{"id":"068b9814-8dd7-4e44-808f-f0d0972a8be5","arxiv_id":"2501.04205","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"For s>5/2, a polynomial derivative Schrödinger nonlinearity is locally well-posed in H^s on the torus exactly when the imaginary part of the derivative of the nonlinearity with respect to ∂_x u has zero mean for every datum.","lead":"This paper proves a complete if-and-only-if condition for when a broad class of semi-linear Schrödinger equations with derivative nonlinearities on the torus admit unique short-time solutions in Sobolev spaces. The condition, a nonlinear analogue of the Mizohata condition, tells researchers whether a given polynomial nonlinearity is well-posed or admits data with no solution at all.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified: the iff characterization for polynomial nonlinearities is internally consistent and the polynomial restriction is an explicit scope limitation, not a correctness risk.","rationale":"The central claim is a necessary-and-sufficient condition for polynomial derivative NLS on T. I stress-tested the three load-bearing steps: (1) the energy estimates for the parabolic regularization, where the gauge transformation must turn the problematic Fβ∂xw term into a skew-symmetric operator with real mean; (2) the Bona-Smith approximation for H^s-continuous dependence; (3) the non-existence proof, where rough data are constructed and the a priori estimate forces P±φ smooth. In each case the estimates are consistent with the stated assumptions. The polynomial restriction, flagged by the reader, is explicitly stated in the abstract and Theorem 1.3; Remark 1.8 discusses a partial C^7 extension but does not claim it as a theorem. Disagreement with consensus is not present: the threshold s>5/2 is acknowledged non-optimal, and the overlap with Tsugawa's independent work is disclosed. I therefore see no change to the ACCEPT verdict; if anything, the paper's internal consistency supports it.","tokens_in":27033,"tokens_out":34005,"duration_ms":305105,"concrete_test":"Independently re-derive the passage from (3.7) to (3.12) and then (3.25), tracking overbars: confirm the coefficient of ∂xW is Fβ (whose P0 part is real by (1.4)) and the coefficient of ∂x\\bar W is F_{\\barβ}e^{-Λ+\\barΛ}; if either assignment is wrong, the energy estimate for W in Proposition 3.3 collapses. This single algebraic check settles whether the central well-posedness estimate is sound.","verdict_should_be":"UNCHANGED","load_bearing_attack":"I read Theorem 1.3 as a statement about polynomial F, and within that scope the argument holds together. The well-posedness proof does not secretly require smooth F: the parabolic regularization u^ε is smooth for t>0, and the polynomial composition estimates in Proposition 2.2 and Lemma 3.2 cover the nonlinear terms. The ill-posedness construction in Lemma 4.2 is valid: in Case 1 the H^{s+δ}-rough perturbation has coefficients k^{-s-δ}, whose H^{s+δ} norm diverges while its H^s norm converges because δ>0, and the sign of the integral is preserved by the H^2 continuity of polynomial composition. The gauge-transformed energy estimate is the most delicate spot; the terms in (3.12) and (3.25) are consistent once the F_{\\barβ} contribution is recognized as carrying ∂x\\bar W, and condition (1.4) is used exactly to make the zero mode P0Fβ real so the skew-symmetric term integrates to zero. No internal inconsistency or unsupported leap in the central argument was found.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the Cauchy problem for the semi-linear Schrödinger equation ∂_t u + i∂_x^2 u = F(u, ∂_x u, u̅, ∂_x u̅) on the torus, where F is a polynomial in its four arguments. Theorem 1.3 claims that for s > 5/2 the problem is well-posed in H^s(T) if and only if condition (1.4) holds, namely ∫_T Im F_β(ψ, ∂_x ψ, ψ̅, ∂_x ψ̅) dx = 0 for every ψ ∈ H^s(T). The well-posedness part is proved by parabolic regularization, a gauge transformation tailored to F_β, energy estimates, and a Bona–Smith approximation argument. The ill-posedness part is proved by a contrapositive regularity statement (Theorem 4.1) and a construction of rough initial data (Lemma 4.2). The polynomial restriction is explicitly stated and used in the composition estimates and in the smoothness of the regularized solutions.","tokens_in":1563,"tokens_out":1810,"duration_ms":169656,"significance":"If the proof is completed as written, Theorem 1.3 is a substantial result: it gives a complete nonlinear analogue of the Mizohata condition for polynomial derivative NLS on the torus, and it upgrades Chihara's earlier conditional well-posedness to strong H^s well-posedness while adding a sharp non-existence statement. The paper is carefully structured, the main estimates are stated explicitly, and the polynomial scope is honestly delimited. The well-posedness proof is genuinely constructive (parabolic approximation plus Bona–Smith), and the ill-posedness mechanism via Fourier-side regularity of P_± φ is a clean and falsifiable criterion. The paper also provides a useful sufficient condition for non-existence in Theorem 4.1 that is of independent interest.","major_comments":[{"comment":"The text states that ∂_x Λ^ε, ∂_x^2 Λ^ε, and ∂_t Λ^ε are polynomials in u^ε, v^ε, w^ε and their conjugates, but this is not correct as written. Equation (3.11) contains the nonlocal term ∂_x^{-1} R_3^ε, which is not a polynomial in the field variables. Since the residual R^ε in (3.12) is subsequently estimated in Proposition 3.3 using polynomial composition estimates from Lemma 3.2, the proof of the key energy estimate is incomplete at this point. The gap is repairable: one needs a separate estimate for ∂_x^{-1} R_3^ε, for instance ‖∂_x^{-1} R_3^ε‖_{H^{r-1}} ≤ C(E_s^ε)(1+E_r^ε), and the residual in (3.12) should be rewritten so that this non-polynomial term is treated explicitly rather than absorbed into a polynomial class.","section":"Section 3.1, Eqs. (3.11)–(3.12)"},{"comment":"In the derivation of the difference estimate, the term coming from (3.32) is (ε_1 − ε_2) e^{-Λ^{ε_1}} ∂_x^2 v^{ε_2}. The displayed I_1 in (3.38), however, is −(ε_1 − ε_2) Re(∂_x v^{ε_2}, ∂_x V̆)_{H^{s-2}}. This is not a direct identity: the factor e^{-Λ^{ε_1}} is dropped, and the commutator of e^{-Λ^{ε_1}} with ⟨∂_x⟩^{s-2} is not accounted for. Since the bound |I_1| ≤ C|ε_1−ε_2| enters the Gronwall inequality (3.35) that yields convergence in C([0,T];H^{s-1}), the missing justification should be supplied.","section":"Section 3.1, Eq. (3.38)"}],"minor_comments":[{"comment":"After the definition of A ∼ B, the line “A /greaterorsimilarB” appears to be a typographical corruption of A ≳ B; please correct it.","section":"Section 1, notation"},{"comment":"Several expressions such as e^{-Λ^ε+Λ^ε} are ambiguous in the current typesetting. If the second Λ carries an overline, it should be typeset as e^{-Λ^ε+Λ̅^ε}; as printed the expression appears to be e^0, which is confusing and makes the gauge-transformation identities hard to verify.","section":"Section 3, Eq. (3.12) and elsewhere"},{"comment":"In Cases 2-2 and 2-3 the construction is clear, but it would help to state explicitly that the added Fourier series is supported on even integers so that the non-smoothness of P_+ψ^(ℓ) and P_−ψ^(ℓ) is preserved after adding the H^{s+δ} term ψ.","section":"Section 4, Lemma 4.2"},{"comment":"Reference [4] is listed only as “preprint”; if an updated published or arXiv version is available, it should be cited with full bibliographic data.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The theorem and the overall strategy are sound and well matched to the journal's scope. The two issues I raise in Section 3 are local and repairable with standard commutator and nonlocal-residual estimates; if the authors supply those details, I would be happy to support acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Plainly: this is a solid pure-PDE paper that proves a clean necessary-and-sufficient condition for local well-posedness of polynomial derivative NLS on the torus, and the proof holds up under scrutiny. The main theorem says that for s>5/2, well-posedness in H^s(T) is equivalent to the mean of Im F_beta being zero for every psi. That is a natural nonlinear analogue of the Mizohata condition, and it organizes several known examples under one criterion.\n\nWhat's new: the characterization itself, plus a distinct proof. The authors use parabolic regularization, a gauge transformation to remove the derivative term, and a Bona-Smith argument to get continuity in H^s, where Chihara only had convergence in H^{s'}. The ill-posedness side shows that when the mean is nonzero, the solution itself cannot exist for generic data—a stronger statement than mere norm inflation. They also disclose that Tsugawa independently obtained similar results, which is the right call.\n\nThe paper does several things well. The energy estimates are written out in detail; the zero-mode term is handled exactly by condition (1.4), and the commutator estimates in Section 2 are standard and correctly applied. The integration-by-parts decomposition in Proposition 4.3 is delicate but the bounds are consistent. I checked the spots the stress test flagged—the sign of the imaginary part in Lemma 4.2 and the gauge-transformed energy identity—and they are fine.\n\nSoft spots, proportionately. The theorem is only for polynomial F, and the authors say a smooth analogue would require more work; Remark 1.8 sketches a C^7 version for H^3 but it is not a full theorem. So readers needing the characterization for smooth nonlinearities are out of luck. The regularity threshold s>5/2 is not optimal; known equations like (1.2) are well-posed in much lower regularity. The ill-posedness is non-existence, not failure of the flow map, which is a legitimate form but weaker than some might expect. None of this undermines the stated scope.\n\nWho it's for: PDE analysts working on derivative NLS and well-posedness theory. It deserves a serious referee. I'd recommend accept after the usual checks, and I'd cite it.","headline":"A complete iff characterization for polynomial derivative NLS on the torus, with a solid proof; caveats are the polynomial-only scope and non-optimal regularity threshold.","tokens_in":27755,"tokens_out":2466,"would_cite":true,"duration_ms":23486,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35Q55"],"pacs":[],"model":"deepseek-v4-flash","headline":"Well-posedness of polynomial derivative Schrödinger equations on the torus is equivalent to a single integral condition on the nonlinearity.","keywords":["Schrödinger equation","well-posedness","ill-posedness","Mizohata condition","gauge transformation","energy estimate","torus","derivative nonlinear Schrödinger equation"],"falsifier":"Take $F(\\alpha,\\beta,\\bar\\alpha,\\bar\\beta)=i(2|\\alpha|^2\\beta+\\alpha^2\\bar\\beta)$ and $\\psi\\equiv 1$, so $\\int_{\\mathbb T}\\operatorname{Im} F_\\beta\\,dx=2\\neq 0$; the theorem predicts that some $H^s$ initial data admit no solution. Exhibiting a solution for every $H^s$ initial data for this equation, for any fixed $s>5/2$, would refute Theorem 1.4, while verifying the predicted loss on the $P_\\pm$ side would corroborate it.","tokens_in":26746,"feed_emoji":"🌊","tokens_out":9808,"duration_ms":79805,"temperature":0.7,"pith_summary":"This paper asks when a polynomial derivative nonlinear Schrödinger equation on the circle, $\\partial_t u + i\\partial_x^2 u = F(u,\\partial_x u,\\bar u,\\partial_x \\bar u)$, has a unique solution that depends continuously on its initial data in the Sobolev space $H^s(\\mathbb T)$ for $s>5/2$. The answer, Theorem 1.3, is that well-posedness holds exactly when a single condition is met: the average of the imaginary part of $F_\\beta$ (the derivative of $F$ in the direction of $\\partial_x u$) vanishes on every profile $\\psi$. This is the nonlinear analogue of the Mizohata condition known for linear equations, and it is sharp: when the integral fails for some $\\psi$, the authors construct initial data for which no solution exists in either time direction. The proof splits into an energy-estimate and gauge-transformation argument for the well-posed half and an integration-by-parts and non-existence argument for the ill-posed half.","feed_headline":"One integral decides well-posedness for derivative Schrödinger equations","feed_subtitle":"If this average is nonzero, some initial data admit no solution at all.","key_machinery":"The central object is the gauge-transformed second derivative $W = \\exp\\bigl(-\\tfrac{i}{2}\\partial_x^{-1} F_\\beta(u,\\partial_x u,\\bar u,\\partial_x\\bar u)\\bigr)\\,\\partial_x^2 u$, where $\\partial_x^{-1}$ is the zero-mean inverse derivative on the torus. Applying this transformation removes the most dangerous term $F_\\beta \\,\\partial_x w$ from the equation for $w=\\partial_x^2 u$, leaving a term whose coefficient is the spatial average of $F_\\beta$ times $\\partial_x W$ plus lower-order polynomial terms. When condition (1.4) holds, that average is real, so the offending term becomes a pure derivative and contributes nothing to the $H^s$ energy; when the average has a nonzero imaginary part, a first-order Cauchy-Riemann-type elliptic operator survives, and the integration-by-parts argument shows that a solution would force one Fourier projection of the initial data to be smoother than $H^s$. The threshold $s>5/2$ arises from the commutator estimates needed for the energy method.","core_discovery":"Under the assumption that $F$ is a polynomial in $\\alpha,\\beta,\\bar\\alpha,\\bar\\beta$, the Cauchy problem is locally well-posed in $H^s(\\mathbb T)$ for $s>5/2$ if and only if $\\int_{\\mathbb T} \\operatorname{Im} F_\\beta(\\psi,\\partial_x\\psi,\\bar\\psi,\\partial_x\\bar\\psi)\\,dx = 0$ for every $\\psi\\in H^s(\\mathbb T)$. Here $F_\\beta$ is the partial derivative of $F$ with respect to its second argument, the slot occupied by $\\partial_x u$. The well-posedness part upgrades Chihara's earlier existence and uniqueness result from weak to strong continuity in $H^s$. The ill-posedness part is proved as non-existence: if the integral is nonzero for some $\\psi$, there is a $\\varphi\\in H^s(\\mathbb T)$ such that no solution exists in $C([0,T];H^s(\\mathbb T))$ or $C([-T,0];H^s(\\mathbb T))$ for any $T>0$; in fact, any solution would force one of the projections $P_+\\varphi$ or $P_-\\varphi$ to gain regularity, depending on the sign of the integral.","pith_inferences":["The sign-dependent half-line smoothing effect in Theorem 4.1 suggests that when (1.5) holds, initial data with one Fourier half-line sufficiently smooth may still evolve, while data rough on the wrong side are forbidden; the paper explicitly leaves such conditional existence questions open.","One expects the same dichotomy for smooth, non-polynomial nonlinearities by approximation, but the paper's $C^7$ remark covers only the $H^3$ well-posedness side, so a full smooth analogue would require fractional composition estimates that are not provided.","The mechanism connecting a nonzero Mizohata mean to a first-order elliptic operator and forced half-line regularity links this failure to Fourier half-space well-posedness, where the solution space itself imposes the one-sided smoothness that the obstruction demands.","A concrete numerical check is possible: run the parabolic regularization for $F=i(2|\\alpha|^2\\beta+\\alpha^2\\bar\\beta)$ with rough data whose Fourier modes are nonzero on both sides; the predicted non-existence should appear as catastrophic growth as the regularization parameter tends to zero, in contrast to the condition-holding model $\\partial_x(|u|^2u)$."],"forward_implications":["For every polynomial $F$ and every $s>5/2$, well-posedness in $H^s(\\mathbb T)$ is decided by the one condition (1.4); no other structure of the nonlinearity matters.","When the condition fails, the Cauchy problem is ill-posed in the strongest sense: some $H^s$ initial data admit no solution at all, not merely a discontinuous solution map.","The criterion reproduces the known split between conjugate-derivative models such as $\\partial_x(\\bar u^m)$, for which $F_\\beta=0$ and the condition holds, and models such as $\\partial_x(u^m)$ with $m\\ge 2$, for which the condition fails and the problem is ill-posed.","The well-posedness conclusion gives continuity of the solution map in the full $H^s$ norm, improving Chihara's earlier weak-continuity result, and yields persistence of regularity for smoother initial data on the same time interval."],"supporting_citations":[{"why":"Established existence and uniqueness under an equivalent condition and weak continuity; the present well-posedness proof builds on and upgrades it to strong $H^s$ continuity.","marker":"[3]"},{"why":"Supplies the Bona-Smith approximation scheme, namely truncation of initial data, used to pass from smooth solutions to continuous dependence in $H^s$.","marker":"[2]"},{"why":"Proved well-posedness for the conjugate-derivative model, one of the benchmark examples the new criterion must reproduce.","marker":"[7]"},{"why":"Proved ill-posedness for $\\partial_x(u^m)$ on the torus and on the line, providing the baseline for the non-existence conclusion.","marker":"[4]"},{"why":"The integration-by-parts technique combined with a gauge transformation that the ill-posedness proof adapts to derive the $P_\\pm$ regularity obstruction.","marker":"[14]"},{"why":"An independent simultaneous proof of similar results, cited for comparison of the energy method and the linear variable-coefficient approach.","marker":"[24]"},{"why":"Provides the multilinear and bilinear estimates used to control products of $H^s$ functions in the energy argument.","marker":"[21]"},{"why":"Supplies the composition estimate (Proposition 2.2) that makes the nonlinear term well-defined for $s>3/2$.","marker":"[11]"}],"fun_headline_variants":["Integral condition decides well-posedness for Schrödinger on torus","Nonzero integral forbids solutions for some initial data","Well-posedness iff one integral over torus vanishes","Schrödinger Cauchy problem: solvability pinned by a single average","Vanishing integral ensures well-posedness; nonzero breaks it"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof treats $F$ as a polynomial nonlinearity: the composition estimates and the regularity of the parabolic smoothed solutions are established only in that setting, so a reader wanting the same equivalence for merely smooth nonlinearities would need a new argument.","fun_headline_variants_meta":{"raw":{"variants":["Integral condition decides well-posedness for Schrödinger on torus","Nonzero integral forbids solutions for some initial data","Well-posedness iff one integral over torus vanishes","Schrödinger Cauchy problem: solvability pinned by a single average","Vanishing integral ensures well-posedness; nonzero breaks it"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000222,"raw_usage":{"total_tokens":1424,"prompt_tokens":887,"completion_tokens":537,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":503,"completion_tokens_details":{"reasoning_tokens":450}},"tokens_in":503,"tokens_out":537,"duration_ms":5069,"temperature":1.0,"reasoning_tokens":450,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T21:40:24.493907+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take $F(\\alpha,\\beta,\\bar\\alpha,\\bar\\beta)=i(2|\\alpha|^2\\beta+\\alpha^2\\bar\\beta)$ and $\\psi\\equiv 1$, so $\\int_{\\mathbb T}\\operatorname{Im} F_\\beta\\,dx=2\\neq 0$; the theorem predicts that some $H^s$ initial data admit no solution. Exhibiting a solution for every $H^s$ initial data for this equation, for any fixed $s>5/2$, would refute Theorem 1.4, while verifying the predicted loss on the $P_\\pm$ side would corroborate it.","supporting_citations":[{"cited_title":"Chihara, The initial value problem for Schr¨ odinger equations on the torus, Int","cited_arxiv_id":null,"evidence_quote":"Established existence and uniqueness under an equivalent condition and weak continuity; the present well-posedness proof builds on and upgrades it to strong $H^s$ continuity."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Bona-Smith approximation scheme, namely truncation of initial data, used to pass from smooth solutions to continuous dependence in $H^s$."},{"cited_title":"Christ, Illposedness of a Schr¨ odinger equation with derivative no nlinearity, preprint","cited_arxiv_id":null,"evidence_quote":"Proved ill-posedness for $\\partial_x(u^m)$ on the torus and on the line, providing the baseline for the non-existence conclusion."},{"cited_title":"Kishimoto, Y","cited_arxiv_id":null,"evidence_quote":"The integration-by-parts technique combined with a gauge transformation that the ill-posedness proof adapts to derive the $P_\\pm$ regularity obstruction."},{"cited_title":"Tsugawa, Local well-posedness of derivative Schr¨ odinger equation s on the torus , preprint","cited_arxiv_id":null,"evidence_quote":"An independent simultaneous proof of similar results, cited for comparison of the energy method and the linear variable-coefficient approach."},{"cited_title":"Tao, Multilinear weighted convolution of L2-functions, and applications to nonlin- ear dispersive equations , Amer","cited_arxiv_id":null,"evidence_quote":"Provides the multilinear and bilinear estimates used to control products of $H^s$ functions in the energy argument."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the composition estimate (Proposition 2.2) that makes the nonlinear term well-defined for $s>3/2$."}],"review_version":1}