{"id":"77156aa7-9551-4865-8f4b-16e278708798","arxiv_id":"2512.15620","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Small-BV data for u_t + A(u)u_x = ε(B(u)u_x)_x with commuting A, B yield a uniform-in-ε total-variation bound, and the conservative case has a unique Liu-admissible vanishing-viscosity limit.","lead":"The paper proves global-in-time total-variation bounds for viscous approximations of n×n hyperbolic conservation laws when the flux matrix A and the state-dependent viscosity matrix B commute, extending the Bianchini-Bressan result for B = I to non-constant viscosity. Uniform BV bounds are the compactness ingredient that selects the physically admissible shock solution as smoothing vanishes, so the result pushes multi-equation vanishing-viscosity limits toward the scalar-case","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 2.1 rests on announced estimates (7.50), (8.36), (9.75) and the omitted proof of Proposition 4.4; without these the BV bootstrap cannot be checked.","rationale":"The reader’s weakest-assumption choice was commutativity AB=BA. That is indeed a restrictive hypothesis and is unverified for the promised NSK application, but it is an explicit assumption of the theorem; if the theorem is read as stated, commutativity is not an internal flaw. The more load-bearing soft spot is that the proof’s decisive estimates are deferred, not demonstrated. Sections 11 and 12 are promised to prove (7.50) and (8.36), and Lemma 9.7 depends on (9.75) from those sections; the reader cannot check whether the many O(1) constants and remainder terms really are small. Proposition 4.4 is explicitly omitted. These are the points where an actual error would most directly destroy Theorem 2.1 and hence Corollary 2.2. Since the reader already assigned CONDITIONAL and cited the deferred pieces, my stress-test does not move the verdict; it sharpens the reason: the theorem should be accepted only after the omitted proof and the deferred estimates are supplied and verified. The commutativity concern remains important for the announced NSK application but is secondary to the internal proof gap.","tokens_in":119028,"tokens_out":14504,"duration_ms":141766,"concrete_test":"Independently re-derive the estimate (9.75) from the explicit formulas in Sections 11.1–11.2, using Lemma 10.9 for the term (v_{j,x} − μ_j^{-1}(λ̃_j − λ_j^* + θ_j)v_j)_{xx}(B − μ_j I_n)[·]. Then substitute (9.75) into (9.74)–(9.77) and check that the coefficient of |v_m v_{m,xxx}| + |v_m w_{m,xxx}| on the right side is O(δ_0^2) with a constant < 1, so the bootstrap closes with L^1([ˆt,T*]×R) norm O(δ_0^2). In parallel, fill in the omitted proof of Proposition 4.4 following [5, Proposition 2.3] and [26, Proposition 3.6] for the non-constant B(u) case, and verify the claimed ||u_x(t)||_{L^1} ≤ δ_0/2 on [0, ˆt].","verdict_should_be":"UNCHANGED","load_bearing_attack":"The global BV conclusion is not currently supported by a fully checkable proof. Proposition 4.4, which supplies the initial interval [0, ˆt] with the small-L1 bound used to enter the wave-decomposition machinery, is explicitly stated with “We omit the proof here.” More centrally, the two estimates that make the Section 9 maximal-time bootstrap close are only announced: (7.50) bounds φ_i, ψ_i by the Λ^1–Λ^{6,1} terms plus an ε-dependent remainder, and (8.36) bounds Φ_i, Ψ_i by Λ^1–Λ^8 plus a remainder; their proofs are deferred to Sections 11–12. The specific subestimate (9.75), |ψ_{j,x}| = O(Σ Λ^l) + R̃_ε, is then used inside Lemma 9.7 to control the third-order terms Λ^2_j = (|v_{j,xxx}|+|w_{j,xxx}|)(|w_j|+|v_j|)|v_k|. Without (9.75), the bootstrap (9.76)–(9.78) cannot close, and the subsequent bound (9.80) for Λ^2 fails. This is not a demonstrated contradiction, but it is the decisive unverified point: the theorem’s correctness is conditional on dozens of pages of O(1) estimates and an omitted base-interval proof that the text does not make available to the reader.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a proof of global-in-time uniform BV bounds for the 1-D parabolic-hyperbolic system u_t + A(u)u_x = (B(u)u_x)_x, for strictly hyperbolic A and non-singular viscosity B, under the commutativity assumption AB=BA and small BV initial data. The proof follows the Bianchini–Bressan template: parabolic regularization on a short time interval, decomposition of u_x and u_t in a modified traveling-wave basis, derivation of advection-diffusion equations with remainder terms, introduction of auxiliary effective fluxes z_i and hat z_i to control the new second-order terms w_{i,xx}v_i - v_{i,xx}w_i, and a maximal-time bootstrap combining transversal interaction estimates, shortening-curve functionals, energy estimates, and higher-derivative estimates. In the conservative case A=Df, a corollary asserts L^1_loc convergence, as ε→0, to the unique Liu-admissible weak solution of the limiting conservation law.","tokens_in":119397,"tokens_out":3659,"duration_ms":41424,"significance":"If the proof is correct, this is a substantial extension of the Bianchini–Bressan theory to a genuinely nonlinear, non-constant viscosity matrix for the commuting class AB=BA, and it would provide a path toward the stated Navier–Stokes–Korteweg application. The paper contains many explicit, labor-intensive computations, a careful construction of the modified wave basis, and a clear articulation of the main technical obstacles. However, the central claims are currently conditional on several announced estimates and on an omitted base-interval proof; the reader cannot verify the maximal-time bootstrap from the text as written. The paper is honest about these gaps, but the gaps are load-bearing rather than cosmetic.","major_comments":[{"comment":"Proposition 4.4 supplies the initial small-L^1 interval (4.17) that is used to enter the wave-decomposition machinery on [t̂,T]. Its proof is explicitly omitted: 'We omit the proof here.' This is not a technical footnote: without (4.17), Lemma 6.9 and all subsequent decompositions have no stated starting point. The reference to [5, Proposition 2.3] and [26, Proposition 3.6] is not enough, because the present system has non-constant B and higher-order regularity is needed. A complete proof, or a precise reduction to the cited results, is required.","section":"§4, Proposition 4.4"},{"comment":"The two central remainder bounds that make the bootstrap close are only announced. Eq. (7.50) bounds φ_i, ψ_i by the Λ^l terms plus an ε-dependent remainder; Eq. (8.36) bounds Φ_i, Ψ_i similarly; both proofs are deferred to Sections 11–12. More critically, the pointwise estimate (9.75), |ψ_{j,x}| = O(Σ Λ^l) + R̃_ε, is used inside Lemma 9.7 to control the third-order terms Λ²_i, and then (9.76)–(9.78) bootstrap to obtain (9.80). Without an independent proof of (9.75), Lemma 9.7 is circular and the estimate (9.81) for the bootstrap is unsupported. These are not merely deferred technicalities; they are the decisive estimates of the paper.","section":"§7.3, §8, §9; Eqs. (7.50), (8.36), (9.75)"},{"comment":"The commutativity assumption AB=BA is load-bearing from the very start of the traveling-wave construction. In (5.7) it is used to obtain Z² in diagonal form, forcing the center subspace N_i to have dimension n+2. Without this, the invariant manifold M_i and the modified basis r̃_i of Section 6 do not exist in the form used, and the gradient decomposition (6.12)–(6.13) breaks down. The paper should state clearly that Theorem 2.1 is a theorem for the commuting subclass, and it should verify that the advertised NSK application satisfies AB=BA; this verification is absent.","section":"§5, Eq. (5.7) and §6, Eqs. (6.12)–(6.13)"},{"comment":"Corollary 2.2 is proved only by a sketch. The convergence to a weak solution uses Helly's theorem and the uniform BV bound, but the identification of the limit as the unique Liu-admissible solution relies on [31, Theorem 2.1] and the Bressan–De Lellis uniqueness class [11]. The paper does not verify that, under the present assumptions (including AB=BA and the non-constant B), the limit satisfies the required Liu admissibility criterion and lies in the uniqueness class. This is a load-bearing step for the vanishing-viscosity application, not a routine consequence of the BV estimate.","section":"§9, §9.8, Corollary 2.2"},{"comment":"The claim that G_i ∈ L^1([t̂,T],L^1(R)) is asserted before the bootstrap. The text argues that Λ²_j is integrable by invoking Lemma 9.7, but Lemma 9.7 itself depends on the announced estimate (9.75). Thus the proof of integrability of the forcing terms used to define T* is part of the deferred material. This should be made explicit, and the dependencies should be reorganized so that the maximal-time bootstrap is not circular.","section":"§9, definition of T* and G_i"}],"minor_comments":[{"comment":"Several displayed estimates contain obvious typos or garbled symbols, e.g. 'co' instead of c_0 in (4.7), and missing parentheses in (4.9)–(4.10). These make an already technical lemma harder to check.","section":"§4, Eq. (4.7), (4.9), (4.10)"},{"comment":"The proof ends with 'This completes the proof Lemma 6.5', but the statement is Lemma 6.4. Please correct the cross-reference.","section":"§6, Lemma 6.4 proof"},{"comment":"The notation eφ_i, φ_i, ψ_i, Φ_i, Ψ_i, R_ε, R̃_ε is introduced in different places and sometimes without a consistent subscript. A table summarizing the notation and the exact location of each announced estimate would aid readability.","section":"General notation"},{"comment":"Remark 6.2 says the estimates are uniform in ε on [0,C(1/ε)] with C(1/ε)→∞, while the proof of Theorem 2.1 fixes ε after assuming a finite maximal time. The logical link is not spelled out; please clarify how the fixed-ε proof yields the ε-uniform statement needed for Corollary 2.2.","section":"§9.8 and Remark 6.2"},{"comment":"The abstract and introduction promise an application to the Navier–Stokes–Korteweg system, but the commutativity condition AB=BA is not checked for that system. Either perform the check or explicitly state that the application is conditional on the commuting subclass.","section":"Abstract and Introduction"}],"recommendation":"major_revision","confidential_remarks":"The paper is a serious contribution with a coherent architecture, and the authors are candid about the hard steps. However, as currently written, the proof is not verifiable: Proposition 4.4 is omitted, and the key estimates (7.50), (8.36), (9.75) are only announced. I would recommend requesting a revised version in which these estimates are proved in full, or at least the dependency chain is restructured so that the bootstrap does not rely on circular or deferred claims. The commutativity restriction should also be addressed explicitly in the advertised NSK application."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe paper is a real attempt to extend the Bianchini-Bressan vanishing-viscosity framework from B=I to state-dependent B(u) under the commutativity assumption AB=BA. The main new devices—the effective fluxes z_i and z-hat_i of (8.4)-(8.5), the cut-off-dependent travelling-wave basis r̃_i in (6.12), and the third-order bootstrap—are substantive, and the overall architecture (smoothing window → wave decomposition → remainder estimates → maximal-time bootstrap) is coherent. The paper is honest about its hard steps and does not fit constants to force the conclusion: all quantities are shown to exist, and the remainder bounds are derived from the PDEs.\n\nThat said, as written the proof cannot be checked end-to-end. Proposition 4.4, which provides the small-L1 bound on the initial interval, has its proof omitted. More importantly, the estimates that make the Section 9 bootstrap close—(7.50), (8.36), and (9.75)—are only announced in the text, with proofs deferred to Sections 11-12; and Lemma 9.7 uses (9.75) as an input to control the third-order terms Λ_j^2. Unless Sections 11-12 deliver those exact bounds, the maximal-time argument doesn't close. This is not a demonstrated contradiction, but it is the load-bearing uncertainty.\n\nThere are also two scope issues. The abstract promises a concrete application to the Navier-Stokes-Korteweg system, but the body contains no such application; and the commutativity AB=BA, while flagged as decisive on p.5, is an ad hoc restriction that the NSK system is not checked against. The conservative-case corollary also relies on 'similar arguments' to [5,6] for Liu admissibility and uniqueness, so that part is a sketch rather than a proof.\n\nProportionately: the paper deserves serious refereeing, and the strategy may well work. But what is currently in front of the referee is an incomplete proof, not a finished one. The right outcome is major revision: either include the deferred estimates, or clearly mark them as part II and state that the present paper is conditional on that sequel; and either add the NSK application or delete the promise.\n\nWho gets value: specialists in hyperbolic conservation laws and vanishing viscosity, particularly those working with center-manifold and travelling-wave decompositions. It is worth citing for the new technique, with a caveat about the gaps. I'd send it to peer review rather than desk reject, but I would want a referee with patience and a hard requirement that the announced estimates actually appear before acceptance.","headline":"Serious extension of Bianchini-Bressan to non-constant viscosity, but the main theorem is currently conditional on several deferred estimates and an omitted proof; send to peer review but require the missing pieces.","tokens_in":120065,"tokens_out":2640,"would_cite":true,"duration_ms":28304,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35L65","35K55"],"pacs":[],"model":"deepseek-v4-flash","headline":"Under the commutativity AB=BA, the parabolic approximation admits global solutions with total variation uniformly bounded in the viscosity parameter; in the conservative case the vanishing-viscosity limit is the unique Liu-admissible weak s","keywords":["hyperbolic conservation laws","vanishing viscosity limit","uniform BV estimates","viscous travelling waves","center manifold","commutativity AB=BA","effective fluxes","Liu admissibility"],"falsifier":"Take a 2×2 system with A(u)=diag(λ1(u),λ2(u)) and constant B=[[1,1],[0,2]] (so AB≠BA at a state where A is not scalar). Compute the linearization Z in (5.5): Z² is not diagonalizable, so the center subspace N_i in (5.8) has dimension less than n+2 and the manifold M_i underlying the basis (6.12)–(6.13) is unavailable — this would show the theorem's method does not extend without commutativity. Alternatively, for a commuting example B(u)=diag(1+u₂, 2−u₁), expand the remainder φ₁ in (7.42) and check that every term containing v_{2,xx} cancels and appears only in the combination w_{2,xx}v₂−w₂v_{2","tokens_in":118757,"feed_emoji":"🌊","tokens_out":11780,"duration_ms":102902,"temperature":0.7,"pith_summary":"This paper proves that the parabolic regularization u_t + A(u)u_x = (B(u)u_x)_x has global solutions whose total variation remains bounded for all time, uniformly in the viscosity parameter, when the initial data have small total variation and the flux matrix A commutes with the viscosity matrix B. The result extends the identity-viscosity theory to nonlinear, non-constant viscosity by decomposing the gradient along travelling-wave directions, with careful cutoffs so that the troublesome ratios |v_i,x|/|v_i| stay controlled. In the conservative case A = Df, the epsilon-family converges to a global weak solution of the limiting hyperbolic conservation law that satisfies the Liu admissibility condition and is unique. This supplies a main ingredient for proving that physically relevant solutions of strictly hyperbolic systems are selected by the vanishing-viscosity process.","feed_headline":"Commuting viscosity yields global BV bounds for hyperbolic systems","feed_subtitle":"In the conservative case the vanishing viscosity limit is a unique Liu-admissible weak solution.","key_machinery":"The central objects are the viscous travelling-wave manifolds M_i — center manifolds of the travelling-wave ODE of dimension n+2, whose dimension and the absence of non-diagonal second-derivative terms depend on the commutativity AB=BA (used to diagonalize Z² in (5.7)) — and the associated gradient basis \\tilde r_i(u,\\bar v_i^ε ξ_i, σ_i). The basis is engineered with cutoffs so that \\tilde r_i reduces to the eigenvector r_i(u) of A whenever |v_i| is tiny, |w_i/v_i| is large, or another wave dominates, which prevents the ratios |v_{i,x}|/|v_i| from becoming singular. The workhorse identities are the coupling formula (6.51), expressing µ_i v_{i,x} − (\\tilde λ_i − λ_i^*)v_i − w_i as a sum of in","core_discovery":"The central claim is Theorem 2.1: for a strictly hyperbolic matrix A with n distinct real eigenvalues and a smooth invertible viscosity matrix B with positive eigenvalues, if A and B commute and the initial data have sufficiently small total variation, the Cauchy problem u_t + A(u)u_x = (B(u)u_x)_x has a unique global solution with TV(u(t)) ≤ L1 TV(ū) and the time-continuity estimate ||u(t)−u(s)||_{L¹} ≤ L2(|t−s|+|√t−√s|). The proof rescales to viscosity coefficient 1 and decomposes u_x into a basis of viscous travelling waves, reducing the system to nearly independent scalar advection-diffusion equations for the wave strengths v_i, w_i. New effective-flux variables z_i and \\hat z_i are intr","pith_inferences":["The abstract promises an application to the Navier–Stokes–Korteweg visco-dispersive limit, but the body never verifies that the NSK viscosity/diffusion matrices commute with the acoustic matrix; if they do not, the advertised application is not a corollary of this theorem.","The commutativity AB=BA is not merely a convenience: the dimension of the manifolds M_i and the absence of v_{j,xx} terms in the equation for v_i both rest on joint diagonalizability. Systems with non-commuting B are outside the theorem's scope and would need a genuinely different analysis.","A testable refinement would be to weaken AB=BA to simultaneous symmetrizability or to B being a polynomial in A; the proof's Section 5 suggests the center manifold would still have the required dimension under such hypotheses, but the paper does not explore this.","The √t rate in the time-continuity estimate (2.5) mirrors parabolic smoothing; for a concrete 2×2 commuting example one could check numerically whether the constant L2 is independent of the wave amplitudes in the way the theorem asserts."],"forward_implications":["For every system satisfying (HA), (HB) and AB=BA, solutions of the ε-equation satisfy TV(u^ε(t)) ≤ L1 TV(ū) uniformly in ε and t, so bounded-variation compactness applies to the whole family.","In the conservative case A=Df, the family converges in L¹_loc to u^∞, which is a global weak solution of u_t + (f(u))_x = 0 and satisfies the Liu admissibility condition at shocks.","The limit u^∞ lies in the L¹ uniqueness class for hyperbolic conservation laws, so the entire family converges (not just a subsequence) and the vanishing-viscosity process selects a unique physical solution.","For non-conservative systems, the same uniform BV estimates are the compactness ingredient needed to define a vanishing-viscosity solution of u_t + A(u)u_x = 0, where the usual weak-solution framework is unavailable."],"fun_headline_variants":["Commuting viscosity yields global BV bounds for hyperbolic systems","Global BV estimates for conservation laws when viscosity commutes","Unique weak limit from vanishing viscosity with commuting matrices","Commuting viscosity ensures global BV bounds and unique limit"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The entire proof depends on the flux matrix A(u) and the viscosity matrix B(u) commuting (AB=BA), because this joint diagonalizability is what gives the travelling-wave center manifolds dimension n+2 and prevents non-diagonal second-derivative terms v_{j,xx} (j≠i) from appearing in the scalar equation for each wave amplitude v_i.","fun_headline_variants_meta":{"raw":{"variants":["Commuting viscosity yields global BV bounds for hyperbolic systems","Global BV estimates for conservation laws when viscosity commutes","Unique weak limit from vanishing viscosity with commuting matrices","Commuting viscosity ensures global BV bounds and unique limit"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000547,"raw_usage":{"total_tokens":2447,"prompt_tokens":734,"completion_tokens":1713,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":478,"completion_tokens_details":{"reasoning_tokens":1651}},"tokens_in":478,"tokens_out":1713,"duration_ms":11556,"temperature":1.0,"reasoning_tokens":1651,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T16:43:20.854492+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a 2×2 system with A(u)=diag(λ1(u),λ2(u)) and constant B=[[1,1],[0,2]] (so AB≠BA at a state where A is not scalar). Compute the linearization Z in (5.5): Z² is not diagonalizable, so the center subspace N_i in (5.8) has dimension less than n+2 and the manifold M_i underlying the basis (6.12)–(6.13) is unavailable — this would show the theorem's method does not extend without commutativity. Alternatively, for a commuting example B(u)=diag(1+u₂, 2−u₁), expand the remainder φ₁ in (7.42) and check that every term containing v_{2,xx} cancels and appears only in the combination w_{2,xx}v₂−w₂v_{2","supporting_citations":[],"review_version":1}