommendation":"major_revision","confidential_remarks":"The mathematical core appears sound and the paper is a serious contribution. The main issue is the gap between the characteristic-horizon theorems and the physical black-hole horizon advertised in the Abstract; this is explicitly acknowledged by the authors but needs to be resolved in framing. I would also ask the authors to check the [23] extension argument in Theorem B carefully; if it is valid, a few explanatory sentences would help."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The genuinely new thing here is the converse direction: for stationary (including rotating) black hole solutions in the general parity-symmetric four-derivative scalar-tensor EFT, a horizon defined by the fastest (quartic) propagation cone is shown to be a metric-null Killing horizon, with all three cones touching at the horizon. That extends the earlier static, spherically symmetric Horndeski result of Benkel et al. to a much broader setting, and it is what would rule out the nested-cone perpetual motion scenarios from Dubovsky-Sibiryakov and Jacobson-Wall. If the theorems hold, the thermodynamic framework survives intact.\n\nThe paper does this properly. The two theorems are conditional and the conditions are stated up front: Theorem A assumes bicharacteristic generators are collinear with a Killing field and β′≠0; Theorem B uses a t-ϕ reflection symmetry and V≥0. The proof is a detailed order-by-order induction in the EFT scale using generalized Gaussian null coordinates. I spot-checked the key lemmas in Appendix D, including the base case and the sign argument that forces k(1) to be a Killing vector of the horizon metric. The logic hangs together, and the factor-of-two discrepancy the reader flagged between (D39a) and (D48) is indeed harmless: both sides are nonnegative/nonpositive, so the conclusion is unaffected. The authors also disclose every technical assumption and identify which steps are skipped with 'can be verified analogously.'\n\nThe soft spot is the one the stress test highlights, and it is real. The physical reading of the abstract depends on identifying the actual event horizon of the nonlinear theory with a characteristic surface of the linearized quartic cone. That identification is an explicit assumption (iii)/(iii′) in the theorems. For a quasi-linear system this would be standard, but these equations are not quasi-linear, and the authors themselves say in Sec. II.B that it is plausible and leave it as an open problem. If that premise fails, the theorems prove something about Q-characteristic surfaces, not necessarily about the horizon of a stationary black hole. This is not a flaw in the proof; it is a gap between the mathematics and the advertised conclusion, and the paper is honest about it. The abstract could be read as overstating what is proven.\n\nTheorem A's assumption (v) is also close in spirit to the conclusion, as the authors admit. It narrows the result but is not circular. The analyticity, small-ℓ, and symmetry restrictions are all stated.\n\nWho is this for? People working on causal structure and black hole thermodynamics in modified gravity / EFT. It deserves a serious referee. My recommendation: send it to peer review, and require the authors to either prove or visibly flag the nonlinear characteristic-surface identification as an open condition in the main theorems, and soften the abstract accordingly. As is, it is a solid conditional result, not a full resolution of the paradoxes.\n\nA serious editor should not desk reject it. It earns referee time.","headline":"Two careful converse theorems showing cone-defined stationary horizons in a broad scalar-tensor EFT are Killing horizons; physically compelling, but the bridge from the linearized propagation cone to the real nonlinear event horizon is assumed, not proved.","tokens_in":35902,"tokens_out":1924,"would_cite":true,"duration_ms":21293,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"pith_extraction":{"msc":["83C57","83D05"],"pacs":["04.50.Kd","04.70.Bw"],"model":"deepseek-v4-flash","headline":"Two theorems show that in a broad scalar-tensor EFT, the horizon of a stationary black hole—defined by the fastest propagation cone—is still a Killing horizon with respect to the metric.","keywords":["effective field theory","scalar-tensor gravity","propagation cone","characteristic surface","Killing horizon","black hole thermodynamics","Gauss-Bonnet","higher-derivative gravity"],"falsifier":"Construct a stationary, analytic, asymptotically flat solution of the EFT (4) with arbitrarily small nonzero ℓ whose fastest-cone horizon is not null with respect to the metric, or show that the domain-of-dependence boundary of the full nonlinear system is not Q-characteristic; either would invalidate the theorems' interpretation as statements about real black hole horizons.","tokens_in":34695,"feed_emoji":"🕳️","tokens_out":6139,"duration_ms":52186,"temperature":0.7,"pith_summary":"In an effective field theory of gravity with higher-derivative terms, different polarizations of the fields can travel on different, nested propagation cones, so the black hole horizon is usually tied to the fastest cone rather than to the metric. The paper establishes two theorems for the most general parity-symmetric scalar-tensor EFT in four dimensions with up to fourth-derivative terms: under stated assumptions, the horizon of a stationary (possibly rotating) black hole is null with respect to the metric, and all three propagation cones touch there. This means the Killing-horizon framework—surface gravity, temperature, and the laws of black hole mechanics—keeps its status in this class of theories, and the nested-cone thermodynamic paradoxes discussed in the literature are evaded. The authors emphasize that the results are conditional on smallness of the EFT length scale and on technical assumptions, one of which they flag as the essential nontrivial premise.","feed_headline":"Black hole horizons stay Killing when propagation cones split","feed_subtitle":"Two theorems show all three polarization cones touch the horizon, saving black hole thermodynamics.","key_machinery":"The proof runs order by order in the EFT length scale ℓ around a known ℓ=0 Einstein-scalar black hole, using generalized Gaussian null coordinates adapted to a candidate Killing field. The load-bearing object is the characteristic cone of the linearized principal symbol, whose quartic sheet Q(ξ)=0 defines the fastest (Garding) cone and hence the horizon, and whose quadratic sheet (C^{-1})^{ab}ξ_aξ_b=0 is the effective-metric cone. The key identities evaluate the contractions of Q and C^{-1} on the horizon normal; at each order they force the gGNC functions f and k_a—which measure whether H+ is null—to vanish, after redefining the Killing field by absorbing constant rotations ω_j. A final ana","core_discovery":"The central claim is a converse to a known observation. A metric-null Killing horizon automatically makes the Killing field a triple direction of the characteristic cone, so all propagation sheets touch there; the paper proves the reverse statement for the parity-symmetric scalar-tensor EFT with a Lagrangian containing the Ricci scalar, a scalar kinetic term, an X^2 term, and a topological quadratic curvature term at order ℓ^2. Theorem III.1 says that, under analyticity, non-degeneracy at ℓ=0, bicharacteristic ruling, and an assumed Killing symmetry aligned with the bicharacteristics, the horizon H+ is an ordinary Killing horizon with constant surface gravity for sufficiently small |ℓ|. Theo","pith_inferences":["One extension the authors do not claim: if a future proof establishes that the boundary of the domain of dependence of the full nonlinear EFT is Q-characteristic, then the theorems would automatically apply to any asymptotically flat stationary black hole without the explicit characteristic-surface assumption.","A natural next step would be to generalize the same order-by-order mechanism to multiple scalar fields and abelian vector fields at the same derivative order; the quartic-cone structure may remain the fastest, so the core identities could survive.","The alignment condition in Theorem A (bicharacteristics collinear with a Killing field) is close to the conclusion; replacing it by global assumptions such as asymptotic flatness and analyticity is the open direction the authors themselves identify.","One could numerically evolve the full nonlinear system with small ℓ on a stationary rotating background to test whether the fastest-cone horizon remains null beyond the linearized approximation."],"forward_implications":["If Theorem III.1 is correct, any stationary, real-analytic black hole in the EFT class that satisfies the bicharacteristic-alignment assumption has a metric-null horizon with constant surface gravity, so the zeroth law of black hole mechanics applies.","If Theorem IV.1 is correct, rotating black holes with a t-φ reflection symmetry and V≥0 also have compactly generated Killing horizons, extending the conclusion beyond the aligned-bicharacteristic case.","Corollary III.2 implies that the fastest and slower polarizations all travel at the same effective speed along the horizon direction, removing the distinction between propagation-cone horizon and metric Killing horizon at the horizon.","The small-ℓ theorem family indicates that the nested-cone 'perpetual motion' scenarios are obstructed for stationary black holes in this theory, because the outer cone horizon coincides with the Killing horizon.","The results support the usual thermodynamic assignment of temperature and entropy to stationary black holes in this EFT, since those quantities are defined on Killing horizons."],"fun_headline_variants":["Split cones don't break Killing horizons for stationary black holes","Theorems show black hole horizons remain Killing despite cone splitting","Even with three propagation cones, black hole horizons stay Killing","Killing horizons persist in EFT with multiple propagation cones","New theorems: Black hole horizons stay metric-null when cones split"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The load-bearing premise is that the physical event horizon coincides with a characteristic surface of the fastest quartic propagation cone built from the linearized equations; the authors explicitly flag this as plausible but unproved for the full nonlinear EFT.","fun_headline_variants_meta":{"raw":{"variants":["Split cones don't break Killing horizons for stationary black holes","Theorems show black hole horizons remain Killing despite cone splitting","Even with three propagation cones, black hole horizons stay Killing","Killing horizons persist in EFT with multiple propagation cones","New theorems: Black hole horizons stay metric-null when cones split"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000167,"raw_usage":{"total_tokens":1073,"prompt_tokens":701,"completion_tokens":372,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":445,"completion_tokens_details":{"reasoning_tokens":288}},"tokens_in":445,"tokens_out":372,"duration_ms":4112,"temperature":1.0,"reasoning_tokens":288,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T23:44:20.676096+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Construct a stationary, analytic, asymptotically flat solution of the EFT (4) with arbitrarily small nonzero ℓ whose fastest-cone horizon is not null with respect to the metric, or show that the domain-of-dependence boundary of the full nonlinear system is not Q-characteristic; either would invalidate the theorems' interpretation as statements about real black hole horizons.","supporting_citations":[],"review_version":1}