{"id":"f95d5138-45c7-47dc-82ec-89ed99e22ed3","arxiv_id":"2501.09985","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":11,"one_line_summary":"A constructed two-field DHOST bouncing cosmology that avoids ghost, gradient, and superluminality problems and produces nearly scale-invariant curvature perturbations.","lead":"This paper proposes a concrete mathematical model of a bouncing universe, built from modified gravity with an extra scalar field, and claims it avoids all known instabilities while matching the observed pattern of early-universe fluctuations. If correct, it provides a testable alternative to inflation for the beginning of the cosmos.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 'fully viable' claim rests on an uncomputed final curvature spectrum and non-Gaussianity: conversion is checked at amplitude level for one mode, and f_NL is only 'expected', so observational compatibility is not established.","rationale":"The reader's weakest_assumption identifies the same load-bearing gap: the paper computes n_s for δχ before conversion (Eq. 4.11), checks conversion efficiency (Fig. 10), but never computes the final curvature power spectrum or bispectrum. My stress-test sharpens this: the conversion efficiency is an amplitude ratio for a single k, so even the scale dependence of the linear transfer function is unchecked, and the non-Gaussianity statement in Section 5 is an explicit deferral ('we expect', 'we plan to pursue'). Because the abstract claims compatibility with observations, this is the pivotal unsupported step. I do not see an internal inconsistency in the algebraic construction; the instability and superluminality checks are plausible for the displayed parameters. A secondary concern is that the stability plots in Section 3 use ε = 5 (Eq. 3.22) while the final parameter set (4.26) uses ε = 10; this reproducibility mismatch should be fixed, but it is not the deciding issue. The reader's CONDITIONAL verdict already reflects the missing calculation, so no verdict adjustment is needed.","tokens_in":26383,"tokens_out":7038,"duration_ms":75061,"concrete_test":"Numerically solve the full linear system (4.15) for the final Lagrangian functions (5.1)-(5.3) with parameters (4.26), evolving Fourier modes over k spanning at least 10^-4 to 1 Mpc^-1 in rescaled units from Bunch-Davies initial conditions in the ekpyrotic phase; extract the late-time ζ_k power spectrum and fit n_s and As at the pivot scale k = 0.05 Mpc^-1. Then compute the leading bispectrum f_NL for the same model (e.g., via the δN formalism across the conversion or from the third-order action) and compare with Planck: require |f_NL| < O(10) and n_s within roughly 0.965 ± 0.01. If both hold, the observational-viability claim is supported; if not, the 'fully viable' assertion fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim requires that the entropy-to-curvature conversion preserve the scale-invariant δχ spectrum and produce small non-Gaussianities. This is not demonstrated. Section 4.3 defines conversion efficiency by the amplitude ratio |ζ(t_end)/δχ(t_begin)| for what appears to be a single mode (Eq. 4.20, Fig. 10), and Eq. (4.32) sets the amplitude As from the δχ power spectrum before conversion. No computation of the final ζ_k power spectrum over the observed k-range is given, so a k-dependent transfer function during the conversion (around t_c = 10τ) could alter n_s or introduce running/oscillatory features. More importantly, Section 5 explicitly says only that 'we expect' sufficiently small non-Gaussianities and defers confirmation to future work; no bispectrum or trispectrum calculation appears. Since Planck bounds on f_NL are part of observational compatibility, this expectation is not evidence. The cited earlier works (Refs. 56, 61, 62) show that even smooth, efficient conversion can produce sizeable non-Gaussianities depending on details; the present parameters are not shown to lie in the safe region. Thus the headline 'fully viable ... compatible with observations' outstrips the computed support.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper constructs a two-field DHOST Ia bouncing cosmology by an inverse method: it first chooses a background with ekpyrotic contraction and kinetic-dominated expansion, then reconstructs the Lagrangian functions so that tensor modes are luminal and the scalar sector satisfies sufficient stability and non-superluminality conditions. An extra scalar field is added with luminal sound speed, and its entropic perturbations acquire a nearly scale-invariant spectrum during ekpyrotic contraction; the paper then numerically demonstrates an approximately 98% conversion of entropy perturbations into curvature perturbations and claims compatibility with Planck constraints on the scalar spectral index and amplitude, with a negligible tensor-to-scalar ratio. The abstract and conclusion assert that this is the first fully viable two-field DHOST bounce, free of BKL, ghost, gradient, and superluminality problems and compatible with observations.","tokens_in":26686,"tokens_out":7591,"duration_ms":82302,"significance":"If fully established, the model would be a significant explicit realization of a bouncing cosmology in the exceptional DHOST Ia subclass, circumventing known superluminality and no-go issues while providing a concrete Lagrangian and background. The paper's strengths are its detailed algebraic derivation of the sufficient stability conditions in Eq. (2.21), the explicit reconstruction in Eqs. (5.1)-(5.3), and the numerical verification of the local no-pathology conditions for the parameter set (4.26). However, the headline claim of observational viability currently rests on uncomputed quantities: the final curvature power spectrum after conversion and the non-Gaussianity parameter. The manuscript itself states in Section 5 that small non-Gaussianities are only 'expected,' not demonstrated. The significance is therefore conditional on completing or substantially revising that part of the claim.","major_comments":[{"comment":"The claim that the model predicts a nearly scale-invariant curvature power spectrum compatible with Planck is not supported by the computations shown. Section 4.3 characterizes the conversion by the amplitude ratio |ζ(t_end)/δχ(t_begin)| in Eq. (4.20) and Figure 10, but Section 4.4 then assumes in Eqs. (4.27) and (4.32) that the δχ power spectrum is simply inherited by ζ. No computation of the final ζ_k power spectrum over the observed k-range, no k-dependent transfer function, and no bispectrum computation are given. Section 5 states only that non-Gaussianities are 'expected' to be small, citing Refs. [56,61,62], which in fact show that high efficiency and smoothness do not by themselves guarantee small f_NL. Since Planck bounds on non-Gaussianity are part of observational compatibility, the 'fully viable ... compatible with observations' claim outstrips the computed support. Please compute the converted curvature spectrum and f_NL (or a controlled estimate), or explicitly revise the claim to a conditional one.","section":"§4.3–§4.4, §5"},{"comment":"The conversion demonstration is made for what appears to be a single Fourier mode with a single set of initial conditions: Eq. (4.20) uses a subscript k, but no k value is specified, and Eq. (4.25) fixes ζ and δχ amplitudes to 10^-5 without justifying that the amplitude ratio is independent of the mode and of the relative initial phase. A k-dependent conversion efficiency would change the final spectral index or introduce running/oscillatory features; the paper should either show the k-independence explicitly or at least state the representative mode and explain why the result is representative for the observable window.","section":"§4.3, Eq. (4.25)"},{"comment":"The no-ghost, no-gradient, and non-superluminality conditions (2.21) and (2.23) are verified numerically for the single parameter set (4.26), and the figures do not display the entire time range used for the checks. In particular, Figure 11 shows only 20 ≤ t ≤ 200, while the text claims the constraints hold 'throughout the whole time evolution.' The asymptotic behavior of the model functions is used to argue that the conditions hold in the far past and future, but the precise interval and resolution of the numerical verification should be stated. This is a curable gap, but it matters because the paper's central claim is that the model is 'completely free' of pathologies.","section":"§3.2, §4.3"}],"minor_comments":[{"comment":"The text states that P04 > 0, but the function defined in Eq. (4.21) vanishes exactly at t = t_c because sech(0) = 1. Please clarify the behavior at that instant and explain why the scalar potential remains bounded from below there, referring to the numerical check in Figure 9.","section":"§4.1, Eq. (4.21)"},{"comment":"The amplitude matching in Eq. (4.32) is only used to quote μ ~ O(10^-2) M_pl. The sensitivity of A_s to the parameters τ, ε, b, and μ should be stated, and the precise central value of μ used for the matching should be given rather than only an order of magnitude.","section":"§4.4, Eq. (4.32)"},{"comment":"Figure 10 should state the value of the wavenumber k (or k̄) used in the numerical integration, as well as the initial time t_i and the numerical method, so that the claim of 98% conversion is reproducible.","section":"§4.3, Fig. 10"},{"comment":"The sentence 'We also expect our models to have sufficiently small non-Gaussianities' is an explicit admission that a central part of observational viability is not computed. This should be moved to a clearly qualified statement in the abstract and conclusion, or removed until the computation is performed.","section":"Abstract, §5"}],"recommendation":"major_revision","confidential_remarks":"The paper is technically careful in the DHOST sector and the inverse-method construction is explicit and reproducible. The main issue is not the local stability algebra but the gap between the computed entropic perturbation spectrum and the claimed final curvature spectrum with small non-Gaussianities. I would recommend requesting a concrete computation of the converted ζ power spectrum and at least an estimate of f_NL, or a substantial softening of the 'fully viable' claim. With that completed, the paper would be a strong contribution to the bouncing-cosmology literature."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nI read the paper carefully. The central construction is real: they build an explicit two-field DHOST model with GR asymptotics, no ghost/gradient instability, no superluminality, and a roughly 98% efficient entropy-to-curvature conversion. The new sufficient condition (2.21) and the explicit Lagrangian functions (5.1)–(5.3) are not in the cited literature. The stability algebra in Secs. 2–3 is careful and internally consistent; the inverse method is standard for constructions like this, and choosing b and μ to match Planck does not make the model circular, it makes the perturbation amplitudes consistent by design.\n\nThat said, the headline claim “fully viable” outstrips the computed support. Three gaps, in decreasing order of severity. First, non-Gaussianity is explicitly deferred. Section 5 says only “we expect” small f_NL and promises future work. Planck bounds on f_NL are part of observational compatibility, and the cited references (56, 61, 62) show that even smooth, efficient conversion can generate large non-Gaussianities depending on the trajectory. The authors do not show their parameters lie in the safe region. Second, conversion is checked at the amplitude level for what looks like a single mode: Eq. (4.20) and Fig. 10 give |ζ/δχ| at one k, not the full ζ_k power spectrum over the observed range. A k-dependent transfer function during the conversion could alter n_s or add running, so the final curvature spectrum is not established. Third, there is a parameter mismatch: the stability plots in Sec. 3 (Figs. 4, 5) use ε = 5, while the final parameters in (4.26) use ε = 10 plus the new conversion parameters t_c, q, λ. Fig. 11 checks one key constraint for the full set, but the complete stability scan is not shown for the exact parameter set that produces the observables.\n\nNone of this destroys the construction. A stable, non-superluminal two-field DHOST bounce likely exists along these lines; what is missing is the demonstration of observational viability, which is exactly what the title promises. The fix is clear: compute the final curvature power spectrum and at least f_NL for the full parameter set, rerun the stability plots for those same parameters, and ideally provide code or data. This is refereeable, and a good referee would ask for exactly that before publication.\n\nI would send it to peer review rather than desk-reject, with the explicit request for the missing non-Gaussianity computation. Once those pieces are filled, this will likely become a benchmark model in the bouncing-cosmology literature.","headline":"A serious two-field DHOST bounce construction whose 'fully viable' label is one missing computation away from being earned.","tokens_in":27193,"tokens_out":2921,"would_cite":true,"duration_ms":30322,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["83F05","83D05"],"pacs":["98.80.-k","98.80.Cq","04.50.Kd"],"model":"deepseek-v4-flash","headline":"This paper constructs the first two-field DHOST bouncing cosmology that claims to be free of ghost, gradient, and BKL instabilities and superluminality, while matching the observed nearly scale-invariant scalar spectrum and a negligible…","keywords":["DHOST theories","bouncing cosmology","ekpyrotic contraction","entropic perturbations","superluminality","non-Gaussianities","primordial perturbations","modified gravity"],"falsifier":"Solve the coupled perturbation equations (4.17) past the conversion end for a range of wavenumbers and compute the final $\\zeta$ power spectrum and bispectrum: if the resulting spectral index shifts from $n_s \\simeq 0.965$ by more than the Planck error bars, or if $f_{\\rm NL}$ exceeds the current observational bound, the 'fully viable' claim is refuted. A second, independent test: since the model gives a deep-blue tensor spectrum with negligible $r$ at the pivot scale, a primordial tensor signal near the inflation-scale upper bound would rule it out.","tokens_in":26176,"feed_emoji":"🌌","tokens_out":7742,"duration_ms":75244,"temperature":0.7,"pith_summary":"The paper aims to show that a non-singular bouncing universe can be a fully viable alternative to inflation, despite the no-go results that block healthy bounces in simpler scalar-tensor theories. It builds a two-field model whose adiabatic field is an exceptional subclass of Degenerate Higher-Order Scalar-Tensor (DHOST) theories and whose extra scalar supplies nearly scale-invariant perturbations through the entropic mechanism. The claim is that this model is free of BKL, ghost, and gradient instabilities, has no superluminal propagation, reduces to General Relativity plus a canonical scalar in the distant past and future, and produces a scalar spectrum with $n_s \\simeq 0.965$, an amplitude matching Planck, and a negligible tensor-to-scalar ratio. If correct, it removes the last standard objections to bouncing cosmology as a competitor to inflation.","feed_headline":"Two-field bounce cosmos passes all stability tests","feed_subtitle":"A DHOST bounce with an extra scalar yields a near-flat spectrum, tiny tensor ratio, and light-speed gravity waves.","key_machinery":"The load-bearing object is the exceptional subclass of DHOST theories (Degenerate Higher-Order Scalar-Tensor gravity, a higher-derivative scalar-tensor class that avoids Ostrogradski ghosts) with a luminal extra scalar: the condition (2.17), $A_3 = \\frac{2(XA_1 - 2F_2)(A_1 - 2F_{2X})}{X(3XA_1 - 4F_2)}$, makes the coupling function $f$ equal to $g$, which is necessary and, together with (2.20), sufficient for the sound speed of the coupled two-scalar system to stay at or below unity. The inverse-method reconstruction then turns the two-inequality set (2.22) and (2.23), namely $G_S \\geq F_S > Y P_Y g^2$ together with $G_T \\geq F_T > 0$, into a concrete Lagrangian by choosing the scale factor (3.5), the functions $g_1$ and $a_1$ in (3.19)-(3.20), $\\Sigma$ in (3.27), $P_1$ in (4.14), $\\chi$ in (4.19), and $P_{04}$ in (4.21). The conversion from entropic to curvature perturbations is driven by the coefficient $\\Psi_1$ in (4.15), whose size is controlled by the acceleration of $\\chi$ rather than its velocity, which is what keeps the trajectory inside the stable region.","core_discovery":"The central discovery is an explicit two-field construction, presented as the first of its kind, in the exceptional subclass of DHOST Ia theories with an extra scalar field $\\chi$. In the DHOST sector the Lagrangian functions are fixed by the inverse method: a chosen scale factor (3.5) with ekpyrotic contraction ($\\epsilon > 3$) and kinetic-dominated expansion, and the ansatz (3.7), with $g_0 = -g_1$ and $a_0 = -a_1$ so that the tensor sector has $G_T = F_T = 1$ and luminal gravitational-wave speed at all times. The extra scalar has luminal sound speed and $P(\\chi,Y,\\varphi) = P_1(\\varphi)Y - P_{02}(\\varphi)\\chi^2 - P_{04}(\\varphi)\\chi^4$, with $P_1$ chosen as (4.14) so that $\\chi$ develops a nearly scale-invariant spectrum during contraction, and a sech-shaped background trajectory (4.19) that bends sharply after the bounce and converts about 98% of $\\delta\\chi$ into curvature perturbations. The model functions (5.1)-(5.3) with parameters (4.26) satisfy the stability and non-superluminality conditions (2.22) and (2.23) throughout the whole background trajectory, and $\\mu \\sim O(10^{-2}) M_{\\rm Pl}$ sets the observed scalar amplitude.","pith_inferences":["Editorial inference: the missing computation that would close the 'fully viable' claim is the final $\\zeta$ power spectrum and bispectrum; the paper only checks the entropic spectrum and the conversion efficiency, and the size of $f_{\\rm NL}$ from the nonlinear conversion remains open.","Editorial inference: because the model is fixed by parameterized template functions rather than derived from a symmetry, the same inequalities might be satisfied by other shapes of $\\chi(t)$; a scan over $\\lambda$, $q$, and $\\chi_0$ with the same $\\Psi_1$ mechanism could test the robustness of the roughly 98% efficiency.","Editorial inference: the deep-blue tensor spectrum (spectral index $4 - 2\\nu_1$) is a generic ekpyrotic signature; a future detection of primordial B-modes at the level expected in slow-roll inflation would distinguish this class from inflation and falsify this particular model."],"forward_implications":["Bounces no longer need to invoke strong gravity in the past: this model has GR-like asymptotics, so the strong-coupling worry and the associated large non-Gaussianities of earlier scenarios do not arise.","The predicted scalar spectral index $n_s \\simeq 0.965$ and scalar amplitude $A_s = 2.105 \\times 10^{-9}$ at the pivot scale match Planck data, while the tensor-to-scalar ratio is negligible.","Gravitational-wave speed equals the speed of light throughout the whole evolution, since $G_T = F_T = 1$.","Any future model with a luminal extra scalar can avoid superluminality by staying in the exceptional subclass and satisfying $F_S > Y P_Y g^2$; this is a checkable sufficient condition.","The nearly scale-invariant curvature perturbation is generated by the entropic field, not by the ekpyrotic adiabatic mode, which would otherwise be blue-tilted."],"supporting_citations":[{"why":"Supplies the inverse method, the background ansatz, and the parameter choices $\\tau = 10$, $u = 1/10$, $w = 2$ that the DHOST sector builds on.","marker":"[27]"},{"why":"Identifies the exceptional subclass of DHOST Ia theories in which a luminal extra scalar avoids superluminality and gives the condition (2.17).","marker":"[36]"},{"why":"Provides the two-field DHOST quadratic action and the stability and non-superluminality conditions used in (2.12) and (2.21).","marker":"[37]"},{"why":"Introduces the non-minimal kinetic coupling $P_1 \\sim \\varphi^{-2(1+b)}$ that generates a nearly scale-invariant entropic spectrum.","marker":"[54]"},{"why":"Establishes the entropic mechanism with small non-Gaussian corrections, supporting the expectation that $\\delta\\chi$ perturbations have negligible non-Gaussianities.","marker":"[56]"},{"why":"Analyzes the ekpyrotic trispectrum and highlights the difficulty of efficient conversion, providing the baseline for the paper's near-98% efficiency claim.","marker":"[61]"},{"why":"Defines the conversion-efficiency and smoothness criteria (one e-fold change of comoving Hubble radius) used to judge the numerical conversion.","marker":"[62]"},{"why":"Supplies the Planck 2018 observational targets for $n_s$ and $A_s$ against which the model's scalar spectrum is compared.","marker":"[50]"}],"fun_headline_variants":["DHOST bounce with extra scalar: first fully viable model","First viable two-field DHOST bounce: stable and luminal","Bounce with extra scalar: no instability, no superluminality","DHOST bounce passes all viability tests","No ghosts, no superluminality: first viable DHOST bounce"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the roughly 98% conversion of $\\delta\\chi$ into $\\zeta$ preserves the nearly scale-invariant spectrum and produces only small non-Gaussianities, a step the paper does not compute directly: it derives $n_s$ for $\\delta\\chi$ before conversion and checks conversion efficiency, then says it expects small non-Gaussianities.","fun_headline_variants_meta":{"raw":{"variants":["DHOST bounce with extra scalar: first fully viable model","First viable two-field DHOST bounce: stable and luminal","Bounce with extra scalar: no instability, no superluminality","DHOST bounce passes all viability tests","No ghosts, no superluminality: first viable DHOST bounce"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000679,"raw_usage":{"total_tokens":3117,"prompt_tokens":1009,"completion_tokens":2108,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":625,"completion_tokens_details":{"reasoning_tokens":2024}},"tokens_in":625,"tokens_out":2108,"duration_ms":15221,"temperature":1.0,"reasoning_tokens":2024,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T19:29:26.436265+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Solve the coupled perturbation equations (4.17) past the conversion end for a range of wavenumbers and compute the final $\\zeta$ power spectrum and bispectrum: if the resulting spectral index shifts from $n_s \\simeq 0.965$ by more than the Planck error bars, or if $f_{\\rm NL}$ exceeds the current observational bound, the 'fully viable' claim is refuted. A second, independent test: since the model gives a deep-blue tensor spectrum with negligible $r$ at the pivot scale, a primordial tensor signal near the inflation-scale upper bound would rule it out.","supporting_citations":[],"review_version":1}