{"id":"3b6d8574-a293-41e9-9c46-84ac2a506a82","arxiv_id":"1908.08644","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For a c_T=1 beyond-Horndeski EFT of cosmological perturbations, positivity bounds including H^2/Λ^2 corrections are derived and applied to slow-roll inflation.","lead":"A theory paper derives constraints on a cosmological effective field theory by requiring that a hidden high-energy completion is unitary, causal and local. The bounds shift when the expanding universe is included, and can be stronger or weaker than the usual flat-space versions.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The positivity bounds rest on an assumed amplitude decomposition whose neglected cosmological term is parametrically larger than the retained H^2/Λ^2 corrections.","rationale":"The reader's weakest_assumption identifies the same central issue: the unproven amplitude decomposition (20) and the neglect of A_cos. My stress-test sharpens this by noting the parametric hierarchy: A_cos is O(H^2/M^2), and since the EFT requires M^2 ≪ Λ^2, this correction is larger than the H^2/Λ^2 terms that the bounds claim to incorporate. This makes the dropped term potentially the dominant correction, so the argument 'δ(0) dominates ρ(0)' is not sufficient to validate (22) and (23). The paper is honest in stating that Eq. (20) is assumed, and the algebraic derivation of the Goldstone Lagrangian and amplitude appears internally consistent, so the appropriate verdict is CONDITIONAL: the central bounds are conditional on a nontrivial unproven assumption, and the abstract's 'derive' language is too strong. No independent verification of the heavy algebra in Appendices B and C was performed, consistent with the reader's moderate confidence. The proposed concrete test would settle whether the neglected A_cos indeed contaminates the dispersion relation at the relevant order.","tokens_in":17624,"tokens_out":3022,"duration_ms":34949,"concrete_test":"Use the explicit form of the correction term in Eq. (19) — the incomplete gamma function Γ(0,−i|p1+p2−k0|τi) — to model ρ(ΔE). Insert the full amplitude (20) into the Cauchy integral (12) with a finite-time regulator, and compute the correction to A''(µ²) at fixed t. If the A_cos contribution is O(H^2/M^2) and not suppressed relative to the retained H^2/Λ^2 terms for M ≪ Λ, then the bounds (22) and (23) are not justified. Alternatively, repeat the positivity derivation in a simple UV-complete toy model (e.g., a massive scalar with a φ^4 interaction in a fixed de Sitter background) using exact mode functions; if the resulting amplitude violates (14) at order H^2/M^2, the assumed decomposition fails.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central bounds (22) and (23) are obtained by applying Minkowski dispersion relations to the A_min part of Eq. (20), while dropping A_cos with the argument that δ(0) dominates ρ(ΔE=0). The paper explicitly states that Eq. (20) is assumed, not derived. The load-bearing problem is not merely that the decomposition is unproven, but that A_cos is estimated as O(H^2/M^2), where M is the scattering energy satisfying M^2 ≪ Λ^2. Hence H^2/M^2 ≫ H^2/Λ^2: the neglected term is parametrically larger than the very H^2/Λ^2 corrections the bounds purport to include. The dominance claim about δ(0) versus ρ is not a derivation: in any regulated finite-volume treatment, δ(0) is a large but finite volume factor, and ρ(ΔE) is a nontrivial distribution whose contribution to the Cauchy integrals (12)–(13) can affect A''(s) at the same or larger order. Since ρ has support at nonzero ΔE and may introduce branch cuts or other non-analyticities beyond those of A_min, the inequalities (22) and (23) do not follow from unitarity, causality, and locality unless Eq. (20) is justified and the A_cos contribution to the dispersion integral is shown to be negligible. The abstract's 'we derive such bounds' overstates what the body concedes is an assumed extraction.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper aims to extend positivity bounds from flat-space EFTs to the EFT of cosmological perturbations. The authors consider a shift-symmetric, c_T=1 beyond-Horndeski theory on an FRW background, derive the Goldstone Lagrangian (11) and the tree-level 2→2 amplitude (21), and then propose a decomposition of the cosmological amplitude into a Minkowski-like piece A_min and a cosmological correction A_cos (Eq. (20)). Applying standard Minkowski dispersion relations to A_min yields the inequalities (22) and (23), which contain O(H^2/Λ^2) corrections. The paper discusses applications to slow-roll inflation and the GR/Galileon limit, and concludes that the cosmological bounds can be either stronger or weaker than their flat-space counterparts. Crucially, the authors explicitly state that Eq. (20) is assumed, not derived, and that the positivity argument is applied only to A_min.","tokens_in":17925,"tokens_out":5066,"duration_ms":46275,"significance":"The goal of obtaining positivity constraints on the EFT of cosmological perturbations is important and timely, and the explicit computation of the Goldstone Lagrangian and tree-level amplitude for a c_T=1 beyond-Horndeski theory is a useful technical contribution. The paper is also unusually candid about the assumptions that underpin its central result. However, as it stands, the claimed derivation of the bounds is conditional on an unproven amplitude decomposition and on neglecting a cosmological term that is parametrically larger than the retained corrections. If the assumptions (18)–(20) were justified, the bounds could provide nontrivial constraints on beyond-Horndeski models; but the paper does not supply that justification, so the central claim is not established in its current form.","major_comments":[{"comment":"The decomposition iA = iA_min δ^(4)(Σp) + iA_cos ρ(ΔE) δ^(3)(Σp) is explicitly assumed rather than derived, as the authors state. The positivity argument is then applied only to A_min, and the inequalities (22) and (23) are extracted from it. However, the Cauchy integrals in (12)–(13) require analyticity and boundedness of the full amplitude, not just of A_min. If A_cos has branch cuts, poles, or other non-analyticities in the complex s-plane, its contribution to the dispersion integral will not cancel and can alter A''(s) at the values used for the bounds. Therefore, until (20) is justified and the A_cos contribution to (12)–(13) is shown to be negligible, the bounds (22) and (23) do not follow from unitarity, causality, and locality alone.","section":"Section III-B, Eq. (20)"},{"comment":"The neglect of A_cos is justified by the claim that δ(0) dominates ρ(ΔE = 0). This is not a sufficient argument. A_cos is estimated to be of order H^2/M^2, where M is the scattering energy with M^2 ≪ Λ^2, whereas the cosmological corrections explicitly retained in (22) and (23) are of order H^2/Λ^2. Since H^2/M^2 ≫ H^2/Λ^2, the neglected contribution is parametrically larger than the terms the bounds claim to compute. In a finite-volume regularization, δ(0) is a large volume factor and ρ(ΔE) is a nontrivial distribution whose contribution to the dispersion integrals can affect A''(s) at the same or larger order. The paper does not demonstrate that A_cos is harmless, so the O(H^2/Λ^2) coefficients in (22) and (23) are not protected from O(H^2/M^2) corrections.","section":"Section III-B, around Eq. (20) and Section III-C"},{"comment":"The abstract states that the paper derives cosmological positivity bounds, and the conclusion claims that 'the leading cosmological correction to positivity bounds indeed comes at H^2/Λ^2'. The body of the paper, however, concedes that the key decomposition (20) is assumed, and the H^2/Λ^2 corrections are computed in the Lagrangian coefficients and in A_min, not in the full amplitude or in the positivity proof itself. The claims in the abstract and conclusion overstate what has been demonstrated. If the assumptions are to be retained, the results should be presented as conditional bounds—valid provided (18)–(20) hold and provided the A_cos contributions to the dispersion integrals are negligible—and the abstract and conclusion should be revised accordingly.","section":"Abstract and Section IV Conclusion"}],"minor_comments":[{"comment":"Typo: 'cosmolgical' should read 'cosmological'.","section":"Section III-D"},{"comment":"Typo: 'compatifying' should read 'compactifying'.","section":"Section III-A"},{"comment":"Typo: 'guage' should read 'gauge'.","section":"Appendix A"},{"comment":"The notation in the integrand, in particular the factor '1/2' and the treatment of the branch of the gamma function, is not fully defined; please specify the assumptions on the integration contour and the principal branch used.","section":"Eq. (19)"},{"comment":"The statement that the coefficients of the final Lagrangian do not contain B(X) or Q(X), followed by a parenthetical reference to Eq. (B12) which contains B(X), is confusing and should be clarified or rephrased.","section":"Section II, after Eq. (11)"}],"recommendation":"major_revision","confidential_remarks":"The paper is transparent about its main gap, which is commendable, but the gap is central. The referee report focuses on the assumed decomposition (20) and the parametric hierarchy H^2/M^2 versus H^2/Λ^2. Before acceptance, the authors should either provide a derivation of (20) and a controlled estimate of the A_cos contribution to the dispersion integrals, or substantially reframe the paper as proposing conditional positivity bounds under explicitly stated assumptions. As written, the abstract's claim to 'derive such bounds' is not supported by the body of the paper."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know two things about arXiv:1908.08644. First, it does something new: it works out the Goldstone Lagrangian for a c_T=1 beyond-Horndeski EFT on an FRW background and writes down positivity inequalities that include H^2/Λ^2 corrections to the Wilson coefficients. Those terms aren't in the earlier Horndeski positivity papers [40,41]. Second, the key move that gets these bounds is an assumed amplitude decomposition, Eq. (20), which the authors openly state they did not derive. That assumption is load-bearing, and the abstract's \"we derive such bounds\" goes beyond what the body concedes.\n\nWhat's good: the algebraic derivation of the Goldstone action from the covariant Lagrangian is careful, and the final coefficients in Appendix C are explicit and reproducible. The application to slow-roll inflation, concluding that B_X is suppressed, is interesting if the bounds hold. The paper is honest about the obstructions in curved spacetime; section III.B is a fair account of why a robust positivity bound is hard.\n\nThe soft spot: the central inequalities (22) and (23) are only as solid as Eq. (20). The authors note that A_cos is O(H^2/M^2) and then drop it because δ(0) dominates ρ(0). But in a regulated finite-volume treatment δ(0) is a finite volume factor, and the analytic structure of ρ is unknown. More importantly, the retained H^2/Λ^2 corrections are parametrically smaller than the neglected H^2/M^2 piece, since M^2 ≪ Λ^2. So the paper's \"cosmological correction of order H^2/Λ^2\" is not the leading cosmological correction to the amplitude; the uncomputed one is larger. Without a derivation of (20) or a controlled treatment of A_cos, the bounds don't follow from unitarity, causality, and locality. This is not a minor caveat; it directly affects the headline result.\n\nThat said, the paper deserves a serious referee. The underlying question—whether positivity bounds can be adapted to cosmological backgrounds—is important, and the Goldstone Lagrangian calculation is useful in its own right. A referee should push the authors to either justify (20) or present the results as conditional on an explicit assumption, and to correct the abstract. If that revision happens, the bounds could be a useful consistency check for beyond-Horndeski model builders. As it stands, I'd be cautious about citing the bounds as derived.","headline":"This paper has a useful EFT derivation but its advertised H^2/Λ^2 positivity bounds rest on an unproven and likely subleading amplitude decomposition.","tokens_in":18417,"tokens_out":2755,"would_cite":false,"duration_ms":23880,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper claims that a $c_T=1$ beyond-Horndeski EFT in an expanding universe obeys positivity bounds with leading $O(H^2/\\Lambda^2)$ corrections, which can be stronger or weaker than their flat-space counterparts.","keywords":["positivity bounds","EFT of cosmological perturbations","beyond-Horndeski","slow-roll inflation","Goldstone scattering","unitarity and causality","c_T=1","cosmological corrections"],"falsifier":"Compute the exact tree-level $2\\to 2$ amplitude with the Hankel mode functions (16) without imposing the decomposition (20): if the cosmological term $\\rho(\\Delta E=0)$ contributes at the same order as the Minkowski delta function, the inequalities (22) and (23) do not follow.","tokens_in":17350,"feed_emoji":"🌌","tokens_out":12781,"duration_ms":114344,"temperature":0.7,"pith_summary":"The paper sets out to extend positivity bounds — the inequalities that a low-energy effective theory must satisfy if it is to come from a unitary, causal, local ultraviolet completion — from flat spacetime to the effective field theory of cosmological perturbations. Using a shift-symmetric beyond-Horndeski theory with gravitational-wave speed $c_T=1$ as a concrete example, it derives two coefficient inequalities, equations (22) and (23), that incorporate the leading cosmological corrections of order $H^2/\\Lambda^2$. Because these corrections can enter with either sign, the cosmological bounds can be stronger or weaker than the corresponding Minkowski bounds. The application to slow-roll inflation shows that potential-driven inflation suppresses the beyond-Horndeski coupling $B_X$, effectively pushing the theory back to general relativity, while kinetically driven inflation can keep it alive. If valid, the bounds give model-independent consistency conditions on inflation models with beyond-Horndeski operators.","feed_headline":"Cosmology tightens or loosens inflation's positivity bounds","feed_subtitle":"At the leading cosmological order, the bounds change which inflation models can have a healthy UV completion.","key_machinery":"The load-bearing mechanism is the assumed amplitude decomposition (20), $A = A_{\\min}\\,\\delta^{(4)}(\\Sigma p) + A_{\\cos}\\,\\rho(\\Delta E)\\,\\delta^{(3)}(\\Sigma p)$, which isolates a Minkowski-like part that obeys ordinary energy-momentum conservation and a cosmological remainder of order $H^2/M^2$ whose effect is argued to be subdominant at zero energy mismatch. Positivity is then imported from the standard Cauchy-integral argument of flat-space dispersion relations: analyticity, crossing symmetry, the Froissart-Martin bound, and the optical theorem yield $A''(s\\to 0)\\ge 0$ and positivity of $t$-derivatives (equations (14) and (15)). Applied to the tree amplitude (21) of the Goldstone Lagrangian (11), these inequalities become the coefficient bounds (22) and (23). A second ingredient is the Goldstone action itself: the decoupling-limit Lagrangian (11) with cubic coefficients $\\alpha_i$ and quartic coefficients $\\beta_i$ encodes the beyond-Horndeski couplings $B(X)$, $G_2(X)$, $G_3(X)$ in specific combinations, and the $H^2/\\Lambda^2$ corrections in those coefficients are what turn the flat-space bound into a cosmological one.","core_discovery":"The central claim is that, for a $c_T=1$ shift-symmetric beyond-Horndeski EFT around a slowly varying FRW background, the tree-level $2\\to 2$ Goldstone amplitude (21) satisfies the positivity inequalities (22) and (23), which are the standard flat-space bounds corrected at leading order by $O(H^2/\\Lambda^2)$. These inequalities constrain $B_X$, $G_{2XX}$, $G_{3X}$, and higher coefficient derivatives; in the slow-roll limit they force $B_X$ to vanish up to slow-roll-suppressed corrections, implying the potential-driven de Sitter limit of the theory reduces to general relativity. The authors deliberately separate the amplitude into an explicitly Minkowski part, which conserves energy and momentum, and a cosmological correction of order $H^2/M^2$, which they argue can be neglected in the positivity argument because the delta-function peak dominates. They state clearly that this decomposition (20) is assumed, not derived, and that the bounds are trustworthy only if the assumptions of section III B hold. The paper also shows how the $H^2/\\Lambda^2$ corrections to the bounds can be either positive or negative, so the cosmological positivity windows can be wider or narrower than flat space.","pith_inferences":["One extension left implicit is that the same decomposition and positivity argument should carry over to other single-field EFTs of cosmological perturbations, yielding inequality constraints on whatever derivative couplings appear once the Goldstone action is constructed.","The bounds may eventually arbitrate between nonsingular bounce models built from beyond-Horndeski operators: the paper notes such models evade the bounds only because $\\varphi$ dependence and $\\ddot{\\varphi}$ violate its assumptions, so a version of the argument that keeps those terms could decide whether these models admit unitary, causal, local UV completions.","A concrete testable prediction of the assumed decomposition is that the subleading $H^2/M^2$ correction must be peaked at zero energy mismatch; a more complete computation of the exact cosmological propagator could confirm or falsify this, which would sharpen or overturn the inequalities.","Combined with future cosmological parameter measurements, bounds like (22) and (23) could be turned around to map out which regions of scalar-tensor theory space are UV-completable, effectively using data as a probe of the ultraviolet completion."],"forward_implications":["For slow-roll inflation driven by a potential, the corrected bounds imply $B_X$ must vanish at leading order, so the $c_T=1$ beyond-Horndeski EFT effectively reduces to general relativity.","In kinetically driven inflation, where the slow-roll parameter $\\epsilon$ need not be small, the same bounds leave room for nonzero $B_X$ and hence for genuine beyond-Horndeski dynamics.","The bounds can be stronger than their flat-space counterparts in some regions of the parameter space and weaker in others, so future data or theory that fixes the coefficients could indicate which case nature realizes.","Because the leading correction is quadratic in $H/\\Lambda$, the bounds do not distinguish between an expanding and a contracting universe at this order.","The $H^2/\\Lambda^2$ corrections carry coefficients of order $10^2$, so even $H/\\Lambda\\sim 0.1$ can substantially modify the allowed parameter space compared with the Minkowski limit."],"supporting_citations":[{"why":"Establishes the foundational flat-space positivity bound from analyticity, unitarity, locality, and causality that the paper adapts.","marker":"[23]"},{"why":"Supplies the modern forward-limit positivity inequalities on EFT coefficients, including the second-derivative bound $Y_{(2,1)}\\ge 0$ behind $B_X\\le 0$.","marker":"[27]"},{"why":"Computes the flat-space tree-level $2\\to 2$ Goldstone amplitude for $\\beta_4=0$, which the paper extends with the $\\beta_4$ term and cosmological corrections.","marker":"[41]"},{"why":"Regulates the forward-limit singularity from massless graviton exchange so positivity arguments can be applied to gravitational EFTs.","marker":"[39]"},{"why":"Provides the EFT-of-inflation decoupling limit used to derive the effective Goldstone Lagrangian.","marker":"[1]"},{"why":"Defines the beyond-Horndeski theory (1) whose coefficients are bounded by the inequalities.","marker":"[9]"},{"why":"Motivates the $c_T=1$ choice through the observed gravitational-wave speed constraint on modified gravity at cosmological scales.","marker":"[49]"},{"why":"Gives earlier positivity bounds on covariant shift-symmetric Horndeski theory, providing the comparison point for the cosmological bounds.","marker":"[40]"}],"fun_headline_variants":["Cosmic corrections can tighten or loosen inflation's positivity bounds","Cosmological positivity bounds can be stronger or weaker than flat space","Curved-space positivity bounds: stronger or weaker than flat","Slow-roll inflation's UV bounds bend with cosmic expansion","Beyond-Horndeski inflation's positivity bounds shift with H^2/Λ^2"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise, stated by the authors rather than derived, is that the amplitude separates into a Minkowski-like part conserving energy and momentum and a cosmological correction that can be ignored in the positivity argument, with Minkowski-like asymptotic states and flat-space external legs assumed throughout.","fun_headline_variants_meta":{"raw":{"variants":["Cosmic corrections can tighten or loosen inflation's positivity bounds","Cosmological positivity bounds can be stronger or weaker than flat space","Curved-space positivity bounds: stronger or weaker than flat","Slow-roll inflation's UV bounds bend with cosmic expansion","Beyond-Horndeski inflation's positivity bounds shift with H^2/Λ^2"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001039,"raw_usage":{"total_tokens":4356,"prompt_tokens":914,"completion_tokens":3442,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":530,"completion_tokens_details":{"reasoning_tokens":3351}},"tokens_in":530,"tokens_out":3442,"duration_ms":26052,"temperature":1.0,"reasoning_tokens":3351,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:33:20.792689+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the exact tree-level $2\\to 2$ amplitude with the Hankel mode functions (16) without imposing the decomposition (20): if the cosmological term $\\rho(\\Delta E=0)$ contributes at the same order as the Minkowski delta function, the inequalities (22) and (23) do not follow.","supporting_citations":[],"review_version":1}