{"id":"2ebed9a9-ac63-4e6a-8efb-8eb49b021c24","arxiv_id":"2505.20386","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Full S-matrix unitarity calculations show the cut-off in single-field nonminimal inflation rises to about 20 M_Pl/ξ for small couplings, while multifield kinetic interactions keep the cut-off near M_Pl/ξ.","lead":"This paper computes precise unitarity cut-off scales for Higgs-like inflation models with nonminimal coupling to gravity, using full S-matrix scattering amplitudes with phase space factors. It finds that near-critical models with small self-coupling have higher cut-offs, while multifield kinetic interactions lower the cut-off back to around M_Pl/ξ.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Quoted Emax values are tree-level estimates; if resummations ('self-healing') restore unitarity, as the authors explicitly leave open, the single-field cutoffs could rise and the central near-criticality claim would be weakened.","rationale":"The reader's weakest assumption and my concern coincide. The paper is careful and the phase-space analysis is a genuine improvement; the numerical factors from Vol_N are real. But the central claim is an upper bound on the scale of new physics inferred from unitarity, and the bound is obtained at fixed order. The exact S-matrix satisfies Eq. (4.6); a truncated tree-level S-matrix need not. The specific danger is not just 'higher orders could matter'—it is that for non-polynomial effective potentials, whole classes of diagrams can reorganize into a unitary amplitude (self-healing), and references [23,24] were written precisely for Higgs inflation. If those authors are right, the conclusion that near-criticality raises Emax to O(20) MPl/xi would be replaced by a much weaker (or no) cutoff. The paper's explicit deferral does not make the central claim robust. I also considered the constant-lambda approximation and the neglect of other channels; both are secondary, because the tree-level assumption enters every numerical result. Consequently I would keep the reader's CONDITIONAL verdict: the calculation is valuable and likely correct as a tree-level estimate, but the central claim is not settled until the resummation question is addressed.","tokens_in":30004,"tokens_out":11252,"duration_ms":127818,"concrete_test":"Compute the full tree-level (or resummed) 2 to N scattering amplitudes in the single-field metric model from the exact Einstein-frame potential V(chi) = lambda MPl^4/(4 xi^2) tanh^4(sqrt(xi) chi / MPl), using Berends-Giele recursion, and evaluate the unitarity sum in Eq. (4.6) without expanding the potential in powers of chi. Then include the class of self-healing diagrams of Refs. [23,24] and check whether |M|^2 <= 1 at E = 20 MPl/xi. If the exact amplitudes remain unitary above the quoted Emax, Eq. (4.16) is not the true cutoff; if they violate unitarity, the tree-level concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative results—especially the single-field near-critical cutoff Emax ~ O(20) MPl/xi—are derived from tree-level, single-insertion amplitudes summed over final states, and the authors explicitly defer resummation effects (Introduction, p. 3: 'we leave possible implications of resummations beyond tree-level for future work'). Equation (4.6) is an exact unitarity constraint on the S-matrix, but the calculation inserts only the first term of the perturbation series. In exactly this model, Refs. [23,24] argue that classes of diagrams can be resummed ('self-healing') so that unitarity is restored below the apparent cutoff; if that mechanism operates, the numerical values in Eqs. (4.16), (4.23), (4.27), and (4.33) are not actual unitarity bounds. The multifield derivative-sector bounds may be less affected, but the headline single-field claim depends entirely on this assumption. The acknowledgement is honest, but it means the paper establishes a tree-level estimate, not a 'precise evaluation' of the true unitarity cutoff.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper revisits perturbative unitarity in scalar-field inflation with nonminimal coupling to gravity, focusing on Higgs-inflation-type models. The authors perform a tree-level S-matrix analysis of 2-to-N scalar scattering, including exact phase-space volume factors, to compute the unitarity cutoff Emax around the vacuum. For single-field models they find that near-critical small self-coupling raises Emax to about 20 MPl/xi in metric gravity and higher in Palatini gravity. In multifield models, momentum-dependent interactions from the noncanonical kinetic sector yield Emax approximately 2.05 MPl/xi (metric) and 4.94 MPl/sqrt(xi) (Palatini), independent of lambda. A canonical-kinetic toy model gives Emax approximately 38 MPl/sqrt(xi) for lambda=1e-5. The paper also derives CMB-normalization relations between xi and lambda, compares thermal energies during reheating with the cutoff, and computes in an appendix that the number of on-shell quanta with E~Emax is exponentially suppressed. The paper explicitly states that all results are tree-level and defers resummation ('self-healing') effects to future work.","tokens_in":30227,"tokens_out":13167,"duration_ms":125561,"significance":"The paper's methodological contribution is the inclusion of full phase-space volumes and sums over final-state multiplicities, going beyond the usual order-of-magnitude estimates in the Higgs-inflation unitarity literature. If the tree-level results are taken as indicative, they sharpen the case that near-critical Higgs inflation has a higher unitarity scale than previously estimated and clearly identify the kinetic-sector derivative interactions as the dominant source of low cutoffs in multifield models. The analytic formulas and appendices are explicit, and the CMB-normalization relations and reheating estimates provide concrete, testable predictions. However, the main quantitative claims are conditional on the assumed tree-level truncation; because the paper acknowledges that resummations could alter the conclusions, the results are most accurately described as tree-level estimates rather than definitive evaluations of the true unitarity cutoff.","major_comments":[{"comment":"The central claim of a 'precise evaluation' of the cut-off scale is based on inserting tree-level amplitudes into the exact unitarity inequality (4.6). The paper explicitly defers resummations ('self-healing', Refs. [23,24]) to future work (p.3). Since these references argue that classes of diagrams can restore unitarity below the apparent cutoff in exactly this model, the single-field near-critical result Emax ~ O(20) MPl/xi is not established as a true upper bound; it is a tree-level estimate. The authors should either assess the potential effect of such resummations on the key numerical values in Eqs. (4.16), (4.23), (4.27), and (4.33), or consistently qualify the results and the abstract as tree-level estimates.","section":"Sec. 4.2 (and abstract)"},{"comment":"Inequality (4.6) is derived for inelastic channels (N not equal to 2). In Eq. (4.16) the elastic 2-to-2 contribution is included in the sum over N, with the remark that the elastic process is taken into account, but no derivation is given for the unitarity bound that applies to the elastic element S22. Please state the criterion used (e.g., |M22|^2 <= 1 versus the optical-theorem constraint) and show that including N=2 in the same inequality is justified; this affects the interpretation of Fig. 2.","section":"Sec. 4.1, Eq. (4.16)"},{"comment":"The abstract labels the canonical-kinetic model 'phenomenologically viable', but Eq. (5.58) predicts ns about 0.973, which lies approximately two standard deviations above the Planck value ns = 0.9649 +/- 0.0042 quoted in Sec. 5.1. This claim should be supported by a quantitative comparison to the data; otherwise the model should be described as an illustrative toy model rather than phenomenologically viable.","section":"Sec. 5.3 and abstract"}],"minor_comments":[{"comment":"The condition 'N in 4+2N' appears to be a typo; based on the subsequent sums it should read 'N in 2+2N'.","section":"Eq. (4.12)"},{"comment":"The phrase 'the term that responsible for lowering Emax' should be 'the term responsible for lowering Emax'.","section":"Sec. 6.1, after Eq. (6.7)"},{"comment":"The subscript notation C_V,M is defined only by context; please define it explicitly at first use.","section":"Eq. (3.11)"},{"comment":"The statement that 'the volume factors cause these elements to vanish for large numbers of outgoing particles' is not immediately evident from Eq. (4.26); consider adding one sentence on the N-to-infinity asymptotic behavior of the matrix element.","section":"Sec. 4.3.2"}],"recommendation":"major_revision","confidential_remarks":"The paper reports a careful tree-level S-matrix calculation and is transparent about the resummation caveat. However, the title and abstract's 'precision' language may overstate the robustness of the quoted cutoffs, since the relevant resummations could change the central single-field result. If the authors reframe the conclusions as tree-level estimates and clarify the elastic-channel criterion, the paper would be publishable after major revision. The 'phenomenologically viable' claim for the canonical-kinetic model should also be corrected. The paper is within the scope of the journal and of interest to the hep-ph community."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is a solid, worthwhile paper, and the central calculation is likely right at tree level. The one thing to know before reading is that \"precise\" here means precise within a tree-level EFT computation; the authors explicitly leave resummation effects open, and that caveat matters most for the single-field near-critical claim.\n\nWhat is actually new: the paper applies the full S-matrix/phase-space method of Refs. [40,41] to nonminimal inflation, and that changes numbers, not just prefactors. The λ-dependent single-field cutoff (Emax ~ O(20) MPl/ξ for λ ~ 10^-5), the multifield metric bound Emax ≈ 2.05 MPl/ξ, the Palatini bound ≈ 4.94 MPl/√ξ, and the canonical-kinetic toy model at ≈ 38 MPl/√ξ are all new quantitative results. The derivation in Secs. 3–4 and Appendix A is careful and internally consistent. The point that phase-space volume factors push the dominant contribution to low N, so that weak couplings really do raise the cutoff, is a genuine improvement over order-of-magnitude estimates. The background-field analysis in Sec. 6, especially the Goldstone-scattering bound Emax ≈ 4√π MPl/√ξ, is a useful addition.\n\nThe soft spots, in proportion: (1) The tree-level truncation is the main one. Equation (4.6) is exact, but only the first term of the perturbation series is inserted. The stress-test concern is real for the single-field numbers: if the self-healing resummations of Refs. [23,24] operate, the near-critical cutoff could rise and the headline \"near-criticality raises the cutoff\" claim weakens. The multifield momentum-dependent bounds are less exposed—those derivative operators are not the ones the self-healing arguments target—but the single-field claim is conditional. The authors say this almost verbatim, so it is an honest limitation, not a hidden one. (2) The constant-λ, quartic-only potential and the neglect of some channels are acknowledged and are probably minor for the central comparison, but they make \"precision\" softer than the word implies. (3) The canonical-kinetic model is called \"phenomenologically viable,\" and it is an interesting toy model, but its ns ≈ 0.973 sits about 2σ from the Planck central value; that label is a bit generous.\n\nWho this is for: people working on Higgs inflation, critical Higgs inflation, or EFT validity in nonminimally coupled scalar theories. It is a serious contribution, not a desk-reject. I would send it to referees, and I would cite it for the numerical cutoffs and the λ-dependence. My own verdict is conditional acceptance with the tree-level caveat stated prominently.","headline":"Solid full S-matrix treatment of unitarity cutoffs in nonminimal inflation; the single-field near-critical numbers are tree-level and explicitly conditional on no self-healing.","tokens_in":30815,"tokens_out":3214,"would_cite":true,"duration_ms":33215,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["98.80.Cq","11.55.-m"],"model":"deepseek-v4-flash","headline":"Full S-matrix sums show near-critical single-field inflation stays unitary to about $E_{\\rm max} \\sim 20\\,M_{\\rm Pl}/\\xi$, while multifield kinetic interactions pull the cut-off down to about $2\\,M_{\\rm Pl}/\\xi$.","keywords":["Higgs inflation","perturbative unitarity","nonminimal coupling","critical Higgs inflation","S-matrix unitarity bounds","phase-space factors","Palatini gravity","multifield inflation"],"falsifier":"Compute the resummed (all-orders) $2 \\to N$ amplitudes in the single-field Einstein-frame theory: if the resummed partial waves remain unitary up to $E \\sim M_{\\rm Pl}/\\sqrt{\\xi}$ or beyond, the claimed $\\lambda$-dependent rise to $\\sim 20\\,M_{\\rm Pl}/\\xi$ is not the true unitarity scale and the paper's central single-field claim fails. A softer but direct check is to repeat the same phase-space-weighted S-matrix sums at one loop and see whether the cut-off shifts by an $\\mathcal{O}(1)$ factor or parametrically, since a parametric shift would signal that the tree-level identification of $E_{\\rm max}$ is not stable.","tokens_in":29794,"feed_emoji":"⚛️","tokens_out":18280,"duration_ms":171106,"temperature":0.7,"pith_summary":"This paper asks whether the unitarity problem of Higgs-style inflation is really as severe as the standard order-of-magnitude estimates suggest, and it answers that the truth depends sharply on how many scalar fields are involved. Computing the full S-matrix, including the exact phase-space volume factors that earlier estimates dropped, the authors find that in single-field models a small self-coupling $\\lambda$ (the near-critical regime) raises the cut-off to $E_{\\rm max} \\sim \\mathcal{O}(20)\\,M_{\\rm Pl}/\\xi$, far above the old $M_{\\rm Pl}/\\xi$ bound. In multifield models, however, momentum-dependent interactions from the curved field-space metric dominate and pull the cut-off back down to $E_{\\rm max} \\simeq 2.05\\,M_{\\rm Pl}/\\xi$ (metric gravity) or $4.94\\,M_{\\rm Pl}/\\sqrt{\\xi}$ (Palatini), with no dependence on $\\lambda$. A canonically-kinetic variant restores unitarity up to $E_{\\rm max} \\simeq 38\\,M_{\\rm Pl}/\\sqrt{\\xi}$ at $\\lambda = 10^{-5}$, and background-field effects during inflation raise the bound only to about $0.6\\,M_{\\rm Pl}$ in metric gravity, still below the field values the model needs. If right, near-critical single-field inflation is far more self-consistent than previously thought, while the realistic multifield Higgs case still needs new physics below the scales sampled during inflation.","feed_headline":"Near-critical Higgs inflation escapes the old unitarity bound","feed_subtitle":"Full S-matrix sums with exact phase-space factors lift the safe energy to ~20 MPl/ξ for small self-couplings.","key_machinery":"The load-bearing identity is the S-matrix unitarity bound from the optical theorem, $\\sum_N |\\mathcal{M}_{2 \\to N}|^2 \\leq 1$, combined with the exact Lorentz-invariant phase-space volume $\\mathrm{Vol}_N = (E/4\\pi)^{2(N-2)}/[8\\pi (N-1)!(N-2)!]$. The volume factors suppress large-$N$ final states, so the most stringent bound comes from the smallest multiplicity, where the Wilson coefficient, proportional to $\\lambda$ in the single-field case, still controls the amplitude; that is the mechanism by which weak coupling raises $E_{\\rm max}$. The second key object is the Einstein-frame field-space metric $G_{IJ}$ produced by the conformal transformation $\\Omega^2 = 1 + \\xi|\\Phi|^2/M_{\\rm Pl}^2$: in metric gravity its non-diagonal, momentum-dependent pieces survive for multiple fields and supply the $\\lambda$-independent elastic amplitudes that dominate the multifield cut-off, whereas in Palatini gravity those pieces vanish ($\\sigma = 0$) and only $\\sqrt{\\xi}/M_{\\rm Pl}$-suppressed diagonal terms remain. In the single-field case the nonminimal coupling is instead absorbed by a canonical field redefinition $\\phi \\to \\chi$, converting the entire effect into a tower of potential operators with coefficients $C_{N+2} \\propto \\lambda$ suppressed by $(\\xi/M_{\\rm Pl})^2$ per field pair in metric gravity and by $\\xi/M_{\\rm Pl}^2$ in Palatini gravity.","core_discovery":"The central claim is that the unitarity cut-off of a nonminimally coupled scalar theory is not a fixed scale but is set by whichever class of higher-dimensional operators dominates the S-matrix sum, so its value changes between single-field and multifield models and depends on the background. Using the optical-theorem bound $\\sum_N |\\mathcal{M}_{2 \\to N}|^2 \\leq 1$ evaluated with exact $N$-particle phase-space volumes, the authors show that in the single-field Einstein-frame theory every higher-dimensional operator comes from the conformally rescaled potential and carries Wilson coefficients proportional to $\\lambda$; hence a small self-coupling genuinely raises the cut-off, to $E_{\\rm max} \\sim \\mathcal{O}(20)\\,M_{\\rm Pl}/\\xi$ for $\\lambda \\sim 10^{-5}$ in metric gravity, and higher in Palatini gravity. In multifield models the non-diagonal field-space metric generates momentum-dependent $2 \\to 2$ interactions that cannot be removed by field redefinitions and are suppressed only by $\\xi/M_{\\rm Pl}$, yielding $E_{\\rm max} \\simeq 2.05\\,M_{\\rm Pl}/\\xi$ (metric) and $4.94\\,M_{\\rm Pl}/\\sqrt{\\xi}$ (Palatini), independent of $\\lambda$. The authors also show that expanding around the inflationary condensate raises the metric-gravity multifield bound to about $4\\sqrt{\\pi}\\,M_{\\rm Pl}/\\sqrt{\\xi} \\simeq 0.6\\,M_{\\rm Pl}$ for $\\lambda = 10^{-5}$, marginally below the field value $\\chi_c \\simeq 5.3\\,M_{\\rm Pl}$ required for 55 e-folds, and that on-shell quanta at $E \\sim E_{\\rm max}$ are exponentially rare inside a Hubble sphere.","pith_inferences":["If the 'self-healing' resummations proposed in earlier work do restore unitarity, the numbers here become conservative upper bounds on the cut-off; the single-field $\\lambda$-dependence is the part most vulnerable, while the multifield kinetic-sector bounds, tied to momentum-dependent operators rather than the potential, are the most likely to survive an all-orders treatment.","The exponential scarcity of quanta at $E \\sim E_{\\rm max}$ in a Hubble volume suggests the formal unitarity violation may be practically inert for ordinary inflationary and reheating observables: the cut-off constrains exotic ultra-energetic processes, not the ambient field dynamics that actually drives inflation.","The mechanism producing the $\\lambda$-dependence is generic: any EFT whose tower of higher-dimension Wilson coefficients shares a single small coupling should show the same pattern, with phase-space factors pushing the strongest bound to the lowest final-state multiplicity; that prediction could be tested in a controlled setting with a known UV completion, where tree-level unitarity bounds could b","A natural extension is to apply the same phase-space-weighted S-matrix sums to the fermion and gauge sectors: the authors estimate Yukawa corrections are small for ordinary couplings, but a full treatment could shift the Palatini multifield bound by $\\mathcal{O}(1)$ factors and should be checked before those numbers are used as benchmarks."],"forward_implications":["For near-critical single-field models such as critical Higgs inflation ($\\lambda \\sim 10^{-5}$), the tree-level cut-off rises to $E_{\\rm max} \\sim \\mathcal{O}(20)\\,M_{\\rm Pl}/\\xi$, roughly an order of magnitude above the traditional $M_{\\rm Pl}/\\xi$ estimate, and it grows as $\\lambda$ shrinks toward the free-field limit where the nonminimal coupling causes no unitarity violation.","Because the multifield kinetic-sector bound $E_{\\rm max} \\simeq 2.05\\,M_{\\rm Pl}/\\xi$ does not depend on $\\lambda$, realistic Higgs inflation with its four real scalar degrees of freedom still faces new physics at scales far below the field values during inflation, so its CMB predictions remain sensitive to the unknown UV completion.","In Palatini gravity both the kinetic-sector and potential operators are suppressed by $\\sqrt{\\xi}/M_{\\rm Pl}$ with small numerical coefficients, giving $E_{\\rm max} \\simeq 4.94\\,M_{\\rm Pl}/\\sqrt{\\xi}$ for the multifield case and making Palatini Higgs inflation considerably less constrained than the metric version.","The canonical-kinetic toy model, with $E_{\\rm max} \\simeq 38\\,M_{\\rm Pl}/\\sqrt{\\xi}$ at $\\lambda = 10^{-5}$, shows that the low multifield cut-off is caused by the curved field-space metric and not by the conformally rescaled potential itself.","Expanding around the inflationary background raises the metric-gravity multifield cut-off to roughly $0.6\\,M_{\\rm Pl}$ for $\\lambda = 10^{-5}$, still below the required field value $\\chi_c \\simeq 5.3\\,M_{\\rm Pl}$, so even the most favorable background-field treatment leaves conventional metric Higgs inflation in the new-physics-sensitive regime."],"supporting_citations":[{"why":"Supplies the model under study: the Standard Model Higgs boson as the inflaton with a nonminimal coupling $\\xi H^\\dagger H R$.","marker":"[3]"},{"why":"The earlier order-of-magnitude unitarity estimates ($E_{\\rm max} \\sim M_{\\rm Pl}/\\xi$ expanded about the global minimum) that this paper refines.","marker":"[8–14]"},{"why":"Establishes the free-field ($\\lambda \\to 0$) limit in which the nonminimal coupling causes no unitarity problem, the limit the single-field results must smoothly approach; also the reference for the multifield elastic $2\\to2$ process.","marker":"[11]"},{"why":"Introduces critical Higgs inflation, the near-critical small-$\\lambda$ scenario whose cut-off this paper recomputes precisely.","marker":"[19]"},{"why":"Source of the background-field estimate $E_{\\rm max} \\sim M_{\\rm Pl}/\\sqrt{\\xi}$ during inflation, which the paper compares against its single-field and multifield bounds.","marker":"[20–24]"},{"why":"The self-healing resummation argument that could restore unitarity beyond tree level; the caveat the authors explicitly leave to future work.","marker":"[23, 24]"},{"why":"Derives the field-space metric under conformal transformations with multiple scalars, the origin of the momentum-dependent operators that dominate the multifield cut-off.","marker":"[38]"},{"why":"Supplies the S-matrix formalism with exact $N$-particle phase-space volumes on which the whole precision calculation rests.","marker":"[40]"},{"why":"Provides the state and amplitude conventions and the sum-over-states unitarity bounds used to extract $E_{\\rm max}$ from the phase-space-weighted amplitudes.","marker":"[41]"}],"fun_headline_variants":["Precision S-matrix redefines Higgs inflation unitarity","Near-critical inflation cut-off rises to 20 MPl/ξ","Multifield inflation unitarity bound tighter than single-field","New canonical model raises inflation cut-off significantly","Exact phase-space sums refine inflation unitarity bounds"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole calculation assumes that tree-level $2 \\to N$ scattering amplitudes in the Einstein-frame EFT fix the true unitarity cut-off, with the authors themselves explicitly leaving to future work the possibility, raised by the 'self-healing' resummation arguments of refs. [23, 24], that summing all orders restores unitarity and lifts the quoted $E_{\\rm max}$ values.","fun_headline_variants_meta":{"raw":{"variants":["Precision S-matrix redefines Higgs inflation unitarity","Near-critical inflation cut-off rises to 20 MPl/ξ","Multifield inflation unitarity bound tighter than single-field","New canonical model raises inflation cut-off significantly","Exact phase-space sums refine inflation unitarity bounds"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000396,"raw_usage":{"total_tokens":2195,"prompt_tokens":1185,"completion_tokens":1010,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":801,"completion_tokens_details":{"reasoning_tokens":930}},"tokens_in":801,"tokens_out":1010,"duration_ms":11401,"temperature":1.0,"reasoning_tokens":930,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T13:56:07.765425+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the resummed (all-orders) $2 \\to N$ amplitudes in the single-field Einstein-frame theory: if the resummed partial waves remain unitary up to $E \\sim M_{\\rm Pl}/\\sqrt{\\xi}$ or beyond, the claimed $\\lambda$-dependent rise to $\\sim 20\\,M_{\\rm Pl}/\\xi$ is not the true unitarity scale and the paper's central single-field claim fails. A softer but direct check is to repeat the same phase-space-weighted S-matrix sums at one loop and see whether the cut-off shifts by an $\\mathcal{O}(1)$ factor or parametrically, since a parametric shift would signal that the tree-level identification of $E_{\\rm max}$ is not stable.","supporting_citations":[],"review_version":1}