{"id":"6476213b-1a7e-487a-9173-0f8453c3ff49","arxiv_id":"2501.01537","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"a-contraction, which holds uniformly in amplitude for inviscid shocks, fails for large viscous shocks in certain scalar and Navier-Stokes conservation laws.","lead":"This paper proves that a-contraction, a modern L2-based stability criterion for shock waves, fails for sufficiently large viscous shocks in a class of scalar conservation laws and in the barotropic Navier-Stokes system. This is a negative result: it shows the criterion is stronger than classical nonlinear stability and cannot be extended uniformly in shock amplitude.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The H-space approximation in §5.2 is invalid as written: the weight (y−v+)(y−v−) is negative on (v+,v−), so the 'norm' is not positive definite, and F contains a term in b_y that this H does not control.","rationale":"The paper's central claim—failure of a-contraction for large viscous shocks—rests on two constructions. The scalar construction (Theorem 1.1) appears to work: the functional F is controlled on a genuinely positive H, the sign bookkeeping in Sections 3.1–3.2 is consistent under the paper's integration convention, and the approximation of C^1 shifts by Lipschitz shifts is standard modulo minor mollifier details. The Navier-Stokes construction (Theorem 4.1), however, depends on an approximation space that is mis-specified. The integrand (y−v+)(y−v−) is negative on (v+,v−), so the 'norm' is not a norm; if one repairs the sign, the functional F still involves b_y, which the stated norm does not control. This is not a mere omission of a proof but a false statement in the text ('F is continuous on H'). Because Theorem 4.1 is a principal advertised extension (the abstract highlights the barotropic Navier-Stokes system), this is a load-bearing gap. The reader's conditional verdict is not changed, since the gap is localized and plausibly repairable by redefining H to include a weighted b_y term and arguing with the positive diffusion term as a lower bound; but the paper as written does not establish Theorem 4.1. Agreement with the reader is only partial: I do not view Lemma 5.1 as the weakest point—the short-time regularity there is standard for hyperbolic-parabolic systems—whereas the H-space flaw affects the core approximation argument.","tokens_in":1059,"tokens_out":1617,"duration_ms":537435,"concrete_test":"Evaluate the quadratic form defining H on the single test function (b,m)=(0,y) with v+=1, v−=2: ∫_1^2 (y−1)(y−2) dy = −1/6 < 0. If this computation is confirmed, the 'norm' in §5.2 is not positive definite, so the approximation step in Theorem 4.1 has no valid underlying space. A secondary check: take b_n(y) = n^(-1) sin(n(y−v+)) supported away from the boundary; b_n → 0 in L^2 but (p′b_n)_y does not vanish, so F(b_n,0) does not converge to F(0,0), disproving the asserted continuity of F on any topology that only controls m_y.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In Section 5.2, the authors define the approximation space H for the Navier-Stokes trial profiles by ‖(b,m)‖_H^2 = ‖b‖_2^2 + ‖m‖_2^2 + ∫_{v+}^{v−}(y−v+)(y−v−)|m_y|^2 dy. Since v+<v−, the weight (y−v+)(y−v−) is negative on the whole interval, so the quadratic form is not positive definite. For example, with v+=1, v−=2 and m(y)=y, the integral is −1/6, giving ‖(0,m)‖_H^2 < 0. Thus H is not a normed space, Lemma 3.1 cannot be invoked, and the density argument producing compactly supported C_c^∞ trial profiles with F>0 and the linear shift condition collapses. Even after correcting the sign to (y−v+)(v−−y), F is not continuous on H: it contains the term ∫ a(b(y))((p′(y)b(y))_y)^2 q(y) dy, which depends on b_y, whereas the stated H norm controls only m_y. The claim 'F is continuous on H' in §5.2 is therefore false. This is the step converting the non-compactly supported pair (w,g)=(1/p′, α1_[v+,2v+]) into an admissible initial perturbation, so Theorem 4.1 is not established as written. The scalar Theorem 1.1 is unaffected.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies whether the a-contraction property with shifts holds uniformly in shock amplitude for viscous conservation laws. It presents two negative results: Theorem 1.1 for the semilinear scalar equation u_t + A(u)_x = u_xx with strictly convex A, and Theorem 1.2 (via Theorem 4.1) for the barotropic Navier-Stokes system with p(v)=v^{-γ}, μ(v)=γv^{-γ}. In each case the authors construct a trial perturbation that makes the time derivative of the weighted relative entropy positive at t=0, then argue by approximation in a functional topology and a short-time continuity argument that there exist compactly supported smooth initial data for which the entropy increases initially under every Lipschitz shift. The paper thus claims a 'viscous destabilization' effect: a-contraction holds for the inviscid limit but fails for large viscous shocks, and the a-contraction property is stronger than classical nonlinear stability.","tokens_in":29852,"tokens_out":8545,"duration_ms":82883,"significance":"If the results were established, they would be a significant contribution to the theory of a-contraction for viscous shocks: they would show that uniform-in-amplitude a-contraction fails for natural classes of fluxes and for the barotropic Navier-Stokes system, contrasting with the inviscid theory and with small-shock viscous results. The paper is constructive, builds on prior work of Kang, Vasseur, Stokols, and others, and its main strategy is transparent. It does not appear to contain circular reasoning or parameter fitting; the claims are falsifiable counterexamples. However, as written, the proof has load-bearing gaps in the approximation steps, and those gaps affect both main theorems.","major_comments":[{"comment":"The quadratic form used to define the space H is not positive definite, so H is not a normed space. In Section 3.2 the norm is defined by ‖f‖_H^2 = ‖f‖_{L^2}^2 + ∫_0^{-K} y(K+y)|f'|^2 dy, and on the interval (-K,0) the weight y(K+y) is strictly negative. In Section 5.2 the analogous weight is (y-v+)(y-v−), which is strictly negative on (v+,v−) because v+<v−. Therefore the 'closure of C_0^∞ in this topology' is not a Banach space, Lemma 3.1 cannot be invoked, and the density argument that converts the trial functions w or (w,g) into compactly supported initial data collapses. This step is essential for both Theorem 1.1 and Theorem 4.1, since without it the initial perturbation is not in C_c^∞ as required by the theorem statements.","section":"Section 3.2 and Section 5.2"},{"comment":"Even after correcting the sign of the weight in the H norm, the functional F is not continuous on H in the system case. The functional defined in Section 5.1.1 contains the term ∫_{v+}^{v−} \\tilde a(b(y)) ((p'(y) w(y))_y)^2 q(y) dy, which depends on w_y (i.e. b_y in the notation of Section 5.2). The proposed H norm controls m_y through the weighted derivative term and only b in L^2; it does not control b_y. Thus the assertion 'F is continuous on H equipped with this topology' in Section 5.2 is not justified, and the approximation of the pair (w,g) = (1/p', α 1_{[v+,2v+]}) cannot be performed as written.","section":"Section 5.2"},{"comment":"The Navier-Stokes short-time argument depends on Lemma 5.1, which asserts that the perturbed solution U satisfies U - \\tilde U ∈ C([0,T]; H^2)^2. The proof is only a sketch, and the iterative viscous approximation contains several unproved uniformity claims: the positivity of inf_{k,ν} T^*_{k,ν}, the ν-uniform contractivity in C([0,T];L^2), and the passage to the limit yielding a solution in C([0,T];H^2). Since Lemma 3.5 and the conclusion that the entropy derivative remains positive for a short time require exactly this regularity, Lemma 5.1 is load-bearing for Theorem 4.1. The references to existing well-posedness results in [21] are not sufficient, because the precise equivalence with the transformed system (26) is not detailed.","section":"Lemma 5.1"}],"minor_comments":[{"comment":"In the proof of Lemma 4.2, the displayed derivative D_λ f(0,λ*) is written as ∫ \\tilde a'(b(y)) p'(y) φ(y) dy, but the formula for f shows it should be ∫ \\tilde a(b(y)) p'(y) φ(y) dy, matching the denominator of λ*.","section":"Lemma 4.2"},{"comment":"The sentence 'Hence, A′(x) goes to −∞ as x goes to ∞' should read 'as x goes to −∞', since the preceding argument concerns the behavior on R^-.","section":"Proof of Lemma 2.3"},{"comment":"The phrase 'with \\tilde U − (u0,u0) ∈ (C_0^∞(R))^2' appears to contain a typo; the intended condition is that the initial data (v0,h0) satisfy (v0,h0) − \\tilde U ∈ (C_0^∞(R))^2.","section":"Theorem 4.1"},{"comment":"The existence of the non-increasing smooth function \\tilde w satisfying properties (1)-(4) is asserted without proof. A short construction (e.g., a smooth monotone transition from 1 on [-1,-1/2] to -C on [-1/4,0], with C chosen by continuity of the integral) would make the argument self-contained.","section":"Section 3.1.1"},{"comment":"The pair (b,m) is introduced in the definition of ‖(b,m)‖_H, but the functional F is defined on (w,g); the notational mismatch should be clarified so that the reader can see which component carries the derivative weight.","section":"Section 5.2"}],"recommendation":"major_revision","confidential_remarks":"The two main theorems are not established as written because the key approximation step uses an indefinite 'norm' — this is a genuine mathematical error in both the scalar and system sections, not a mere presentation issue. In addition, the system case has a separate structural mismatch: F depends on w_y while the proposed H norm controls only m_y. These problems are potentially fixable by introducing a genuinely positive norm and re-proving continuity of F, but this requires nontrivial work and may affect the constructions. The scalar theorem may be salvageable if the weight sign is corrected and the continuity of F is proved with respect to a positive norm; if that cannot be done, the scalar result would need substantial revision as well. I recommend major revision rather than rejection because the central idea is plausible and the contribution would be valuable if the approximation gaps are closed."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read Blochas-Cheng on viscous destabilization. My take: the scalar result is probably right and worth a paper, but the Navier-Stokes half has a load-bearing gap in the approximation argument.\n\nWhat's actually new: Theorem 1.1 gives a class of super-polynomial fluxes where a-contraction fails uniformly in shock amplitude, contrasting with Kang's small-shock results. That's a clean negative result, and the proof strategy—constructing a perturbation that makes the time derivative of the weighted relative entropy positive at t=0—is coherent. The paper writes out the relative entropy identities carefully, and the scalar estimates in Section 3.1 are detailed enough to follow.\n\nThe soft spot is Section 5.2. The stress-test note is correct: the H-norm there is not a norm. The weight (y−v+)(y−v−) is negative on (v+,v−), so the quadratic form is indefinite. That alone is a simple sign typo—should be (y−v+)(v−−y). But even after fixing the sign, the claim that F is continuous on H is false: F contains a term involving b_y, but the H norm controls only m_y. Also, the trial profile g = α 1_{[v+,2v+]} is a step function, which is not in the closure of C_c^∞ under a norm that includes a weighted L2 norm of the derivative. So Theorem 4.1 is not established as written.\n\nNote the same sign issue appears in the scalar H-norm in Section 3.2: y(K+y) is negative on (-K,0). There it's probably a typo, and since the scalar w is smooth and the functional depends on w_y with a coefficient comparable to the corrected weight, the scalar proof likely survives once the sign is fixed. But the Navier-Stokes issues are not just typos.\n\nIn sum: the paper deserves a serious referee, but the NS part needs substantial revision. I'd suggest the authors split the paper or fix the approximation topology properly. The scalar theorem alone is a solid contribution.\n\nRecommendation: send to peer review, but flag the NS gap as major. I would not cite the paper as-is; I'd revisit once the NS half is repaired.","headline":"The scalar counterexample is likely right, but the Navier-Stokes half rests on a broken approximation space; the paper deserves review but needs real revision.","tokens_in":30328,"tokens_out":6867,"would_cite":false,"duration_ms":68415,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35B35","35L65","76N15","35L67"],"pacs":[],"model":"deepseek-v4-flash","headline":"Large viscous shocks of certain conservation laws fail a-contraction even for arbitrarily small perturbations, although the inviscid waves contract and the viscous waves are nonlinearly stable.","keywords":["Stability","Viscous conservation law","Relative entropy","Shock wave","Contraction","a-contraction","Navier-Stokes","Scalar conservation law"],"falsifier":"Compute the scalar quadratic form $F(w)$ defined in Section 3.1.1 for $A(u)=e^{-u}+u-1$, $\\tilde a=1$, and $w(y)=\\tilde w(y/K)$ at $K=2^m$. The theorem predicts $F(w)>0$ for all sufficiently large $K$ with $Y(w)=0$; any such $K$ with $F(w)\\le0$ refutes Theorem 1.1. For the Navier-Stokes theorem, run the transformed system (26) with $p(v)=v^{-\\gamma}$, $v_+=10^{-3}$, and initial perturbation $(1/p'(y), \\alpha\\,1_{[v_+,2v_+]})$, and check whether the weighted relative entropy derivative at $t=0$ is positive; a non-positive value refutes Theorem 4.1.","tokens_in":29248,"feed_emoji":"🌊","tokens_out":15058,"duration_ms":120494,"temperature":0.7,"pith_summary":"This paper asks whether the $a$-contraction property --- a weighted $L^2$ contraction up to shift that holds uniformly in shock amplitude for inviscid shocks and for small viscous shocks --- remains valid for large viscous shocks. The authors establish that it does not: for scalar conservation laws with strictly convex fluxes whose derivative goes to $-\\infty$ faster than any polynomial, and for $1$-shocks of a barotropic Navier-Stokes system as the post-shock specific volume tends to zero, arbitrarily small smooth perturbations can increase the weighted $L^2$ distance for a short time under every Lipschitz shift. Because ordinary nonlinear orbital stability of these viscous shocks is known, the failure shows that $a$-contraction is strictly stronger than asymptotic stability. The inviscid versions of the same models are contractive up to shift, so the paper calls this effect viscous destabilization.","feed_headline":"Large viscous shocks defeat a-contraction","feed_subtitle":"Viscosity can destroy a stability property that holds for the inviscid equation, in scalar laws and Navier-Stokes.","key_machinery":"The engine is the relative entropy method in shock variables. With $y=S(x)$ and $w(y,t)=u^{-X}(S^{-1}(y),t)-y$, the time derivative of the weighted $L^2$ distance splits as $2\\dot X(t)Y(w)+Z(w)+R_1(w)$, where $Y$ is a linear functional controlling the shift, $Z$ is a quadratic form involving weighted integrals of $w^2$, $|w_y|^2$, and the relative flux $A(w+y|y)$, and $R_1$ is a cubic remainder. The construction chooses $w$ with $Y(w)=0$ and $Z(w)>0$. The flux-growth hypothesis enters through Lemma 2.4: for arbitrarily large $K$, the ratio $|\\sigma/A'(-K)|$ can be made as small as desired, so the positive term $\\int \\tilde a A'' w^2$ dominates all error terms and $Z(w)\\ge c|A'(-K)|$. Compact support and Lipschitz shifts are handled by approximating the profile in a weighted $H$ norm and by a continuity lemma (Lemma 3.5) that preserves the sign of the time derivative over a short interval. In the Navier-Stokes case, the effective velocity $h=u+p(v)_x$ transforms the system into a tractable form, and the pair $w=1/p'(y)$, $g=\\alpha\\,1_{[v_+,2v_+]}$ with $\\alpha$ from Lemma 4.3 plays the same role: it solves the linear shift condition and leaves a positive quadratic form.","core_discovery":"The central claim is a pair of counterexamples. Theorem 1.1: for any smooth strictly convex flux $A$ with $A'(0)=0$ and with $A'(x)$ going to $-\\infty$ faster than every polynomial as $x\\to-\\infty$, and for any fixed weight profile $\\tilde a\\in W^{2,\\infty}([0,1])$, there are shock amplitudes $K_n\\to\\infty$ such that the viscous shock $S$ joining $0$ to $-K_n$ admits compactly supported smooth initial data $u_0$ with $u_0-S\\in C_c^\\infty$ for which, for every Lipschitz shift $X$ and all sufficiently small $t>0$, the weighted $L^2$ distance $\\int_{\\mathbb R} \\tilde a(-S(x)/K_n)\\,|u(x+X(t),t)-S(x)|^2\\,dx$ is strictly larger than at $t=0$. Theorem 1.2 transfers the same conclusion to the barotropic Navier-Stokes system with $p(v)=v^{-\\gamma}$, $\\mu(v)=\\gamma v^{-\\gamma}$, for $1$-shocks with $v_+$ sufficiently small, with perturbations arbitrarily small in $H^s$, using the relative entropy $\\eta=h^2/2+Q(v)$ and weights $\\tilde a(b(\\tilde v(x)))$. The underlying discovery is that $a$-contraction fails uniformly in the amplitude even though each viscous wave is nonlinearly stable, and even though the inviscid scalar model with the same flux satisfies $L^2$-contraction up to shift.","pith_inferences":["The proof's reliance on $\\sigma/A'(-K)\\to0$ suggests the same failure should occur for any strictly convex flux whose derivative goes to $-\\infty$ faster than linearly, not only super-polynomially; the super-polynomial hypothesis is one sufficient route to that ratio.","For polynomial fluxes, where $\\sigma/A'(-K)$ stays bounded, uniform $a$-contraction may still be possible, so the growth threshold separating these counterexamples from the cubic-flux large-shock result could be characterized by the decay rate of $A'$.","A testable quantitative prediction is that the short-time entropy increase scales roughly like $|A'(-K)|$ in the scalar case and like $|\\sigma|$ in the fluid case, so $T^*$ should shrink as the amplitude grows; direct numerical evaluation of $F(w)$ for $A(u)=e^{-u}+u-1$ at $K=10^n$ would check this.","For the Navier-Stokes system, the same construction may extend to $2$-shocks or to other pressure laws with $p'(v)\\to-\\infty$ at vacuum, because the proof only uses $[p]\\to\\infty$ and the bound of Lemma 4.4."],"forward_implications":["Uniform-in-amplitude $a$-contraction for viscous scalar conservation laws cannot hold for the standard weight family when the flux is super-polynomial; any such theorem must restrict the flux growth or enlarge the weight class.","Viscous destabilization is real: the inviscid scalar equation with the same flux is $L^2$-contractive up to shift for shocks of every size, while the viscous equation is not, even for arbitrarily small perturbations.","$a$-contraction is strictly stronger than nonlinear stability: the counterexample perturbations are arbitrarily small in $H^s$ but still increase the weighted distance, despite the underlying shock being asymptotically stable.","For the Navier-Stokes system the failure is not a large-perturbation artifact; Theorem 1.2 produces counterexamples with initial perturbation norm below any prescribed $\\delta>0$.","The mechanism is amplitude-driven: the relevant small parameter is the ratio $\\sigma/A'(-K)$ for scalar fluxes, or the post-shock specific volume $v_+$ for the fluid system, not the size of the perturbation."],"supporting_citations":[{"why":"Supplies the shock-variable change of coordinates and the expansion strategy: the paper fixes the flux and varies shock amplitude, reversing [20]'s setup.","marker":"[20]"},{"why":"Establishes a-contraction for small shock profiles of the barotropic Navier-Stokes system and provides the relative entropy formula that Lemma 4.1 imports.","marker":"[18]"},{"why":"Proves a-contraction with nontrivial weights for small scalar viscous shocks; the weight family whose uniform-in-K contraction is shown to fail is the one used here.","marker":"[34]"},{"why":"Proves L2-contraction up to shift for inviscid scalar shocks, providing the contrast that makes the failure a viscous destabilization.","marker":"[24]"},{"why":"Foundation of the relative entropy method: the time-derivative representation for the entropy of one solution relative to another is the framework the paper extends.","marker":"[5]"},{"why":"Establishes asymptotic stability of viscous scalar shocks without smallness, the classical result that the paper shows does not imply a-contraction.","marker":"[15]"}],"fun_headline_variants":["Viscosity kills a-contraction for big shocks","Large shocks defeat a-contraction when viscous","Viscous destabilization: a-contraction fails at large amplitude","a-contraction fails uniformly for large viscous shocks","Large viscous shocks lose a-contraction despite stability"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"For the Navier-Stokes half, the proof needs the perturbed solution to differ from the shock profile by a function with two square-integrable spatial derivatives, continuously in time on a short interval; this is asserted in Lemma 5.1 with only a sketch, and the fluid counterexample collapses if it fails. The scalar half does not rely on this lemma.","fun_headline_variants_meta":{"raw":{"variants":["Viscosity kills a-contraction for big shocks","Large shocks defeat a-contraction when viscous","Viscous destabilization: a-contraction fails at large amplitude","a-contraction fails uniformly for large viscous shocks","Large viscous shocks lose a-contraction despite stability"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000499,"raw_usage":{"total_tokens":2490,"prompt_tokens":1036,"completion_tokens":1454,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":652,"completion_tokens_details":{"reasoning_tokens":1393}},"tokens_in":652,"tokens_out":1454,"duration_ms":10124,"temperature":1.0,"reasoning_tokens":1393,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T22:26:23.386781+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the scalar quadratic form $F(w)$ defined in Section 3.1.1 for $A(u)=e^{-u}+u-1$, $\\tilde a=1$, and $w(y)=\\tilde w(y/K)$ at $K=2^m$. The theorem predicts $F(w)>0$ for all sufficiently large $K$ with $Y(w)=0$; any such $K$ with $F(w)\\le0$ refutes Theorem 1.1. For the Navier-Stokes theorem, run the transformed system (26) with $p(v)=v^{-\\gamma}$, $v_+=10^{-3}$, and initial perturbation $(1/p'(y), \\alpha\\,1_{[v_+,2v_+]})$, and check whether the weighted relative entropy derivative at $t=0$ is positive; a non-positive value refutes Theorem 4.1.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the shock-variable change of coordinates and the expansion strategy: the paper fixes the flux and varies shock amplitude, reversing [20]'s setup."},{"cited_title":"Contraction propert y for large perturbations of shocks of the barotropic Navier-Stokes system","cited_arxiv_id":null,"evidence_quote":"Establishes a-contraction for small shock profiles of the barotropic Navier-Stokes system and provides the relative entropy formula that Lemma 4.1 imports."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Proves a-contraction with nontrivial weights for small scalar viscous shocks; the weight family whose uniform-in-K contraction is shown to fail is the one used here."},{"cited_title":"Dafermos","cited_arxiv_id":null,"evidence_quote":"Foundation of the relative entropy method: the time-derivative representation for the entropy of one solution relative to another is the framework the paper extends."},{"cited_title":"Il’in and Olga A","cited_arxiv_id":null,"evidence_quote":"Establishes asymptotic stability of viscous scalar shocks without smallness, the classical result that the paper shows does not imply a-contraction."}],"review_version":1}