{"id":"81e4fe4d-e42b-4524-8291-05b54c543f18","arxiv_id":"1909.00881","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"In the standard effective field theory of gravity, gravitational waves on cosmological backgrounds propagate at a speed differing from unity, and analyticity arguments favor superluminal speed relative to matter.","lead":"This paper calculates that gravitational waves should travel at a very slightly different speed from light on expanding cosmological backgrounds, with the sign depending on quantum corrections from massive particles. The result matters because many cosmological models assume gravitational waves and light travel at exactly the same speed, and that assumption is used to constrain new physics.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The superluminality claim rests on the unproven sign of CW2: positivity bounds require neglecting the t-channel pole and convergent spectral integrals, which Section 4.3.3 concedes may fail; without settling this, Eq. (3.36) does not enforce cs^2 > 1.","rationale":"The reader's weakest_assumption identifies exactly the place where the central claim is least secure: the sign of CW2 is constrained only by applying positivity bounds after neglecting the massless graviton t-channel pole and assuming no subtractions spoil the inequality. The paper is candid about this in Section 4.3.3, saying that with subtractions it is not possible to conclude positivity of the LHS. Since Eq. (3.36) is linear in CW2, the superluminality conclusion is fully controlled by this unresolved sign. The concern is not about internal inconsistency: the calculation of the speed correction itself is well-motivated and carefully defined, and the paper's honest hedging in the bullet and Section 4 justifies a conditional rather than unconditional acceptance. Thus the reader's CONDITIONAL verdict is appropriate, and my stress-test does not move it.","tokens_in":45910,"tokens_out":2243,"duration_ms":23963,"concrete_test":"Take a concrete UV completion with a finite or Kaluza-Klein tower of massive spin-2 modes (e.g., a 5D flat or Randall-Sundrum extra dimension) and compute the full TT spectral function including the massless graviton pole. Check whether the integrals in Eq. (4.37) converge without subtractions and whether the pole-subtracted positivity bound of [27] actually yields CW2 > 0. If subtractions are required, compute the sign of the subtraction constants; a negative CW2 would reverse the sign in Eq. (3.36).","verdict_should_be":"UNCHANGED","load_bearing_attack":"The strongest bullet conclusion is that positivity bounds 'enforce' superluminal gravitational waves, but the derivation depends entirely on CW2 > 0 in Eq. (3.36). That sign is not established. Section 4.2 states the bounds apply only 'provided we argue or assume that the contribution of the graviton exchange t-channel pole can be neglected,' and Section 4.3.3 identifies the real obstruction: the Kallen-Lehmann integrals in (4.37) may not converge, in which case subtractions are needed and 'it is hence not possible to conclude positivity of the LHS' for CW2. The paper's own bullet includes the qualifier 'if the contribution from the massless graviton t-channel pole can be ignored,' so the unconditional phrasing in the abstract and discussion overshoots the derivation. If the spectral integrals require subtractions, or if the t-channel-pole subtraction procedure of [27] fails, CW2 could be negative, reversing Eq. (3.36) to subluminal waves on NEC-preserving backgrounds. This is the single load-bearing assumption: the sign of the Weyl-squared coefficient is what converts a well-computed small correction to the speed into a general superluminality theorem.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the low-energy speed of gravitational waves in the standard effective field theory of general relativity coupled to matter, on backgrounds that spontaneously break Lorentz invariance such as FLRW. The central technical result is Eq. (3.36): on NEC-preserving FLRW backgrounds, the tensor-mode sound speed is c_s^2 = 1 + 16 C_W2 (-\\(\\dot H\\))/M_Pl^2 to leading order in the curvature-squared EFT, so the sign of the Weyl-squared coefficient C_W2 controls whether gravitational waves are superluminal. The paper argues that positivity bounds, applied to matter scattering amplitudes after a field redefinition, select C_W2 > 0 and hence superluminal gravitational waves. It also analyzes dimension-6 and dimension-8 curvature operators: after tuning the IR curvature-squared coefficients to zero, finite one-loop corrections from massive scalars, spinors, and vectors produce an epoch- and species-dependent modification of the speed, given in Eq. (5.10), with superluminal behavior for scalar-dominated contributions over most of standard cosmological history and subluminal behavior in the radiation era. The appendices reproduce the one-loop effective action for a massive scalar and provide the detailed FLRW reduction for dimension-6 operators.","tokens_in":46261,"tokens_out":6291,"duration_ms":78640,"significance":"If the sign assumption on C_W2 is eventually justified, the paper establishes a concrete, potentially observable consequence of a standard gravitational EFT: a tiny but nonzero difference between the gravitational-wave speed and the matter lightcone, scaling as |\\(\\dot H\\)|/\\Lambda^2. The derivation of the speed formula is careful, the field-frame invariance argument (Section 2.2) is a useful clarification, and the explicit one-loop coefficients from Avramidi are reproduced rather than fitted. The dimension-6 result is especially valuable because it is finite, calculable, and predicts a sign that depends on the spin of the lightest integrated-out field, including the interesting statement that radiation-era gravitational waves are subluminal for all three spin species considered. The paper is also honest in its main bullet by including the qualifier that the t-channel-pole contribution is ignored; the problem is that this qualifier is not consistently carried through the abstract and discussion.","major_comments":[{"comment":"The paper's headline claim that positivity bounds 'enforce' superluminal gravitational waves is not established, because the sign of C_W2 is never proven. The speed formula (3.36) gives c_s^2 = 1 + 16 C_W2(-\\dot H)/M_Pl^2, so on NEC backgrounds (\\dot H < 0) the conclusion c_s^2 > 1 requires C_W2 > 0. The arguments for C_W2 > 0 are explicitly conditional: Section 4.2 applies positivity bounds only 'provided we argue or assume that the contribution of the graviton exchange t-channel pole can be neglected', and Section 4.3.3 states that if the Kallen-Lehmann integrals in (4.37) require subtractions, 'it is hence not possible to conclude positivity of the LHS'. No proof of convergence of (4.37), and no independent justification of the t-channel-pole prescription beyond citation to [27], is supplied. The valid conclusion is therefore a conditional theorem, not the generic statement in the abstract and Section 6 that the speed is 'in general superluminal'. The abstract, introduction, and discussion should be revised so that the logical dependence on the unproven t-channel/subtraction assumption is explicit in every statement of the main result.","section":"§4.2, §4.3.3; Eq. (3.36)"},{"comment":"The phenomenological discussion of scalar dark matter in Section 5.1.4 applies the dimension-6 result to 'the whole standard cosmological history', but this application is only valid on the tuned subspace C_IR^{W2} = C_IR^{R2} = 0 introduced at the start of Section 5. If positivity bounds do select C_W2 > 0, the dimension-4 contribution generically dominates the dimension-6 one, and the epoch-dependent conclusions, including the subluminal radiation-era statement, would not hold. The paper partially acknowledges this in the 'Discriminator Redux' paragraph, but the abstract's claim that finite loop corrections 'lead to an epoch dependent modification' should be framed as a statement about a deliberately tuned EFT, not about the generic EFT. The section would benefit from a clear summary stating which conclusions are generic and which require the tuning.","section":"§5.1.4; Eq. (5.11)"}],"minor_comments":[{"comment":"Equation (4.46) contains an apparent typo: the inequality '|1-c_s^2(M_2)| < |1-c_s^2(M_2)|' for M_2 > M_1 is trivially false; the right-hand side should presumably refer to M_1. The surrounding text makes the intended RG monotonicity clear.","section":"§4.3.5; Eq. (4.46)"},{"comment":"The axes of Figure 3 are not fully labeled; the approximate numerical boundaries (-1.8, 0.2, 1.2, -1.5) are given only in the caption. Adding explicit axis labels and marking the NEC boundary \\omega = -1 directly on the figure would improve readability.","section":"Fig. 3"},{"comment":"The effective numbers N_*^s weight fields by M^2/M_i^2, so the statement that the effect is 'determined by the lightest particle' is only true when there is no large multiplicity of heavier fields. This is stated implicitly, but a one-sentence caveat at Eq. (5.2) would prevent a common misreading.","section":"§5.1.2, Eq. (5.2)"},{"comment":"The discussion says that the magnitude of the effect is of order |\\dot H|/\\Lambda^2, which is correct for the dimension-4 case, but the dimension-6 contributions scale as |\\dot H| H^2/(M^2 M_Pl^2) and are suppressed by an additional H^2/M^2 factor. A sentence distinguishing the two parametric regimes would avoid confusion.","section":"§6"}],"recommendation":"major_revision","confidential_remarks":"The paper's central calculation is sound and the conditional statements are mostly internally consistent, but the abstract and discussion oversell the main theorem by omitting the t-channel/subtraction caveat that the authors themselves identify in Section 4.3.3. Since the sign of C_W2 is the single load-bearing input, and the paper concedes that the positivity argument can fail if subtractions are required, the headline should be recast as a conditional result. I would ask the authors to either prove convergence of the spectral integrals (4.37) in a meaningful class of UV completions, or to systematically state every main conclusion with the qualifier. The phenomenological section on scalar dark matter should also be clearly scoped to the tuned C_IR = 0 subspace; otherwise the 'discriminator' language is stronger than the calculation supports."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nTwo things up front. The paper contains real, checkable physics: it derives the correction to the gravitational wave sound speed on FLRW from heavy loops and higher-spin tree effects, including the clean result c_s^2 = 1 + 16 C_W2 (−Hdot)/M_Pl^2, and—more durably—a finite, epoch- and spin-dependent dimension-6 contribution that survives even if the curvature-squared story collapses. But the headline claim, that GWs are in general superluminal, is conditional on the sign of C_W2, and the authors themselves concede the sign is not established. Section 4.3.3: if the Kallen-Lehmann integrals in (4.37) need subtractions, \"it is hence not possible to conclude positivity of the LHS.\" The abstract states the conclusion flatly; the intro bullet carries the qualifier about the t-channel pole.\n\nCredit where due. The speed derivation is careful, especially the discussion in Section 3.1 of why you cannot read the speed off a naive WKB dispersion relation. The frame-independence argument—the ratio of tensor to matter speed is invariant under metric field redefinitions—is clean and should reset how people phrase subluminality demands in cosmological EFTs. The dimension-6 one-loop coefficients are taken from Avramidi, reproduced in Appendix A, finite, and genuinely new in their FLRW application: the sign of the correction depends on the spin of the lightest integrated-out particle and on the equation of state. That part does not rely on positivity bounds and is the paper's most robust contribution.\n\nThe soft spots, in proportion. The load-bearing assumption is C_W2 > 0. Getting there requires discarding the massless graviton t-channel pole (following [27]) and assuming the spectral integrals converge. Neither is proven; the paper says so. If either fails, the sign of C_W2 is unconstrained and Eq. (3.36) does not enforce superluminality—indeed a negative C_W2 would give subluminal GWs on NEC-preserving backgrounds. So the general claim is a conditional statement, not a theorem. The authors are honest about this, which is why the paper still warrants refereeing, but the abstract and discussion should carry the same caveat as the intro bullet. The axion/fuzzy-dark-matter speculation is minor; it is hedged and observationally remote anyway.\n\nWho benefits: anyone doing EFT model building in cosmology, especially work that assumes tensor modes must be subluminal. It deserves a serious referee. I would send it out, and instruct the referee to focus on the t-channel/subtraction question and on aligning the abstract with the qualifier.\n\nBest,\n\n[you]","headline":"A carefully derived EFT calculation of gravitational wave speed corrections on FLRW whose superluminality headline is honestly conditional on the unproven sign of C_W2; the finite dimension-6 part is the most durable piece.","tokens_in":46742,"tokens_out":3281,"would_cite":true,"duration_ms":35012,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Gravitational waves slightly outrun light on cosmological backgrounds, this paper argues.","keywords":["gravitational waves","effective field theory","speed of gravitational waves","superluminality","positivity bounds","null energy condition","curvature-squared operators","cosmological backgrounds"],"falsifier":"In an explicit weakly coupled UV completion containing an infinite tower of massive spin-2 states, compute the full two-to-two amplitude without dropping the massless graviton t-channel pole; if the twice-subtracted forward amplitude does not force the Weyl-squared coefficient to be positive, the sign conclusion fails. A complementary observation would be a cosmological multi-messenger event whose gravitational-wave and photon arrival times imply a subluminal tensor speed in the matter frame.","tokens_in":45712,"feed_emoji":"🌌","tokens_out":13811,"duration_ms":145404,"temperature":0.7,"pith_summary":"Within the standard effective field theory of General Relativity, this paper tries to establish that the low-energy speed of gravitational waves on a cosmological background is generally not equal to the speed of light. Integrating out massive fields generates curvature corrections, and on an FLRW spacetime the leading correction is fixed by the Weyl-squared coefficient $C_{W^2}$: the tensor speed becomes $c_s^2 = 1 + 16 C_{W^2}(-\\dot H)/M_{\\rm Pl}^2$, which exceeds one when $C_{W^2}>0$ and matter satisfies the null energy condition, so $\\dot H<0$. The paper argues that S-matrix locality, unitarity and analyticity—through positivity bounds applied to matter scattering after the massless graviton t-channel pole is neglected—select this positive sign, so gravitational waves are superluminal relative to the metric to which photons and Standard Model fields are minimally coupled. If the curvature-squared terms are tuned to zero, finite one-loop contributions from known particles still shift the speed, making it epoch-dependent and sensitive to the spin of the lightest particles above the Hubble scale. A sympathetic reader should care because this changes how the causal structure of cosmological effective field theories is to be assessed.","feed_headline":"Gravitational waves may slightly outrun light","feed_subtitle":"Loop corrections from known particles would make gravitational waves slightly faster than light, if standard quantum constraints apply.","key_machinery":"The central object is the low-energy effective action for gravity organized in curvature operators, truncated at the Weyl-squared (dimension-4) and curvature-cubed (dimension-6) level. The speed is defined through the lightcone of the second-order hyperbolic equation obtained after perturbatively reducing the higher-derivative terms; it is the coefficient of $k^2$ in the tensor dispersion relation. At leading order the Weyl-squared coefficient $C_{W^2}$, the coupling of the conformal curvature-squared operator, carries the whole effect via $c_s^2 = 1 + 16 C_{W^2}(-\\dot H)/M_{\\rm Pl}^2$. Positivity bounds on matter amplitudes, together with the Källén–Lehmann spectral representation of the two-point function of the stress tensor, are the machinery that fixes $C_{W^2}>0$; the assumption of neglecting the massless graviton t-channel pole is what lets those bounds go through. At next order the finite one-loop coefficients for fields of spin 0, 1/2, and 1 supply the species-dependent terms that make the speed epoch-dependent.","core_discovery":"The paper's central claim is that in the effective field theory of gravity the sound speed of tensor modes on a spontaneously Lorentz-breaking background is modified by irrelevant curvature operators, and for the signs selected by positivity bounds the modification is generically superluminal. On FLRW spacetime the Weyl-squared interaction gives $c_s^2 = 1 + 16 C_{W^2}(-\\dot H)/M_{\\rm Pl}^2$; since ordinary matter obeys the null energy condition and hence $\\dot H<0$, a positive $C_{W^2}$ makes gravitational waves outpace the lightcone of the metric to which Standard Model fields are minimally coupled. A positive $C_{W^2}$ is what follows from Källén–Lehmann spectral positivity of the stress-tensor two-point function and from weakly coupled tree-level completions with massive spin-2 states, provided the massless graviton t-channel pole in the relevant amplitudes can be neglected. When the leading curvature-squared terms are set to zero, the finite one-loop dimension-6 effective action produces a speed shift of order $H^4/(M_{\\rm Pl}^2 M^2)$ whose sign depends on the spin and mass of the lightest integrated-out particles: scalar-dominated content gives superluminal gravitational waves for most of standard cosmological history, while the radiation era is subluminal for all spins.","pith_inferences":["One testable extension: the predicted speed shift is frequency-dependent and grows in the IR, so a multi-messenger gravitational-wave event at cosmological distance could in principle measure the sign of the effect and thereby test $C_{W^2}>0$ in the matter frame.","If positivity bounds are invalidated by the t-channel pole, the leading-order sign is unknown, but the finite dimension-6 loop effect remains; the same measurement strategy could still extract the spin and mass content of the lightest states above the Hubble scale.","The lightcone ordering implied by the paper (matter cone inside gravity cone) reverses for NEC-violating sources, so these sign arguments could sharpen consistency conditions on phantom dark energy or other negative-energy cosmological models."],"forward_implications":["If the central claim survives, low-frequency gravitational waves on a cosmological background arrive slightly earlier than photons emitted from the same source, with the departure growing toward the infrared.","No field redefinition can make both sectors luminal at low energies: the ratio of tensor to matter sound speeds is frame invariant, so a frame with luminal gravity leaves matter fluctuations subluminal through gravitationally induced $T\\bar T$-type interactions.","The front velocity is luminal in the high-frequency limit, so the low-energy superluminal group velocity does not imply propagation of information outside the lightcone.","Setting the curvature-squared corrections to zero does not restore luminality; the finite dimension-6 loop corrections still shift the speed, and scalar-dominated heavy spectra make gravitational waves superluminal through most of standard cosmological history.","Cosmological model builders should not impose subluminality of all fluctuations as a consistency criterion; the operative causality criterion becomes S-matrix analyticity, which for this system selects superluminal gravitational waves."],"supporting_citations":[{"why":"Key assumption: the massless graviton t-channel pole can be neglected, allowing positivity bounds to fix the sign of the Weyl-squared coefficient.","marker":"[27]"},{"why":"Supplies the forward-limit S-matrix positivity bounds used to constrain the coefficients of matter interactions after field redefinitions.","marker":"[10]"},{"why":"Represents the contrasting causality constraint demanding subluminality of higher-curvature corrections, which this paper argues is inverted once positivity bounds are applied.","marker":"[11]"},{"why":"Shows that low-energy superluminal photon speeds from QED loops are compatible with causality, the analogue used to interpret the gravitational result.","marker":"[30]"},{"why":"Source of the one-loop dimension-6 effective action coefficients for spin-0, spin-1/2 and spin-1 fields that determine the next-to-leading speed correction.","marker":"[80]"},{"why":"Offers a complementary positivity argument based on graviton pseudo-amplitudes that also sets the Weyl-squared coefficient positive while neglecting the massless graviton.","marker":"[28]"},{"why":"Provides the Källén–Lehmann spectral representation of the stress-tensor two-point function whose positive spectral densities imply positive contributions to the curvature-squared coefficients.","marker":"[74]"},{"why":"Computes the covariant one-loop effective action from which the dimension-6 operator coefficients for all spins are drawn.","marker":"[52]"}],"fun_headline_variants":["Gravitational waves may outpace light, says new theory","Loop effects in gravity could make waves faster than light","Speed of gravity: tiny deviations from light speed","Gravity waves might travel marginally faster than light","Quantum correction: gravity waves slightly superluminal"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument assumes that standard positivity bounds apply to matter scattering once the massless graviton t-channel pole is neglected, and that the Källén–Lehmann spectral integrals need at most one subtraction; if either premise fails, the sign of $C_{W^2}$ is unconstrained and the generic superluminality conclusion does not follow.","fun_headline_variants_meta":{"raw":{"variants":["Gravitational waves may outpace light, says new theory","Loop effects in gravity could make waves faster than light","Speed of gravity: tiny deviations from light speed","Gravity waves might travel marginally faster than light","Quantum correction: gravity waves slightly superluminal"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000448,"raw_usage":{"total_tokens":2311,"prompt_tokens":1046,"completion_tokens":1265,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":662,"completion_tokens_details":{"reasoning_tokens":1190}},"tokens_in":662,"tokens_out":1265,"duration_ms":10720,"temperature":1.0,"reasoning_tokens":1190,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T05:33:41.924275+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"In an explicit weakly coupled UV completion containing an infinite tower of massive spin-2 states, compute the full two-to-two amplitude without dropping the massless graviton t-channel pole; if the twice-subtracted forward amplitude does not force the Weyl-squared coefficient to be positive, the sign conclusion fails. A complementary observation would be a cosmological multi-messenger event whose gravitational-wave and photon arrival times imply a subluminal tensor speed in the matter frame.","supporting_citations":[{"cited_title":"Covariant methods for the calculation of the effective action in quantum field theory and investigation of higher-derivative quantum gravity","cited_arxiv_id":"hep-th/9510140","evidence_quote":"Source of the one-loop dimension-6 effective action coefficients for spin-0, spin-1/2 and spin-1 fields that determine the next-to-leading speed correction."},{"cited_title":"A Weak Gravity Theorem","cited_arxiv_id":"1905.02736","evidence_quote":"Offers a complementary positivity argument based on graviton pseudo-amplitudes that also sets the Weyl-squared coefficient positive while neglecting the massless graviton."}],"review_version":1}